add 17

17/paper.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8b0fc065482b1eb4516d89a5be07f984a421548fd1c966979f01941928ee03a
+size 1797770

17/replication_package/code/LICENSE-code.md

@@ -0,0 +1,11 @@
+MIT License
+
+Copyright (c) 2021 Hunt Allcott, Matthew Gentzkow, and Lena Song
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+

17/replication_package/code/LICENSE-data.md

@@ -0,0 +1,297 @@
+
+[Creative Commons Attribution 4.0 International Public License](https://creativecommons.org/licenses/by/4.0/)
+- applies to databases, images, tables, text, and any other objects
+
+
+#### Creative Commons Attribution 4.0 International Public License
+By exercising the Licensed Rights (defined below), You accept and agree
+to be bound by the terms and conditions of this Creative Commons
+Attribution 4.0 International Public License ("Public License"). To the
+extent this Public License may be interpreted as a contract, You are
+granted the Licensed Rights in consideration of Your acceptance of these
+terms and conditions, and the Licensor grants You such rights in
+consideration of benefits the Licensor receives from making the Licensed
+Material available under these terms and conditions.
+
+Section 1 – Definitions.
+
+(a)    Adapted Material means material subject to Copyright and Similar
+Rights that is derived from or based upon the Licensed Material and in
+which the Licensed Material is translated, altered, arranged,
+transformed, or otherwise modified in a manner requiring permission
+under the Copyright and Similar Rights held by the Licensor. For
+purposes of this Public License, where the Licensed Material is a
+musical work, performance, or sound recording, Adapted Material is
+always produced where the Licensed Material is synched in timed relation
+with a moving image.
+
+(b)   Adapter's License means the license You apply to Your Copyright
+and Similar Rights in Your contributions to Adapted Material in
+accordance with the terms and conditions of this Public License.
+
+(c)  Copyright and Similar Rights means copyright and/or similar rights
+closely related to copyright including, without limitation, performance,
+broadcast, sound recording, and Sui Generis Database Rights, without
+regard to how the rights are labeled or categorized. For purposes of
+this Public License, the rights specified in Section 2(b)(1)-(2) are not
+Copyright and Similar Rights.
+    
+(d) Effective Technological Measures means those measures that, in the
+absence of proper authority, may not be circumvented under laws
+fulfilling obligations under Article 11 of the WIPO Copyright Treaty
+adopted on December 20, 1996, and/or similar international agreements.
+  
+(e) Exceptions and Limitations means fair use, fair dealing, and/or any
+other exception or limitation to Copyright and Similar Rights that
+applies to Your use of the Licensed Material.
+
+(f) Licensed Material means the artistic or literary work, database, or
+other material to which the Licensor applied this Public License.
+
+(g) Licensed Rights means the rights granted to You subject to the terms
+and conditions of this Public License, which are limited to all
+Copyright and Similar Rights that apply to Your use of the Licensed
+Material and that the Licensor has authority to license.
+
+(h) Licensor means the individual(s) or entity(ies) granting rights
+under this Public License.
+
+(i) Share means to provide material to the public by any means or
+process that requires permission under the Licensed Rights, such as
+reproduction, public display, public performance, distribution,
+dissemination, communication, or importation, and to make material
+available to the public including in ways that members of the public may
+access the material from a place and at a time individually chosen by
+them.
+
+(j) Sui Generis Database Rights means rights other than copyright
+resulting from Directive 96/9/EC of the European Parliament and of the
+Council of 11 March 1996 on the legal protection of databases, as
+amended and/or succeeded, as well as other essentially equivalent rights
+anywhere in the world.
+
+(k) You means the individual or entity exercising the Licensed Rights
+under this Public License. Your has a corresponding meaning.
+
+Section 2 – Scope.
+
+(a) License grant.
+
+  1. Subject to the terms and conditions of this Public License, the
+  Licensor hereby grants You a worldwide, royalty-free,
+  non-sublicensable, non-exclusive, irrevocable license to exercise the
+  Licensed Rights in the Licensed Material to:
+
+    A. reproduce and Share the Licensed Material, in whole or in part;
+    and
+
+    B. produce, reproduce, and Share Adapted Material.
+
+  2. Exceptions and Limitations. For the avoidance of doubt, where
+  Exceptions and Limitations apply to Your use, this Public License does
+  not apply, and You do not need to comply with its terms and
+  conditions.
+        
+  3. Term. The term of this Public License is specified in Section 6(a).
+  
+  4. Media and formats; technical modifications allowed. The Licensor
+  authorizes You to exercise the Licensed Rights in all media and
+  formats whether now known or hereafter created, and to make technical
+  modifications necessary to do so. The Licensor waives and/or agrees
+  not to assert any right or authority to forbid You from making
+  technical modifications necessary to exercise the Licensed Rights,
+  including technical modifications necessary to circumvent Effective
+  Technological Measures. For purposes of this Public License, simply
+  making modifications authorized by this Section 2(a)(4) never produces
+  Adapted Material.
+  
+  5. Downstream recipients.
+      
+      A. Offer from the Licensor – Licensed Material. Every recipient of
+      the Licensed Material automatically receives an offer from the
+      Licensor to exercise the Licensed Rights under the terms and
+      conditions of this Public License.
+      
+      B. No downstream restrictions. You may not offer or impose any
+      additional or different terms or conditions on, or apply any
+      Effective Technological Measures to, the Licensed Material if
+      doing so restricts exercise of the Licensed Rights by any
+      recipient of the Licensed Material.
+        
+  6. No endorsement. Nothing in this Public License constitutes or may
+  be construed as permission to assert or imply that You are, or that
+  Your use of the Licensed Material is, connected with, or sponsored,
+  endorsed, or granted official status by, the Licensor or others
+  designated to receive attribution as provided in Section
+  3(a)(1)(A)(i).
+
+(b) Other rights.
+  
+  1. Moral rights, such as the right of integrity, are not licensed
+  under this Public License, nor are publicity, privacy, and/or other
+  similar personality rights; however, to the extent possible, the
+  Licensor waives and/or agrees not to assert any such rights held by
+  the Licensor to the limited extent necessary to allow You to exercise
+  the Licensed Rights, but not otherwise.
+  
+  2. Patent and trademark rights are not licensed under this Public
+  License.
+  
+  3. To the extent possible, the Licensor waives any right to collect
+  royalties from You for the exercise of the Licensed Rights, whether
+  directly or through a collecting society under any voluntary or
+  waivable statutory or compulsory licensing scheme. In all other cases
+  the Licensor expressly reserves any right to collect such royalties.
+
+Section 3 – License Conditions.
+
+Your exercise of the Licensed Rights is expressly made subject to the
+following conditions.
+
+(a) Attribution.
+
+  1. If You Share the Licensed Material (including in modified form),
+  You must:
+      
+    A. retain the following if it is supplied by the Licensor with the
+    Licensed Material:
+      
+      i.   identification of the creator(s) of the Licensed Material and
+      any others designated to receive attribution, in any reasonable
+      manner requested by the Licensor (including by pseudonym if
+      designated);
+      
+      ii.  a copyright notice;
+      
+      iii. a notice that refers to this Public License;
+      
+      iv.  a notice that refers to the disclaimer of warranties;
+      
+      v.   a URI or hyperlink to the Licensed Material to the extent
+      reasonably practicable;
+    
+    B. indicate if You modified the Licensed Material and retain an
+    indication of any previous modifications; and
+    
+    C. indicate the Licensed Material is licensed under this Public
+    License, and include the text of, or the URI or hyperlink to, this
+    Public License.
+  
+  2. You may satisfy the conditions in Section 3(a)(1) in any reasonable
+  manner based on the medium, means, and context in which You Share the
+  Licensed Material. For example, it may be reasonable to satisfy the
+  conditions by providing a URI or hyperlink to a resource that includes
+  the required information.
+  
+  3. If requested by the Licensor, You must remove any of the
+  information required by Section 3(a)(1)(A) to the extent reasonably
+  practicable.
+  
+  4. If You Share Adapted Material You produce, the Adapter's License
+  You apply must not prevent recipients of the Adapted Material from
+  complying with this Public License.
+
+Section 4 – Sui Generis Database Rights.
+
+Where the Licensed Rights include Sui Generis Database Rights that apply
+to Your use of the Licensed Material:
+
+(a) for the avoidance of doubt, Section 2(a)(1) grants You the right to
+extract, reuse, reproduce, and Share all or a substantial portion of the
+contents of the database;
+
+(b) if You include all or a substantial portion of the database contents
+in a database in which You have Sui Generis Database Rights, then the
+database in which You have Sui Generis Database Rights (but not its
+individual contents) is Adapted Material; and
+
+(c) You must comply with the conditions in Section 3(a) if You Share all
+or a substantial portion of the contents of the database.
+
+For the avoidance of doubt, this Section 4 supplements and does not
+replace Your obligations under this Public License where the Licensed
+Rights include other Copyright and Similar Rights.
+
+Section 5 – Disclaimer of Warranties and Limitation of Liability.
+
+(a) Unless otherwise separately undertaken by the Licensor, to the
+extent possible, the Licensor offers the Licensed Material as-is and
+as-available, and makes no representations or warranties of any kind
+concerning the Licensed Material, whether express, implied, statutory,
+or other. This includes, without limitation, warranties of title,
+merchantability, fitness for a particular purpose, non-infringement,
+absence of latent or other defects, accuracy, or the presence or absence
+of errors, whether or not known or discoverable. Where disclaimers of
+warranties are not allowed in full or in part, this disclaimer may not
+apply to You.
+
+(b) To the extent possible, in no event will the Licensor be liable to
+You on any legal theory (including, without limitation, negligence) or
+otherwise for any direct, special, indirect, incidental, consequential,
+punitive, exemplary, or other losses, costs, expenses, or damages
+arising out of this Public License or use of the Licensed Material, even
+if the Licensor has been advised of the possibility of such losses,
+costs, expenses, or damages. Where a limitation of liability is not
+allowed in full or in part, this limitation may not apply to You.
+
+(c) The disclaimer of warranties and limitation of liability provided
+above shall be interpreted in a manner that, to the extent possible,
+most closely approximates an absolute disclaimer and waiver of all
+liability.
+
+Section 6 – Term and Termination.
+
+(a) This Public License applies for the term of the Copyright and
+Similar Rights licensed here. However, if You fail to comply with this
+Public License, then Your rights under this Public License terminate
+automatically.
+
+(b) Where Your right to use the Licensed Material has terminated under
+Section 6(a), it reinstates:
+  
+  1. automatically as of the date the violation is cured, provided it is
+  cured within 30 days of Your discovery of the violation; or
+  
+  2. upon express reinstatement by the Licensor.
+
+  For the avoidance of doubt, this Section 6(b) does not affect any
+  right the Licensor may have to seek remedies for Your violations of
+  this Public License.
+
+(c) For the avoidance of doubt, the Licensor may also offer the Licensed
+Material under separate terms or conditions or stop distributing the
+Licensed Material at any time; however, doing so will not terminate this
+Public License.
+
+(d) Sections 1, 5, 6, 7, and 8 survive termination of this Public License.
+
+Section 7 – Other Terms and Conditions.
+
+(a) The Licensor shall not be bound by any additional or different terms
+or conditions communicated by You unless expressly agreed.
+
+(b) Any arrangements, understandings, or agreements regarding the
+Licensed Material not stated herein are separate from and independent of
+the terms and conditions of this Public License.
+
+Section 8 – Interpretation.
+
+(a) For the avoidance of doubt, this Public License does not, and shall
+not be interpreted to, reduce, limit, restrict, or impose conditions on
+any use of the Licensed Material that could lawfully be made without
+permission under this Public License.
+
+(b) To the extent possible, if any provision of this Public License is
+deemed unenforceable, it shall be automatically reformed to the minimum
+extent necessary to make it enforceable. If the provision cannot be
+reformed, it shall be severed from this Public License without affecting
+the enforceability of the remaining terms and conditions.
+
+(c) No term or condition of this Public License will be waived and no
+failure to comply consented to unless expressly agreed to by the
+Licensor.
+
+(d) Nothing in this Public License constitutes or may be interpreted as
+a limitation upon, or waiver of, any privileges and immunities that
+apply to the Licensor or You, including from the legal processes of any
+jurisdiction or authority.

17/replication_package/code/README.md

@@ -0,0 +1,303 @@
+# README
+
+## Overview
+The code in this replication package constructs the analysis tables, figures and scalars found in our paper using Stata and R. 
+The results presented in our paper are obtained in three steps. 
+In the first step, all original raw data is processed from our server. 
+In the second step, the raw data is stripped of any PII elements and all anonymized datasets are merged together to create a dataset named `final_data_sample.dta`. 
+All of the analysis presented in the paper is based on this anonymized data. 
+In the third step, all descriptive tables and figures as well as all regression outputs are produced.
+
+In this replication archive, we provide the code necessary to carry out all three steps. We provide the anonymized dataset  `final_data_sample.dta` in a separate archive (Allcott, Gentzkow, and Song, 2023): https://doi.org/10.7910/DVN/GN636M.
+
+Under access to that separate archive, please download  `final_data_sample.dta`. Then, manually add the dataset to this archive under the folder `data/temptation/output`. This will allow you  to run the `/analysis/` module (third step) which constructs the tables, figures and scalars found in the paper. The module `/data/` relies on confidential data which is not provided, and therefore will not properly run. 
+
+This replication archive contains additional files to help replication in the `/docs/` folder. 
+The first is called `DescriptionOfSteps.pdf` and it describes which modules are included in either steps as well as how they relate to each other. The file  `Step1_Step2_DAG.pdf` illustrates how step 1 and 2 are carried via a directed-acyclic graph.
+The third is called `MappingsTablesAndFigures.pdf` and it provides a mapping of all the tables
+and figures to their corresponding program.
+
+The replication routine can be run by following the instructions in the **Instructions to replicators** section of this README. 
+Support by the authors for replication is provided if necessary. The replicator should expect the code to run for about 40 minutes.
+
+
+## Data Availability and Provenance Statements
+This archive includes data that was collected from an Android application and from surveys as detailed in Section 3 of the paper. 
+The folder `experiment_design` contains the questionnaires of all 5 surveys (recruitment survey and the next 4 surveys administered to our sample). It also contains a subfolder `AppScreenshots` that has various screenshots of our application Phone Dashboard.
+In the separate archive, we provide the anonymized dataset  `final_data_sample.dta` which gathers aggregated usage data from the application and survey data. Each individual in our final sample is assigned a user ID. Variables that correspond to information coming from the application start with `PD_P1`, `PD_P2`, `PD_P3`, `PD_P4`, `PD_P5` (depending on which period 1-5 they were collected). Variables that correspond to information coming from surveys start with either `S1_`, `S2_`, `S3_` or `S4_` (depending on which survey 1-4 they were collected). 
+The `codebook.xlsx` file at the root of the repository is the codebook for `final_data_sample.dta`. It lists all the variables found in the dataset along with their labels, units and values (if applicable). 
+
+### Statement about Rights
+We certify that the author(s) of the manuscript have legitimate access to and permission to use the data used in this manuscript. 
+
+###  License for Data
+
+All databases, images, tables, text, and any other objects are available under a Creative Commons Attribution 4.0 International Public License. Please refer to the document `LICENSE-data.md` at the root of the repository.
+
+### Summary of Availability
+
+Some data **cannot be made** publicly available.
+
+
+### Details on each Data Source
+
+The raw data for this project are confidential and were collected by the authors. The authors will assist with any reasonable replication attempts and can be contacted by email.  This paper uses data obtained from an Android application Phone Dashboard  and surveys. 
+
+The dataset  `final_data_sample.dta`, provided in the separate archive, combines the data from both our surveys and our Phone Dashboard application. It is derived after processing all the raw confidential data from the Phone Dashboard application. This dataset aggregates usage data at the user level and combines it with variables obtained from our surveys. All variables in this dataset have corresponding value labels. One can also refer to the provided codebook, `codebook.xlsx`, at the root of the repository, for more information on each variable. 
+
+## Dataset list
+- As detailed in the graph `doc/Step1_Step2_DAG.pdf`, our pipeline processes the raw data from our Phone Dashboard application as well as from our surveys. The code for this data-processing is provided in the `/data/` folder.  Multiple intermediate files are generated through this pipeline. These files are not provided as part of this replication archive for confidentiality reasons. 
+
+- The file `final_data_sample.dta`, provided in the separate archive,  is obtained at the end of the data-processing pipeline. It combines data from the application and the surveys. It serves as input for the analysis figures and tables. This file is provided in this replication archive.
+
+
+
+## Computational requirements
+All requirements must be installed and set up for command line usage. For further detail, see the **Command Line Usage** section below.
+
+We manage Python and R installations using conda or miniconda. 
+To build the repository as-is, the following applications are additionally required:
+* LyX 2.3.5.2
+* R 3.6.3
+* Stata 16.1
+* Python 3.7
+
+These software are used by the scripts contained in the repository in the `setup` folder. Instructions to set up the environment are found below in the section `Instructions to run the repository`. 
+
+### Software requirements
+The file `setup/conda_env.yaml` will install all the R and Python dependencies. Please refer to the section `Instructions to run the repository` for detailed steps on how to install the required environment and run the scripts. 
+Below we list  the softwares and packages required to run the repository with the version used.
+
+ - Python 3.7
+ - `pyyaml` (5.3.1)
+    - `numpy` (1.16.2)
+    - `pandas` (0.25.0)
+    - `matplotlib` (3.0.3)
+    - `gitpython` (2.1.15)
+    - `termcolor` (1.1.0)
+    - `colorama` (0.4.3)
+    - `jupyter` (4.6.3)
+    - `future` (0.17.1)
+    - `linearmodels` (4.17)
+    - `patsy` (0.5.1)
+    - `stochatreat` (0.0.8)
+    - `pympler` (0.9)
+    - `memory_profiler`
+    - `dask`(1.2.1)
+    - `openpyxl` (2.6.4)
+    - `requests` (2.24.0)
+    - `pip` (19)
+    
+    - R 3.6
+     - `yaml` (2.2.1)
+     - `haven` (2.3.1)
+     - `tidyverse` (1.3.1)
+     - `r.utils` (4.0.3)
+     - `plm` (2.6.1)
+     - `janitor` (2.1.0)
+     - `rio` (0.5.26)
+     - `lubridate` (1.7.10)
+     - `magrittr` (2.0.1)
+     - `stargazer` (5.2.2)
+     - `rootSolve` (1.8.2.1)
+     - `rlist` (0.4.6.1)
+     - `ebal` (0.1.6)
+     - `latex2exp` (0.5.0)
+     - `estimatr` (0.30.2)
+     
+### Controlled Randomness
+We control randomness by setting random seeds. 
+1. For the data-processing: The program`/data/source/clean_master/cleaner.py` has its own random seed set on line 24. The program `data/source/build_master/builder.py` calls the `/lib/` file  `lib/data_helpers/clean_survey.py` that sets a random seed on line 21.The program `lib/experiment_specs/study_config.py` contains parameters used by `data/source/clean_master/management/earnings.py` and `data/source/clean_master/management/midline_prep.py` which include a random seed set on line 459. 
+2. For the analysis: The program `lib/ModelFunctions.R` contains parameters used by `structural/code/StructuralModel.R` and `treatment_effects/code/ModelHeterogeneity.R` which include a random seed set on line 48. 
+
+### Memory and Runtime Requirements
+The folder `/data/` is responsible for all the data-processing using the raw Phone Dashboard data as well as the survey data. At the end of this data-processing, the file `final_data_sample.dta` is created. In the presence of the raw confidential data (which is not provided with this replication archive), this whole process normally takes around 60 hours on 20 CPUs and 12GB memory per CPU. 
+
+The folder `/analysis/` is responsible for the construction of all the figures, plots and scalars used in the paper using the `final_data_sample.dta`  dataset provided in the separate archive. The replicator will be able to run all scripts in this folder. The whole analysis takes around 40 minutes to run on a computer with 4 cores and 16GB of memory. Most files within `analysis` take less than 5 minutes to run. However, the file `analysis/code/StructuralModel.R` takes around 20 minutes to run. 
+
+#### Summary
+Approximate time needed to reproduce the analyses on a standard (2022) desktop machine is <1 hour. 
+
+#### Details
+The `analysis` code was last run on a **4-core Intel-based laptop with MacOS version 10.15.5**. 
+
+The `data` code was last run on a **an Intel server with 20 CPUs and 12GB of memory per CPU**. Computation took 60 hours. 
+
+
+## Description of programs/code 
+In this replication archive : 
+- The folder `/data/source/`  is responsible for all the data processing of our Phone Dashboard application and our surveys.
+The subfolders `/data/source/build_master/`, `/data/source/clean_master/` and `/data/source/exporters/`  contains Python files that define the  classes and auxiliary functions called in the main script `/data/run.py`. This main script generates the master files gathering all information at the user-level or at the user-app-level. 
+
+- The folder `/data/temptation/`  is responsible for cleaning the master files produced as output of `/data/source/`.
+It outputs the anonymized dataset `final_data_sample.dta` which contains all the information at the user level. This dataset is used throughout the analysis of the paper. 
+
+- The folder `/analysis/` contains all the programs generating the tables, figures and scalars in the paper. The programs in the `/analysis/` folder has been categorised under three subfolders:
+
+   1. `/analysis/descriptive/` produces tables and charts of descriptives statistics. It contains the below programs: 
+        * `code/CommitmentDemand.do` (willingness-to-pay and limit tightness plots)
+        * `code/COVIDResponse.do` (survey stats on response to COVID)
+        * `code/DataDescriptive.do` (sample demographics and attrition tables)
+        * `code/HeatmapPlots.R` (predicted vs. actual FITSBY usage)
+        * `code/QualitativeEvidence.do` (descriptive plots for addiction scale, interest in bonus/limit)
+        * `code/SampleStatistics.do` (statistics about completion rates for study)
+        * `code/Scalars.do` (statistics about MPL and ideal usage reduction)
+        * `code/Temptation.do` (plots desired usage change for various tempting activities)
+    
+    2. `/analysis/structural/` estimates parameters and generates plots for our structural model. It contains the below program: 
+        * `code/StructuralModel.R`
+    
+    3. `/analysis/treatment_effects/` produces model-free estimates of treatment effects. It contains the below programs : 
+        *  `code/Beliefs.do` (compares actual treatment effect with predicted treatment effect)
+        *  `code/CommitmentResponse.do` (plots how treatment effect differs by SMS addiction scale and other survey indicators)
+        *  `code/FDRTable.do` (estimates how treatment effect differs by SMS addiction scale and other indicators, adjusted for false-discovery rate. Also plots some descriptive statistics)
+        *  `code/HabitFormation.do` (compares actual and predicted usage)
+        *  `code/Heterogeneity.do` (plots heterogeneous treatment effects)
+        *  `code/HeterogeneityInstrumental.do`  (plots heterogeneous treatment effects)
+        *  `code/ModelHeterogeneity.R` (generates other heterogeneity plots, some temptation plots)
+        *  `code/SurveyValidation.do` (plots effect of rewarding accurate usage prediction on usage prediction accuracy)
+        Most of the programs in the analysis folder rely on the dataset `final_data_sample.dta`. However, some programs further require the datasets `final_data.dta` and `AnalysisUser.dta` to compte certain scalars mentioned in the paper. These programs are `/analysis/descriptive/code/DataDescriptive.do`, `/analysis/descriptive/code/SampleStatistics.do`,  `/analysis/descriptive/code/Scalars.do` and `/analysis/treatment_effects/code/ModelHeterogeneity.R`. Since these two datasets are not provided with the replication archive for confidentiality reasons, the portions of the code requiring them have been commented out in the relevant programs.
+                
+        - The folder `/lib/` contains auxiliary functions and helpers.
+        - The folder `/paper_slides/` contains all the input and files necessary to the compiling of the paper. The subfolder `/paper_slides/figures/` contains screenshots and other figures that are not derived from programs. The subfolder `/paper_slides/figures/` contains the paper Lyx file, the bibliography as well as the `motivation_correlation.lyx` Lyx table. 
+        - The folder `setup` contains files to setup the conda environment as well as to install the R, Python and Stata dependencies. 
+        - The folder `experiment_design` contains the questionnaires to our surveys as well as screenshots from the Phone Dashboard application.
+        - The folder `/docs/` contains additional documents to guide the replicator. The file  `docs/DescriptionOfSteps.pdf` gives a high-level overview of the steps involved in the data processing from our
+        application Phone Dashboard to the analysis in the paper. It splits the data-processing into three steps : 
+                        1) Processing the Raw Data from PhoneDashboard (done by the `/data/source/` folder)
+                        2) Cleaning the Original Data from PhoneDashboard  (done by the `/data/temptation/` folder)
+                        3) Analyze the Anonymized Data  (done by the `/analysis/` folder)
+        Since the data inputs for step 1 and 2 are not provided with this replication archive, we include a further document  `docs/Step1_Step2_DAG.pdf` that illustrate how we carried them internally via a
+        directed-acyclic graph. Finally, the file `docs/MappingsTablesAndFigures.pdf` provides a mapping of all the tables and figures to their corresponding program.
+        
+        Note that the modules or portions of programs that cannot be run due to unshared data have been commented out in the relevant main run scripts. 
+
+### License for code 
+
+All code is available under a MIT License. Please refer to the document `LICENSE-code.md` at the root of the repository.
+
+## Instructions to replicators
+
+### Setup 
+
+1. Create a `config_user.yaml` file in the root directory. A template can be found in the `setup` subdirectory. See the **User Configuration** section below for further detail. If you do not have any external paths you wish to specify, and wish to use the default executable names you can skip this step and the default `config_user.yaml` will be copied over in step 4.
+
+2. If you already have conda setup on your local machine, feel free to skip this step. If not, this will install a lightweight version of conda that will not interfere with your current python and R installations.
+Install miniconda and jdk to be used to manage the R/Python virtual environment, if you have not already done this. You can install these programs from their websites [here for miniconda](https://docs.conda.io/en/latest/miniconda.html) and [here for jdk](https://www.oracle.com/java/technologies/javase-downloads.html). If you use homebrew (which can be download [here](https://brew.sh/)) these two programs can be downloaded as follows:
+   ```
+   brew install --cask miniconda
+   brew install --cask oracle-jdk
+   ```
+Once you have done this you need to initialize conda by running the following lines and restarting your terminal:
+   ```
+   conda config --set auto_activate_base false
+   conda init $(echo $0 | cut -d'-' -f 2)
+   ```
+
+3. Create conda environment with the command:
+   ```
+   conda env create -f setup/conda_env.yaml
+   ```
+
+4. Run the `check_setup.py` file. One way to do this is to run the following bash command in a terminal from the `setup` subdirectory:
+   ```
+   python3 check_setup.py
+   ```
+
+5. Install R dependencies that cannot be managed using conda with the `setup_r.r` file. One way to do this is to run the following bash command in a terminal from the `setup` subdirectory:
+   ```
+   Rscript setup_r.r
+   ```
+   
+### Usage
+
+Once you have succesfully completed the **Setup** section above, each time that you run any analysis make sure the virtual environment associated with this project is activated, using the command below (replacing with the name of this project).
+```
+   conda activate PROJECT_NAME
+``` 
+If you wish to return to your base installation of python and R you can easily deactivate this virtual environment using the command below:
+```
+   conda deactivate
+``` 
+
+### Adding Packages
+#### Python
+Add any required packages to `setup/conda_env.yaml`. If possible add the package version number. If there is a package that is not available from `conda` add this to the `pip` section of the `yaml` file. In order to not re-run the entire environment setup you can download these individual files from `conda` with the command
+
+```
+   conda install -c conda-forge <PACKAGE>
+```
+
+#### R
+Add any required packages that are available via CRAN to `setup/conda_env.yaml`. These must be prepended with `r-`. If there is a package that is only available from GitHub and not from CRAN, add this package to `setup/setup_r.r`. These individual packages can be added in the same way as Python packages above (with the `r-` prepend).
+
+#### Stata
+
+Install Stata dependencies using `setup/download_stata_ado.do`. We keep all non-base Stata ado files in the `lib` subdirectory, so most non-base Stata ado files will be versioned. To add additional stata dependencies, use the following bash command from the `setup` subdirectory:
+```
+stata-mp -e download_stata_ado.do
+```
+
+### Build
+
+1. Follow the *Setup* instructions above.
+
+2. From the root of repository, run the following bash command:
+   ```
+   python run_all.py
+   ```
+
+### Command Line Usage
+
+For specific instructions on how to set up command line usage for an application, refer to the [RA manual](https://github.com/gentzkow/template/wiki/Command-Line-Usage).
+
+By default, the repository assumes the following executable names for the following applications:
+
+```
+application : executable
+python      : python
+lyx         : lyx
+r           : Rscript
+stata       : statamp (will need to be updated if using a version of Stata that is not Stata-MP)
+```
+
+Default executable names can be updated in `config_user.yaml`. For further detail, see the **User Configuration** section below.
+
+## User Configuration
+`config_user.yaml` contains settings and metadata such as local paths that are specific to an individual user and thus should not be committed to Git. For this repository, this includes local paths to [external dependencies](https://github.com/gentzkow/template/wiki/External-Dependencies) as well as executable names for locally installed software.
+
+Required applications may be set up for command line usage on your computer with a different executable name from the default. If so, specify the correct executable name in `config_user.yaml`. This configuration step is explained further in the [RA manual](https://github.com/gentzkow/template/wiki/Repository-Structure#Configuration-Files).
+
+## Windows Differences
+The instructions above are for Linux and Mac users. However, with just a handful of small tweaks, this repo can also work on Windows. 
+
+If you are using Windows, you may need to run certain bash commands in administrator mode due to permission errors. To do so, open your terminal by right clicking and selecting `Run as administrator`. To set administrator mode on permanently, refer to the [RA manual](https://github.com/gentzkow/template/wiki/Repository-Usage#Administrator-Mode).
+
+The executable names are likely to differ on your computer if you are using Windows. Executable names for Windows will typically look like the following:
+
+```
+application : executable
+python      : python
+lyx         : LyX#.# (where #.# refers to the version number)
+r           : Rscript
+stata       : StataMP-64 (will need to be updated if using a version of Stata that is not Stata-MP or 64-bit)
+```
+
+To download additional `ado` files on Windows, you will likely have to adjust this bash command:
+```
+stata_executable -e download_stata_ado.do
+```
+
+`stata_executable` refers to the name of your Stata executable. For example, if your Stata executable was located in `C:\Program Files\Stata15\StataMP-64.exe`, you would want to use the following bash command:
+
+```
+StataMP-64 -e download_stata_ado.do
+```
+
+
+## List of tables and programs
+The file  `docs/MappingsTablesAndFigures.pdf` provides a mapping of all the tables and figures to their corresponding program. 
+
+## References
+Allcott, Hunt, Matthew Gentzkow, and Lena Song. “Data for: Digital Addiction.” Harvard Dataverse, 2023. https://doi.org/10.7910/DVN/GN636M.
+Allcott, Hunt, Matthew Gentzkow, and Lena Song. “Digital Addiction.” American Economic Review 112, no. 7 (July 2022): 2424–63. https://doi.org/10.1257/aer.20210867.

17/replication_package/code/README.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0accfd826dda5929fe257e41b3c25bde9551fa0ef29d46c999fa967245c811a0
+size 92738

17/replication_package/code/analysis/descriptive/README.md

@@ -0,0 +1,21 @@
+# README
+
+This module produces tables and charts of descriptives statistics. 
+
+`/code/` contains the below files : 
+
+* CommitmentDemand.do (willingness-to-pay and limit tightness plots)
+
+* COVIDResponse.do (survey stats on response to COVID)
+
+* DataDescriptive.do (sample demographics and attrition tables)
+
+* HeatmapPlots.R (predicted vs. actual FITSBY usage)
+
+* QualitativeEvidence.do (descriptive plots for addiction scale, interest in bonus/limit)
+
+* SampleStatistics.do (statistics about completion rates for study)
+
+* Scalars.do (statistics about MPL and ideal usage reduction)
+
+* Temptation.do (plots desired usage change for various tempting activities)

17/replication_package/code/analysis/descriptive/code/COVIDResponse.do

@@ -0,0 +1,168 @@
+// Baseline qualitative evidence
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program define_plot_settings
+    global HIST_SETTINGS ///
+        xtitle(" " "Fraction of sample") ///
+        bcolor(maroon) graphregion(color(white)) ///
+        xsize(6.5) ysize(4.5)   
+    
+    global HIST_DISCRETE_SETTINGS ///
+        gap(50) ylabel(, valuelabel noticks angle(horizontal)) ///
+        $HIST_SETTINGS
+        
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon gray) ///
+        cicolor(maroon gray) ///
+
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5)
+        
+    global CISPIKE_STACKED_GRAPHOPTS ///
+        row(2) ///
+        graphregion(color(white)) ///
+        xsize(5.5) ysize(8)
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+    
+    plot_hist_covid
+    plot_cispike_covid
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program plot_hist_covid
+    twoway hist S1_CovidChangesFreeTime, frac discrete horizontal ///
+        $HIST_DISCRETE_SETTINGS ///
+        ytitle("Change in free time" " ") ///
+        ylabel(1(1)7) 
+ 
+    graph export "output/hist_covid.pdf", replace 
+    
+    recode S1_CovidChangeReason ///
+        (1 = 4 "Increased phone usage") ///
+        (2 = 4 "Increased phone usage") ///
+        (3 = 3 "No change") ///
+        (4 = 4 "Increased phone usage") ///
+        (5 = 2 "Decreased phone usage") ///
+        (6 = 1 "Other"), ///
+        gen(S1_CovidChangeReason_recode)
+        
+    twoway hist S1_CovidChangeReason_recode, ///
+        frac discrete horizontal ///
+        $HIST_DISCRETE_SETTINGS ///
+        ytitle("Effect of COVID-19 on phone use" " ") 
+ 
+    graph export "output/hist_covid_reason.pdf", replace 
+end
+
+program plot_cispike_covid
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S1_PhoneUseChange* S1_LifeBetter*
+    rename S1_PhoneUseChange* S1_PhoneUseChange_*
+    rename S1_LifeBetter* S1_LifeBetter_*
+    rename_but, varlist(UserID) prefix(outcome)
+    reshape long outcome, i(UserID) j(measure) string
+
+    split measure, p("_")
+    drop measure measure1
+    rename (measure2 measure3) (measure time)
+    replace time = "2020" if time == ""
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+    encode time, generate(time_encode) 
+    
+    recode measure_encode ///
+        (1 = 1 "Phone use makes life better") ///
+        (2 = 2 "Ideal use change"), ///
+        gen(measure_recode)
+    
+    recode time_encode ///
+        (1 = 1 "2019") ///
+        (2 = 2 "Now"), ///
+        gen(time_recode)
+        
+    * Plot data
+    gen dummy = 1
+
+    
+    ttest outcome if measure_encode == 1, by(time_recode)
+    local diff : display %9.3fc `r(mu_2)' - `r(mu_1)'
+    local diff = subinstr("`diff'", " ", "", .)
+    local se : display %9.3fc `r(se)' 
+    local se = subinstr("`se'", " ", "", .) 
+    
+    ciquartile outcome if measure_encode == 1, ///
+        over1(dummy) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Phone use makes life better" " ") ///
+                  ysc(r(-1)) ///
+                  legend(off) ///
+                  text(-0.75 0 "Difference in means = `diff' (`se')", place(e)))
+    
+    graph save "output/cispike_covid_life.gph", replace
+    
+    ttest outcome if measure_encode == 2, by(time_recode)
+    local diff : display %9.3fc `r(mu_2)' - `r(mu_1)'
+    local diff = subinstr("`diff'", " ", "", .)
+    local se : display %9.3fc `r(se)' 
+    local se = subinstr("`se'", " ", "", .) 
+    
+    ciquartile outcome if measure_encode == 2, ///
+        over1(dummy) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Ideal use change" " ") ///
+                  ysc(r(-40)) ///
+                  legend(off) ///
+                  text(-37.5 0 "Difference in means = `diff' (`se')", place(e)))
+                  
+    graph save "output/cispike_covid_ideal.gph", replace    
+    
+    graph combine ///
+        "output/cispike_covid_ideal.gph" ///
+        "output/cispike_covid_life.gph", ///
+        $CISPIKE_STACKED_GRAPHOPTS
+        
+    graph export "output/cispike_covid.pdf", replace
+    
+    * Restore data
+    restore
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/CommitmentDemand.do

@@ -0,0 +1,457 @@
+// Demand for commitment, moderated by demand for flexibility
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program define_plot_settings
+    global HIST_SETTINGS ///
+        bcolor(maroon) graphregion(color(white)) ///
+        xsize(6.5) ysize(4.5)
+		
+	global HIST_DISCRETE_SETTINGS ///
+        gap(50) xlabel(, valuelabel noticks) ///
+        $HIST_SETTINGS
+
+    global HIST_SNOOZE_SETTINGS ///
+        gap(50) ylabel(1(1)10, valuelabel noticks angle(horizontal) labsize(small)) ///
+        xtitle(" " "Fraction of sample") ///
+        $HIST_SETTINGS 
+
+    global HIST_CONTINUOUS_SETTINGS ///
+        $HIST_SETTINGS
+            
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon black gray) ///
+        cicolor(maroon black gray) 
+
+    global CISPIKE_SETTINGS4 ///
+        spikecolor(maroon black gray navy) ///
+        cicolor(maroon black gray navy) 
+		
+    global CISPIKE_VERTICAL_LARGE_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(8) ysize(4.5) ///
+		legend(cols(4))
+        
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5) ///
+		legend(cols(4))
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+    
+    plot_midline_demand    
+    plot_wtp_for_rsi
+    plot_wtp_for_limit
+    plot_wtp_for_limit_by_limit
+    plot_wtp_for_limit_by_bonus
+	plot_limit_tight
+    plot_limit_tight, fitsby
+	plot_limit_tight_by_limit
+    plot_limit_tight_by_limit, fitsby
+    plot_limit_tight_dist
+    plot_preferred_snooze
+    plot_motivation_by_reason
+    plot_motivation_bar
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program plot_midline_demand        
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_PredictUseInitialEarn S2_PredictUseBonusEarn S2_MPL
+    rename_but, varlist(UserID) prefix(dollar)
+    reshape long dollar, i(UserID) j(measure) string
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+	
+    * Plot data
+    gen dummy = 1
+    
+    cispike dollar, ///
+        over1(dummy) over2(measure_encode) ///
+        $CISPIKE_SETTINGS gap2(100) ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Dollars" " ") ///
+                  /// Labels too long for encode
+                  xlabel(0.5 " " ///
+                         1 `" "Valuation of" "Bonus" "' ///
+                         3 `" "Expected earnings" "at predicted usage" "with Bonus" "' ///
+                         5 `" "Expected earnings" "at predicted usage" "without Bonus" "' ///
+                         5.5 " ") ///
+                  legend(off)) 
+        
+    graph export "output/cispike_midline_demand.pdf", replace
+        
+    * Restore data
+    restore
+end
+
+program plot_wtp_for_rsi
+    hist S2_MPL, frac discrete ///
+        xtitle(" " "Valuation of bonus ($)") ///
+        ytitle("Fraction of sample" " ") ///
+        $HIST_DISCRETE_SETTINGS
+        
+    graph export "output/hist_rsi_wtp.pdf", replace    
+end
+
+program plot_wtp_for_limit
+    hist S3_MPLLimit, frac ///
+        xtitle(" " "Valuation of limit functionality ($)") ///
+        ytitle("Fraction of sample" " ") ///
+        $HIST_DISCRETE_SETTINGS
+        
+    graph export "output/hist_limit_wtp.pdf", replace    
+end
+
+program plot_wtp_for_limit_by_limit
+    * Preserve data
+    preserve
+    
+    * Add average
+    tempfile temp
+    save `temp', replace
+    keep if inlist(S2_LimitType, 1, 2, 3, 4, 5)
+    replace S2_LimitType = 6
+    append using `temp'
+    
+    * Recode data
+    recode S2_LimitType ///
+        (0 = .) ///
+        (1 = 2 "Snooze 0") ///
+        (2 = 3 "Snooze 2") ///
+        (3 = 4 "Snooze 5") ///
+        (4 = 5 "Snooze 20") ///
+        (5 = 6 "No snooze") ///
+        (6 = 1 "All limits"), ///
+        gen(S2_LimitType_recode)
+        
+    * Plot data
+    gen dummy = 1
+    
+    cispike S3_MPLLimit, ///
+        over1(dummy) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_LARGE_GRAPHOPTS ///
+                  ytitle("Willingness-to-pay for limit (dollars)" " ") ///
+                  legend(off))
+                  
+    graph export "output/cispike_limit_wtp.pdf", replace
+
+    * Restore data
+    restore
+end
+
+program plot_wtp_for_limit_by_bonus
+    * Preserve data
+    preserve
+    
+    * Recode data
+	recode S3_Bonus ///
+        (0 = 0 "Control") ///
+        (1 = 1 "Bonus"), ///
+        gen(S3_Bonus_recode)
+        
+    * Plot data
+    gen dummy = 1
+
+	cispike S3_MPLLimit, ///
+        over1(dummy) over2(S3_Bonus_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_LARGE_GRAPHOPTS ///
+                  ytitle("Willingness-to-pay for Limit (dollars)" " ") ///
+				  legend(off))
+			
+	graph export "output/cispike_limit_wtp_by_bonus.pdf", replace
+	
+    * Restore data
+    restore
+end
+
+program plot_limit_tight
+    syntax, [fitsby]
+    
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_LimitType *LimitTight`fitsby'
+    rename_but, varlist(UserID S2_LimitType) prefix(tight)
+    reshape long tight, i(UserID S2_LimitType) j(measure) string
+     
+    * Recode data
+    sort measure
+	encode measure, generate(measure_encode) 
+        
+    recode measure_encode ///
+        (1 = 1 "Period 2") ///
+		(2 = 2 "Period 3") ///
+		(5 = 3 "Period 4") ///
+        (7 = 4 "Period 5") ///
+		(4 = 5 "Periods 3 & 4") ///
+		(3 = 6 "Periods 2 to 4") ///
+        (6 = 7 "Periods 2 to 5"), ///
+        gen(measure_recode)
+        
+    * Plot data
+	gen dummy = 1
+    
+	cispike tight if measure_recode <= 4, ///
+        over1(dummy) over2(measure_recode) ///
+        $CISPIKE_SETTINGS4 ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Limit tightness (minutes/day)" " ") ///
+                  legend(off))  
+                  
+    graph export "output/cispike_limit_tight`suffix'.pdf", replace
+    
+    * Restore data
+    restore
+end
+
+program plot_limit_tight_by_limit
+    syntax, [fitsby]
+    
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_LimitType *LimitTight`fitsby'
+    rename_but, varlist(UserID S2_LimitType) prefix(tight)
+    reshape long tight, i(UserID S2_LimitType) j(measure) string
+     
+    * Recode data
+    sort measure
+    encode measure, generate(measure_encode) 
+
+    recode measure_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (5 = 3 "Period 4") ///
+        (7 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (3 = 6 "Periods 2 to 4") ///
+        (6 = 7 "Periods 2 to 5"), ///
+        gen(measure_recode)
+    
+    recode S2_LimitType ///
+        (0 = .) ///
+        (1 = 1 "Snooze 0") ///
+        (2 = 2 "Snooze 2") ///
+        (3 = 3 "Snooze 5") ///
+        (4 = 4 "Snooze 20") ///
+        (5 = 5 "No snooze"), ///
+        gen(S2_LimitType_recode)
+        
+	* Plot data (all periods together) 2 - 5
+    gen dummy = 1
+	
+    cispike tight if measure_recode == 7, ///
+        over1(dummy) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_LARGE_GRAPHOPTS ///
+                  ytitle("Limit tightness (minutes/day)" " ") ///
+                  xlabel(, labsize(medlarge)) xtitle(, size(medlarge)) ///
+                  ylabel(, labsize(medlarge)) ytitle(, size(medlarge)) ///
+                  legend(off))  
+                  
+    graph export "output/cispike_limit_tight_combined_by_limit`suffix'.pdf", replace
+
+    * Plot data (by period)
+    cispike tight if measure_recode <= 4, ///
+        over1(measure_recode) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS4 ///
+        graphopts($CISPIKE_VERTICAL_LARGE_GRAPHOPTS ///
+                  ytitle("Limit tightness (minutes/day)" " ") ///
+                  xlabel(, labsize(medlarge)) xtitle(, size(medlarge)) ///
+                  ylabel(, labsize(medlarge)) ytitle(, size(medlarge)) ///
+                  legend(size(medlarge)))  
+                  
+    graph export "output/cispike_limit_tight_by_limit`suffix'.pdf", replace
+	
+    * Restore data
+    restore
+end
+
+program plot_limit_tight_dist
+    * Preserve data
+    preserve
+    
+    * Plot data (by period)
+    hist PD_P2_LimitTight, frac ///
+        xtitle(" " "Period 2 limit tightness (minutes/day)") ///
+        ytitle("Fraction of sample" " ") ///
+        $HIST_CONTINUOUS_SETTINGS
+    
+    graph export "output/hist_limit_tight_p2.pdf", replace    
+
+	* Plot data (all periods together)
+    hist PD_P5432_LimitTight, frac ///
+        xtitle(" " "Periods 2 to 5 limit tightness (minutes/day)") ///
+        ytitle("Fraction of sample" " ") ///
+        $HIST_CONTINUOUS_SETTINGS
+    
+    graph export "output/hist_limit_tight.pdf", replace    
+	
+    * Reshape data
+    keep UserID PD_P2_LimitTight_*
+	drop *Other
+    reshape long PD_P2, i(UserID) j(measure) string
+    
+    * Recode data
+    sort measure
+    encode measure, generate(measure_encode) 
+    
+    recode measure_encode ///
+		(2 = 1 "Facebook") ///
+        (3 = 2 "Instagram") ///
+        (5 = 3 "Twitter") ///
+        (4 = 4 "Snapchat") ///
+        (1 = 5 "Browser") ///
+        (6 = 6 "YouTube"), ///
+        gen(measure_recode)
+    
+    * Plot data (by app)
+    local app_1 "Facebook"
+    local app_2 "Instagram"
+    local app_3 "Twitter"
+    local app_4 "Snapchat"
+    local app_5 "Browser"
+    local app_6 "YouTube"
+    
+    foreach num of numlist 1/6 {
+        hist PD_P2 if measure_encode == `num', frac ///
+            xtitle(" " "Period 2 limit tightness for `app_`num'' (minutes/day)") ///
+            ytitle("Fraction of sample" " ") ///
+            $HIST_CONTINUOUS_SETTINGS ///
+            xlabel(, labsize(large)) xtitle(, size(large)) ///
+            ylabel(, labsize(large)) ytitle(, size(large)) ///
+            legend(size(large))
+            
+        graph export "output/hist_limit_tight_`num'.pdf", replace    
+    }
+
+    * Restore data
+    restore
+end
+
+program plot_preferred_snooze
+    recode S4_PreferredSnooze ///
+        (1  = 1  "No delay") ///
+        (2  = 2  "1 minute") ///
+        (3  = 3  "2 minutes") ///
+        (4  = 4  "3-4 minutes") ///
+        (5  = 5  "5 minutes") ///
+        (6  = 6  "10 minutes") ///
+        (7  = 7  "20 minutes") ///
+        (8  = 8  "30 minutes+") ///
+        (9  = 9  "Prefer no snooze") ///
+        (10 = 10 "Does not matter"), ///
+        gen(S4_PreferredSnooze_short_names)
+
+    twoway hist S4_PreferredSnooze_short_names, ///
+        frac discrete horizontal ///
+        $HIST_SNOOZE_SETTINGS ///
+        ytitle("Preferred Snooze Length (minutes)" " ") 
+
+        
+    graph export "output/hist_preferred_snooze.pdf", replace    
+end
+
+program plot_motivation_by_reason
+    preserve 
+    * Plot data
+    gen dummy = 1
+    
+    cispike S2_Motivation, ///
+        over1(dummy) over2(S2_MPLReasoning) ///
+        $CISPIKE_SETTINGS4 gap2(100) ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Behavior change premium ($)" " ") ///
+                  /// Labels too long for encode
+                  xlabel(0.5 " " ///
+                         1 `" "Only wanted" "to maximize" "earnings" "' ///
+                         3 `" "Wanted incentive" "to use phone" "less" "' ///
+                         5 `" "Don't want pressure" "to use phone" "less" "' ///
+                         7 `" "Other" "' ///
+                         7.5 " ") ///
+                  legend(off)) 
+        
+    graph export "output/cispike_motivation_reason.pdf", replace
+        
+    * Restore data
+    restore 
+end
+
+program plot_motivation_bar
+    preserve 
+    * Plot data    
+    twoway hist S2_MPLReasoning, frac discrete ///
+                $HIST_DISCRETE_SETTINGS ///
+                xlabel(1 `" "Only wanted" "to maximize" "earnings" "' ///
+                       2 `" "Wanted incentive" "to use phone" "less" "' ///
+                       3 `" "Don't want pressure" "to use phone" "less" "' ///
+                       4 `" "Other" "') ///
+                ytitle("Fraction of sample" " ") ///
+                xtitle("")
+
+        
+    graph export "output/hist_motivation_mpl.pdf", replace
+        
+    * Restore data
+    restore 
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/DataDescriptives.do

@@ -0,0 +1,668 @@
+// Description of data
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program define_settings
+    global DESCRIPTIVE_TAB ///
+        collabels(none) nodepvars noobs replace 
+    
+	global DESCRIPTIVE_TAB_DETAILED ///
+		nomtitle nonumbers noobs compress label replace ///
+		cells((mean(fmt(%8.1fc)) ///
+		       sd(fmt(%8.1fc)) ///
+			   min(fmt(%8.0fc)) ///
+			   max(fmt(%8.0fc)))) ///
+		collabels("\shortstack{Mean}" ///
+		          "\shortstack{Standard\\deviation}" ///
+				  "\shortstack{Minimum\\value}" ///
+				  "\shortstack{Maximum\\value}") ///		 
+
+    global BALANCE_TAB ///
+        order(1 0) grplabels(1 Treatment @ 0 Control) ///
+        pftest pttest ftest fmissok vce(robust) stdev ///
+        rowvarlabel onenrow tblnonote format(%8.2fc) replace
+        
+    global HIST_CONTINUOUS_SETTINGS ///
+        bcolor(maroon) graphregion(color(white)) ///
+        xsize(6.5) ysize(4)
+        
+    global BAR_SETTINGS ///
+        region(lcolor(white))) graphregion(color(white)) ///
+        xsize(6.5) ysize(4)
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_settings
+    import_data
+    clean_data
+    
+  sample_demographics_balance_all
+	sample_demographics
+    sample_demographics_balance
+    * limit_attrition
+    * bonus_attrition
+	balance
+	historical_use
+    historical_use, fitsby
+	summary_welfare
+	share_use_by_app  
+    addiction_plot
+end
+
+program import_data
+	use "input/final_data_sample.dta", clear
+    
+	foreach time in S3 S4 {
+		replace `time'_Finished = 0 if `time'_Finished == .
+	}
+end
+
+program clean_data
+    * Demographics
+    recode S1_Income ///
+        (1 = 5) ///
+        (2 = 15) ///
+		(3 = 25) ///
+		(4 = 35) ///
+		(5 = 45) ///
+		(6 = 55) ///
+		(7 = 67) ///
+		(8 = 87.5) ///
+		(9 = 112.5) ///
+		(10 = 137.5) ///
+		(11 = 150) ///
+		(12 = .), ///
+		gen(income)
+		
+    gen college = (S1_Education >= 5)
+    gen male = (S0_Gender == 1)
+    gen white = (S1_Race == 5)
+    
+    * Limit treatment
+    gen limit_T = 1 if S2_LimitType > 0 & S2_LimitType != .
+    replace limit_T = 0 if S2_LimitType == 0
+    
+    * Labels
+    label var college "College"
+    label var male "Male"
+    label var white "White"
+
+    label var income "Income (\\$000s)"
+    label var S0_Age "Age"
+    label var PD_P1_UsageFITSBY "Period 1 FITSBY use (minutes/day)"
+end
+
+program sample_demographics
+    local varset income college male white S0_Age PD_P1_Usage PD_P1_UsageFITSBY
+    
+    * Sample demographics
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store sample_col
+
+    * Preserve data
+    preserve
+    
+    * US demographics
+    replace income = 43.01
+    replace college = 0.3009
+    replace male = 0.4867
+    replace white = 0.73581
+    replace S0_Age = 47.6
+    replace PD_P1_Usage = . 
+	replace PD_P1_UsageFITSBY = .
+	
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store us_col
+    
+    * Restore data
+    restore
+    
+    * Export table
+    esttab sample_col us_col using "output/sample_demographics.tex", ///
+		mtitle("\shortstack{Analysis\\sample}" ///
+			   "\shortstack{U.S.\\adults}")  ///
+		coeflabels(income "Income (\\$000s)" ///
+				   college "College" ///
+				   male "Male" ///
+				   white "White" ///
+				   S0_Age "Age" ///
+				   PD_P1_Usage "Period 1 phone use (minutes/day)" ///
+				   PD_P1_UsageFITSBY "Period 1 FITSBY use (minutes/day)") ///
+		$DESCRIPTIVE_TAB ///
+		cells(mean(fmt(%9.1fc %9.2fc %9.2fc %9.2fc %9.1fc %9.1fc %9.1fc)))
+
+    est clear
+end
+
+program sample_demographics_balance
+    local varset balance_income balance_college balance_male balance_white balance_age ///
+                 PD_P1_Usage PD_P1_UsageFITSBY
+    
+    * Sample demographics
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store sample_col
+
+    * Preserve data
+    preserve
+
+    local income  43.01
+    local college 0.3009
+    local male    0.4867
+    local white   0.73581
+    local age     47.6
+
+    ebalance balance_income balance_college balance_male balance_white balance_age, ///
+             manualtargets(`income' `college' `male' `white' `age') generate(weight)
+
+    * Winsorize weights
+    gen weight2 = weight 
+    replace weight2 = 2 if weight2 > 2
+    replace weight2 = 1/2 if weight2 < 1/2
+
+    estpost tabstat `varset' [weight=weight2], statistics(mean) columns(statistics)
+    est store sample_col_w2
+
+    * US demographics
+    replace balance_income = `income'
+    replace balance_college = `college'
+    replace balance_male = `male'
+    replace balance_white = `white'
+    replace balance_age = `age'
+    replace PD_P1_Usage = . 
+    replace PD_P1_UsageFITSBY = .
+    
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store us_col
+    
+    * Restore data
+    restore
+    
+    * Export table
+    esttab sample_col sample_col_w2 us_col ///
+        using "output/sample_demographics_balance.tex", ///
+        mtitle("\shortstack{Analysis\\sample}" ///
+               "\shortstack{Balanced\\sample}" ///
+               "\shortstack{U.S.\\adults}" ///
+               )  ///
+        coeflabels(balance_income "Income (\\$000s)" ///
+                   balance_college "College" ///
+                   balance_male "Male" ///
+                   balance_white "White" ///
+                   balance_age "Age" ///
+                   PD_P1_Usage "Period 1 phone use (minutes/day)" ///
+                   PD_P1_UsageFITSBY "Period 1 FITSBY use (minutes/day)") ///
+        $DESCRIPTIVE_TAB ///
+        cells(mean(fmt(%9.1fc %9.2fc %9.2fc %9.2fc %9.1fc %9.1fc %9.1fc)))
+
+    est clear
+end
+
+program sample_demographics_balance_all
+    local varset balance_income balance_college balance_male balance_white balance_age ///
+                 PD_P1_Usage PD_P1_UsageFITSBY
+    
+    * Sample demographics
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store sample_col
+
+    * Preserve data
+    preserve
+
+    local income  43.01
+    local college 0.3009
+    local male    0.4867
+    local white   0.73581
+    local age     47.6
+
+    ebalance balance_income balance_college balance_male balance_white balance_age, ///
+             manualtargets(`income' `college' `male' `white' `age') generate(weight)
+
+    * Winsorize weights
+    gen weight2 = weight 
+    replace weight2 = 2 if weight2 > 2
+    replace weight2 = 1/2 if weight2 < 1/2
+
+    gen weight3 = weight 
+    replace weight3 = 3 if weight3 > 3
+    replace weight3 = 1/3 if weight3 < 1/3
+
+    gen weight4 = weight 
+    replace weight4 = 4 if weight4 > 4
+    replace weight4 = 1/4 if weight4 < 1/4
+
+    gen weight5 = weight 
+    replace weight5 = 5 if weight5 > 5
+    replace weight5 = 1/5 if weight5 < 1/5
+
+    estpost tabstat `varset' [weight=weight2], statistics(mean) columns(statistics)
+    est store sample_col_w2
+
+    estpost tabstat `varset' [weight=weight3], statistics(mean) columns(statistics)
+    est store sample_col_w3
+
+    estpost tabstat `varset' [weight=weight4], statistics(mean) columns(statistics)
+    est store sample_col_w4
+
+    estpost tabstat `varset' [weight=weight5], statistics(mean) columns(statistics)
+    est store sample_col_w5
+
+    * US demographics
+    replace balance_income = `income'
+    replace balance_college = `college'
+    replace balance_male = `male'
+    replace balance_white = `white'
+    replace balance_age = `age'
+    replace PD_P1_Usage = . 
+    replace PD_P1_UsageFITSBY = .
+    
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store us_col
+    
+    * Restore data
+    restore
+    
+    * Export table
+    esttab us_col sample_col  sample_col_w2 sample_col_w3 sample_col_w4 sample_col_w5  ///
+        using "output/sample_demographics_balance_all.tex", ///
+        mtitle("\shortstack{U.S.\\adults}" ///
+               "\shortstack{Analysis\\sample}" ///
+               "\shortstack{(w=2)}" ///
+               "\shortstack{(w=3)}" ///
+               "\shortstack{(w=4)}" ///
+               "\shortstack{(w=5)}" ///
+               )  ///
+        coeflabels(balance_income "Income (\\$000s)" ///
+                   balance_college "College" ///
+                   balance_male "Male" ///
+                   balance_white "White" ///
+                   balance_age "Age" ///
+                   PD_P1_Usage "Period 1 use (min/day)" ///
+                   PD_P1_UsageFITSBY "Period 1 FITSBY use (min/day)") ///
+        $DESCRIPTIVE_TAB ///
+        cells(mean(fmt(%9.1fc %9.2fc %9.2fc %9.2fc %9.1fc %9.1fc %9.1fc)))
+
+    est clear
+end
+
+
+program limit_attrition    
+    local varset ///
+		S3_Finished ///
+		S4_Finished ///
+		I_P2_Usage ///
+		I_P3_Usage ///
+		I_P4_Usage ///
+		I_P5_Usage
+    
+    * Preserve data
+    preserve
+
+    * Use old sample definition
+    use "input/final_data.dta", clear
+    keep if S2_RevealConfirm == 1 & S3_Bonus <= 1
+    foreach time in S3 S4 {
+        replace `time'_Finished = 0 if `time'_Finished == .
+    }
+    
+	* Create usage indicators
+	foreach time in P2 P3 P4 P5 {
+		gen I_`time'_Usage = 0
+		replace I_`time'_Usage = 1 if PD_`time'_Usage != .
+	}
+	
+    * Attrition by limit group
+    forvalues i = 0/5 {
+        local if if S2_LimitType == `i'
+        estpost tabstat `varset' `if', statistics(mean) columns(statistics)
+        est store attrition_b`i'
+    }
+    
+	* Attrition for limit groups
+	local if if S2_LimitType != 0
+    estpost tabstat `varset' `if', statistics(mean) columns(statistics)
+    est store attrition_b
+	
+    * F-test for limit groups
+    foreach var of varlist `varset' {
+        reg `var' i.S2_LimitType
+        local fvalue = Ftail(e(df_m), e(df_r), e(F))
+        replace `var' = `fvalue'
+    }
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store fval_b
+
+    * Export limit attrition table
+    esttab attrition_b0 attrition_b ///
+	       attrition_b1 attrition_b2 ///
+		   attrition_b3 attrition_b4 ///
+		   attrition_b5 fval_b ///
+        using "output/attrition_limit.tex", ///
+        mtitle("\shortstack{Control}" ///
+			   "\shortstack{All\\limits}" ///
+               "\shortstack{Snooze\\0}" ///
+               "\shortstack{Snooze\\2}" ///
+               "\shortstack{Snooze\\5}" ///
+               "\shortstack{Snooze\\20}" ///
+               "\shortstack{No\\snooze}" ///
+               "\shortstack{F-test\\p-value}") ///
+        coeflabels(S3_Finished "Completed survey 3" ///
+				   S4_Finished "Completed survey 4" ///
+				   I_P2_Usage "Have period 2 usage" ///
+				   I_P3_Usage "Have period 3 usage" ///
+				   I_P4_Usage "Have period 4 usage" ///
+				   I_P5_Usage "Have period 5 usage") ///
+        $DESCRIPTIVE_TAB ///
+		cells(mean(fmt(%9.2fc)))
+		
+    est clear
+    
+    * Restore data
+    restore
+end
+        
+program bonus_attrition
+    local varset ///
+		S3_Finished ///
+		S4_Finished ///
+		I_P2_Usage ///
+		I_P3_Usage ///
+		I_P4_Usage ///
+		I_P5_Usage
+		
+    * Preserve data
+    preserve
+    
+    * Use old sample definition
+    use "input/final_data.dta", clear
+    keep if S2_RevealConfirm == 1 & S3_Bonus <= 1
+    foreach time in S3 S4 {
+        replace `time'_Finished = 0 if `time'_Finished == .
+    }
+
+    keep if S3_Bonus != 2
+    
+	* Create usage indicators
+	foreach time in P2 P3 P4 P5 {
+		gen I_`time'_Usage = 0
+		replace I_`time'_Usage = 1 if PD_`time'_Usage != .
+	}
+	
+    * Attrition by bonus group
+    forvalues i = 0 / 1 {
+        local if if S3_Bonus == `i'
+        estpost tabstat `varset' `if', statistics(mean) columns(statistics)
+        est store attrition_bonus`i'
+    }
+
+    * T-test for bonus groups
+	foreach var of varlist `varset' {
+        capture prtest `var', by(S3_Bonus)
+		
+		if _rc == 0 { 
+			local diff = -1 * r(P_diff)
+			local pval = r(p)
+			gen `var'_d = `diff'
+			gen `var'_p = `pval'
+        } 
+		else { 
+			gen `var'_d = 0
+			gen `var'_p = .
+        } 
+    }
+    
+    * Append bonus differences
+    foreach var of varlist `varset' {
+        replace `var' = `var'_d
+    }
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store diff_bonus
+
+    * Append bonus p-values
+    foreach var of varlist `varset' {
+        replace `var' = `var'_p
+    }
+    estpost tabstat `varset', statistics(mean) columns(statistics)
+    est store pval_bonus
+
+    display("here")
+        
+    * Export Bonus attrition table
+    esttab attrition_bonus0 attrition_bonus1 pval_bonus using "output/attrition_bonus.tex", ///
+        mtitle("\shortstack{Control}" ///
+               "\shortstack{Treatment}" ///
+               "\shortstack{t-test\\p-value}") ///     
+        coeflabels(S3_Finished "Completed survey 3" ///
+				   S4_Finished "Completed survey 4" ///
+				   I_P2_Usage "Have period 2 usage" ///
+				   I_P3_Usage "Have period 3 usage" ///
+				   I_P4_Usage "Have period 4 usage" ///
+				   I_P5_Usage "Have period 5 usage") ///
+        $DESCRIPTIVE_TAB ///
+		cells(mean(fmt(%9.2fc)))
+        
+    est clear
+    
+    * Restore data
+    restore
+end
+
+program balance
+    local varset income college male white S0_Age PD_P1_UsageFITSBY
+
+    iebaltab_edit `varset', /// 
+        grpvar(limit_T) ///
+        savetex("output/balance_limit.tex") ///
+        $BALANCE_TAB
+
+    iebaltab_edit `varset', /// 
+        grpvar(S3_Bonus) ///
+        savetex("output/balance_bonus.tex") ///
+        $BALANCE_TAB
+
+  *   panelcombine, ///
+  *       use(output/balance_limit.tex ///
+  *           output/balance_bonus.tex) ///
+		* paneltitles("Limit Treatment" ///
+  *                   "Bonus Treatment") ///
+  *       columncount(4) ///
+  *       save("output/balance.tex") cleanup
+end
+
+program historical_use
+    syntax, [fitsby]
+    
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+		local word "FITSBY"
+    } 
+    
+    else {
+        local fitsby ""
+        local suffix ""
+		local word "phone"
+    }
+    
+	local var PD_P1_Usage`fitsby'
+    label var PD_P1_Usage`fitsby' "Period 1 `word' use (minutes/day)"
+    
+    local label : var label `var'
+    sum `var', d
+    
+    twoway histogram `var', frac ///
+		   ytitle("Fraction of sample" " ") ///
+		   xtitle(" " "`label'") ///
+           $HIST_CONTINUOUS_SETTINGS
+           
+    graph export "output/hist_baseline_usage`suffix'.pdf", replace
+end
+
+program summary_welfare
+    local varset ///
+        S1_PhoneUseChange ///
+        S1_AddictionIndex ///
+        S1_SMSIndex ///
+		S1_LifeBetter ///
+		S1_SWBIndex
+
+
+	estpost tabstat `varset', ///
+		statistics(mean, sd, max, min) columns(statistics)
+
+	est store baseline
+	
+	esttab baseline using "output/baseline_welfare.tex", ///
+		$DESCRIPTIVE_TAB_DETAILED ///     
+        coeflabels(S1_PhoneUseChange "Ideal use change" ///
+				   S1_AddictionIndex "Addiction scale x (-1)" ///
+                   S1_SMSIndex "SMS addiction scale x (-1)" ///
+				   S1_LifeBetter "Phone makes life better" ///
+				   S1_SWBIndex  "Subjective well-being")
+end
+
+program share_use_by_app
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID PD_P1_Usage_* PD_P1_Installed_*
+	drop *Other *_H*
+    reshape long PD_P1_Usage_ PD_P1_Installed_ , i(UserID) j(app) s
+    replace PD_P1_Usage_ = 0 if PD_P1_Usage_ == .
+	
+	* Collapse data
+	collapse (mean) PD_P1_Usage_ PD_P1_Installed_, by(app)
+    gsort -PD_P1_Usage_  
+    gen order = _n
+
+    cap drop appname1 appname2
+    gen appname1 = _n - 0.2
+    gen appname2 = _n + 0.2
+
+    local N = _N
+    forvalues i = 1/`N' {
+        local t`i' = app[`i']
+    }
+    
+    * Plot data
+    twoway bar PD_P1_Installed_ appname1, ///
+               fintensity(inten50) barw(0.35) ///
+               yaxis(1) yscale(axis(1) range(0)) ylabel(0(0.2)1, axis(1)) ///
+               xlabel(1  "`t1'"  2  "`t2'"  3  "`t3'"  4  "`t4'"  5  "`t5'" ///
+                      6  "`t6'"  7  "`t7'"  8  "`t8'"  9  "`t9'"  10 "`t10'" ///
+                      11 "`t11'" 12 "`t12'" 13 "`t13'" 14 "`t14'", ///
+                      valuelabel angle(45)) || ///
+           bar PD_P1_Usage_ appname2, ///
+               fintensity(inten100) barw(0.35) ///
+               yaxis(2) yscale(axis(2) range(0)) ylabel(#5, axis(2)) ///
+           xtitle("") ytitle("Share of users", axis(1)) ytitle("Minutes/day", axis(2)) ///
+           legend(label(1 "Users at baseline") ///
+                  label(2 "Period 1 use") ///
+           $BAR_SETTINGS
+           
+    graph export "output/bar_share_use_by_app.pdf", replace 
+
+    * Restore data
+    restore
+end
+
+program addiction_plot
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID *_Addiction_*
+    keep if S3_Addiction_1 != .
+
+    foreach i in 3 {
+        forvalues j = 1/16 {
+            gen S`i'_Addiction_Binary_`j' = S`i'_Addiction_`j' > 0.5
+
+        }
+    }
+
+    keep UserID S3_Addiction_Binary_*
+
+    reshape long  S3_Addiction_Binary_ , i(UserID) j(question)
+
+    rename S3_Addiction_Binary_ S3_Addiction
+    
+    * Collapse data
+    collapse (mean) S3_Addiction , by(question)
+
+    gen order = _n
+
+    cap drop qname
+    gen qname = _N - _n + 1
+
+    gen category = qname < 9 
+
+    
+    * Plot data
+    twoway bar S3_Addiction qname, ///
+               fintensity(inten100) barw(0.6) bcolor(maroon) ///
+               yaxis(1) yscale(axis(1) range(0)) xlabel(0(0.2)1, axis(1)) ///
+               ylabel(1  "Procrastinate by using phone" 2  "Prefer phone to human interaction" ///
+                      3  "Lose sleep from use" 4  "Harms school/work performance" ///
+                      5  "Annoyed at interruption in use" 6  "Difficult to put down phone" ///
+                      7  "Feel anxious without phone" 8  "Others are concerned about use" ///
+                      9  "Try and fail to reduce use" 10 "Use to relax to go to sleep" ///
+                      11 "Use to distract from anxiety/etc."  12 "Use to distract from personal issues" ///
+                      13 "Tell yourself just a few more minutes" 14 "Use longer than intended" ///
+                      15 "Wake up, check phone immediately" 16 "Fear missing out online", ///
+                      valuelabel angle(0)) horizontal ///
+           ytitle("      relapse, withdrawal, conflict           salience, tolerance, mood", size(small)) ///
+           xtitle(`"Share of people who "often" or "always""', axis(1)) ///
+           legend(label(1 "Survey 3") ///
+           $BAR_SETTINGS 
+           
+    graph export "output/addiction.pdf", replace 
+
+  
+    * Plot data
+    twoway bar S3_Addiction qname, ///
+               fintensity(inten100) barw(0.75) bcolor(maroon) ///
+               yaxis(1) yscale(axis(1) range(0)) xlabel(0(0.2)0.8, axis(1)) ///
+               xlabel(, labsize(large)) ///
+               ylabel(1  "Procrastinate by using phone" 2  "Prefer phone to human interaction" ///
+                      3  "Lose sleep from use" 4  "Harms school/work performance" ///
+                      5  "Annoyed at interruption in use" 6  "Difficult to put down phone" ///
+                      7  "Feel anxious without phone" 8  "Others are concerned about use" ///
+                      9  "Try and fail to reduce use" 10 "Use to relax to go to sleep" ///
+                      11 "Use to distract from anxiety/etc."  12 "Use to distract from personal issues" ///
+                      13 "Tell yourself just a few more minutes" 14 "Use longer than intended" ///
+                      15 "Wake up, check phone immediately" 16 "Fear missing out online", ///
+                      valuelabel angle(0) labsize(large)) horizontal ///
+           ytitle(, size(zero)) ///
+           xtitle(`"Share of people who "often" or "always""', axis(1) justification(right) size(large)) ///
+           legend(label(1 "Survey 3") ///
+           region(lcolor(white))) graphregion(color(white)) ///
+           xsize(6.5) ysize(4.5 )
+
+    graph export "output/addiction_large.pdf", replace 
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/HeatmapPlots.R

@@ -0,0 +1,144 @@
+library(ggplot2)
+library(tidyverse)
+library(haven)
+
+maroon <- '#94343c'
+grey   <- '#848484'
+
+low_grey <- "grey90"
+
+plot_wtp_prediction <- function(df){
+
+  # Tally the bins. Create bins centered at 5, 15, 25, etc.
+  counted <- df %>% 
+              mutate(S2_PredictUseBonusEarnBin = S2_PredictUseBonusEarn - (S2_PredictUseBonusEarn %% 10) + 5) %>%
+              select(UserID, S2_PredictUseBonusEarnBin, S2_MPL) %>%
+              group_by(S2_MPL, S2_PredictUseBonusEarnBin) %>%
+              count(name="Count")
+
+  # Create an empty dataframe of all of the index combinations
+  mpls <- unique(counted$S2_MPL)
+  pred <- unique(counted$S2_PredictUseBonusEarnBin)
+
+  S2_MPL                    <- rep(mpls, length(pred))
+  S2_PredictUseBonusEarnBin <- rep(pred, each=length(mpls))
+
+  empty <- data.frame(S2_MPL, S2_PredictUseBonusEarnBin)
+
+  # replaces the non-missing 
+  full <- empty %>% 
+          left_join(counted, by= c('S2_MPL', 'S2_PredictUseBonusEarnBin')) %>%
+          mutate(Count=ifelse(is.na(Count), 0, Count))
+
+  #plots
+  a <- full %>% 
+        ggplot(aes(S2_MPL, S2_PredictUseBonusEarnBin, fill= Count)) + 
+        geom_tile() +
+        scale_fill_gradient(low = low_grey, high = maroon) +
+        theme_classic() + 
+        labs(x= "Valuation of bonus ($)", y = "Predicted earnings from bonus ($)") + 
+        geom_abline(intercept = 0, slope=1)
+
+  ggsave('output/heatmap_wtp_prediction.pdf', plot=a, width=6.5, height=4.5, units="in")
+}
+
+plot_predicted_actual <- function(df, period){
+  bin_size <- 20
+
+  # filter to just control
+  data <- df %>%
+            filter(B == 0 & L == 0)
+
+  #rename
+  data %<>% mutate(Predicted = !!sym(paste0('S', period, '_PredictUseNext_1'))) %>%
+            mutate(Actual    = !!sym(paste0('PD_P', period, '_UsageFITSBY'))) %>%
+            filter(!is.na(Predicted) & !is.na(Actual))
+
+  counts <- data %>% 
+              mutate(PredictedBin = Predicted - (Predicted %% bin_size) + (bin_size/2)) %>%
+              mutate(ActualBin    = Actual - (Actual %% bin_size) + (bin_size/2)) %>%
+              select(PredictedBin, ActualBin) %>%
+              group_by(PredictedBin, ActualBin) %>%
+              count(name="Count")
+
+  #plots
+  a <- counts %>% 
+        ggplot(aes(PredictedBin, ActualBin, fill= Count)) + 
+        geom_tile() +
+        scale_fill_gradient(low = low_grey, high = maroon) +
+        theme_classic() + 
+        labs(x= "Predicted FITSBY use (minutes/day)", y = "Actual FITSBY use (minutes/day)") + 
+        geom_abline(intercept = 0, slope=1) +
+        xlim(0, 500) + ylim(0, 500)
+
+  ggsave(sprintf('output/heatmap_usage_P%s.pdf', period), plot=a, width=6.5, height=4.5, units="in")
+
+}
+
+plot_predicted_actual_all <- function(df){
+  bin_size <- 20
+
+  # filter to just control
+  data <- df %>%
+            filter(B == 0 & L == 0)
+
+  #rename
+  p2 <- data %>% mutate(Predicted = S2_PredictUseNext_1) %>%
+                 mutate(Actual    = PD_P2_UsageFITSBY) %>%
+                 filter(!is.na(Predicted) & !is.na(Actual)) %>%
+                 select(Predicted, Actual)
+
+  p3 <- data %>% mutate(Predicted = S3_PredictUseNext_1) %>%
+                 mutate(Actual    = PD_P3_UsageFITSBY) %>%
+                 filter(!is.na(Predicted) & !is.na(Actual)) %>%
+                 select(Predicted, Actual)
+
+  p4 <- data %>% mutate(Predicted = S4_PredictUseNext_1) %>%
+                 mutate(Actual    = PD_P4_UsageFITSBY) %>%
+                 filter(!is.na(Predicted) & !is.na(Actual)) %>%
+                 select(Predicted, Actual)
+
+  all_periods <- rbind(p2, p3, p4)
+
+  counts <- all_periods %>% 
+              mutate(PredictedBin = Predicted - (Predicted %% bin_size) + (bin_size/2)) %>%
+              mutate(ActualBin    = Actual - (Actual %% bin_size) + (bin_size/2)) %>%
+              select(PredictedBin, ActualBin) %>%
+              group_by(PredictedBin, ActualBin) %>%
+              count(name="Count")
+
+  #plots
+  a <- counts %>% 
+        ggplot(aes(PredictedBin, ActualBin, fill= Count)) + 
+        geom_tile() +
+        scale_fill_gradient(low = low_grey, high = maroon) +
+        theme_classic() + 
+        labs(x= "Predicted FITSBY use (minutes/day)", y = "Actual FITSBY use (minutes/day)") + 
+        geom_abline(intercept = 0, slope=1) +
+        xlim(0, 500) + ylim(0, 500)
+
+  ggsave('output/heatmap_usage.pdf', plot=a, width=6.5, height=4.5, units="in")
+
+}
+
+main <- function(){
+  df <- read_dta('input/final_data_sample.dta')
+
+  # clean data
+  df %<>%
+    mutate(L = ifelse(S2_LimitType != 0, 1, 0)) %>%
+    mutate(B = ifelse(S3_Bonus == 1, 1, 0)) %>%
+    mutate(S = as.character(Stratifier))
+
+  plot_wtp_prediction(df)
+
+  plot_predicted_actual(df, 2)
+  plot_predicted_actual(df, 3)
+  plot_predicted_actual(df, 4)
+
+  plot_predicted_actual_all(df)
+
+}
+
+
+main()

17/replication_package/code/analysis/descriptive/code/QualitativeEvidence.do

@@ -0,0 +1,152 @@
+// Baseline qualitative evidence
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program define_plot_settings
+    global HIST_SETTINGS ///
+        xlabel(, labsize(large)) ///
+		ylabel(, labsize(large)) ///
+        ytitle("Fraction of sample" " ", size(large)) ///
+        bcolor(maroon) graphregion(color(white)) ///
+        xsize(6.5) ysize(4.5)	
+	
+	global HIST_DISCRETE_SETTINGS ///
+        gap(50) xlabel(, valuelabel noticks) ///
+		$HIST_SETTINGS
+
+    global HIST_CONTINUOUS_SETTINGS ///
+		$HIST_SETTINGS
+        
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5)
+        
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon black gray) ///
+        cicolor(maroon black gray) 
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+    
+	plot_self_control
+    plot_self_control_by_age
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program plot_self_control
+	twoway hist S1_InterestInLimits, frac discrete ///
+        $HIST_DISCRETE_SETTINGS ///
+		xtitle(" " "Interest in limits", size(large))
+ 
+    graph export "output/hist_limits_interest.pdf", replace 
+	
+    twoway hist S1_PhoneUseChange, frac ///
+        $HIST_CONTINUOUS_SETTINGS ///
+        width(5) start(-102.5) ///
+		xtitle(" " "Ideal use change (percent)", size(large))
+
+    graph export "output/hist_phone_use.pdf", replace
+            
+    twoway hist S1_LifeBetter, frac discrete ///
+        $HIST_CONTINUOUS_SETTINGS ///
+		xtitle(" " "Phone use makes life worse (left) or better (right)", size(large)) ///
+		xtick(-5(2.5)5) xlabel(-5(5)5) 
+        
+    graph export "output/hist_life_betterworse.pdf", replace
+            
+    hist S1_AddictionIndex, frac ///
+        $HIST_CONTINUOUS_SETTINGS ///
+        xtitle(" " "Addiction scale", size(large)) 
+
+    graph export "output/hist_addiction_index.pdf", replace
+	
+    
+	hist S1_SMSIndex, frac ///
+        $HIST_CONTINUOUS_SETTINGS ///
+		xtitle(" " "SMS addiction scale", size(large))
+        
+    graph export "output/hist_sms_index.pdf", replace
+    
+end
+
+program plot_self_control_by_age
+    * Preserve data
+    preserve
+
+    * Reshape data
+    keep UserID AgeGroup PD_P1_UsageFITSBY  Strat*Index
+    rename_but, varlist(UserID AgeGroup) prefix(index)
+    reshape long index, i(UserID AgeGroup) j(measure) string
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+        
+    recode measure_encode ///
+        (2 = 1 "Addiction index") ///
+        (3 = 2 "Restriction index") ///
+		(1 = 3 "Period 1 FITSBY Usage"), ///
+        gen(measure_recode)
+    
+	* Define plot settings
+	
+	// - When creating multiple y-axis plots, Stata unfortunately makes no 
+	//   attempt to align the different y-axes
+	// - Manually adjust the follaowing options to properly align the y-axes
+	// - Note that values for legend order are also manually specified 
+	//   (but do not need to be adjusted) as including multiple y-axes jumbles
+	//   the legend order expected by the cispike command
+	local ylabel1 -.4(.2).6
+    local ylabel2 100(10)200
+    local yrange2 range(100, 200)
+	
+	* Plot data
+	
+    cispike index, ///
+        over1(measure_recode) over2(AgeGroup) ///
+        $CISPIKE_SETTINGS ///
+		spike( yaxis(1) || yaxis(1) || yaxis(2)) ci( yaxis(1) || yaxis(1) || yaxis(2)) ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+				  ytitle("Standard deviations" " ", axis(1)) ///
+				  ytitle(" " "Usage (minutes/day)", axis(2)) ///
+				  ylabel(`ylabel1', axis(1)) ///
+				  ylabel(`ylabel2', axis(2)) ///
+				  yscale(`yrange2' axis(2)) ///
+				  legend(order(11 "Addiction index" ///
+		                       16 "Restriction index" ///
+					           26 "Period 1 FITSBY Usage"))) 
+		
+    graph export "output/cispike_self_control_index_by_age.pdf", replace
+    
+    * Restore data
+    restore
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/SampleStatistics.do

@@ -0,0 +1,138 @@
+// Sample statistics
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program latex
+   syntax, name(str) value(str)
+   
+   local command = "\newcommand{\\`name'}{`value'}"
+   
+   file open scalars using "output/scalars.tex", write append
+   file write scalars `"`command'"' _n
+   file close scalars
+end
+
+program latex_integer
+   syntax, name(str) value(str)
+   
+   local value : display %8.0gc `value'
+   local value = trim("`value'")
+   
+   latex, name(`name') value(`value')
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+   define_constants
+   import_data
+   
+   get_samples
+end
+
+program import_data
+    use "input/final_data.dta", clear
+end
+
+program get_samples
+   cap sencode UserID, replace
+   
+   * Shown ad
+   latex, name(shownad) value("3,271,165") 
+
+   * Clicked on ad
+   sum UserID if S0_Finished != .
+   latex_integer, name(clickedonad) value(`r(N)')
+
+   * Passed pre-screen
+   sum UserID if S0_Android == 1 & S0_Country == 1 & S0_Age >= 18 & S0_Age < 65 & ///
+         S0_PhoneCount == 1 & S0_Android == 1
+   latex_integer, name(passedprescreen) value(`r(N)')
+ 
+   * Consented
+   sum UserID if S0_Consent == 1
+   latex_integer, name(consented) value(`r(N)')
+ 
+   * Finished intake
+   sum UserID if S0_Finished == 1 & S0_Consent == 1
+   latex_integer, name(finishedintake) value(`r(N)')
+   
+   * Began baseline
+   sum UserID if S1_Finished != .
+   latex_integer, name(beganbaseline) value(`r(N)')
+
+   * Finished baseline
+   sum UserID if S1_Finished == 1
+   latex_integer, name(finishedbaseline) value(`r(N)')
+   local finishedbaseline `r(N)'
+   
+   * Randomized
+   sum UserID if S1_Finished == 1 & Randomize == 1
+   latex_integer, name(randomized) value(`r(N)')
+   local randomized `r(N)'
+   
+   * Dropped from baseline
+   local dropped = `finishedbaseline' - `randomized'
+   latex_integer, name(droppedbaseline) value(`dropped')
+   
+   * Began midline
+   sum UserID if S2_Finished != .
+   latex_integer, name(beganmidline) value(`r(N)')
+
+   * Informed of treatment
+   sum UserID if S2_RevealConfirm == 1
+   latex_integer, name(informedtreat) value(`r(N)')
+   
+   * Finished midline
+   sum UserID if S2_Finished == 1 & S2_RevealConfirm == 1
+   latex_integer, name(finishedmidline) value(`r(N)')
+   
+   * Began endline
+   sum UserID if S3_Finished != .
+   latex_integer, name(beganendline) value(`r(N)')
+   
+   * Finished endline
+   sum UserID if S3_Finished == 1
+   latex_integer, name(finishedendline) value(`r(N)')
+   
+   * Began post-endline
+   sum UserID if S4_Finished != .
+   latex_integer, name(beganpostendline) value(`r(N)')
+   
+   * Finished endline
+   sum UserID if S4_Finished == 1
+   latex_integer, name(finishedpostendline) value(`r(N)')
+
+   sum UserID if S4_Finished == 1 & PD_P5_Usage != .
+   latex_integer, name(kepttoend) value(`r(N)')
+
+   * Analytical sizes
+   sum UserID if S2_RevealConfirm == 1 & S3_Bonus <= 1
+   latex_integer, name(informedtreatanalysis) value(`r(N)')
+
+   sum UserID if S2_RevealConfirm == 1 & S3_Bonus <= 1 & PD_P5_Usage != . & S4_Finished == 1
+   latex_integer, name(kepttoendanalysis) value(`r(N)')
+
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/Scalars.do

@@ -0,0 +1,625 @@
+// Ad hoc scalars for text of main paper
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+    yaml global STRATA = YAML.metadata.strata
+end
+
+program latex
+   syntax, name(str) value(str)
+   
+   local command = "\newcommand{\\`name'}{`value'}"
+   
+   file open scalars using "output/scalars.tex", write append
+   file write scalars `"`command'"' _n
+   file close scalars
+end
+
+program latex_rounded
+   syntax, name(str) value(str) digits(str)
+   
+   local value : display %8.`digits'fc `value'
+   local value = trim("`value'")
+   
+   latex, name(`name') value(`value')
+end
+
+program latex_precision
+   syntax, name(str) value(str) digits(str)
+   
+   autofmt, input(`value') dec(`digits') strict
+   local value = r(output1)
+   
+   latex, name(`name') value(`value')
+end
+
+program reshape_swb
+	* Reshape wide to long
+	keep UserID S3_Bonus S2_LimitType Stratifier S*_SWBIndex_N 
+	
+	local indep UserID S3_Bonus S2_LimitType Stratifier S1_*
+	rename_but, varlist(`indep') prefix(outcome)
+    reshape long outcome, i(`indep') j(measure) string
+	
+	split measure, p(_)
+	replace measure = measure2 + "_" + measure3
+	rename measure1 survey
+    drop measure2 measure3 
+	
+	* Reshape long to wide
+	reshape wide outcome, i(UserID survey) j(measure) string
+	rename outcome* *
+	
+	* Recode data
+	encode survey, gen(S)
+	
+	* Label data
+	label var SWBIndex "Subjective well-being"
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    import_sample_data
+	
+	get_usage_info_open
+    get_percent_fitsby
+	get_percent_limit
+	get_ideal_use
+	get_life_worse
+	get_addict
+	get_bonus_effect
+	get_limit_effect
+	get_valuations
+	get_baseline_usage
+	get_compare2019
+	get_substitution
+	get_swb_pvalues
+	get_bonus_desire
+	get_pd_usage
+	get_medians
+	get_bound_use
+	
+
+	* import_data
+	* get_other_blocker_use
+
+end
+
+program import_sample_data
+    use "input/final_data_sample.dta", clear
+end
+
+program import_data
+    use "input/final_data.dta", clear
+end
+
+program tab_percent
+	syntax, var(str) key(str) name(str) digits(str)
+	
+	* Generate dummy
+	cap drop dummy
+	gen dummy = 0
+	replace dummy = 1 if inlist(`var', `key')
+	
+	* Tabulate dummy
+	sum dummy
+	local perc = `r(mean)' * 100
+	latex_rounded, name(`name') value(`perc') digits(`digits')
+end
+
+program get_usage_info_open
+    latex, name(usageinfoopen) value("XXX") // WIP
+end
+
+program get_percent_fitsby
+	* Preserve data
+	preserve
+	  
+	* Reshape data
+	keep UserID PD_*_Usage_* PD_*_Installed_*
+	keep UserID *Facebook *Instagram *Twitter *Snapchat *Browser *YouTube
+    rename_but, varlist(UserID) prefix(use)
+    reshape long use, i(UserID) j(j) string
+	
+	split j, p(_)
+    rename j4 app
+
+	* Get apps used
+	collapse (sum) use, by(UserID app)
+    replace use = 1 if use > 0 & use != .
+	
+	* Get number of apps used
+	collapse (sum) use, by(UserID)
+	
+	* Get percent all apps used
+	tab_percent, ///
+		var(use) key(6) ///
+		name(percentfitsby) digits(1)
+		
+	* Restore data
+	restore
+end
+
+program get_percent_limit
+	* Get percent moderately or very interested
+	tab_percent, ///
+		var(S1_InterestInLimits) key(3, 4) ///
+		name(percentlimitinterested) digits(0)
+	
+	* Get percent not at all interested
+	tab_percent, ///
+		var(S1_InterestInLimits) key(1) ///
+		name(percentlimitnot) digits(0)
+end
+
+program get_ideal_use
+	* Get percent just right
+	tab_percent, ///
+		var(S1_PhoneUseFeel) key(2) ///
+		name(percentuseright) digits(0)
+		
+	* Get percent too little
+	tab_percent, ///
+		var(S1_PhoneUseFeel) key(3) ///
+		name(percentuselittle) digits(1)
+			
+	* Get mean total ideal reduction
+	sum S1_PhoneUseReduce
+	local mean = r(mean)
+	latex_rounded, name(idealreduction) value(`mean') digits(0)
+	
+	* Get mean Facebook ideal reduction
+	recode S1_IdealApp_Facebook ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_Facebook_recode)
+	
+	sum S1_IdealApp_Facebook_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductionfacebook) value(`mean') digits(0)
+
+	* Get mean Instagram ideal reduction
+	recode S1_IdealApp_Instagram ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_Instagram_recode)
+	
+	sum S1_IdealApp_Instagram_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductioninsta) value(`mean') digits(0)
+
+	* Get mean Twitter ideal reduction
+	recode S1_IdealApp_Twitter ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_Twitter_recode)
+	
+	sum S1_IdealApp_Twitter_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductiontwitter) value(`mean') digits(0)
+
+	* Get mean Snapchat ideal reduction
+	recode S1_IdealApp_Snapchat ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_Snapchat_recode)
+	
+	sum S1_IdealApp_Snapchat_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductionsnap) value(`mean') digits(0)
+
+	* Get mean Browser ideal reduction
+	recode S1_IdealApp_Browser ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_Browser_recode)
+	
+	sum S1_IdealApp_Browser_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductionbrowser) value(`mean') digits(0)
+
+	* Get mean YouTube ideal reduction
+	recode S1_IdealApp_YouTube ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_YouTube_recode)
+	
+	sum S1_IdealApp_YouTube_recode
+	local mean = r(mean) * -1
+	latex_rounded, name(idealreductionyoutube) value(`mean') digits(0)
+
+end
+
+program get_life_worse
+	* Get percent life worse
+	tab_percent, ///
+		var(S1_LifeBetter) key(-5, -4, -3, -2, -1) ///
+		name(percentlifeworse) digits(0)
+end
+
+program get_addict		
+	* Get mean addiction index
+	sum S1_AddictionIndex
+	local mean = r(mean) * -1
+	latex_rounded, name(scaleaddict) value(`mean') digits(1)
+end
+
+program get_bonus_effect	
+	preserve
+	local baseline PD_P1_UsageFITSBY
+	local yvar PD_P2_UsageFITSBY
+	gen_treatment, simple
+    reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+	local treatment = -_b[B]
+	latex_precision, name(bonustwo) value(`treatment') digits(2)
+
+	local baseline PD_P1_UsageFITSBY
+	local yvar PD_P3_UsageFITSBY
+	gen_treatment, simple
+    reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+	local treatment = -_b[B]
+	latex_precision, name(bonusthree) value(`treatment') digits(2)
+
+	sum PD_P3_UsageFITSBY if B == 0 & L == 0 
+	local reduction = (`treatment'/r(mean))*100
+	latex_precision, name(bonusthreepct) value(`reduction') digits(2)
+
+	local baseline PD_P1_UsageFITSBY
+	local yvar PD_P4_UsageFITSBY
+	gen_treatment, simple
+    reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+	local treatment4 = -_b[B]
+	latex_precision, name(bonusfour) value(`treatment4') digits(2)
+
+	local baseline PD_P1_UsageFITSBY
+	local yvar PD_P5_UsageFITSBY
+	gen_treatment, simple
+    reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+	local treatment5 = -_b[B]
+	latex_precision, name(bonusfive) value(`treatment5') digits(2)
+	restore
+end
+
+program get_limit_effect	
+	preserve
+	
+	local baseline PD_P1_UsageFITSBY
+	local yvar PD_P5432_UsageFITSBY
+	gen_treatment, simple
+    reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+	local treatment = -_b[L]
+	latex_precision, name(limiteffectstataadhoc) value(`treatment') digits(2)
+
+	sum PD_P5432_UsageFITSBY if B == 0 & L == 0 
+	local reduction = (`treatment'/r(mean))*100
+
+	latex_precision, name(limiteffectpct) value(`reduction') digits(2)
+	restore
+end
+
+program get_valuations	
+	preserve
+
+	sum S2_MPL
+	local vb = r(mean)
+	latex_precision, name(valuebonus) value(`vb') digits(2)
+
+	sum S3_MPLLimit
+	local vl = r(mean)
+	local numlimit = r(N)
+	latex_precision, name(valuelimit) value(`vl') digits(3)
+
+
+	sum S3_MPLLimit if S3_MPLLimit  > 0
+	local numpaylimit = r(N)
+	local positivelimit = (`numpaylimit' / `numlimit') * 100
+	latex_precision, name(positivelimit) value(`positivelimit') digits(2)
+
+	sum S3_MPLLimit if S3_MPLLimit  > 10
+	local numpayten = r(N)
+	local tenlimit = (`numpayten' / `numlimit') * 100
+	latex_precision, name(tenlimit) value(`tenlimit') digits(2)
+
+	restore
+end
+
+program get_baseline_usage	
+	preserve
+
+	sum PD_P1_Usage
+	local avg_all = r(mean)
+	latex_precision, name(avgOverall) value(`avg_all') digits(2)
+
+	sum PD_P1_UsageFITSBY
+	local avg_fitsby = r(mean)
+	latex_precision, name(avgFITSBY) value(`avg_fitsby') digits(2)
+
+	local avg_fitsby_pct = (`avg_fitsby' / `avg_all') * 100
+	latex_precision, name(avgFITSBYpct) value(`avg_fitsby_pct') digits(2)
+
+	sum PD_P1_Usage_Facebook
+	local avg_fb = r(mean)
+	latex_precision, name(avgFB) value(`avg_fb') digits(2)
+
+	sum PD_P1_Usage_Browser
+	local avg_br = r(mean)
+	latex_precision, name(avgBR) value(`avg_br') digits(2)
+
+	sum PD_P1_Usage_YouTube
+	local avg_yt = r(mean)
+	latex_precision, name(avgYT) value(`avg_yt') digits(2)
+
+	sum PD_P1_Usage_Instagram
+	local avg_in = r(mean)
+	latex_precision, name(avgIN) value(`avg_in') digits(2)
+
+	sum PD_P1_Usage_Snapchat
+	local avg_sc = r(mean)
+	latex_precision, name(avgSC) value(`avg_sc') digits(2)
+
+	sum PD_P1_Usage_Twitter
+	local avg_tw = r(mean)
+	latex_precision, name(avgTW) value(`avg_tw') digits(2)
+
+	restore
+end
+
+program get_compare2019
+	preserve
+
+	sum S1_CovidChangesFreeTime
+	local ss = r(N)
+
+	sum S1_CovidChangesFreeTime if S1_CovidChangesFreeTime > 4
+	local num_worse = r(N)
+
+	local covidfree = 100 * `num_worse'/`ss'
+	latex_precision, name(covidmorefree) value(`covidfree') digits(2)
+
+    recode S1_CovidChangeReason ///
+        (1 = 4 "Increased phone usage") ///
+        (2 = 4 "Increased phone usage") ///
+        (3 = 3 "No change") ///
+        (4 = 4 "Increased phone usage") ///
+        (5 = 2 "Decreased phone usage") ///
+        (6 = 1 "Other"), ///
+        gen(S1_CovidChangeReason_recode)
+
+	sum S1_CovidChangesFreeTime
+	local ss2 = r(N)
+
+	sum S1_CovidChangeReason_recode if S1_CovidChangeReason_recode == 4
+	local num_more_phone = r(N)
+
+	local morephoneuse = 100 * `num_more_phone'/`ss2'
+	latex_precision, name(morephoneuse) value(`morephoneuse') digits(2)
+
+	restore
+end
+
+program get_substitution
+	preserve
+
+    gen_treatment, simple
+    reg_treatment, yvar(S4_Substitution_W) indep($STRATA) simple
+
+	local bsub = -_b[B]
+	latex_precision, name(bonussubstitution) value(`bsub') digits(2)
+
+	local lsub = _b[L]
+	latex_precision, name(limitsubstitution) value(`lsub') digits(2)
+
+	gen avg_overall = (PD_P3_Usage + PD_P4_Usage + PD_P5_Usage)/3
+	gen avg_fitsby  = (PD_P3_UsageFITSBY + PD_P4_UsageFITSBY + PD_P5_UsageFITSBY)/3
+
+	gen avg_non_fitsby = avg_overall - avg_fitsby
+    reg_treatment, yvar(avg_non_fitsby) indep($STRATA) simple
+
+	local bsub = -_b[B]
+	latex_precision, name(bonusnonfitsby) value(`bsub') digits(2)
+
+	local lsub = _b[L]
+	latex_precision, name(limitnonfitsby) value(`lsub') digits(2)
+	restore
+end
+
+program get_swb_pvalues 
+    est clear
+     
+	* Preserve data
+	preserve
+
+	* Reshape data
+	reshape_swb
+	
+    * Specify regression 
+    local yvar SWBIndex_N 
+		
+    * Run regressions
+	local baseline = "S1_`yvar'"
+	
+	* Treatment indicators
+	gen_treatment, simple
+	cap drop B3
+	cap drop B4
+	gen B3 = B * (S == 1)
+	gen B4 = B * (S == 2)
+	
+	* Specify regression
+	local indep i.S i.S#$STRATA i.S#c.`baseline'
+	
+	reg `yvar' L B3 B4 `indep', robust cluster(UserID)
+	
+	local lprob = _P[L]
+	local lcoef = _b[L]
+	latex_precision, name(limitSWBpval) value(`lprob') digits(2)
+	latex_precision, name(limitSWBcoef) value(`lcoef') digits(1)
+
+ 	local bprob = _P[B4]
+	local bcoef = _b[B4]
+	latex_precision, name(bonusSWBpval) value(`bprob') digits(2)
+	latex_precision, name(bonusSWBcoef) value(`bcoef') digits(1)
+
+	* Restore data
+	restore
+end
+
+program get_bonus_desire
+	sum S2_PredictUseInitial_W
+	local avg_prediction = r(mean)
+	latex_precision, name(MPLprediction) value(`avg_prediction') digits(2)
+
+	sum S2_PredictUseBonus 
+	local avg_reduction_pct = r(mean)
+	latex_precision, name(MPLreductionpct) value(`avg_reduction_pct') digits(2)
+
+	gen reduction = S2_PredictUseInitial_W * (S2_PredictUseBonus / 100)
+	sum reduction
+	local avg_reduction_mins = r(mean)
+	latex_precision, name(MPLreductionmins) value(`avg_reduction_mins') digits(2)
+
+	gen value = (reduction/60)*50
+	sum value
+	local avg_bonus_earnings = r(mean)
+	latex_precision, name(MPLearnings) value(`avg_bonus_earnings') digits(2)
+
+	sum S2_MPL
+	local avg_value_bonus = r(mean)
+	latex_precision, name(MPLvalue) value(`avg_value_bonus') digits(2)
+
+	gen premium = S2_MPL - value
+	sum premium
+	local avg_premium = r(mean)
+	latex_precision, name(MPLpremium) value(`avg_premium') digits(2)
+
+	sum S2_MPLReasoning
+	local total_respondents = r(N)
+
+	sum S2_MPLReasoning if S2_MPLReasoning == 2
+	local wish_reduce = r(N)
+	local wish_reduce_pct = (`wish_reduce' / `total_respondents') * 100
+	latex_precision, name(MPLwishreduce) value(`wish_reduce_pct') digits(2)
+
+	sum S2_MPLReasoning if S2_MPLReasoning == 1
+	local maximize = r(N)
+	local maximize_pct = (`maximize' / `total_respondents') * 100
+	latex_precision, name(MPLmaximize) value(`maximize_pct') digits(2)
+
+	sum S2_MPLReasoning if S2_MPLReasoning == 3
+	local no_pressure = r(N)
+	local no_pressure_pct = (`no_pressure' / `total_respondents') * 100
+	latex_precision, name(MPLnopressure) value(`no_pressure_pct') digits(2)
+
+	sum premium if S2_MPLReasoning == 2
+	local premium_reduce = r(mean)
+
+	sum premium if S2_MPLReasoning == 3
+	local premium_no_pressure = r(mean)	
+
+	local premium_difference = `premium_reduce' - `premium_no_pressure'
+	latex_precision, name(MPLpremiumdifference) value(`premium_difference') digits(2)
+end
+
+program get_pd_usage
+	gen_treatment, simple
+
+	sum PD_P5432_UsageMinutesPD  if B == 1
+	local mins_bonus = r(mean) / 84
+	latex_precision, name(BonusPDmins) value(`mins_bonus') digits(2)
+
+	sum PD_P5432_UsageMinutesPD  if L == 1
+	local mins_limit = r(mean) / 84
+	latex_precision, name(LimitPDmins) value(`mins_limit') digits(2)	
+
+	sum PD_P5432_UsageMinutesPD  if B == 0 & L == 0
+	local mins_control = r(mean) / 84
+	latex_precision, name(ControlPDmins) value(`mins_control') digits(1)	 
+
+end
+
+program get_medians
+	sum S0_Age, detail
+	local med_age = r(p50)
+	latex_precision, name(MedianAge) value(`med_age') digits(2)
+
+	sum PD_P1_UsageFITSBY, detail
+	local med_use = r(p50)
+	latex_precision, name(MedianFITSBYUsage) value(`med_use') digits(2)
+end
+
+program get_bound_use
+	gen baseline = ceil(PD_P1_UsageFITSBY/60)*60
+	gen exceeds  = PD_P3_UsageFITSBY > baseline
+	sum exceeds if S3_Bonus == 1
+	local pct_exceed = r(mean) * 100
+	latex_precision, name(PercentExceedBonus) value(`pct_exceed') digits(2)
+
+	gen huge_drop  = PD_P3_UsageFITSBY < (baseline - 180)
+	sum huge_drop if S3_Bonus == 1
+	local pct_huge_drop = r(mean) * 100
+	latex_rounded, name(PercentBoundDrop) value(`pct_huge_drop') digits(0)
+end
+
+program get_other_blocker_use
+	sum S1_OtherLimitUse if S1_Finished == 1
+	local pct_other_blocker_use = r(mean) * 100
+	latex_rounded, name(OtherBlockerUse) value(`pct_other_blocker_use') digits(0)
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/code/Temptation.do

@@ -0,0 +1,100 @@
+// Figure 1
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+	yaml global STRATA = YAML.metadata.strata
+
+    global app_list Facebook Instagram Twitter Snapchat Browser YouTube Other
+end
+
+
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    import_data
+    plot_figure_1
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+
+program plot_figure_1
+	* Preserve data
+    preserve
+
+	* Drop unnecessary columns
+	keep UserID S4_Temptation_*
+	* Pivot the columns into a new variable
+    reshape long S4_ , i(UserID) j(control) string
+	* Assign values to too little (-1) the right amount (0) too much (1)
+	recode  S4_ (1 = -1) (2 = 0 ) (3 = 1), gen(S4_N)
+
+	* Relabel
+	replace control="Exercise" if control=="Temptation_1"
+	replace control="{bf:Use smartphone" if control=="Temptation_2"
+	replace control="Eat unhealthy food" if control=="Temptation_3"
+	replace control="{bf:Check email" if control=="Temptation_4"
+	replace control="{bf:Play video games" if control=="Temptation_5"
+	replace control="Watch TV" if control=="Temptation_6"
+	replace control="Work" if control=="Temptation_7"
+	replace control="{bf:Browse social media" if control=="Temptation_8"
+	replace control="Smoke cigarettes" if control=="Temptation_9"
+	replace control="{bf:Read online news" if control=="Temptation_10"
+	replace control="Drink alcohol" if control=="Temptation_11"
+	replace control="Sleep" if control=="Temptation_12"
+	replace control="Save for retirement" if control=="Temptation_13"
+
+	* Collapse to values of interest
+	drop UserID
+	collapse (mean)  S4_m = S4_N  (semean) S4_se=S4_N (count) S4_count = S4_N, by(control)
+	* Change label for - values, take absolute, and sort.
+	replace control=control+" (-1)" if S4_m<0
+	replace control=control+"}" if strpos(control,"bf")>0
+
+	replace S4_m=abs(S4_m)
+	gsort -S4_m
+
+	* Create 95% CI bands
+	gen S4_m_lb = S4_m - 1.96*S4_se
+	gen S4_m_ub = S4_m + 1.96*S4_se
+
+	* Plot
+	gen axis = _n
+	labmask axis, val(control)
+
+	twoway (rcap S4_m_lb S4_m_ub axis, lcolor(maroon)) (scatter S4_m axis, msize(small)), ///
+	xlabel(1(1)13,valuelabel angle(45) labsize(small)) ///
+	ytitle("absolute value of" "(share “too much” – share “too little”)", ///
+	size(small)) xtitle("{&larr} more perceived self-control problems | less perceived self-control problems {&rarr}") legend(off) graphregion(color(white))
+
+	graph export "output/online_and_offline_temptation_scatter.pdf", replace
+
+	* Restore data
+    restore
+
+
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/descriptive/input.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1aacb5e3a47846afaf251dbe069f5cee136e1255fe4ec43721a033fafc1d837d
+size 812

17/replication_package/code/analysis/descriptive/make.py

@@ -0,0 +1,75 @@
+###################
+### ENVIRONMENT ###
+###################
+import git
+import imp
+import os
+
+### SET DEFAULT PATHS
+ROOT = '../..'
+
+PATHS = {
+    'root'             : ROOT,
+    'lib'              : os.path.join(ROOT, 'lib'),
+    'config'           : os.path.join(ROOT, 'config.yaml'),
+    'config_user'      : os.path.join(ROOT, 'config_user.yaml'),
+    'input_dir'        : 'input',
+    'external_dir'     : 'external',
+    'output_dir'       : 'output',
+    'output_local_dir' : 'output_local',
+    'makelog'          : 'log/make.log',
+    'output_statslog'  : 'log/output_stats.log',
+    'source_maplog'    : 'log/source_map.log',
+    'source_statslog'  : 'log/source_stats.log',
+}
+
+### LOAD GSLAB MAKE
+f, path, desc = imp.find_module('gslab_make', [PATHS['lib']])
+gs = imp.load_module('gslab_make', f, path, desc)
+
+### LOAD CONFIG USER
+PATHS = gs.update_paths(PATHS)
+gs.update_executables(PATHS)
+
+############
+### MAKE ###
+############
+
+### START MAKE
+gs.remove_dir(['input', 'external'])
+gs.clear_dir(['output', 'log', 'temp'])
+gs.start_makelog(PATHS)
+
+### GET INPUT FILES
+inputs = gs.link_inputs(PATHS, ['input.txt'])
+# gs.write_source_logs(PATHS, inputs + externals)
+# gs.get_modified_sources(PATHS, inputs + externals)
+
+### RUN SCRIPTS
+"""
+Critical
+--------
+Many of the Stata analysis scripts recode variables using
+the `recode` command. Double-check all `recode` commands
+to confirm recoding is correct, especially when reusing
+code for a different experiment version.
+"""
+
+gs.run_stata(PATHS, program = 'code/Scalars.do')
+#gs.run_stata(PATHS, program = 'code/SampleStatistics.do')
+gs.run_stata(PATHS, program = 'code/DataDescriptives.do')
+gs.run_stata(PATHS, program = 'code/QualitativeEvidence.do')
+gs.run_stata(PATHS, program = 'code/CommitmentDemand.do')
+gs.run_stata(PATHS, program = 'code/COVIDResponse.do')
+gs.run_stata(PATHS, program = 'code/Temptation.do')
+
+gs.run_r(PATHS,     program = 'code/HeatmapPlots.R')
+
+### LOG OUTPUTS
+gs.log_files_in_output(PATHS)
+
+### CHECK FILE SIZES
+#gs.check_module_size(PATHS)
+
+### END MAKE
+gs.end_makelog(PATHS)

17/replication_package/code/analysis/structural/.RData

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7292ca6438e8b7e80608a5015f136523305b10e1497fa96808904f0c51ab72cd
+size 3785816

17/replication_package/code/analysis/structural/.Rhistory

@@ -0,0 +1,512 @@
+install.packages("optimx")
+library("optimx")
+library("stats")
+library("tidyverse")
+square_function <- function(x){
+y = x^2
+return(y)
+}
+square_function(x=3)
+square_function(x=0)
+intial_values <- c(-2,1,2)
+minimise_function <- optimr(intial_values, square_function)
+intial_values <- c(-2)
+minimise_function <- optimr(intial_values, square_function)
+minimise_function <- optimr(intial_values, square_function, method = "Brent")
+minimise_function <- optimr(intial_values, square_function)
+minimise_function <- optimr(par = intial_values, fn=square_function, method = "Brent")
+install.packages("Brent")
+minimise_function <- optimize(f=square_function, lower = -10, upper=10)
+minimise_function$par
+minimise_function
+minimise_function <- optimize(f=square_function, lower = -10000000, upper=100000000)
+minimise_function
+cube_function <- function(x){
+y = x^3
+return(y)
+}
+minimise_function        <- optimize(f=cube_function, lower = -10000000, upper=100000000)
+minimise_function
+sinus_function <- function(x){
+y = sin(x)
+return(y)
+}
+minimise_function        <- optimize(f=sinus_function, lower = -10000000, upper=100000000)
+minimise_function
+bivariate_function <- function(x,y){
+z <- 2*x*(y**2)+2*(x**2)*y+x*y
+return(z)
+}
+# 1. First try a few values of x, y and see how it affect z
+x<- seq(-0.5,0.5, len=200)
+y<- seq(-0.5,0.5, len=200)
+z <- outer(x,y,bivariate_function)
+persp(x,y,z, theta=-30,phi=15,ticktype="detailed")
+image(x,y,z)
+bivariate_function_vector <- function(vec){
+x <- vec[1]
+y <- vec[2]
+z <- 2*x*(y**2)+2*(x**2)*y+x*y
+return(z)
+}
+minimise_function_bivariate <- optimr(par = c(0.5,0.5),  bivariate_function_vector, control=list(fnscale=-1))
+minimise_function_bivariate$par
+minimise_function_bivariate <- optimr(par = c(0.5,0.5),  bivariate_function_vector)
+minimise_function_bivariate$par
+minimise_function_bivariate$par
+minimise_function_bivariate <- optimr(par = c(0.5,0.5),  bivariate_function)
+minimise_function_bivariate <- optimr(par = c(0.5,0.5),  bivariate_function_vector)
+minimise_function_bivariate$par
+bivariate_function_vector <- function(vec){
+x <- vec[1]
+y <- vec[2]
+z <-  (1-x)^2 + 100*(y-x^2)
+return(z)
+}
+minimise_function_bivariate <- optimr(par = c(0,0),  bivariate_function_vector)
+minimise_function_bivariate$par
+bivariate_function_vector <- function(vec){
+x <- vec[1]
+y <- vec[2]
+z <-  (1-x)^2 + 100*(y-x^2)^2
+return(z)
+}
+minimise_function_bivariate <- optimr(par = c(0,0),  bivariate_function_vector)
+minimise_function_bivariate$par
+remvove(list=ls())
+remove(list=ls())
+getwd()
+setwd("/Users/houdanaitelbarj/Desktop/PhoneAddiction/analysis/structural")
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Setup
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Import plotting functions and constants from lib file
+source('input/lib/r/ModelFunctions.R')
+# Import data
+df <- import_data()
+param %<>%
+list.merge(
+#get_opt(df),
+get_taus(df, winsorize=winsorize, full=full),
+get_mispredict(df),
+get_ideal(df),
+get_predict(df),
+get_wtp(df),
+get_avg_use(df),
+get_fb(df),
+get_limit_last_week(df)
+)
+param <- param_initial
+param %<>%
+list.merge(
+#get_opt(df),
+get_taus(df, winsorize=winsorize, full=full),
+get_mispredict(df),
+get_ideal(df),
+get_predict(df),
+get_wtp(df),
+get_avg_use(df),
+get_fb(df),
+get_limit_last_week(df)
+)
+winsorize=F
+full=F
+param %<>%
+list.merge(
+#get_opt(df),
+get_taus(df, winsorize=winsorize, full=full),
+get_mispredict(df),
+get_ideal(df),
+get_predict(df),
+get_wtp(df),
+get_avg_use(df),
+get_fb(df),
+get_limit_last_week(df)
+)
+View(param)
+param %<>%
+solve_sys_eq_1 %>%
+as.list %>%
+list.merge(param)
+# Solve system of equations #2
+param %<>%
+solve_sys_eq_2(display_warning=display_warning) %>%
+as.list %>%
+list.merge(param)
+display_warning=FALSE
+# Solve system of equation #1
+param %<>%
+solve_sys_eq_1 %>%
+as.list %>%
+list.merge(param)
+# Solve system of equations #2
+param %<>%
+solve_sys_eq_2(display_warning=display_warning) %>%
+as.list %>%
+list.merge(param)
+param %<>%
+solve_sys_eq_3 %>%
+as.list %>%
+list.merge(param)
+# Solve for individual effects
+tau_L_2_spec   <- find_tau_L2_spec(df)
+tau_tilde_spec <- find_tau_L3_spec(df)
+x_ss_i_data    <- calculate_x_ss_i_spec(df)
+param %<>%
+solve_effects_individual(x_ss_i_data= x_ss_i_data, tau_tilde_L=tau_tilde_spec, tau_L_2=tau_L_2_spec, w=df$w)%>%
+as.list %>%
+list.merge(param)
+rho                               <- param[['rho']]
+lambda                            <- param[['lambda']]
+rho_res                               <- param[['rho_res']]
+lambda_res                            <- param[['lambda_res']]
+delta                             <- param[['delta']]
+alpha                             <- param[['alpha']]
+omega                             <- param[['omega']]
+omega_est                          <- param[['omega_est']]
+mispredict                        <- param[['mispredict']]
+d_L                               <- param[['d_L']]
+d_CL                              <- param[['d_CL']]
+eta                               <- param[['eta']]
+zeta                              <- param[['zeta']]
+naivete                           <- param[['naivete']]
+gamma_L_effect                    <- param[['gamma_L_effect']]
+gamma_tilde_L_effect              <- param[['gamma_tilde_L_effect']]
+gamma_tilde_L_effect_omega        <- param[['gamma_tilde_L_effect_omega']]
+gamma_L_effect_omega              <- param[['gamma_L_effect_omega']]
+gamma_L_effect_multiple           <- param[['gamma_L_effect_multiple']]
+gamma_tilde_L_effect_multiple     <- param[['gamma_tilde_L_effect_multiple']]
+gamma_L                           <- param[['gamma_L']]
+gamma_tilde_L                     <- param[['gamma_tilde_L']]
+gamma_tilde_L_omega               <- param[['gamma_tilde_L_omega']]
+gamma_L_omega                     <- param[['gamma_L_omega']]
+gamma_tilde_L_multiple            <- param[['gamma_tilde_L_multiple']]
+gamma_L_multiple                  <- param[['gamma_L_multiple']]
+gamma_B                           <- param[['gamma_B']]
+gamma_tilde_B                     <- param[['gamma_tilde_B']]
+gamma_tilde_B_multiple            <- param[['gamma_tilde_B_multiple']]
+gamma_B_multiple                  <- param[['gamma_B_multiple']]
+eta_res                               <- param[['eta_res']]
+zeta_res                              <- param[['zeta_res']]
+naivete_res                           <- param[['naivete_res']]
+gamma_L_effect_res                    <- param[['gamma_L_effect_res']]
+gamma_tilde_L_effect_res              <- param[['gamma_tilde_L_effect_res']]
+gamma_tilde_L_effect_omega_res        <- param[['gamma_tilde_L_effect_omega_res']]
+gamma_L_effect_omega_res              <- param[['gamma_L_effect_omega_res']]
+gamma_tilde_L_effect_multiple_res     <- param[['gamma_tilde_L_effect_multiple_res']]
+gamma_L_res                           <- param[['gamma_L_res']]
+gamma_L_omega_res                     <- param[['gamma_L_omega_res']]
+gamma_L_multiple_res                   <- param[['gamma_L_multiple_res']]
+gamma_B_res                            <- param[['gamma_B_res']]
+gamma_B_multiple_res                   <- param[['gamma_B_multiple_res']]
+tau_L_2_signed                    <- param[['tau_L_2']]*-1
+# Gamma-spec
+term1                <- (1-alpha)*delta*rho
+term2                <- term1*(1+lambda)
+term3                <- (eta*lambda + zeta*(1 - lambda))*(rho*tau_L_2/omega)
+num                  <- eta*tau_L_2/omega - term1*term3 - term2*naivete
+denom                <- 1 - term2
+num_omega            <- eta*tau_L_2/omega_est - term1*term3 - term2*naivete
+gamma_spec                <- num/denom
+gamma_spec_omega          <- num_omega/denom
+gamma_tilde_spec          <- gamma_spec - naivete
+gamma_tilde_spec_omega    <- gamma_spec_omega - naivete
+tau_L_2 <- param[['tau_L_2']]
+# Gamma-spec
+term1                <- (1-alpha)*delta*rho
+term2                <- term1*(1+lambda)
+term3                <- (eta*lambda + zeta*(1 - lambda))*(rho*tau_L_2/omega)
+num                  <- eta*tau_L_2/omega - term1*term3 - term2*naivete
+denom                <- 1 - term2
+num_omega            <- eta*tau_L_2/omega_est - term1*term3 - term2*naivete
+gamma_spec                <- num/denom
+gamma_spec_omega          <- num_omega/denom
+gamma_tilde_spec          <- gamma_spec - naivete
+gamma_tilde_spec_omega    <- gamma_spec_omega - naivete
+intercept_spec                <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec,           gamma_spec,      alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_L_effect        <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_B               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_B,              gamma_B,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L,              gamma_L,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_spec_omega          <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec_omega,     gamma_spec_omega,      alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_omega  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect_omega, gamma_L_effect_omega,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_omega         <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_omega,        gamma_L_omega,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_multiple  <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_effect_multiple, gamma_L_effect_multiple,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_B_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_B_multiple,        gamma_B_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_multiple,        gamma_L_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_eta_high  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,  eta, zeta, eta_scale=1.1)
+intercept_het_L_effect_eta_low  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,   eta, zeta, eta_scale=0.9)
+x_ss_spec                        <- calculate_steady_state(param,                gamma_tilde_spec, gamma_spec,  alpha, rho,  lambda, mispredict, eta, zeta, intercept_spec)
+x_ss_zero_un                     <- calculate_steady_state(param,                     0,              0,        alpha, rho,  lambda,  0,          eta, zeta, intercept_spec)
+x_ss_zero                        <- ifelse(x_ss_zero_un<0, 0, x_ss_zero_un)
+delta_x                          <- x_ss_spec - x_ss_zero
+x_ss_spec_w                   <- weighted.mean(x_ss_spec, w, na.rm=T)
+w=df$w
+x_ss_spec_w                   <- weighted.mean(x_ss_spec, w, na.rm=T)
+rho                               <- param[['rho']]
+lambda                            <- param[['lambda']]
+rho_res                               <- param[['rho_res']]
+lambda_res                            <- param[['lambda_res']]
+delta                             <- param[['delta']]
+alpha                             <- param[['alpha']]
+omega                             <- param[['omega']]
+omega_est                          <- param[['omega_est']]
+mispredict                        <- param[['mispredict']]
+d_L                               <- param[['d_L']]
+d_CL                              <- param[['d_CL']]
+eta                               <- param[['eta']]
+zeta                              <- param[['zeta']]
+naivete                           <- param[['naivete']]
+gamma_L_effect                    <- param[['gamma_L_effect']]
+gamma_tilde_L_effect              <- param[['gamma_tilde_L_effect']]
+gamma_tilde_L_effect_omega        <- param[['gamma_tilde_L_effect_omega']]
+gamma_L_effect_omega              <- param[['gamma_L_effect_omega']]
+gamma_L_effect_multiple           <- param[['gamma_L_effect_multiple']]
+gamma_tilde_L_effect_multiple     <- param[['gamma_tilde_L_effect_multiple']]
+gamma_L                           <- param[['gamma_L']]
+gamma_tilde_L                     <- param[['gamma_tilde_L']]
+gamma_tilde_L_omega               <- param[['gamma_tilde_L_omega']]
+gamma_L_omega                     <- param[['gamma_L_omega']]
+gamma_tilde_L_multiple            <- param[['gamma_tilde_L_multiple']]
+gamma_L_multiple                  <- param[['gamma_L_multiple']]
+gamma_B                           <- param[['gamma_B']]
+gamma_tilde_B                     <- param[['gamma_tilde_B']]
+gamma_tilde_B_multiple            <- param[['gamma_tilde_B_multiple']]
+gamma_B_multiple                  <- param[['gamma_B_multiple']]
+eta_res                               <- param[['eta_res']]
+zeta_res                              <- param[['zeta_res']]
+naivete_res                           <- param[['naivete_res']]
+gamma_L_effect_res                    <- param[['gamma_L_effect_res']]
+gamma_tilde_L_effect_res              <- param[['gamma_tilde_L_effect_res']]
+gamma_tilde_L_effect_omega_res        <- param[['gamma_tilde_L_effect_omega_res']]
+gamma_L_effect_omega_res              <- param[['gamma_L_effect_omega_res']]
+gamma_tilde_L_effect_multiple_res     <- param[['gamma_tilde_L_effect_multiple_res']]
+gamma_L_res                           <- param[['gamma_L_res']]
+gamma_L_omega_res                     <- param[['gamma_L_omega_res']]
+gamma_L_multiple_res                   <- param[['gamma_L_multiple_res']]
+gamma_B_res                            <- param[['gamma_B_res']]
+gamma_B_multiple_res                   <- param[['gamma_B_multiple_res']]
+tau_L_2 <- param[['tau_L_2']]
+tau_L_2_signed                    <- param[['tau_L_2']]*-1
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual intercepts and steady states under different strategies - Unrestricted alpha
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Gamma-spec
+term1                <- (1-alpha)*delta*rho
+term2                <- term1*(1+lambda)
+term3                <- (eta*lambda + zeta*(1 - lambda))*(rho*tau_L_2/omega)
+num                  <- eta*tau_L_2/omega - term1*term3 - term2*naivete
+denom                <- 1 - term2
+num_omega            <- eta*tau_L_2/omega_est - term1*term3 - term2*naivete
+gamma_spec                <- num/denom
+gamma_spec_omega          <- num_omega/denom
+gamma_tilde_spec          <- gamma_spec - naivete
+gamma_tilde_spec_omega    <- gamma_spec_omega - naivete
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual intercepts and steady states under different strategies
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+intercept_spec                <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec,           gamma_spec,      alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_L_effect        <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_B               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_B,              gamma_B,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L,              gamma_L,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_spec_omega          <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec_omega,     gamma_spec_omega,      alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_omega  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect_omega, gamma_L_effect_omega,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_omega         <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_omega,        gamma_L_omega,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_multiple  <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_effect_multiple, gamma_L_effect_multiple,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_B_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_B_multiple,        gamma_B_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_multiple,        gamma_L_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_eta_high  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,  eta, zeta, eta_scale=1.1)
+intercept_het_L_effect_eta_low  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,   eta, zeta, eta_scale=0.9)
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual counterfactuals
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+x_ss_spec                        <- calculate_steady_state(param,                gamma_tilde_spec, gamma_spec,  alpha, rho,  lambda, mispredict, eta, zeta, intercept_spec)
+x_ss_zero_un                     <- calculate_steady_state(param,                     0,              0,        alpha, rho,  lambda,  0,          eta, zeta, intercept_spec)
+x_ss_zero                        <- ifelse(x_ss_zero_un<0, 0, x_ss_zero_un)
+delta_x                          <- x_ss_spec - x_ss_zero
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual intercepts and steady states under different strategies - Restricted alpha
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Gamma-spec
+alpha_res <- 1
+term1_res                <- (1-alpha_res)*delta*rho_res
+term2_res                <- term1_res*(1+lambda_res)
+term3_res                <- (eta_res*lambda_res + zeta_res*(1 - lambda_res))*(rho_res*tau_L_2/omega)
+num_res                  <- eta_res*tau_L_2/omega - term1_res*term3_res - term2_res*naivete_res
+denom_res                <- 1 - term2_res
+num_omega_res            <- eta_res*tau_L_2/omega_est - term1_res*term3_res - term2_res*naivete_res
+gamma_spec_res                <- num_res/denom_res
+gamma_spec_omega_res          <- num_omega_res/denom_res
+gamma_tilde_spec_res          <- gamma_spec_res - naivete_res
+gamma_tilde_spec_omega_res    <- gamma_spec_omega_res - naivete_res
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual intercepts and steady states under different strategies
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+intercept_spec_res                <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec_res,           gamma_spec_res,      alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_effect_res        <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect_res,       gamma_L_effect_res,  alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_B_res               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_B,              gamma_B_res,         alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_res               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L,              gamma_L_res,         alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_spec_omega_res          <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec_omega_res,     gamma_spec_omega_res,      alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_effect_omega_res  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect_omega_res, gamma_L_effect_omega_res,  alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_omega_res         <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_omega,        gamma_L_omega_res,         alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_effect_multiple_res  <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_effect_multiple, gamma_L_effect_multiple,  alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_B_multiple_res         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_B_multiple,        gamma_B_multiple,         alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_multiple_res         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_multiple,        gamma_L_multiple,         alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res, zeta = zeta_res)
+intercept_het_L_effect_eta_high_res  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res,  zeta =  zeta_res, eta_scale=1.1)
+intercept_het_L_effect_eta_low_res  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha = 1, rho_res, lambda_res, mispredict,   eta = eta_res,  zeta =  zeta_res, eta_scale=0.9)
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Calculate individual counterfactuals
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+x_ss_spec_res                        <- calculate_steady_state(param,                gamma_tilde_spec_res, gamma_spec_res,  alpha = 1, rho_res, lambda_res, mispredict, eta = eta_res,  zeta =  zeta_res, intercept_spec_res)
+x_ss_zero_un_res                     <- calculate_steady_state(param,                     0,              0,        alpha = 1, rho_res, lambda_res, 0,          eta = eta_res,  zeta =  zeta_res, intercept_spec_res)
+x_ss_zero_res                        <- ifelse(x_ss_zero_un_res<0, 0, x_ss_zero_un_res)
+delta_x_res                          <- x_ss_spec_res - x_ss_zero_res
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Compute population averages
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+x_ss_spec_w                   <- weighted.mean(x_ss_spec, w, na.rm=T)
+gamma_tilde_spec_w            <- weighted.mean(gamma_tilde_spec, w, na.rm=T)
+gamma_spec_w                  <- weighted.mean(gamma_spec, w, na.rm=T)
+gamma_spec_omega_w            <- weighted.mean(gamma_spec_omega, w, na.rm=T)
+delta_x_spec                  <- weighted.mean(delta_x, w, na.rm=T)
+x_ss_i_data                   <- weighted.mean(x_ss_i_data, w, na.rm=T)
+remove(list=ls())
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Setup
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Import plotting functions and constants from lib file
+source('input/lib/r/ModelFunctions.R')
+# Import data
+df <- import_data()
+param <- param_initial
+winsorize=F, full=F, display_warning=FALS
+winsorize=F
+full=F
+display_warning=FALSE
+param %<>%
+list.merge(
+#get_opt(df),
+get_taus(df, winsorize=winsorize, full=full),
+get_mispredict(df),
+get_ideal(df),
+get_predict(df),
+get_wtp(df),
+get_avg_use(df),
+get_fb(df),
+get_limit_last_week(df)
+)
+# Solve system of equation #1
+param %<>%
+solve_sys_eq_1 %>%
+as.list %>%
+list.merge(param)
+# Solve system of equations #2
+param %<>%
+solve_sys_eq_2(display_warning=display_warning) %>%
+as.list %>%
+list.merge(param)
+# Solve system of equations #3
+param %<>%
+solve_sys_eq_3 %>%
+as.list %>%
+list.merge(param)
+# Solve for individual effects
+tau_L_2_spec   <- find_tau_L2_spec(df)
+tau_tilde_spec <- find_tau_L3_spec(df)
+x_ss_i_data    <- calculate_x_ss_i_spec(df)
+param %<>%
+solve_effects_individual(x_ss_i_data= x_ss_i_data, tau_tilde_L=tau_tilde_spec, tau_L_2=tau_L_2_spec, w=df$w)%>%
+as.list %>%
+list.merge(param)
+tau_tilde_L=tau_tilde_spec
+tau_L_2=tau_L_2_spec
+w=df$w
+rho                               <- param[['rho']]
+lambda                            <- param[['lambda']]
+rho_res                               <- param[['rho_res']]
+lambda_res                            <- param[['lambda_res']]
+delta                             <- param[['delta']]
+alpha                             <- param[['alpha']]
+omega                             <- param[['omega']]
+omega_est                          <- param[['omega_est']]
+mispredict                        <- param[['mispredict']]
+d_L                               <- param[['d_L']]
+d_CL                              <- param[['d_CL']]
+eta                               <- param[['eta']]
+zeta                              <- param[['zeta']]
+naivete                           <- param[['naivete']]
+gamma_L_effect                    <- param[['gamma_L_effect']]
+gamma_tilde_L_effect              <- param[['gamma_tilde_L_effect']]
+gamma_tilde_L_effect_omega        <- param[['gamma_tilde_L_effect_omega']]
+gamma_L_effect_omega              <- param[['gamma_L_effect_omega']]
+gamma_L_effect_multiple           <- param[['gamma_L_effect_multiple']]
+gamma_tilde_L_effect_multiple     <- param[['gamma_tilde_L_effect_multiple']]
+gamma_L                           <- param[['gamma_L']]
+gamma_tilde_L                     <- param[['gamma_tilde_L']]
+gamma_tilde_L_omega               <- param[['gamma_tilde_L_omega']]
+gamma_L_omega                     <- param[['gamma_L_omega']]
+gamma_tilde_L_multiple            <- param[['gamma_tilde_L_multiple']]
+gamma_L_multiple                  <- param[['gamma_L_multiple']]
+gamma_B                           <- param[['gamma_B']]
+gamma_tilde_B                     <- param[['gamma_tilde_B']]
+gamma_tilde_B_multiple            <- param[['gamma_tilde_B_multiple']]
+gamma_B_multiple                  <- param[['gamma_B_multiple']]
+eta_res                               <- param[['eta_res']]
+zeta_res                              <- param[['zeta_res']]
+naivete_res                           <- param[['naivete_res']]
+gamma_L_effect_res                    <- param[['gamma_L_effect_res']]
+gamma_tilde_L_effect_res              <- param[['gamma_tilde_L_effect_res']]
+gamma_tilde_L_effect_omega_res        <- param[['gamma_tilde_L_effect_omega_res']]
+gamma_L_effect_omega_res              <- param[['gamma_L_effect_omega_res']]
+gamma_tilde_L_effect_multiple_res     <- param[['gamma_tilde_L_effect_multiple_res']]
+gamma_L_res                           <- param[['gamma_L_res']]
+gamma_L_omega_res                     <- param[['gamma_L_omega_res']]
+gamma_L_multiple_res                   <- param[['gamma_L_multiple_res']]
+gamma_B_res                            <- param[['gamma_B_res']]
+gamma_B_multiple_res                   <- param[['gamma_B_multiple_res']]
+tau_L_2_signed                    <- param[['tau_L_2']]*-1
+# Gamma-spec
+num                  <- eta*tau_L_2/omega - (1-alpha)*delta*rho*(((eta-zeta)*tau_tilde_L/omega+zeta*rho*tau_L_2/omega) + (1+lambda)*mispredict*(-eta+(1-alpha)*delta*rho^2*((eta-zeta)*lambda+zeta)))
+denom                <- 1 - (1-alpha)*delta*rho*(1+lambda)
+num_omega            <- eta*tau_L_2/omega_est - (1-alpha)*delta*rho*(((eta-zeta)*tau_tilde_L/omega_est+zeta*rho*tau_L_2/omega) + (1+lambda)*mispredict*(-eta+(1-alpha)*delta*rho^2*((eta-zeta)*lambda+zeta)))
+gamma_spec                <- num/denom
+gamma_spec_omega          <- num_omega/denom
+gamma_tilde_spec          <- gamma_spec - naivete
+gamma_tilde_spec_omega    <- gamma_spec_omega - naivete
+intercept_spec                <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec,           gamma_spec,      alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_L_effect        <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho, lambda, mispredict, eta, zeta)
+intercept_het_B               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_B,              gamma_B,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L               <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L,              gamma_L,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_spec_omega          <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_spec_omega,     gamma_spec_omega,      alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_omega  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect_omega, gamma_L_effect_omega,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_omega         <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_omega,        gamma_L_omega,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_multiple  <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_effect_multiple, gamma_L_effect_multiple,  alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_B_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_B_multiple,        gamma_B_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_multiple         <-  calculate_intercept_spec(x_ss_i_data, param, gamma_tilde_L_multiple,        gamma_L_multiple,         alpha, rho,  lambda, mispredict, eta, zeta)
+intercept_het_L_effect_eta_high  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,  eta, zeta, eta_scale=1.1)
+intercept_het_L_effect_eta_low  <-  calculate_intercept_spec(x_ss_i_data, param,    gamma_tilde_L_effect,       gamma_L_effect,  alpha, rho,  lambda, mispredict,   eta, zeta, eta_scale=0.9)
+x_ss_spec                        <- calculate_steady_state(param,                gamma_tilde_spec, gamma_spec,  alpha, rho,  lambda, mispredict, eta, zeta, intercept_spec)
+x_ss_spec                        <- calculate_steady_state(param,                gamma_tilde_spec, gamma_spec,  alpha, rho,  lambda, mispredict, eta, zeta, intercept_spec)
+calculate_steady_state <- function(param, gamma_tilde, gamma, alpha, rho, lambda, mispredict, eta, zeta, intercept=NA, eta_scale=1){
+# Define
+eta           <- eta * eta_scale
+delta         <- param[['delta']]
+p_B           <- param[['p_B']]
+# Calculate
+p <- 0
+term_pre <- (1 - (1-alpha)*delta*rho)
+term1 <- intercept - p*term_pre
+term2 <- (1-alpha)*delta*rho
+term3 <- (eta - zeta) * mispredict + gamma_tilde*(1+lambda)
+num <- term1 - term2*term3 + gamma
+terma <- term_pre*(-eta - zeta * (rho / (1 - rho)))
+termb <- (1-alpha)*delta*rho*zeta
+denom <- terma + termb
+print(paste0("denom: ", denom))
+x_ss_calc <- num /denom
+return(x_ss_calc)
+}
+x_ss_spec                        <- calculate_steady_state(param,                gamma_tilde_spec, gamma_spec,  alpha, rho,  lambda, mispredict, eta, zeta, intercept_spec)
+x_ss_zero_un                     <- calculate_steady_state(param,                     0,              0,        alpha, rho,  lambda,  0,          eta, zeta, intercept_spec)
+x_ss_zero                        <- ifelse(x_ss_zero_un<0, 0, x_ss_zero_un)
+delta_x                          <- x_ss_spec - x_ss_zero
+x_ss_spec_w                   <- weighted.mean(x_ss_spec, w, na.rm=T)

17/replication_package/code/analysis/structural/README.md

@@ -0,0 +1,6 @@
+# README
+
+This module estimates parameters and generates plots for our structural model.
+
+`/code/` contains the below file: 
+* StructuralModel.R  

17/replication_package/code/analysis/structural/code/StructuralModel.R

@@ -0,0 +1,295 @@
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Setup
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+# Import plotting functions and constants from lib file
+source('input/lib/r/ModelFunctions.R')
+
+# Import data
+df <- import_data()
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Nice scalars
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+nice_scalars <- function(param){
+
+  limiteffectlastweeknice <- signif(param$limit_effect_last_week, digits=2) * -1
+  limiteffect         <- signif(param$tau_L, digits=2) * -1
+  mispredictnice      <- signif(param$mispredict, digits=2)
+
+  tautildenice        <- signif(param$tau_tilde_B, digits=2)* -1
+
+
+ # pctirrationaltwo    <- param$rho_tilde/param$rho
+  #pctirrationaltwo    <- signif(pctirrationaltwo, digits=2) 
+
+  mispredictpct <- param$mispredict/param$x_ss
+  mispredictpct <- signif(mispredictpct, digits=2)*100
+
+  pctreductiontemptation <- param$delta_x_temptation/ param$x_ss
+  pctreductiontemptationres <- param$delta_x_temptation_res/ param$x_ss
+
+  pctreductiontemptation <- signif(pctreductiontemptation, digits=2)*100
+  pctreductiontemptationres <- signif(pctreductiontemptationres, digits=2)*100
+
+
+  dLpercent <- param$d_L/100
+  dLpercent <- signif(dLpercent, digits=3) 
+
+  dCLpercent <- param$d_CL/100
+  dCLpercent <- signif(dCLpercent, digits=2) 
+
+  taubtwonice         <- signif(param$tau_B_2, digits=2) * -1
+  taubtwofullnice      <- signif(param$tau_B_2_full , digits=2) * -1
+  taubthreenice        <- signif(param$tau_B_3, digits=2) * -1
+  taubfournice         <- signif(param$tau_B_4, digits=2) * -1
+  taubfivenice         <- signif(param$tau_B_5, digits=2) * -1
+
+  gammaLeffectnice    <- signif(param$gamma_L_effect, digits=1)
+  gammaLnice          <- signif(param$gamma_L, digits=2)
+  gammaBnice          <- signif(param$gamma_B, digits=2)
+
+  naivetenice         <- signif(param$naivete, digits=2)
+  gammaLeffectresnice <- signif(param$gamma_L_effect_res, digits=1)
+  gammaLresnice       <- signif(param$gamma_L_res, digits=2)
+  gammaBresnice       <- signif(param$gamma_B_res, digits=2)
+
+  naiveteresnice      <- signif(param$naivete_res, digits=2)
+
+  attritionratenice   <- signif(param$attritionrate, digits=2)*100
+
+
+  dLnice         <- signif(param$d_L, digits=2)*-1
+  dCLnice        <- signif(param$d_CL, digits=2)*-1
+
+  underestimatetemp <- format(round(param$underestimatetemp,3), digits=2)
+
+  tautildeBtwothreenice <- signif(param$tau_tilde_B_3_2, digits=2)*-1
+
+  MPLStwonice              <- signif(param$MPL_S2, digits=2)*-1
+
+  tauLtwosigned <- signif(param$tau_L_2)*-1
+
+                   
+
+
+  #Have hourly variables
+  gammaBnicehour             <- gammaBnice*60
+  gammaLnicehour             <- gammaLnice*60
+  gammaLeffectnicehour       <- gammaLeffectnice*60
+  naivetenicehour            <- naivetenice*60
+  gammaBresnicehour          <- gammaBresnice*60
+  gammaLresnicehour          <- gammaLresnice*60
+  gammaLeffectresnicehour    <- gammaLeffectresnice*60
+  naiveteresnicehour         <- naiveteresnice*60
+  taubtwohour                <- taubtwonice*60
+
+
+
+  # Return
+  solution <- list(
+    mispredictnice          = mispredictnice,
+    tautildenice            = tautildenice,
+    taubtwonice             = taubtwonice,
+    gammaLeffectnice        = gammaLeffectnice,
+    gammaLnice              = gammaLnice,
+    gammaBnice              = gammaBnice,
+    naivetenice             = naivetenice,
+    gammaLeffectnicehour    = gammaLeffectnicehour,
+    gammaLnicehour          = gammaLnicehour,
+    gammaBnicehour          = gammaBnicehour,
+    naivetenicehour         = naivetenicehour,
+    taubtwohour             = taubtwohour,
+    gammaLeffectresnice     = gammaLeffectresnice,
+    gammaLresnice           = gammaLresnice,
+    gammaBresnice           = gammaBresnice,
+    naiveteresnice          = naiveteresnice,
+    gammaLeffectresnicehour = gammaLeffectresnicehour,
+    gammaLresnicehour       = gammaLresnicehour,
+    gammaBresnicehour       = gammaBresnicehour,
+    naiveteresnicehour      = naiveteresnicehour,
+    dLnice                  = dLnice, 
+    dCLnice                 = dCLnice,
+    dLpercent               = dLpercent,
+    dCLpercent              = dCLpercent, 
+    underestimatetemp       = underestimatetemp,
+    tautildeBtwothreenice   = tautildeBtwothreenice,
+    limiteffect             = limiteffect,
+    attritionratenice       = attritionratenice,
+    taubthreenice           = taubthreenice,
+    taubfournice            = taubfournice,
+    taubfivenice            = taubfivenice,
+    pctreductiontemptation  = pctreductiontemptation,
+    pctreductiontemptationres = pctreductiontemptationres,
+    MPLStwonice             = MPLStwonice,
+    mispredictpct           = mispredictpct, 
+    taubtwofullnice         = taubtwofullnice,
+    tauLtwosigned           = tauLtwosigned,
+    limiteffectlastweeknice = limiteffectlastweeknice)
+ 
+  return(solution)
+}
+
+
+
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Full model, taub2=full
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+# Define constants
+param <- param_initial
+
+# Estimate model
+param_full <- estimate_model(df, param, full=T, display_warning=F)
+
+# Add some auto-import figures
+param_additional_full_taub2 <-
+  param_full %>%
+  as.list %>%
+  list.merge(param_full)
+
+save_tex(param_additional_full_taub2,  filename="structural_fulltaub2", suffix="fulltaubtwo")
+
+
+df$w <- 1
+
+results <- vector(mode = "list", length = size)
+
+results <- run_boot_procedure(run_boot_iter_full)
+
+# Get bootstrap distribution
+bottom <- lapply(results, find_bottom)
+top    <- lapply(results, find_top)
+
+save_boot_tex_percentile(bottom, top,
+                         suffix="bootfulltaubtwo",
+                         filename="structural_boot_fulltaubtwo")
+
+
+
+
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Full model, taub2 half period
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+# Define constants
+param <- param_initial
+
+# Estimate model
+param_full <- estimate_model(df, param)
+
+print(param_full$eta)
+print(param_full$zeta)
+check_steady_state(param_full)
+
+
+
+
+# Add some auto-import figures
+param_additional <-
+  param_full %>%
+  as.list %>%
+  list.merge(param_full)
+
+save_tex(param_additional,     filename="structural")
+
+# Add some auto-import figures
+param_additional_two <-
+  param_full %>%
+  as.list %>%
+  list.merge(param_full)
+
+
+save_tex2(param_additional_two,     filename="structural_two", suffix="twodigits")
+
+
+param_additional_nice <-
+  param_full %>%
+  nice_scalars %>%
+  as.list %>%
+  list.merge(param_full)
+
+save_tex_nice(param_additional_nice,     filename="structural_nice", suffix="nice")
+
+
+
+
+
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Balanced model
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+
+# Add weights
+df %<>% balance_data(magnitude=3)
+
+# Define constants
+param <- param_initial
+# Estimate model
+param_balanced <- estimate_model(df, param, winsorize=T)
+
+
+# # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# # Bootstrap model no perceived habit formation 
+# # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Revert to unbalanced
+df$w <- 1
+
+results <- vector(mode = "list", length = size)
+
+
+results <- run_boot_procedure(run_boot_iter)
+
+# Get bootstrap distribution
+bottom <- lapply(results, find_bottom)
+top    <- lapply(results, find_top)
+
+plot_time_effects(param_full, bottom, top, filename="structural_time_effects_plot")
+
+save_boot_tex_percentile(bottom, top,
+                         suffix="boot",
+                         filename="structural_boot")
+
+plot_decomposition_boot(param_full, bottom, top,
+                        filename="structural_decomposition_plot_boot")
+
+plot_decomposition_boot_unique(param_full, bottom, top,
+                        filename="structural_decomposition_plot_boot_restricted")
+
+plot_decomposition_boot_etas(param_full, bottom, top,
+                        filename="structural_decomposition_plot_boot_restricted_etas")
+
+plot_time_effects_both_est(param_full, bottom, top,  filename="time_effects_both_est")
+
+
+
+
+
+
+# # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# # Bootstrap balanced model no perceived 
+# # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+results_bal <- vector(mode = "list", length = size)
+
+
+#CHANGE MAGNITUDE OF WEIGHTS ONCE ITS SORTED!
+results_bal <- run_boot_procedure(run_boot_iter_bal)
+
+
+# Get bootstrap distribution
+median_bal <- lapply(results_bal, median, na.rm = T)
+sdevs_bal  <- lapply(results_bal, sd, na.rm = T)
+bottom_bal <- lapply(results_bal, find_bottom)
+top_bal    <- lapply(results_bal, find_top)
+
+save_tex(param_balanced,     filename="balanced_median", suffix="balancedmedian")
+
+#For restricted model 
+
+plot_time_effects_bal(param_full, param_balanced, bottom, top, bottom_bal, top_bal, filename="time_effects_balanced") 
+plot_time_effects_both(param_full, param_balanced, bottom, top, bottom_bal, top_bal, filename="time_effects_both")
+
+
+save_boot_tex_percentile(bottom_bal, top_bal,
+                         suffix="balanced",
+                         filename="balanced_boot")

17/replication_package/code/analysis/structural/input.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1aacb5e3a47846afaf251dbe069f5cee136e1255fe4ec43721a033fafc1d837d
+size 812

17/replication_package/code/analysis/structural/make.py

@@ -0,0 +1,67 @@
+###################
+### ENVIRONMENT ###
+###################
+import git
+import imp
+import os
+
+### SET DEFAULT PATHS
+ROOT = '../..'
+
+PATHS = {
+    'root'             : ROOT,
+    'lib'              : os.path.join(ROOT, 'lib'),
+    'config'           : os.path.join(ROOT, 'config.yaml'),
+    'config_user'      : os.path.join(ROOT, 'config_user.yaml'),
+    'input_dir'        : 'input',
+    'external_dir'     : 'external',
+    'output_dir'       : 'output',
+    'output_local_dir' : 'output_local',
+    'makelog'          : 'log/make.log',
+    'output_statslog'  : 'log/output_stats.log',
+    'source_maplog'    : 'log/source_map.log',
+    'source_statslog'  : 'log/source_stats.log',
+}
+
+### LOAD GSLAB MAKE
+f, path, desc = imp.find_module('gslab_make', [PATHS['lib']])
+gs = imp.load_module('gslab_make', f, path, desc)
+
+### LOAD CONFIG USER
+PATHS = gs.update_paths(PATHS)
+gs.update_executables(PATHS)
+
+############
+### MAKE ###
+############
+
+### START MAKE
+gs.remove_dir(['input', 'external'])
+gs.clear_dir(['output', 'log', 'temp'])
+gs.start_makelog(PATHS)
+
+### GET INPUT FILES
+inputs = gs.link_inputs(PATHS, ['input.txt'])
+# gs.write_source_logs(PATHS, inputs + externals)
+# gs.get_modified_sources(PATHS, inputs + externals)
+
+### RUN SCRIPTS
+"""
+Critical
+--------
+Many of the Stata analysis scripts recode variables using
+the `recode` command. Double-check all `recode` commands
+to confirm recoding is correct, especially when reusing
+code for a different experiment version.
+"""
+
+gs.run_r(PATHS,     program = 'code/StructuralModel.R')
+
+### LOG OUTPUTS
+gs.log_files_in_output(PATHS)
+
+### CHECK FILE SIZES
+#gs.check_module_size(PATHS)
+
+### END MAKE
+gs.end_makelog(PATHS)

17/replication_package/code/analysis/treatment_effects/README.md

@@ -0,0 +1,22 @@
+# README
+
+This module produces model-free estimates of treatment effects.
+
+`/code/` contains the below files : 
+* Beliefs.do (compares actual treatment effect with predicted treatment effect)
+
+* CommitmentResponse.do (plots how treatment effect differs by SMS addiction scale and other survey indicators)
+
+* FDRTable.do (estimates how treatment effect differs by SMS addiction scale and other indicators, adjusted for false-discovery rate. Also plots some descriptive statistics)
+
+* HabitFormation.do (compares actual and predicted usage)
+
+* Heterogeneity.do (plots heterogeneous treatment effects)
+
+* HeterogeneityInstrumental.do  (plots heterogeneous treatment effects)
+
+* ModelHeterogeneity.R (generates other heterogeneity plots, some temptation plots)
+
+* SurveyValidation.do (plots effect of rewarding accurate usage prediction on usage prediction accuracy)
+
+The script `ModelHeterogeneity.R` requires the dataset `AnalysisUser.dta` when calling the function `get_opt()`. This function computes the number of users who opted out of the limit functionality. Since this dataset contains PII, it has been omitted from this replication package. As such, the call to `get_opt()` (l.1396) has been commented out so it does not prevent the user from smoothly running this module.

17/replication_package/code/analysis/treatment_effects/code/Beliefs.do

@@ -0,0 +1,359 @@
+// Naivete about past and future usage
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+end
+
+program define_plot_settings
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5) ///
+        
+    global CISPIKE_HORIZONTAL_GRAPHOPTS ///
+        xlabel(#6) ///
+        xsize(6.5) ysize(8.5) 
+        
+    global CISPIKE_STACKED_GRAPHOPTS ///
+        xcommon row(2) ///
+        graphregion(color(white)) ///
+        xsize(6.5) ysize(8.5)
+        
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon black navy gray) ///
+        cicolor(maroon black navy gray) ///
+        spike(msymbol(O)||msymbol(S)||msymbol(D)||msymbol(T))
+
+    global COEFPLOT_VERTICAL_SETTINGS ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) ///
+        yline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) ///
+        legend(region(lcolor(white))) ///
+        xsize(6.5) ysize(4.5) ///
+        ytitle("Treatment effect (minutes/day)" " ") 
+
+    global COLOR_MAROON ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+        
+    global COLOR_GRAY ///
+        mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+
+    global COLOR_BLACK ///
+        mcolor(black) ciopts(recast(rcap) lcolor(black))
+
+    global COLOR_NAVY ///
+        mcolor(navy) ciopts(recast(rcap) lcolor(navy))
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+    
+	plot_naivete_all
+    plot_naivete_all, sixty
+    plot_naivete_all, hundred
+
+    reg_bonus
+    reg_bonus_S2
+    reg_bonus_new
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program plot_naivete_all
+    syntax, [sixty hundred]
+
+    local suffix ""
+    local winsorization "W0"
+    if ("`sixty'" == "sixty"){
+        local suffix "_W"
+        local winsorization "W60"
+    }
+    if ("`hundred'" == "hundred"){
+        local suffix "_W100"
+        local winsorization "W100"
+    }
+
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    rename PD_*_UsageFITSBY UsageActual_*
+
+    rename S2_PredictUseNext_1`suffix' UsagePredicted0_P2
+
+    rename S3_PredictUseNext_1`suffix' UsagePredicted0_P3
+    rename S2_PredictUseNext_2`suffix' UsagePredicted1_P3
+
+    rename S4_PredictUseNext_1`suffix' UsagePredicted0_P4    
+    rename S3_PredictUseNext_2`suffix' UsagePredicted1_P4    
+    rename S2_PredictUseNext_3`suffix' UsagePredicted2_P4    
+
+    rename S4_PredictUseNext_2`suffix' UsagePredicted1_P5    
+    rename S3_PredictUseNext_3`suffix' UsagePredicted2_P5   
+
+    keep UserID S3_Bonus S2_LimitType UsagePredicted* UsageActual*
+    keep UserID S3_Bonus S2_LimitType *_P2 *_P3 *_P4 *_P5
+    reshape long Usage, i(UserID S3_Bonus S2_LimitType) j(j) string
+    
+    split j, p(_)
+    rename (j1 j2) (measure time)
+    
+    * Recode data
+    encode time, generate(time_encode) 
+    encode measure, generate(measure_encode) 
+    
+    recode time_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (3 = 3 "Period 4") ///
+        (4 = 4 "Period 5"), ///
+        gen(time_recode)
+        
+    recode measure_encode ///
+        (1 = 1 "Actual") ///
+        (2 = 2 "Survey t prediction") ///
+        (3 = 3 "Survey t-1 prediction") ///
+        (4 = 4 "Survey t-2 prediction"), ///
+        gen(measure_recode)
+    
+    * Plot data
+    cispike Usage if S3_Bonus == 0 & S2_LimitType == 0, ///
+        over1(measure_recode) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Usage (minutes/day)" " "))
+        
+    graph export "output/cispike_naivete_BcontrolxLcontrol_`winsorization'.pdf", replace
+
+    cispike Usage if S3_Bonus == 1 & S2_LimitType == 0, ///
+        over1(measure_recode) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Usage (minutes/day)" " "))
+        
+    graph export "output/cispike_naivete_BtreatxLcontrol_`winsorization'.pdf", replace
+
+    cispike Usage if S3_Bonus == 0 & S2_LimitType > 0, ///
+        over1(measure_recode) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Usage (minutes/day)" " "))
+        
+    graph export "output/cispike_naivete_BcontrolxLtreat_`winsorization'.pdf", replace
+
+    cispike Usage if S3_Bonus == 1 & S2_LimitType > 0, ///
+        over1(measure_recode) over2(time_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Usage (minutes/day)" " "))
+        
+    graph export "output/cispike_naivete_BtreatxLtreat_`winsorization'.pdf", replace
+        
+    * Restore data
+    restore
+end
+
+program reg_bonus
+    est clear
+
+    preserve
+
+    gen S1_Usage_FITSBY = PD_P1_UsageFITSBY
+    gen S3_Usage_FITSBY = PD_P3_UsageFITSBY
+    gen S4_Usage_FITSBY = PD_P4_UsageFITSBY
+    gen S5_Usage_FITSBY = PD_P5_UsageFITSBY
+
+    gen S2_Predict_FITSBY = S2_PredictUseNext_1_W
+    gen S3_Predict_FITSBY = S3_PredictUseNext_1_W
+    gen S4_Predict_FITSBY = S3_PredictUseNext_2_W
+    gen S5_Predict_FITSBY = S3_PredictUseNext_3_W
+
+    * Run regressions
+    foreach survey in S3 S4 S5 {
+        local yvar `survey'_Usage_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    * Run regressions
+    foreach survey in S3 S4 S5 {
+        local yvar `survey'_Predict_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    gen S2_reduction = S2_PredictUseInitial_W * - (S2_PredictUseBonus / 100)
+
+    cap drop B_S3
+    gen B_S3 = 1
+    reg S2_reduction B_S3, noconstant
+    est store S2_reduction
+
+    * Plot regressions (by period)
+    coefplot (*Usage*,   label("Actual") $COLOR_MAROON msymbol(O)) ///
+             (S2_reduction,    label("Survey 2 MPL prediction") $COLOR_NAVY msymbol(S)) ///
+             (*Predict*, label("Survey 3 prediction") $COLOR_GRAY msymbol(D)), ///
+             keep(B_*) ///
+             vertical ///
+             $COEFPLOT_VERTICAL_SETTINGS ///
+             xlabel(1 "Period 3" 2 "Period 4" 3 "Period 5", ///
+             valuelabel angle(0))
+    
+    graph export "output/coef_belief_bonus_effect.pdf", replace
+
+    restore
+end
+
+program reg_bonus_new
+    est clear
+
+    preserve
+
+    gen S1_Usage_FITSBY = PD_P1_UsageFITSBY
+    gen S2_Usage_FITSBY = PD_P2_UsageFITSBY
+    gen S3_Usage_FITSBY = PD_P3_UsageFITSBY
+    gen S4_Usage_FITSBY = PD_P4_UsageFITSBY
+    gen S5_Usage_FITSBY = PD_P5_UsageFITSBY
+
+    gen S2_Predict_FITSBY = S2_PredictUseNext_1_W
+    gen S3_Predict_FITSBY = S3_PredictUseNext_1_W
+    gen S4_Predict_FITSBY = S3_PredictUseNext_2_W
+    gen S5_Predict_FITSBY = S3_PredictUseNext_3_W
+
+    * Run regressions
+    foreach survey in S2 S3 S4 S5 {
+        local yvar `survey'_Usage_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    * Run regressions
+    foreach survey in S3 S4 S5 {
+        local yvar `survey'_Predict_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    gen S2_reduction = S2_PredictUseInitial_W * - (S2_PredictUseBonus / 100)
+
+    cap drop B_S3
+    gen B_S3 = 1
+    reg S2_reduction B_S3, noconstant
+    est store S2_reduction
+
+    matrix C = J(3,1,.)
+    matrix rownames C = mean ll ul
+    matrix colnames C = B_S2
+
+    * TODO: make this reproducible
+    matrix C[1,1] = -16.11756 \ -19.64522 \ -12.62825
+    matrix list C
+    coefplot matrix(C), ci((2 3))
+
+    * Plot regressions (by period)
+    coefplot (matrix(C), ci((2 3)) label("Makes {&alpha} = 0") $COLOR_BLACK) ///
+             (S2_reduction,    label("Survey 2 MPL prediction") $COLOR_NAVY) ///
+             (*Usage*,   label("Actual") $COLOR_MAROON) /// 
+             (*Predict*, label("Survey 3 prediction") $COLOR_GRAY), ///
+             keep(B_*) ///
+             vertical ///
+             $COEFPLOT_VERTICAL_SETTINGS ///
+             xlabel(1 "Period 2" 2 "Period 3" 3 "Period 4" 4 "Period 5", ///
+             valuelabel angle(0))
+    
+    graph export "output/coef_belief_bonus_effect_new.pdf", replace
+
+    restore
+end
+
+program reg_bonus_S2
+
+    est clear
+
+    preserve
+
+    gen S1_Usage_FITSBY = PD_P1_UsageFITSBY
+    gen S2_Usage_FITSBY = PD_P2_UsageFITSBY
+    gen S3_Usage_FITSBY = PD_P3_UsageFITSBY
+
+    gen S2_Predict_FITSBY = S2_PredictUseNext_1_W
+    gen S3_Predict_FITSBY = S2_PredictUseNext_2_W
+
+    * Run regressions
+    foreach survey in S2 S3 {
+        local yvar `survey'_Usage_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    * Run regressions
+    foreach survey in S2 S3 {
+        local yvar `survey'_Predict_FITSBY
+        local baseline S1_Usage_FITSBY
+        
+        gen_treatment, suffix(_`survey') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`survey') simple
+        est store `yvar'
+    }
+
+    gen S2_reduction = S2_PredictUseInitial_W * - (S2_PredictUseBonus / 100)
+
+    cap drop B_S2
+    gen B_S2 = 1
+    reg S2_reduction B_S2, noconstant
+    est store S2_reduction
+
+
+    * Plot regressions (by period)
+    coefplot (*Usage*,   label("Actual") $COLOR_MAROON) ///
+             (*Predict*, label("Predicted") $COLOR_GRAY) ///
+             (S2_reduction, label("Bonus Predicted") $COLOR_NAVY), ///
+             keep(B_*) ///
+             vertical ///
+             $COEFPLOT_VERTICAL_SETTINGS ///
+             xlabel(1 "Period 2" 2 "Period 3", ///
+             valuelabel angle(0))
+    
+    graph export "output/coef_belief_bonus_survey2.pdf", replace
+
+    restore
+end
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/code/CommitmentResponse.do

@@ -0,0 +1,1404 @@
+// Response to commitment, moderated by demand for flexibility
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+    yaml global STRATA = YAML.metadata.strata
+end
+
+program define_plot_settings       
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon black gray) ///
+        cicolor(maroon black gray) 
+       
+	global CISPIKE_DOUBLE_SETTINGS ///
+	    spike(yaxis(1) || yaxis(2)) ///
+		ci(yaxis(1) || yaxis(2)) ///
+		spikecolor(maroon gray) ///
+		cicolor(maroon gray)
+		
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5) ///
+		legend(cols(4))
+		
+    global COLOR_MAROON ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+
+    global COLOR_MAROON_LIGHT ///
+        mcolor(maroon*0.9) ciopts(recast(rcap) lcolor(maroon*0.9))
+
+    global COLOR_MAROON_DARK ///
+        mcolor(maroon*1.1) ciopts(recast(rcap) lcolor(maroon*1.1))
+
+
+ global COLOR_GRAY_LIGHT ///
+        mcolor(gray*0.9) ciopts(recast(rcap) lcolor(gray*0.9))
+
+    global COLOR_GRAY_DARK ///
+        mcolor(gray*1.1) ciopts(recast(rcap) lcolor(gray*1.1))
+
+
+    global COLOR_BLUE ///
+        mcolor(edkblue) ciopts(recast(rcap) lcolor(edkblue))
+    
+	global COLOR_BLACK ///
+        mcolor(black) ciopts(recast(rcap) lcolor(black))
+		
+    global COLOR_GRAY ///
+        mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+	
+	global COLOR_NAVY ///
+        mcolor(navy) ciopts(recast(rcap) lcolor(navy))
+
+    global COLOR_NAVY_LIGHT ///
+        mcolor(navy*0.5) ciopts(recast(rcap) lcolor(navy*0.5))
+		
+    global COEFPLOT_SETTINGS_MINUTES ///
+        vertical ///
+		yline(0, lwidth(thin) lcolor(black)) ///
+		bgcolor(white) graphregion(color(white)) ///
+		legend(cols(4) region(lcolor(white))) ///
+		xsize(6.5) ysize(4.5) ///
+		ytitle("Treatment effect (minutes/day)" " ")
+
+    global COEFPLOT_SETTINGS_THIN ///
+        vertical ///
+        yline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) ///
+        legend(cols(4) region(lcolor(white))) ///
+        xsize(4.5) ysize(4.5) ///
+        ytitle("Treatment effect (minutes/day)" " ")
+
+ 	global COEFPLOT_SETTINGS_STD ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(4.5) ///
+		xtitle(" " "Treatment effect (standard deviations per hour/day of use)") 
+	
+    global COEFPLOT_SETTINGS_ITT ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(4.5) ///
+        xtitle(" " "Treatment effect (standard deviations)") 
+
+    global COEFPLOT_LABELS_LIMIT ///
+        coeflabels(L_1 = `"Snooze 0"' ///
+				   L_2 = `"Snooze 2"' ///
+				   L_3 = `"Snooze 5"' ///
+				   L_4 = `"Snooze 20"' ///
+				   L_5 = `"No snooze"' ///
+                   L   = `"Limit"' ///
+				   B   = `"Bonus"') 
+		
+	global COEFPLOT_STACKED_GRAPHOPTS ///
+		ycommon row(2) ///
+		graphregion(color(white)) ///
+        xsize(6.5) ysize(8.5)
+
+    global COEFPLOT_ADDICTION_SETTINGS ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(7) ysize(6.5)
+        
+    global ADDICTION_LABELS ///
+        xlabel(, labsize(small)) ///
+        xtitle(, size(small)) ///
+        ylabel(, labsize(vsmall)) ///
+        ytitle(, size(small)) ///
+        legend(size(small))
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+  
+	reg_usage
+    reg_usage, fitsby
+	reg_usage_simple
+    reg_usage_simple, fitsby
+    reg_usage_simple_balanced
+    reg_usage_simple_balanced, fitsby
+	plot_snooze
+	plot_snooze, fitsby 
+    plot_snooze, minutes
+    plot_snooze, fitsby minutes
+	plot_snooze_by_limit
+    plot_snooze_by_limit, fitsby 
+	plot_snooze_by_limit, minutes
+    plot_snooze_by_limit, fitsby minutes
+    plot_snooze_both
+    plot_snooze_both, fitsby 
+	plot_snooze_both_by_limit
+    plot_snooze_both_by_limit, fitsby
+	plot_phone_use_change
+    plot_phone_use_change_simple
+    reg_usage_interaction
+    reg_usage_interaction, fitsby
+	reg_self_control
+    reg_self_control_null
+	reg_iv_self_control
+    reg_usage_simple_weekly
+    reg_usage_simple_weekly, fitsby
+    reg_usage_simple_daily_p12
+    reg_usage_simple_daily_p12, fitsby
+    reg_addiction_simple
+    reg_sms_addiction_simple
+    reg_swb_simple
+    reg_swb_icw_simple
+    reg_sms_addiction_simple_weekly
+    reg_substitution
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program reg_usage    
+    syntax, [fitsby]
+    
+    est clear
+    
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+
+    * Run regressions
+    foreach yvar in PD_P2_Usage`fitsby' ///
+	                PD_P3_Usage`fitsby' ///
+					PD_P4_Usage`fitsby' ///
+					PD_P5_Usage`fitsby' ///
+                    PD_P432_Usage`fitsby' ///
+                    PD_P5432_Usage`fitsby' {
+        local baseline PD_P1_Usage`fitsby'
+        
+        gen_treatment
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline')
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_MAROON) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_BLACK) ///
+			 (PD_P4_Usage`fitsby', label("Period 4") $COLOR_NAVY) ///
+			 (PD_P5_Usage`fitsby', label("Period 5") $COLOR_GRAY) , ///
+             keep(L_*) order(L_1 L_2 L_3 L_4 L_5) ///
+			 $COEFPLOT_SETTINGS_MINUTES ///
+			 $COEFPLOT_LABELS_LIMIT
+    
+	graph export "output/coef_usage`suffix'.pdf", replace
+	
+	* Plot regressions (all period)
+	coefplot (PD_P5432_Usage`fitsby', label("Period 2 to 5") $COLOR_MAROON), ///
+             keep(L_*) order(L_1 L_2 L_3 L_4 L_5) ///
+			 $COEFPLOT_SETTINGS_MINUTES ///
+			 $COEFPLOT_LABELS_LIMIT ///
+			 legend(off)
+			 
+    graph export "output/coef_usage_combined`suffix'.pdf", replace
+end
+
+program reg_usage_simple   
+    syntax, [fitsby]
+    
+    est clear
+    
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Run regressions
+    foreach yvar in PD_P2_Usage`fitsby' ///
+	                PD_P3_Usage`fitsby' ///
+					PD_P4_Usage`fitsby' ///
+					PD_P5_Usage`fitsby' ///
+                    PD_P432_Usage`fitsby' ///
+                    PD_P5432_Usage`fitsby' {
+        local baseline PD_P1_Usage`fitsby'
+        
+        gen_treatment, simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_MAROON msymbol(O)) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_BLACK msymbol(S)) ///
+			 (PD_P4_Usage`fitsby', label("Period 4") $COLOR_NAVY msymbol(D)) ///
+			 (PD_P5_Usage`fitsby', label("Period 5") $COLOR_GRAY msymbol(T)), ///
+             keep(B L) order(B L) ///
+			 $COEFPLOT_SETTINGS_MINUTES ///
+			 $COEFPLOT_LABELS_LIMIT
+    
+	graph export "output/coef_usage_simple`suffix'.pdf", replace
+
+    * Plot regressions (by period)
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_MAROON msymbol(O)) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_MAROON msymbol(S)) ///
+             (PD_P4_Usage`fitsby', label("Period 4") $COLOR_MAROON msymbol(D)) ///
+             (PD_P5_Usage`fitsby', label("Period 5") $COLOR_MAROON msymbol(T)), ///
+             keep(B) order(B) ///
+             $COEFPLOT_SETTINGS_THIN ///
+             $COEFPLOT_LABELS_LIMIT
+    
+    graph export "output/coef_usage_simple`suffix'_bonus_only.pdf", replace
+
+    * Plot regressions (by period)
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_GRAY msymbol(O)) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_GRAY msymbol(S)) ///
+             (PD_P4_Usage`fitsby', label("Period 4") $COLOR_GRAY msymbol(D)) ///
+             (PD_P5_Usage`fitsby', label("Period 5") $COLOR_GRAY msymbol(T)), ///
+             keep(L) order(L) ///
+             ysc(r(-60 0)) ///
+             ylabel(-60(20)0) ///
+             $COEFPLOT_SETTINGS_THIN ///
+             $COEFPLOT_LABELS_LIMIT // 
+
+    graph export "output/coef_usage_simple`suffix'_limit_only.pdf", replace
+
+	
+	* Plot regressions (all period)
+	coefplot (PD_P5432_Usage`fitsby', label("Period 2 to 5") $COLOR_MAROON), ///
+             keep(B L) order(B L) ///
+			 $COEFPLOT_SETTINGS_MINUTES ///
+			 $COEFPLOT_LABELS_LIMIT ///
+			 legend(off)
+			 
+    graph export "output/coef_usage_combined_simple`suffix'.pdf", replace
+end
+
+program reg_usage_simple_balanced   
+    syntax, [fitsby]
+    
+    est clear
+    
+    preserve
+
+    local income  43.01
+    local college 0.3009
+    local male    0.4867
+    local white   0.73581
+    local age     47.6
+    
+    ebalance balance_income balance_college balance_male balance_white balance_age, ///
+             manualtargets(`income' `college' `male' `white' `age') ///
+             generate(weight)
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Run regressions
+    foreach yvar in PD_P2_Usage`fitsby' ///
+                    PD_P3_Usage`fitsby' ///
+                    PD_P4_Usage`fitsby' ///
+                    PD_P5_Usage`fitsby' ///
+                    PD_P432_Usage`fitsby' ///
+                    PD_P5432_Usage`fitsby' {
+        local baseline PD_P1_Usage`fitsby'
+        
+        gen_treatment, simple
+
+        reg `yvar' B L $STRATA `baseline' [w=weight], robust
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_MAROON) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_BLACK) ///
+             (PD_P4_Usage`fitsby', label("Period 4") $COLOR_NAVY) ///
+             (PD_P5_Usage`fitsby', label("Period 5") $COLOR_GRAY), ///
+             keep(B L) order(B L) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT
+    
+    graph export "output/coef_usage_simple_balanced`suffix'.pdf", replace
+
+    restore
+end
+program reg_substitution       
+    est clear
+        
+    gen_treatment, simple
+    reg_treatment, yvar(S4_Substitution) indep($STRATA) simple
+    est store S4_Substitution
+    
+    * Plot regressions (all period)
+    coefplot (S4_Substitution, $COLOR_MAROON), ///
+             keep(B L) order(B L) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT ///
+             legend(off)
+             
+    graph export "output/coef_self_reported_substitution.pdf", replace
+
+    gen_treatment, simple
+    reg_treatment, yvar(S4_Substitution_W) indep($STRATA) simple
+    est store S4_Substitution_W
+    
+    * Plot regressions (all period)
+    coefplot (S4_Substitution_W, $COLOR_MAROON), ///
+             keep(B L) order(B L) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT ///
+             legend(off)
+             
+    graph export "output/coef_self_reported_substitution_w.pdf", replace
+end
+
+
+program reg_usage_simple_weekly  
+    syntax, [fitsby]
+    
+    est clear
+    
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Run regressions
+    foreach yvar in PD_WeeklyUsage`fitsby'_4  ///
+                    PD_WeeklyUsage`fitsby'_5  ///
+                    PD_WeeklyUsage`fitsby'_6  ///
+                    PD_WeeklyUsage`fitsby'_7  ///
+                    PD_WeeklyUsage`fitsby'_8  ///
+                    PD_WeeklyUsage`fitsby'_9  ///
+                    PD_WeeklyUsage`fitsby'_10 ///
+                    PD_WeeklyUsage`fitsby'_11 ///
+                    PD_WeeklyUsage`fitsby'_12 ///
+                    PD_WeeklyUsage`fitsby'_13 ///
+                    PD_WeeklyUsage`fitsby'_14 ///
+                    PD_WeeklyUsage`fitsby'_15 {
+        local baseline PD_WeeklyUsage`fitsby'_3
+        
+        gen_treatment, simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (PD_WeeklyUsage`fitsby'_4 , label("Week 4")  $COLOR_MAROON msymbol(O)) ///
+             (PD_WeeklyUsage`fitsby'_5 , label("Week 5")  $COLOR_BLACK msymbol(S)) ///
+             (PD_WeeklyUsage`fitsby'_6 , label("Week 6")  $COLOR_GRAY  msymbol(D)) ///
+             (PD_WeeklyUsage`fitsby'_7 , label("Week 7")  $COLOR_MAROON msymbol(O)) ///
+             (PD_WeeklyUsage`fitsby'_8 , label("Week 8")  $COLOR_BLACK msymbol(S)) ///
+             (PD_WeeklyUsage`fitsby'_9 , label("Week 9")  $COLOR_GRAY  msymbol(D)) ///
+             (PD_WeeklyUsage`fitsby'_10, label("Week 10") $COLOR_MAROON msymbol(O)) ///
+             (PD_WeeklyUsage`fitsby'_11, label("Week 11") $COLOR_BLACK msymbol(S)) ///
+             (PD_WeeklyUsage`fitsby'_12, label("Week 12") $COLOR_GRAY  msymbol(D)) ///
+             (PD_WeeklyUsage`fitsby'_13, label("Week 13") $COLOR_MAROON msymbol(O)) ///
+             (PD_WeeklyUsage`fitsby'_14, label("Week 14") $COLOR_BLACK msymbol(S)) ///
+             (PD_WeeklyUsage`fitsby'_15, label("Week 15") $COLOR_GRAY  msymbol(D)), ///
+             keep(B L) order(B L) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT
+    
+    graph export "output/coef_usage_simple_weekly`suffix'.pdf", replace
+end
+
+program reg_usage_simple_daily_p12
+    syntax, [fitsby]
+    
+    est clear
+    
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Run regressions
+    foreach day of numlist 1/42 {
+        local yvar PD_DailyUsage`fitsby'_`day'
+        
+        gen_treatment, suffix(`day') simple
+        reg_treatment, yvar(`yvar') indep($STRATA) suffix(`day') simple
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (PD_DailyUsage`fitsby'_1,  label("Day 1")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_2,  label("Day 2")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_3,  label("Day 3")  $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_4,  label("Day 4")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_5,  label("Day 5")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_6,  label("Day 6")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_7,  label("Day 7")  $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_8,  label("Day 8")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_9,  label("Day 9")  $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_10, label("Day 10") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_11, label("Day 11") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_12, label("Day 12") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_13, label("Day 13") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_14, label("Day 14") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_15, label("Day 15") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_16, label("Day 16") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_17, label("Day 17") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_18, label("Day 18") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_19, label("Day 19") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_20, label("Day 20") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_21, label("Day 21") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_22, label("Day 22") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_23, label("Day 23") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_24, label("Day 24") $COLOR_NAVY   ) ///
+             (PD_DailyUsage`fitsby'_25, label("Day 25") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_26, label("Day 26") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_27, label("Day 27") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_28, label("Day 28") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_29, label("Day 29") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_30, label("Day 30") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_31, label("Day 31") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_32, label("Day 32") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_33, label("Day 33") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_34, label("Day 34") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_35, label("Day 35") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_36, label("Day 36") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_37, label("Day 37") $COLOR_NAVY) ///
+             (PD_DailyUsage`fitsby'_38, label("Day 38") $COLOR_NAVY ) ///
+             (PD_DailyUsage`fitsby'_39, label("Day 39") $COLOR_NAVY  ) ///
+             (PD_DailyUsage`fitsby'_40, label("Day 40") $COLOR_NAVY ) ///    
+             (PD_DailyUsage`fitsby'_41, label("Day 41") $COLOR_NAVY) ///    
+             (PD_DailyUsage`fitsby'_42, label("Day 42") $COLOR_NAVY), ///         
+             keep(B*) xline(22) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT legend(off) ///
+             xlabel(10 "Period 1" 22 "Survey 2" 34 "Period 2") ///
+
+    
+    graph export "output/coef_usage_simple_daily_p12`suffix'.pdf", replace
+end
+
+program reg_sms_addiction_simple_weekly 
+    syntax
+    
+    est clear
+    
+    * Run regressions
+    foreach week of numlist 4/9 {
+        local yvar Week`week'_SMSIndex 
+        local comparison_week = `week' - 3
+        if (`comparison_week' > 3){
+            local comparison_week = `week' - 6
+        }
+
+        local baseline Week`comparison_week'_SMSIndex
+        
+        gen_treatment, simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') simple
+        est store `yvar'
+    }
+
+    * Plot regressions (by period)
+    coefplot (Week4_SMSIndex , label("Week 4")  $COLOR_MAROON) ///
+             (Week5_SMSIndex , label("Week 5")  $COLOR_BLACK ) ///
+             (Week6_SMSIndex , label("Week 6")  $COLOR_GRAY  ) ///
+             (Week7_SMSIndex , label("Week 7")  $COLOR_MAROON) ///
+             (Week8_SMSIndex , label("Week 8")  $COLOR_BLACK ) ///
+             (Week9_SMSIndex , label("Week 9")  $COLOR_GRAY  ), ///
+             keep(B L) order(B L) ///
+             $COEFPLOT_SETTINGS_MINUTES ///
+             $COEFPLOT_LABELS_LIMIT
+    
+    graph export "output/coef_sms_addiction_simple_weekly.pdf", replace
+end
+
+program reg_addiction_simple   
+    syntax
+    
+    est clear
+    
+    * Run regressions for limit
+    foreach num of numlist 1/16 {
+        local baseline S1_Addiction_`num'
+
+        gen S43_Addiction_`num' = (S3_Addiction_`num' + S4_Addiction_`num') / 2
+        local yvar S43_Addiction_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    * Run regressions for bonus
+    foreach num of numlist 1/16 {
+        local baseline S1_Addiction_`num'
+
+        local yvar S4_Addiction_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    coefplot (S4_Addiction_*, keep(B_*) label("Bonus") mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) rename(B_S4_* = *)) ///
+             (S43_Addiction_*, keep(L_*) label("Limit") mcolor(gray) ciopts(recast(rcap) lcolor(gray)) rename(L_S43_* = *)), ///
+             $COEFPLOT_ADDICTION_SETTINGS ///
+             $ADDICTION_LABELS ///
+             yaxis(1) yscale(axis(1) range(0)) xlabel(-0.06(0.02)0.06, axis(1)) ///
+             ylabel(1  "Fear missing what happening online"  2  "Check social media/messages immediately after waking up"  ///
+                    3  "Use longer than intended"  4  "Tell yourself just a few more minutes"  ///
+                    5  "Use to distract from personal issues" 6  "Use to distract from anxiety/depression/etc."  ///
+                    7  "Use to relax to go to sleep"  8  "Try and fail to reduce use"  ///
+                    9  "Others are concerned about use"  10 "Feel anxious without phone" ///
+                    11 "Have difficulty putting down phone " 12 "Annoyed at interruption in use" ///
+                    13 "Use harms school/work performance" 14 "Lose sleep from use" ///
+                    15 "Prefer phone to human interaction" 16 "Procrastinate by using phone", ///
+                    valuelabel angle(0)) horizontal ///
+             ytitle("") xtitle("Treatment effect", axis(1)) 
+    
+    graph export "output/coef_addiction_simple.pdf", replace
+
+end
+
+
+program reg_sms_addiction_simple       
+    est clear
+    
+    preserve
+
+    * Run regressions for limit
+    foreach num of numlist 1/9 {
+        local baseline S1_AddictionText_`num'
+
+        gen S23_AddictionText_`num' = (S2_AddictionText_`num' + S3_AddictionText_`num') / 2
+        local yvar S23_AddictionText_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    * Run regressions for bonus
+    foreach num of numlist 1/9 {
+        local baseline S1_AddictionText_`num'
+
+        local yvar S3_AddictionText_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    coefplot (S3_AddictionText_*, keep(B_*) label("Bonus") mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) rename(B_S3_* = *)) ///
+             (S23_AddictionText_*, keep(L_*) label("Limit") mcolor(gray) ciopts(recast(rcap) lcolor(gray)) rename(L_S23_* = *)), ///
+             $COEFPLOT_ADDICTION_SETTINGS ///
+             $ADDICTION_LABELS ///
+             yaxis(1) yscale(axis(1) range(0)) xlabel(-0.2(0.05)0.2, axis(1)) ///
+             ylabel(1 "Use longer than intended"  2 "Use harms school/work performance" ///
+                    3 "Easy to control screen time x (-1)" 4 "Use mindlessly" ///
+                    5 "Use because felt down" 6 "Use kept from working on something needed" ///
+                    7 "Ideally used phone less" 8 "Lose sleep from use" ///
+                    9 "Check social media/messages immediately after waking up", ///
+                    valuelabel angle(0)) horizontal ///
+             ytitle("") xtitle("Treatment effect", axis(1)) 
+    
+    graph export "output/coef_sms_addiction_simple.pdf", replace
+
+    restore
+end
+
+
+program reg_swb_simple       
+    est clear
+    
+    preserve
+
+    gen S1_WellBeing_8 = (S1_WellBeing_1 + S1_WellBeing_2 + S1_WellBeing_3 + S1_WellBeing_4)/4
+    gen S1_WellBeing_9 = (S1_WellBeing_5 + S1_WellBeing_6 + S1_WellBeing_7)/3
+    gen S3_WellBeing_8 = (S3_WellBeing_1 + S3_WellBeing_2 + S3_WellBeing_3 + S3_WellBeing_4)/4
+    gen S3_WellBeing_9 = (S3_WellBeing_5 + S3_WellBeing_6 + S3_WellBeing_7)/3
+    gen S4_WellBeing_8 = (S4_WellBeing_1 + S4_WellBeing_2 + S4_WellBeing_3 + S4_WellBeing_4)/4
+    gen S4_WellBeing_9 = (S4_WellBeing_5 + S4_WellBeing_6 + S4_WellBeing_7)/3
+
+
+    * Run regressions for limit
+    foreach num of numlist 1/9 {
+        local baseline S1_WellBeing_`num'
+
+        gen S43_WellBeing_`num' = (S4_WellBeing_`num' + S3_WellBeing_`num') / 2
+        local yvar S43_WellBeing_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    * Run regressions for bonus
+    foreach num of numlist 1/9 {
+        local baseline S1_WellBeing_`num'
+
+        local yvar S4_WellBeing_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    coefplot (S4_WellBeing_*, keep(B_*) label("Bonus") mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) rename(B_S4_* = *)) ///
+             (S43_WellBeing_*, keep(L_*) label("Limit") mcolor(gray) ciopts(recast(rcap) lcolor(gray)) rename(L_S43_* = *)), ///
+             $COEFPLOT_ADDICTION_SETTINGS ///
+             $ADDICTION_LABELS ///
+             yaxis(1) yscale(axis(1) range(0)) xlabel(-0.09(0.03)0.09, axis(1)) ///
+             ylabel(1 "Was happy"  2 "Was satisfied with life" ///
+                    3 "Felt anxious x (-1)" 4 "Felt depressed x (-1)" ///
+                    5 "Could concentrate" 6 "Was easily distracted x (-1)" ///
+                    7 "Slept well" 8 "Happy <-> depressed index" ///
+                    9 "Concentrate <-> sleep index", ///
+                    valuelabel angle(0)) horizontal ///
+             ytitle("") xtitle("Treatment effect", axis(1)) 
+    
+    graph export "output/coef_swb_simple.pdf", replace
+
+    restore
+end
+
+program reg_swb_icw_simple      
+    est clear
+    
+    preserve
+
+    * Run regressions for limit
+    foreach num of numlist 1/7 {
+        local baseline S1_WellBeing_`num'
+
+        gen S43_WellBeing_`num' = (S4_WellBeing_`num' + S3_WellBeing_`num') / 2
+        local yvar S43_WellBeing_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    * Run regressions for bonus
+    foreach num of numlist 1/7 {
+        local baseline S1_WellBeing_`num'
+
+        local yvar S3_WellBeing_`num'  
+
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    } 
+
+    foreach idx in HSAD CDS {
+        local baseline S1_index_`idx'
+        gen S43_index_`idx' = (S3_index_`idx' + S4_index_`idx') / 2
+        local yvar S43_index_`idx'  
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+
+        local yvar S3_index_`idx'  
+        gen_treatment, suffix(_`yvar') simple
+        reg_treatment, yvar(`yvar') indep($STRATA `baseline') suffix(_`yvar') simple
+        est store `yvar'
+    }
+
+    coefplot (S3_WellBeing_* S3_index_*, keep(B_*) label("Bonus")  mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) rename(B_S3_* = *)) ///
+             (S43_WellBeing_* S43_index_*, keep(L_*) label("Limit") mcolor(gray)   ciopts(recast(rcap) lcolor(gray))   rename(L_S43_* = *)), ///
+             $COEFPLOT_ADDICTION_SETTINGS ///
+             $ADDICTION_LABELS ///
+             yscale(axis(1) range(0)) xlabel(-0.09(0.03)0.09, axis(1)) ///
+             horizontal ///
+             xtitle("Treatment effect", axis(1)) ///
+             group(*index*="", nolabels) ///
+             ylabel(1 "Was happy"  2 "Was satisfied with life" ///
+                    3 "Felt anxious x (-1)" 4 "Felt depressed x (-1)" ///
+                    5 "Could concentrate" 6 "Was easily distracted x (-1)" ///
+                    7 "Slept well" ///
+                    9 "Happy, satisfied, anxious, depressed index" ///
+                    10 "Concentrate, distracted, sleep index", ///
+                    valuelabel angle(0))
+
+    
+    graph export "output/coef_swb_icw_simple.pdf", replace
+
+    restore
+end
+
+program plot_snooze
+    syntax, [fitsby] [minutes]
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+	
+    * Determine snooze measure
+    if ("`minutes'" == "minutes") {
+        local measure "Min_W"
+        local root "min"
+		local ytitle "(minutes/day)"
+    } 
+    else {
+		local measure "Count"
+		local root "count"
+		local ytitle "(count/day)"
+    }
+	
+    * Preserve data
+    preserve
+	
+    * Reshape data
+    keep UserID PD_*Snooze`measure'`fitsby'
+    rename_but, varlist(UserID) prefix(snooze)
+    reshape long snooze, i(UserID) j(measure) string
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+
+    recode measure_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (5 = 3 "Period 4") ///
+        (7 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (3 = 6 "Periods 2 to 4") ///
+        (6 = 7 "Periods 2 to 5"), ///
+        gen(measure_recode) 
+    
+    * Plot data
+	gen dummy = 1
+	
+    cispike snooze if measure_recode <= 4, ///
+        over1(dummy) over2(measure_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Snooze use `ytitle'" " ") ///
+				  legend(off)) 
+    
+    graph export "output/cispike_snooze_`root'`suffix'.pdf", replace    
+     
+    * Restore data
+    restore
+end
+
+program plot_snooze_both
+    syntax, [fitsby]
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+		local ylabel2 0(8)40
+		local ylabel1 0(.1).5
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+		local ylabel2 0(8)40
+		local ylabel1 0(.1).5
+    }
+	
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID PD_*SnoozeCount`fitsby' *SnoozeMin_W`fitsby'
+	rename PD_*Snooze*`fitsby' **
+	rename_but, varlist(UserID) prefix(snooze)
+    reshape long snooze, i(UserID) j(measure) string
+    
+	split measure, p("_")
+	drop measure
+	rename (measure1 measure2) (time measure)
+	
+    * Recode data
+    encode time, generate(time_encode) 
+	encode measure, generate(measure_encode) 
+
+	recode time_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (3 = 3 "Period 4") ///
+        (6 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (5 = 6 "Periods 2 to 4") ///
+        (7 = 7 "Periods 2 to 5"), ///
+        gen(time_recode)  
+		
+    recode measure_encode ///
+        (1 = 1 "Snoozes per day") ///
+		(2 = 2 "Snooze minutes per day"), ///
+        gen(measure_recode)    
+    
+    * Plot data
+	
+	// Manually set labels and legends for double axis figures
+    cispike snooze if time_recode <= 3, ///
+        over1(measure_recode) over2(time_recode) ///
+		$CISPIKE_DOUBLE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+				  ylabel(`ylabel1', axis(1)) /// 
+				  ylabel(`ylabel2', axis(2)) /// 
+				  ytitle("Snoozes per day" " ", axis(1)) ///
+				  ytitle(" " "Snooze minutes per day", axis(2)) ///
+				  legend(order(4 "Snoozes per day" 10 "Snooze minutes per day")))
+    
+    graph export "output/cispike_snooze_both`suffix'.pdf", replace    
+    
+    * Restore data
+    restore
+end
+
+program plot_snooze_by_limit
+    syntax, [fitsby] [minutes]
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+	
+    * Determine snooze measure
+    if ("`minutes'" == "minutes") {
+        local measure "Min_W"
+        local root "min"
+		local ytitle "(minutes/day)"
+    } 
+    else {
+		local measure "Count"
+		local root "count"
+		local ytitle "(count/day)"
+    }
+	
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_LimitType PD_*Snooze`measure'`fitsby'
+    rename_but, varlist(UserID S2_LimitType) prefix(snooze)
+    reshape long snooze, i(UserID S2_LimitType) j(measure) string
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+
+    recode measure_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (5 = 3 "Period 4") ///
+        (7 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (3 = 6 "Periods 2 to 4") ///
+        (6 = 7 "Periods 2 to 5"), ///
+        gen(measure_recode) 
+    
+    recode S2_LimitType ///
+        (0 = .) ///
+        (1 = 1 "Snooze 0") ///
+        (2 = 2 "Snooze 2") ///
+        (3 = 3 "Snooze 5") ///
+        (4 = 4 "Snooze 20") ///
+        (5 = .), ///
+        gen(S2_LimitType_recode)
+        
+    * Plot data (by period)
+    cispike snooze if measure_recode <= 3, ///
+        over1(measure_recode) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Snooze use `ytitle'" " ")) 
+    
+    graph export "output/cispike_snooze_`root'_by_limit`suffix'.pdf", replace    
+    
+	* Plot data (all periods)
+    cispike snooze if measure_recode == 5, ///
+        over1(measure_recode) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Snooze use `ytitle'" " ") ///
+				  legend(off)) 
+    
+    graph export "output/cispike_snooze_`root'_combined_by_limit`suffix'.pdf", replace    
+    
+			 
+    * Restore data
+    restore
+end
+
+program plot_snooze_both_by_limit
+    syntax, [fitsby]
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+		local ylabel2 0(12)60
+		local ylabel1 0(.3)1.5
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+		local ylabel2 0(12)60
+		local ylabel1 0(.3)1.5
+    }
+	
+    * Preserve data
+    preserve
+    
+    * Reshape data
+	keep UserID S2_LimitType PD_*SnoozeCount`fitsby' *SnoozeMin_W`fitsby'
+	rename PD_*Snooze*`fitsby' **
+	rename_but, varlist(UserID S2_LimitType) prefix(snooze)
+    reshape long snooze, i(UserID S2_LimitType) j(measure) string
+    
+	split measure, p("_")
+	drop measure
+	rename (measure1 measure2) (time measure)
+	
+    * Recode data
+    encode time, generate(time_encode) 
+	encode measure, generate(measure_encode) 
+
+    recode S2_LimitType ///
+        (0 = .) ///
+        (1 = 1 "Snooze 0") ///
+        (2 = 2 "Snooze 2") ///
+        (3 = 3 "Snooze 5") ///
+        (4 = 4 "Snooze 20") ///
+        (5 = .), ///
+        gen(S2_LimitType_recode)
+    
+    recode time_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (3 = 3 "Period 4") ///
+        (6 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (5 = 6 "Periods 2 to 4") ///
+        (7 = 7 "Periods 2 to 5"), ///
+        gen(time_recode)  
+	
+	recode measure_encode ///
+        (1 = 1 "Snoozes per day") ///
+		(2 = 2 "Snooze minutes per day"), ///
+        gen(measure_recode)    
+		
+    * Plot data
+	
+	// Manually set labels and legends for double axis figures
+	    
+	* Plot data (all periods)
+    cispike snooze if time_recode == 6, ///
+        over1(measure_recode) over2(S2_LimitType_recode)  ///
+		$CISPIKE_DOUBLE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+				  ylabel(`ylabel1', axis(1)) /// 
+				  ylabel(`ylabel2', axis(2)) /// 
+				  ytitle("Snoozes per day" " ", axis(1)) ///
+				  ytitle(" " "Snooze minutes per day", axis(2)) ///
+				  legend(order(5 "Snoozes per day" 13 "Snooze minutes per day")))
+    
+    graph export "output/cispike_snooze_both_combined_by_limit`suffix'.pdf", replace    
+    
+    * Restore data
+    restore
+end
+
+program plot_phone_use_change
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_LimitType *PhoneUseChange
+    rename_but, varlist(UserID S2_LimitType) prefix(phone_use)
+    reshape long phone_use, i(UserID S2_LimitType) j(measure) string
+
+    * Recode data
+    encode measure, generate(measure_encode) 
+
+    recode measure_encode ///
+        (1 = 1 "Survey 1") ///
+        (2 = 2 "Survey 3") ///
+        (3 = 3 "Survey 4"), ///
+        gen(measure_recode)
+    
+    recode S2_LimitType ///
+        (0 = 0 "Control") ///
+        (1 = 1 "Snooze 0") ///
+        (2 = 2 "Snooze 2") ///
+        (3 = 3 "Snooze 5") ///
+        (4 = 4 "Snooze 20") ///
+        (5 = 5 "No snooze"), ///
+        gen(S2_LimitType_recode)
+        
+    * Plot data
+    cispike phone_use, ///
+        over1(measure_recode) over2(S2_LimitType_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Phone use change (percent)" " ") ///
+				  yline(0, lwidth(thin) lcolor(black))) 
+    
+    graph export "output/cispike_phone_use.pdf", replace    
+    
+    * Restore data
+    restore
+end
+
+program plot_phone_use_change_simple
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID S2_LimitType S3_Bonus *PhoneUseChange
+    rename_but, varlist(UserID S2_LimitType S3_Bonus) prefix(phone_use)
+    reshape long phone_use, i(UserID S2_LimitType S3_Bonus) j(measure) string
+
+    * Recode data
+    encode measure, generate(measure_encode) 
+
+    recode measure_encode ///
+        (1 = 1 "Survey 1") ///
+        (2 = 2 "Survey 3") ///
+        (3 = 3 "Survey 4"), ///
+        gen(measure_recode)
+    
+	gen treatment = .
+	replace treatment = 0 if S2_LimitType == 0 & S3_Bonus == 0
+	replace treatment = 1 if S2_LimitType == 0 & S3_Bonus == 1
+	replace treatment = 2 if S2_LimitType != 0 & S3_Bonus == 0
+	replace treatment = 3 if S2_LimitType != 0 & S3_Bonus == 1
+    
+	recode treatment ///
+        (0 = 0 "Control") ///
+        (1 = 1 "Bonus only") ///
+        (2 = 2 "Limit only") ///
+        (3 = 3 "Bonus and limit"), ///
+        gen(treatment_recode)
+        
+    * Plot data
+    cispike phone_use, ///
+        over1(measure_recode) over2(treatment_recode) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Phone use change (percent)" " ") ///
+				  yline(0, lwidth(thin) lcolor(black))) 
+    
+    graph export "output/cispike_phone_use_simple.pdf", replace    
+    
+    * Restore data
+    restore
+end
+
+program reg_usage_interaction    
+    syntax, [fitsby]
+    
+    est clear
+    
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+		
+    * Run regressions
+    foreach yvar in PD_P2_Usage`fitsby' ///
+	                PD_P3_Usage`fitsby' ///
+					PD_P4_Usage`fitsby' ///
+					PD_P5_Usage`fitsby' {
+        local baseline PD_P1_Usage`fitsby'
+        
+        gen_interaction
+        reg_interaction, yvar(`yvar') indep($STRATA `baseline')
+        est store `yvar'
+    }
+
+    * Plot regressions
+    coefplot (PD_P2_Usage`fitsby', label("Period 2") $COLOR_MAROON msymbol(O)) ///
+             (PD_P3_Usage`fitsby', label("Period 3") $COLOR_BLACK msymbol(S)) ///
+			 (PD_P4_Usage`fitsby', label("Period 4") $COLOR_NAVY msymbol(D)) ///
+			 (PD_P5_Usage`fitsby', label("Period 5") $COLOR_GRAY msymbol(T)), ///
+             keep(B_* L_*) order(B_1 L_1 B_L_1_1) ///
+			 $COEFPLOT_SETTINGS_MINUTES
+    
+    graph export "output/coef_usage_interaction`suffix'.pdf", replace
+end
+
+program reshape_self_control_outcomes
+	* Reshape wide to long
+	gen S4_Usage_FITSBY = PD_P3_UsageFITSBY
+	gen S3_Usage_FITSBY = PD_P2_UsageFITSBY
+	
+	keep UserID S3_Bonus S2_LimitType Stratifier ///
+		S*_Usage_FITSBY ///
+		S*_PhoneUseChange_N ///
+		S*_AddictionIndex_N ///
+        S*_SMSIndex_N ///
+		S*_SWBIndex_N ///
+		S*_LifeBetter_N ///
+		S*_index_well_N
+	
+	local indep UserID S3_Bonus S2_LimitType Stratifier S1_*
+	rename_but, varlist(`indep') prefix(outcome)
+    reshape long outcome, i(`indep') j(measure) string
+	
+	split measure, p(_)
+	replace measure = measure2 + "_" + measure3 + "_" + measure4 if measure4 != ""
+	replace measure = measure2 + "_" + measure3 if measure4 == ""
+	rename measure1 survey
+    drop measure2 measure3 measure4
+	
+	* Reshape long to wide
+	reshape wide outcome, i(UserID survey) j(measure) string
+	rename outcome* *
+	
+	* Recode data
+	encode survey, gen(S)
+	
+	* Label data
+	label var PhoneUseChange "Ideal use change"
+    label var AddictionIndex "Addiction scale x (-1)"
+    label var SMSIndex "SMS addiction scale x (-1)"
+	label var LifeBetter "Phone makes life better"
+	label var SWBIndex "Subjective well-being"
+	label var index_well "Survey index"
+end
+
+program gen_coefficient
+	syntax, var(str) suffix(str) label_var(str)
+	
+	cap drop C`suffix'
+	gen C`suffix' = `var' 	
+	
+	local vlabel: variable label `label_var'
+    label var C`suffix' "`vlabel'"
+end
+
+program reg_self_control 
+    est clear
+     
+	* Preserve data
+	preserve
+
+	* Reshape data
+	reshape_self_control_outcomes
+	
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+		AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+		SWBIndex_N ///
+        index_well_N 
+		
+    * Run regressions
+    foreach yvar in `yvarset' {
+		local baseline = "S1_`yvar'"
+		
+		* Treatment indicators
+		gen_treatment, suffix(_`yvar') simple
+		cap drop B3_`yvar'
+		cap drop B4_`yvar'
+		gen B3_`yvar' = B_`yvar' * (S == 1)
+		gen B4_`yvar' = B_`yvar' * (S == 2)
+		
+		* Specify regression
+		local indep i.S i.S#$STRATA i.S#c.`baseline'
+		
+		* Limit
+		gen_coefficient, var(L_`yvar') suffix(_`yvar') label_var(`yvar')
+		reg `yvar' C_`yvar' B3_`yvar' B4_`yvar' `indep', robust cluster(UserID)
+        est store L_`yvar'
+		
+		* Bonus
+		gen_coefficient, var(B4_`yvar') suffix(_`yvar') label_var(`yvar')
+		reg `yvar' L_`yvar' B3_`yvar' C_`yvar'  `indep', robust cluster(UserID)
+		est store B_`yvar'
+    }
+		
+    * Plot regressions
+    coefplot (B_*, label("Bonus") $COLOR_MAROON) ///
+             (L_*, label("Limit") $COLOR_GRAY), ///
+			              keep(C_*) ///
+			 $COEFPLOT_SETTINGS_ITT
+			 
+	graph export "output/coef_self_control.pdf", replace	
+	
+	* Restore data
+	restore
+end
+
+program reg_self_control_null
+    est clear
+     
+    * Preserve data
+    preserve
+
+    * Reshape data
+    reshape_self_control_outcomes
+    
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+        AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+        SWBIndex_N ///
+        index_well_N 
+        
+    * Run regressions
+    foreach yvar in `yvarset' {
+        local baseline = "S1_`yvar'"
+        
+        * Treatment indicators
+        gen_treatment, suffix(_`yvar') simple
+        cap drop B3_`yvar'
+        cap drop B4_`yvar'
+        cap drop L3_`yvar'
+        cap drop L4_`yvar'
+        gen B3_`yvar' = B_`yvar' * (S == 1)
+        gen B4_`yvar' = B_`yvar' * (S == 2)
+        gen L3_`yvar' = L_`yvar' * (S == 1)
+        gen L4_`yvar' = L_`yvar' * (S == 2)
+        
+        * Specify regression
+        local indep i.S i.S#$STRATA i.S#c.`baseline'
+        
+        * Limit
+        gen_coefficient, var(L3_`yvar') suffix(_`yvar') label_var(`yvar')
+        reg `yvar' C_`yvar' B3_`yvar' B4_`yvar' L4_`yvar' `indep', robust cluster(UserID)
+        est store L3_`yvar'
+        
+        gen_coefficient, var(L4_`yvar') suffix(_`yvar') label_var(`yvar')
+        reg `yvar' C_`yvar' B3_`yvar' B4_`yvar' L3_`yvar' `indep', robust cluster(UserID)
+        est store L4_`yvar'
+
+        * Bonus
+        gen_coefficient, var(B3_`yvar') suffix(_`yvar') label_var(`yvar')
+        reg `yvar' C_`yvar' L_`yvar' B4_`yvar'  `indep', robust cluster(UserID)
+        est store B3_`yvar'
+
+
+        gen_coefficient, var(B4_`yvar') suffix(_`yvar') label_var(`yvar')
+        reg `yvar' C_`yvar' L_`yvar' B3_`yvar' `indep', robust cluster(UserID)
+        est store B4_`yvar'
+    }
+        
+    * Plot regressions
+    coefplot (B3_*, label("Bonus: Survey 3") $COLOR_MAROON_LIGHT msymbol(o))   ///
+             (B4_*, label("Bonus: Survey 4") $COLOR_MAROON_DARK msymbol(s))   ///
+             (L3_*, label("Limit: Survey 3") $COLOR_GRAY_LIGHT msymbol(o))    ///
+             (L4_*, label("Limit: Survey 4") $COLOR_GRAY_DARK msymbol(s)),   ///
+                          keep(C_*) ///
+             $COEFPLOT_SETTINGS_ITT ///
+             $ADDICTION_LABELS
+             
+    graph export "output/coef_self_control_null.pdf", replace    
+    
+    * Restore data
+    restore
+end
+
+program reg_iv_self_control  
+    est clear
+     
+	* Preserve data
+	preserve
+
+	* Reshape data
+	reshape_self_control_outcomes
+	
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+		AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+		SWBIndex_N ///
+        index_well_N 
+		
+    * Run regressions
+    foreach yvar in `yvarset' {
+		local baseline = "S1_`yvar'"
+		
+		* Treatment indicators
+		gen_treatment, suffix(_`yvar') simple
+
+		* Specify regression
+		local indep i.S i.S#$STRATA i.S#c.`baseline'
+		
+		* Run regression
+		gen_usage_stacked, yvar(`yvar') suffix(_`yvar') var(`yvar') 
+        reg_usage_stacked, yvar(`yvar') suffix(_`yvar') indep(`indep') 
+		est store U_`yvar'
+    }
+		
+    * Plot regressions
+    coefplot (U_*, $COLOR_NAVY), ///
+             keep(U_*) ///
+			 $COEFPLOT_SETTINGS_STD ///
+			 legend(off)
+			 
+	graph export "output/coef_iv_self_control.pdf", replace	
+	
+	* Restore data
+	restore
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/code/FDRTable.do

@@ -0,0 +1,252 @@
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+program main
+    define_constants
+    import_data
+    run_regs
+
+    create_pval_tables
+    create_pval_tables, limit
+end
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+    yaml global STRATA = YAML.metadata.strata
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+    gen_treatment, simple
+end
+
+program latex
+   syntax, name(str) value(str)
+   
+   local command = "\newcommand{\\`name'}{`value'}"
+   
+   file open scalars using "output/scalars.tex", write append
+   file write scalars `"`command'"' _n
+   file close scalars
+end
+
+program latex_precision
+   syntax, name(str) value(str) digits(str)
+   
+   autofmt, input(`value') dec(`digits') strict
+   local value = r(output1)
+   
+   latex, name(`name') value(`value')
+end
+
+program reshape_self_control_outcomes
+  * Reshape wide to long
+  gen S4_Usage_FITSBY = PD_P3_UsageFITSBY
+  gen S3_Usage_FITSBY = PD_P2_UsageFITSBY
+  
+  keep UserID S3_Bonus S2_LimitType Stratifier ///
+       S*_Usage_FITSBY ///
+       S*_PhoneUseChange_N ///
+       S*_AddictionIndex_N ///
+       S*_SMSIndex_N ///
+       S*_SWBIndex_N ///
+       S*_LifeBetter_N ///
+       S*_index_well_N
+
+  local indep UserID S3_Bonus S2_LimitType Stratifier S1_*
+  rename_but, varlist(`indep') prefix(outcome)  
+  reshape long outcome, i(`indep') j(measure) string
+  
+  split measure, p(_)
+  replace measure = measure2 + "_" + measure3 + "_" + measure4 if measure4 != ""
+  replace measure = measure2 + "_" + measure3 if measure4 == ""
+  rename measure1 survey
+  drop measure2 measure3 measure4
+  
+  * Reshape long to wide
+  reshape wide outcome, i(UserID survey) j(measure) string
+  rename outcome* *
+  
+  * Recode data
+  encode survey, gen(S)
+  
+  * Label data
+  label var PhoneUseChange "Ideal use change"
+  label var AddictionIndex "Addiction scale x (-1)"
+  label var SMSIndex "SMS addiction scale x (-1)"
+  label var LifeBetter "Phone makes life better"
+  label var SWBIndex "Subjective well-being"
+  label var index_well "Survey index"
+
+end
+
+program make_treatment_indicators
+    * Hacky way to not have LifeBetter be dropped
+    gen alt_LifeBetter_N = LifeBetter_N
+    * Treatment indicators
+    gen_treatment, simple
+    cap drop LifeBetter_N
+    gen LifeBetter_N = alt_LifeBetter_N
+    label var LifeBetter_N "Phone makes life better"
+end
+
+program run_regs
+    * Reshape data
+    reshape_self_control_outcomes
+
+    local swb_vars ///
+          PhoneUseChange_N ///
+          AddictionIndex_N ///
+          SMSIndex_N ///
+          LifeBetter_N ///
+          SWBIndex_N ///
+          index_well_N 
+
+    make_treatment_indicators
+
+    cap drop B3
+    cap drop B4
+    gen B3 = B * (S == 1)
+    gen B4 = B * (S == 2)
+    replace B = B4
+
+    * Run regressions
+    foreach yvar in `swb_vars' {
+      local baseline = "S1_`yvar'"
+      
+      * Specify regression
+      local indep i.S i.S#$STRATA i.S#c.`baseline'
+      
+      * Limit
+      reg `yvar' B B4 L `indep', robust cluster(UserID)
+      est store `yvar'
+    }
+end
+
+program create_pval_tables    
+    syntax, [limit]
+    
+    if ("`limit'" == "limit") {
+        local T B
+        local Survey "S34"
+        local file_suffix "limit"
+    } 
+    else {
+        local T L
+        local Survey "S3"
+        local file_suffix "bonus"
+    }
+
+    local swb_vars ///
+          PhoneUseChange_N ///
+          AddictionIndex_N ///
+          SMSIndex_N ///
+          LifeBetter_N ///
+          SWBIndex_N ///
+          index_well_N 
+
+    local mat_length = 0
+    foreach var in `swb_vars' {
+      local mat_length = `mat_length' + 1
+    }
+
+    foreach matname in sd count mean Var min max sum range {
+      mat `matname'_swb = J(1,`mat_length',.)
+      mat rownames `matname'_swb = `matname'
+      mat colnames `matname'_swb = `swb_vars'
+    }
+
+    local mat_length = 0
+    foreach var in `swb_vars' {
+      local mat_length = `mat_length' + 1
+    }
+    mat pvalues = J(1,`mat_length',.)
+
+    ** Make descriptive stats and estimate tables
+    local pvalue_counter = 1
+    foreach varset in swb_vars  {
+      local suffix swb
+      local mat_counter = 1
+
+      foreach yvar in ``varset'' {
+        est restore `yvar'
+        mat count_`suffix'[1, `mat_counter'] = e(N)
+        mat mean_`suffix'[1, `mat_counter'] = _b[`T']
+        mat Var_`suffix'[1, `mat_counter'] = _se[`T']
+        local pvalue = 2 * ttail(e(N) - e(df_m), abs(_b[`T']/_se[`T']))
+
+        est restore `yvar'
+        mat min_`suffix'[1, `mat_counter'] = _b[`T']
+        mat max_`suffix'[1, `mat_counter'] = _se[`T']
+        mat sum_`suffix'[1, `mat_counter'] = `pvalue'
+        mat pvalues[1, `pvalue_counter'] = `pvalue'
+        local mat_counter = `mat_counter' + 1
+        local pvalue_counter = `pvalue_counter' + 1
+      }
+    }
+
+    clear
+
+    mat pvalues = pvalues'
+    svmat float pvalues, name(pval)
+
+    do "../../lib/stata/SharpenPValues.do"
+
+    * Note that SWB index is the fifth variable
+    * Save SWB index FDR sharpened q value as a scalar
+    local fdr_val = bky06_qval[5]
+    *latex_precision, name(`file_suffix'SWBfdr) value(`fdr_val') digits(2)
+
+    mkmat bky06_qval, matrix(sharpened_vals)
+    mat sharpened_vals = sharpened_vals'
+
+    import_data
+    reshape_self_control_outcomes
+    make_treatment_indicators
+
+    local pvalue_counter = 1
+    foreach varset in swb_vars {
+      local suffix swb
+
+      local mat_counter = 1
+      foreach yvar in ``varset'' {
+        mat range_`suffix'[1, `mat_counter'] = sharpened_vals[1, `pvalue_counter']
+      	local mat_counter = `mat_counter' + 1
+        local pvalue_counter = `pvalue_counter' + 1
+      }
+
+      estpost tabstat ``varset'' if `T'==0, statistics(mean, sd, max, min, count) columns(statistics)
+      foreach value in count {
+        estadd mat `value' = `value'_`suffix', replace
+      }
+      est store `varset'
+
+      estpost tabstat ``varset'', statistics(mean, Var, max, min, sum, range) columns(statistics)
+      foreach value in mean Var max min sum range {
+        estadd mat `value' = `value'_`suffix', replace
+      }
+      est store `varset'_reg
+
+      esttab `varset' using "output/`varset'_descriptive_stats_`file_suffix'.tex", ///
+      label cells((mean(fmt(%8.2fc)) sd(fmt(%8.2fc)) min(fmt(%8.0fc)) max(fmt(%8.0fc)) count(fmt(%8.0fc)))) ///
+      collabels("\shortstack{Mean}" "\shortstack{Standard\\deviation}" "\shortstack{Minimum\\value}" "\shortstack{Maximum\\value}" "\shortstack{N in\\regression}") ///
+      noobs replace nomtitle nonumbers compress
+
+      esttab `varset'_reg using "output/`varset'_estimates_`file_suffix'.tex", ///
+      label cells((mean(fmt(%8.2fc)) Var(fmt(%8.2fc)) min(fmt(%8.2fc)) max(fmt(%8.2fc)) sum(fmt(%8.2fc)) range(fmt(%8.2fc)))) ///
+      collabels("\shortstack{(1)\\Treatment\\effect\\(original\\units)}" "\shortstack{(2)\\Standard\\error\\(original\\units)}" "\shortstack{(3)\\Treatment\\effect\\(SD units)}" ///
+      "\shortstack{(4)\\Standard\\error\\(SD units)}" "\shortstack{(5)\\P-value}" "\shortstack{(6)\\Sharpened\\FDR-\\adjusted\\q-value}") ///
+      noobs replace nomtitle nonumbers compress
+    }
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/code/HabitFormation.do

@@ -0,0 +1,121 @@
+// Habit formation and naivete
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+	yaml global STRATA = YAML.metadata.strata
+end
+
+program define_plot_settings
+    global COLOR_MAROON ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+    
+	global COLOR_BLACK ///
+        mcolor(black) ciopts(recast(rcap) lcolor(black))
+		
+    global COLOR_GRAY ///
+        mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+	
+	global COLOR_NAVY ///
+        mcolor(navy) ciopts(recast(rcap) lcolor(navy))
+		
+	global COEFPLOT_SETTINGS_MINUTES ///
+        vertical ///
+		yline(0, lwidth(thin) lcolor(black)) ///
+		bgcolor(white) graphregion(color(white)) ///
+		legend(cols(3) region(lcolor(white))) ///
+		xsize(6.5) ysize(4.5) ///
+		ytitle("Treatment effect (minutes/day)" " ") ///
+        coeflabels(B_P3 = `"Period 3"' ///
+				   B_P4 = `"Period 4"' ///
+				   B_P5 = `"Period 5"') 
+
+	global COEFPLOT_SETTINGS_MINUTES_DOUBLE ///
+		vertical ///
+		yline(0, lwidth(thin) lcolor(black)) ///
+		bgcolor(white) graphregion(color(white)) ///
+		xsize(6.5) ysize(4.5) ///
+		ytitle("Treatment effect (minutes/day)" " ") ///
+		ytitle("Treatment effect (ICW index)" " ", axis(2)) ///
+        coeflabels(B_P3 = `"Period 3"' ///
+				   B_P4 = `"Period 4"' ///
+				   B_P5 = `"Period 5"') ///
+		legend(cols(1) region(lcolor(white))) ///
+		ylabel(-60(30)60) ///
+		ylabel(-0.1(0.05)0.1, axis(2))
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+	import_data
+    
+    survey_effects_rsi
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program survey_effects_rsi
+	preserve
+
+	* Clean data
+	rename PD_*_UsageFITSBY UsageActual_*
+    rename S3_PredictUseNext_1_W UsagePredicted_P3
+    rename S3_PredictUseNext_2_W UsagePredicted_P4
+    rename S3_PredictUseNext_3_W UsagePredicted_P5
+	
+    * Run regressions
+    foreach yvar in UsageActual  {
+		foreach survey in P3 P4 P5 {
+			local baseline `yvar'_P1
+        
+			gen_treatment, suffix(_`survey')
+			reg_treatment, yvar(`yvar'_`survey') suffix(_`survey') indep($STRATA `baseline')
+			est store `yvar'_`survey'
+		}
+	}
+
+    foreach yvar in UsagePredicted  {
+		foreach survey in P3 P4 P5 {
+			local baseline UsageActual_P1
+        
+			gen_treatment, suffix(_`survey')
+			reg_treatment, yvar(`yvar'_`survey') suffix(_`survey') indep($STRATA `baseline')
+			est store `yvar'_`survey'
+		}
+	}
+
+    * Plot regressions (by period)
+    coefplot (UsageActual*, label("Actual use") $COLOR_MAROON) ///
+			 (UsagePredicted*, label("Predicted use") $COLOR_GRAY), ///
+             keep(B_*) ///
+			 $COEFPLOT_SETTINGS_MINUTES 
+
+	restore
+    
+	graph export "output/habit_formation_fitsby.pdf", replace
+
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/code/Heterogeneity.do

@@ -0,0 +1,963 @@
+// Heterogeneity
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+	yaml global STRATA = YAML.metadata.strata
+	
+    global app_list Facebook Instagram Twitter Snapchat Browser YouTube Other
+end
+
+program define_plot_settings    
+    global CISPIKE_VERTICAL_GRAPHOPTS ///
+        ylabel(#6) ///
+        xsize(6.5) ysize(4.5) ///
+		legend(cols(3))
+		
+	global CISPIKE_HORIZONTAL_GRAPHOPTS ///
+        xlabel(#6) ///
+        xsize(6.5) ysize(6.5) 
+        
+	global CISPIKE_STACKED_GRAPHOPTS ///
+		xcommon row(2) ///
+		graphregion(color(white)) ///
+        xsize(6.5) ysize(8)
+		
+    global CISPIKE_SETTINGS ///
+        spikecolor(maroon black gray) ///
+        cicolor(maroon black gray) 
+        
+	global COLOR_MAROON ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+
+    global COLOR_LIGHT_RED ///
+        mcolor(maroon*0.7) ciopts(recast(rcap) lcolor(maroon*0.7))
+
+    global COLOR_DARK_RED ///
+        mcolor(maroon*1.3) ciopts(recast(rcap) lcolor(maroon*1.3))
+
+    global COLOR_LIGHT_GREY ///
+        mcolor(gray*0.8) ciopts(recast(rcap) lcolor(gray*0.8))
+
+    global COLOR_DARK_GREY ///
+        mcolor(gray*1.3) ciopts(recast(rcap) lcolor(gray*1.3))
+
+    global COLOR_DARK_GREEN ///
+        mcolor(teal) ciopts(recast(rcap) lcolor(teal))
+
+    global COLOR_LIGHT_GREEN ///
+        mcolor(eltgreen) ciopts(recast(rcap) lcolor(eltgreen))
+    
+	global COLOR_BLACK ///
+        mcolor(black) ciopts(recast(rcap) lcolor(black))
+		
+    global COLOR_GRAY ///
+        mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+		
+    global COEFPLOT_VERTICAL_SETTINGS ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon)) ///
+		yline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) ///
+		legend(rows(1) region(lcolor(white))) ///
+        xsize(8) ysize(4) ///
+        ytitle("Treatment effect (minutes/day)" " ") 
+    
+    global COEFPLOT_HORIZONTAL_HTE_SETTINGS ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(cols(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(6.5)
+
+    global COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(6.5)
+
+	global SMALL_LABELS ///
+		xlabel(, labsize(small)) ///
+		xtitle(, size(small)) ///
+		ylabel(, labsize(small)) ///
+		ytitle(, size(small)) ///
+		legend(size(small))
+		
+	global COEF_SMALL_LABELS ///
+        coeflabels(, labsize(small)) ///
+		$SMALL_LABELS
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+    
+	get_temptation_ranks
+	get_usage_ranks
+    plot_temptation
+    reg_usage_by_app
+    reg_usage_by_app_combined
+	plot_limit_tight_by_app
+	reg_usage_by_time
+    reg_usage_by_time, fitsby
+    reg_usage_by_time_scaled
+    reg_usage_by_time_scaled, fitsby
+    reg_usage_by_person
+    reg_usage_by_person_p3
+    reg_usage_by_person, fitsby
+    reg_usage_by_person_p3, fitsby
+	reg_iv_stacked_by_person
+	plot_wtp_motivation
+    plot_limit_wtp
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+    rename S1_IdealApp_Messenger S1_IdealApp_Messaging
+end
+
+program reshape_ideal_use
+    * Reshape data
+    keep UserID S1_IdealApp_*
+    reshape long S1_IdealApp_, i(UserID) j(app) string
+    
+    * Recode data
+    encode app, generate(app_encode) 
+    
+    recode S1_IdealApp_ ///
+        (1 = -75  ) ///
+        (2 = -37.5) ///
+        (3 = -12.5) ///
+        (4 =  0   ) ///
+        (5 =  12.5) ///
+        (6 =  37.5) ///
+        (7 =  75  ) ///
+        (8 =  0   ), ///
+        gen(S1_IdealApp_recode)
+end
+
+program get_usage_ranks
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    keep UserID PD_P1_Usage_*
+
+    drop PD_P1_Usage_H*
+
+    reshape long PD_P1_Usage_, i(UserID) j(app) s
+    replace PD_P1_Usage_ = 0 if PD_P1_Usage_ == .
+
+    * Get temptation rankings
+    collapse (mean) PD_P1_Usage_, by(app)
+	drop if app == "Other"
+    gsort -PD_P1_Usage_
+    gen app_rank = _n
+
+	* Append other last
+	set obs `=_N+1'
+	replace app = "Other" if app == ""
+	replace app_rank = `=_N' if app_rank == .
+    labmask app_rank, values(app) 
+
+    * Categorize apps
+    gen category = 2
+    replace category = 1 if ///
+        inlist(app, "Facebook", "Instagram", "Twitter", "Snapchat", "Browser", "YouTube")
+        
+    * Save data
+    keep app app_rank category
+    save "temp/app_rank_usage.dta", replace
+    
+    * Restore data
+    restore
+end
+
+program get_temptation_ranks
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    reshape_ideal_use
+        
+    * Get temptation rankings
+    collapse (mean) S1_IdealApp_recode, by(app)
+    gsort +S1_IdealApp_recode
+    gen app_rank = _n
+	
+	* Append other last
+	set obs `=_N+1'
+	replace app = "Other" if app == ""
+	replace app_rank = `=_N' if app_rank == .
+    labmask app_rank, values(app) 
+    
+    * Categorize apps
+    gen category = 2
+    replace category = 1 if ///
+        inlist(app, "Facebook", "Instagram", "Twitter", "Snapchat", "Browser", "YouTube")
+        
+    * Save data
+    keep app app_rank category
+    save "temp/app_rank.dta", replace
+    
+    * Restore data
+    restore
+end
+
+program gen_rank_labels
+    syntax, [prefix(str) suffix(str)]
+     
+    * Preserve data
+    preserve
+
+    * Import ranks
+    use "temp/app_rank.dta", clear
+    
+    global rank_labels ""
+    local N = _N
+    
+    forvalues i = 1/`N' {
+        local app = app[`i']
+        global rank_labels "$rank_labels `prefix'`app'`suffix'"
+    }
+    
+    * Restore data
+    restore
+end
+
+program gen_rank_labels_usage
+    syntax, [prefix(str) suffix(str)]
+     
+    * Preserve data
+    preserve
+
+    * Import ranks
+    use "temp/app_rank_usage.dta", clear
+    
+    global rank_labels_usage ""
+    local N = _N
+    
+    forvalues i = 1/`N' {
+        local app = app[`i']
+        global rank_labels_usage "$rank_labels_usage `prefix'`app'`suffix'"
+    }
+    
+    * Restore data
+    restore
+end
+
+program plot_temptation
+    * Preserve data
+    preserve
+    
+    * Reshape data
+    reshape_ideal_use
+    
+    * Merging in rankings
+    merge m:1 app using "temp/app_rank.dta", nogen assert(2 3) keep(3)
+    
+	* Plot data (app categories together)
+    gen dummy = 1
+
+    //TODO: fix this bug
+	cispike S1_IdealApp_recode, ///
+        over1(dummy) over2(app_rank) ///
+        horizontal missing reverse $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_HORIZONTAL_GRAPHOPTS ///
+                  ylabel(none, axis(2)) ///
+                  xtitle(" " "Ideal use change (percent)") ///
+                  legend(off) ///
+				  $SMALL_LABELS) 
+        
+    graph export "output/overuse_by_app.pdf", replace
+	
+	* Plot data (app categories separately)
+	cispike S1_IdealApp_recode if category == 1, ///
+        over1(dummy) over2(app_rank) over3(category) ///
+        horizontal missing reverse $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_HORIZONTAL_GRAPHOPTS ///
+                  ylabel(none, axis(2)) ///
+				  xtitle("") ///
+				  legend(off) fysize(45))
+    
+	graph save "output/overuse_by_app_fitsby.gph", replace
+		
+    cispike S1_IdealApp_recode if category == 2, ///
+        over1(dummy) over2(app_rank) over3(category) ///
+        horizontal missing reverse $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_HORIZONTAL_GRAPHOPTS ///
+                  ylabel(none, axis(2)) ///
+                  xtitle(" " "Ideal use change (percent)") ///
+                  legend(off))
+				  
+    graph save "output/overuse_by_app_non_fitsby.gph", replace
+    
+	graph combine ///
+		"output/overuse_by_app_fitsby.gph" ///
+		"output/overuse_by_app_non_fitsby.gph", ///
+		$CISPIKE_STACKED_GRAPHOPTS
+		
+    graph export "output/overuse_by_app_stacked.pdf", replace
+		
+	* Restore data
+    restore
+end
+
+program reg_usage_by_app
+    est clear
+	
+	foreach app in $app_list {    
+        * Specify regression
+		cap drop `app'
+        cap gen `app' = PD_P5432_Usage_`app'
+        label var `app' "`app'"
+        local yvar `app'
+        local baseline PD_P1_Usage_`app'
+		
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store Limit_Est_`yvar'
+    }
+
+    foreach app in $app_list {    
+        * Specify regression
+		cap drop `app'
+        cap gen `app' = PD_P3_Usage_`app'
+        label var `app' "`app'"
+        local yvar `app'
+        local baseline PD_P1_Usage_`app'
+		
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store Bonus_Est_`yvar'
+    }
+
+    local app_list_bonus Facebook_B Instagram_B Twitter_B Snapchat_B Browser_B YouTube_B Other_B
+
+    * Plot regressions
+	gen_rank_labels_usage, prefix("")
+    
+    coefplot (Bonus_Est_*, keep(B_*) label("Bonus") $COLOR_MAROON) ///
+	         (Limit_Est_*, keep(L_*) label("Limit") $COLOR_GRAY), ///	         
+        rename(L_* = "" B_* = "") ///
+		order($rank_labels_usage) vertical ///
+        $COEFPLOT_VERTICAL_SETTINGS
+            
+    graph export "output/coef_usage_by_app.pdf", replace
+end
+
+program reg_usage_by_app_combined
+    est clear
+	
+	foreach app in $app_list {  
+        * Specify regression
+		cap drop `app'
+        cap gen `app' = PD_P432_Usage_`app'
+        label var `app' "`app'"
+        local yvar `app'
+        local baseline PD_P1_Usage_`app'
+		
+		* Run regression
+        gen_treatment_combined, suffix(_`yvar') var(`yvar')
+        reg_treatment_combined, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline')
+        est store `yvar'
+    }
+    
+    * Plot regressions
+	gen_rank_labels, prefix("C_")
+    
+    coefplot $app_list, ///
+        keep(C_*) order($rank_labels) vertical ///
+        nooffsets $COEFPLOT_VERTICAL_SETTINGS ///
+		legend(off)
+            
+    graph export "output/coef_usage_by_app_combined.pdf", replace
+end
+
+program plot_limit_tight_by_app
+    * Preserve data
+    preserve
+	
+    * Make zero in areas where not all zeros
+    foreach time in P2 P3 P4 P5 P5432 P432 P43 {
+        foreach category in Facebook Instagram Twitter Snapchat Browser YouTube Other {
+            replace PD_`time'_LimitTight_`category' = 0 if PD_`time'_LimitTight != . & PD_`time'_LimitTight_`category' == .
+        }
+    }
+
+	* Reshape data
+    keep UserID *LimitTight_*
+
+	rename *LimitTight_*  **
+    rename_but, varlist(UserID) prefix(limit)
+    reshape long limit, i(UserID) j(measure) string
+    
+	split measure, p("_")
+	drop measure measure1
+	rename (measure2 measure3) (measure app)
+	
+    * Recode data
+    encode measure, generate(measure_encode) 
+    
+	merge m:1 app using "temp/app_rank_usage.dta", nogen keep(3)
+	
+    recode measure_encode ///
+        (1 = 1 "Period 2") ///
+        (2 = 2 "Period 3") ///
+        (3 = 3 "Period 4") ///
+        (6 = 4 "Period 5") ///
+        (4 = 5 "Periods 3 & 4") ///
+        (5 = 6 "Periods 2 to 4") ///
+        (7 = 7 "Periods 2 to 5"), ///
+        gen(measure_recode)
+	 
+	* Plot data (all periods together)
+    gen dummy = 1
+	
+	cispike limit if measure_recode == 7, ///
+        over1(dummy) over2(app_rank) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Limit tightness (minutes/day)" " ") ///
+				  legend(off))
+	
+	graph export "output/cispike_limit_tight_combined_by_app.pdf", replace
+	
+    * Plot data (by period)
+    cispike limit if measure_recode <= 3, ///
+        over1(measure_recode) over2(app_rank) ///
+        $CISPIKE_SETTINGS ///
+        graphopts($CISPIKE_VERTICAL_GRAPHOPTS ///
+                  ytitle("Limit tightness (minutes/day)" " "))
+	
+	graph export "output/cispike_limit_tight_by_app.pdf", replace
+		
+	* Restore data
+	restore 
+end
+
+program reg_usage_by_time
+    syntax, [fitsby]
+    
+	est clear
+	
+	* Determine FITSBY restriction
+	if ("`fitsby'" == "fitsby") {
+		local fitsby "FITSBY"
+		local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+		local suffix ""
+    }
+	
+	foreach hour of num 1(2)23 {        		
+		* Specify regression
+		cap drop H_`hour'
+		gen H_`hour' = PD_P432_Usage`fitsby'_H`hour'
+		label var H_`hour' "`hour'"
+		local yvar H_`hour'
+		local baseline PD_P1_Usage`fitsby'_H`hour'
+			
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store L`yvar'
+
+        * run bonus regressions separately
+		cap drop H_`hour'
+ 		gen H_`hour' = PD_P3_Usage`fitsby'_H`hour'
+		label var H_`hour' "`hour'"
+		local yvar H_`hour'
+		local baseline PD_P1_Usage`fitsby'_H`hour'
+
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store B`yvar'
+    }
+    
+    * Plot regressions
+    coefplot (BH_*, keep(B_*) label("Bonus") $COLOR_MAROON) ///
+	         (LH_*, keep(L_*) label("Limit") $COLOR_GRAY), ///	         
+        rename(L_* = "" B_* = "") vertical ///
+		xtitle(" " "Hour") ///
+		$COEFPLOT_VERTICAL_SETTINGS ///
+		ytitle("Treatment effect (minutes/hour)" " ") 
+      
+	graph export "output/coef_usage_by_time`suffix'.pdf", replace
+	
+	* Preserve data
+	preserve
+	
+	* Reshape data
+	keep PD_P1_Usage`fitsby'_H*
+	collapse (mean) PD_P1_Usage`fitsby'_H*
+	gen dummy = 1
+	reshape long PD_P1_Usage`fitsby'_H, i(dummy) j(hour)
+	
+	* Recode data
+	replace PD_P1_Usage`fitsby'_H = PD_P1_Usage`fitsby'_H / 2
+	replace hour = (hour + 1) / 2
+	
+	* Label data
+	foreach hour of num 1(2)23 {        		
+		gen H_`hour' = .
+		label var H_`hour' "`hour'"
+    }
+	
+	* Plot regressions (with usage)
+    
+	// Manually set labels and legends for double axis figures
+	coefplot (BH_*, keep(B_*) label("Bonus") $COLOR_MAROON) ///
+	         (LH_*, keep(L_*) label("Limit") $COLOR_BLACK), ///	         
+        rename(L_* = "" B_* = "") vertical ///
+		xtitle(" " "Hour") ///
+		$COEFPLOT_VERTICAL_SETTINGS ///
+		ytitle("Treatment effect (minutes/hour)" " ", axis(1)) ///
+		ytitle(" " "Usage (minutes/hour)", axis(2)) ///
+		ylabel(-4(2)4, axis(1)) yscale(range(-4, 4)) /// 
+		ylabel(0(0.75)3, axis(2)) ///
+		yscale(alt) ///
+		addplot(bar PD_P1_Usage`fitsby'_H hour, ///
+		        below yaxis(2) yscale(alt axis(2)) ///
+				color(gray%50) fintensity(100) barw(0.75))
+		
+    graph export "output/coef_usage_by_time_usage`suffix'.pdf", replace
+	
+	* Restore data
+	restore
+end
+
+program reg_usage_by_time_scaled
+    syntax, [fitsby]
+    
+	est clear
+	
+	* Determine FITSBY restriction
+	if ("`fitsby'" == "fitsby") {
+		local fitsby "FITSBY"
+		local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+		local suffix ""
+    }
+	
+	* Preserve data
+	preserve
+	
+	foreach hour of num 1(2)23 {
+        display(`hour')        		
+		* Normalize usage // ASK ABOUT THIS
+        cap drop H_`hour'
+		sum PD_P432_Usage`fitsby'_H`hour' if S3_Bonus == 0 & S2_LimitType == 0
+		gen H_`hour' = PD_P432_Usage`fitsby'_H`hour' / r(mean)
+		
+		* Specify regression
+		label var H_`hour' "`hour'"
+		local yvar H_`hour'
+		local baseline PD_P1_Usage`fitsby'_H`hour'
+			
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store L`yvar'
+
+        * run bonus regressions separately
+		cap drop H_`hour'
+		sum PD_P3_Usage`fitsby'_H`hour' if S3_Bonus == 0 & S2_LimitType == 0
+		gen H_`hour' = PD_P3_Usage`fitsby'_H`hour' / r(mean)
+		
+		* Specify regression
+		label var H_`hour' "`hour'"
+		local yvar H_`hour'
+		local baseline PD_P1_Usage`fitsby'_H`hour'
+			
+		* Run regression
+        gen_treatment, suffix(_`yvar') var(`yvar') simple
+        reg_treatment, yvar(`yvar') suffix(_`yvar') indep($STRATA `baseline') simple
+        est store B`yvar'
+    }
+    
+    * Plot regressions
+    coefplot (BH_*, keep(B_*) label("Bonus") $COLOR_MAROON) ///
+	         (LH_*, keep(L_*) label("Limit") $COLOR_GRAY), ///	         
+        rename(L_* = "" B_* = "") vertical ///
+		xtitle(" " "Hour") ///
+		$COEFPLOT_VERTICAL_SETTINGS ///
+		ytitle("Treatment effect" "(share of Control group usage)" " ") 
+      
+	graph export "output/coef_usage_by_time_scaled`suffix'.pdf", replace
+	
+	* Restore data
+	restore
+end
+
+program reg_usage_by_person
+    syntax, [fitsby]
+    
+	est clear
+
+    * Determine FITSBY restriction
+	if ("`fitsby'" == "fitsby") {
+		local fitsby "FITSBY"
+		local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+		local suffix ""
+    }
+	
+    * Specify regressions	
+	include "input/lib/stata/define_heterogeneity.do"
+
+
+    local label_E "Education"
+    local label_A "Age"
+    local label_G "Female"
+    local label_U "Baseline usage"
+    local label_R "Restriction index"
+    local label_L "Addiction index"
+	 
+	* Run regressions
+    foreach mod in /*I*/ E A G U R L {
+		foreach group in 0 1 {
+			foreach yvar in PD_P5432_Usage`FITSBY' {
+				local baseline PD_P1_Usage`FITSBY'
+				local if `mod'`group'
+				
+				gen_treatment, suffix(_`mod') simple
+				label var L_`mod' "`label_`mod''" 
+				label var B_`mod' "`label_`mod''" 
+				reg_treatment, yvar(`yvar') suffix(_`mod') indep($STRATA `baseline') if(``if'') simple
+				est store `yvar'_`if'
+			}
+		}
+    }
+	
+	* Plot regressions
+	coefplot (*1, label("Above median") $COLOR_DARK_GREY) ///
+             (*0, label("Below median") $COLOR_LIGHT_GREY), ///
+		keep(L_*) ///
+		$COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+		xtitle(" " "Treatment effect (minutes/day)") ///
+		$COEF_SMALL_LABELS
+		
+	graph export "output/coef_limit_usage_by_heterogeneity`suffix'.pdf", replace
+	
+    coefplot (*1, label("Above median") $COLOR_DARK_RED) ///
+             (*0, label("Below median") $COLOR_LIGHT_RED), ///
+		keep(B_*) ///
+		$COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+		xtitle(" " "Treatment effect (minutes/day)") ///
+		$COEF_SMALL_LABELS
+		
+	graph export "output/coef_bonus_usage_by_heterogeneity`suffix'.pdf", replace
+end
+
+program reg_usage_by_person_p3
+    syntax, [fitsby]
+    
+    est clear
+
+    * Determine FITSBY restriction
+    if ("`fitsby'" == "fitsby") {
+        local fitsby "FITSBY"
+        local suffix "_fitsby"
+    } 
+    else {
+        local fitsby ""
+        local suffix ""
+    }
+    
+    * Specify regressions   
+    include "input/lib/stata/define_heterogeneity.do"
+
+
+    local label_E "Education"
+    local label_A "Age"
+    local label_G "Female"
+    local label_U "Baseline usage"
+    local label_R "Restriction index"
+    local label_L "Addiction index"
+     
+    * Run regressions
+    foreach mod in /*I*/ E A G U R L {
+        foreach group in 0 1 {
+            foreach yvar in PD_P3_Usage`FITSBY' {
+                local baseline PD_P1_Usage`FITSBY'
+                local if `mod'`group'
+                
+                gen_treatment, suffix(_`mod') simple
+                label var L_`mod' "`label_`mod''" 
+                label var B_`mod' "`label_`mod''" 
+                reg_treatment, yvar(`yvar') suffix(_`mod') indep($STRATA `baseline') if(``if'') simple
+                est store `yvar'_`if'
+            }
+        }
+    }
+    
+    * Plot regressions
+    coefplot (*1, label("Above median") $COLOR_DARK_GREY) ///
+             (*0, label("Below median") $COLOR_LIGHT_GREY), ///
+        keep(L_*) ///
+        $COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xtitle(" " "Treatment effect (minutes/day)") ///
+        $COEF_SMALL_LABELS
+        
+    graph export "output/coef_limit_usage_by_heterogeneity_P3`suffix'.pdf", replace
+    
+    coefplot (*1, label("Above median") $COLOR_DARK_RED) ///
+             (*0, label("Below median") $COLOR_LIGHT_RED), ///
+        keep(B_*) ///
+        $COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xtitle(" " "Treatment effect (minutes/day)") ///
+        $COEF_SMALL_LABELS
+        
+    graph export "output/coef_bonus_usage_by_heterogeneity_P3`suffix'.pdf", replace
+end
+
+
+
+
+
+program reshape_self_control_outcomes
+	* Reshape wide to long
+	gen S4_Usage_FITSBY = PD_P3_UsageFITSBY
+	gen S3_Usage_FITSBY = PD_P2_UsageFITSBY
+	
+	keep UserID S3_Bonus S2_LimitType Stratifier ///
+		S1_Income S1_Education S0_Age S0_Gender ///
+		StratWantRestrictionIndex StratAddictionLifeIndex PD_P1_UsageFITSBY ///
+		S*_Usage_FITSBY ///
+		S*_PhoneUseChange_N ///
+		S*_AddictionIndex_N ///
+        S*_SMSIndex_N ///
+		S*_SWBIndex_N ///
+		S*_LifeBetter_N ///
+		S*_index_well_N
+	
+	local indep UserID S3_Bonus S2_LimitType Stratifier S1_* S0_* Strat* PD_*
+	rename_but, varlist(`indep') prefix(outcome)
+    reshape long outcome, i(UserID) j(measure) string
+	
+	split measure, p(_)
+	replace measure = measure2 + "_" + measure3 + "_" + measure4 if measure4 != ""
+	replace measure = measure2 + "_" + measure3 if measure4 == ""
+	rename measure1 survey
+    drop measure2 measure3 measure4
+	
+	* Reshape long to wide
+	reshape wide outcome, i(UserID survey) j(measure) string
+	rename outcome* *
+	
+	* Recode data
+	encode survey, gen(S)
+	
+	* Label data
+	label var PhoneUseChange "Ideal use change"
+    label var AddictionIndex "Addiction scale x (-1)"
+    label var SMSIndex "SMS addiction scale x (-1)"
+	label var LifeBetter "Phone makes life better"
+	label var SWBIndex "Subjective well-being"
+	label var index_well "Survey index"
+end
+
+program reg_iv_stacked_by_person
+    est clear
+    
+	* Preserve data
+	preserve
+	
+	* Reshape data
+	reshape_self_control_outcomes
+	
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+		AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+		SWBIndex_N ///
+        index_well_N 
+		
+	include "input/lib/stata/define_heterogeneity.do"
+	
+	* Run regressions
+    foreach if in /*I0 I1*/ E0 E1 A0 A1 G0 G1 U0 U1 R0 R1 L0 L1 {
+		foreach yvar in `yvarset' {
+			local baseline = "S1_`yvar'"
+		
+			* Treatment indicators
+			gen_treatment, suffix(_`yvar') simple
+
+			* Specify regression
+			local indep i.S i.S#$STRATA i.S#c.`baseline'
+			
+			* Run regression
+			gen_usage_stacked, yvar(`yvar') suffix(_`yvar') var(`yvar') 
+			reg_usage_stacked, yvar(`yvar') suffix(_`yvar') indep(`indep') if(``if'')
+			est store U_`yvar'_`if'
+        }
+    }
+	
+	* Plot regressions
+	foreach mod in /*I*/ E A G U R L {
+		local coef_plot0 ///
+			label("`label_`mod'0'") ///
+			mcolor(edkblue*0.7) ciopts(recast(rcap) lcolor(edkblue*0.7))
+			
+		local coef_plot1 ///
+			label("`label_`mod'1'") ///
+			mcolor(edkblue*1.3) ciopts(recast(rcap) lcolor(edkblue*1.3))
+		
+        coefplot (U_*_`mod'1, `coef_plot1') ///
+                 (U_*_`mod'0, `coef_plot0'), ///
+            keep(U_*) ///
+            $COEFPLOT_HORIZONTAL_HTE_SETTINGS ///
+			xtitle(" " "Treatment effect" "(standard deviations per hour/day of use)") ///
+			$COEF_SMALL_LABELS
+        
+		graph export "output/coef_iv_self_control_by_`suffix_`mod''.pdf", replace
+    }
+	
+	* Restore data
+	restore
+end
+
+program plot_wtp_motivation
+    * Preserve data
+    preserve
+    
+    * Specify groups
+	include "input/lib/stata/define_heterogeneity.do"
+    
+	foreach mod in I E A G U R L {
+		foreach group in 0 1 {
+			gen Motivation_`mod'_`group' = S2_Motivation ``mod'`group''
+		}
+    }
+	
+	* Reshape data
+	keep UserID Motivation_*
+	reshape long Motivation, i(UserID) j(measure) string
+	split measure, p("_")
+	drop measure measure1
+	rename measure2 measure
+	rename measure3 group
+	
+    * Recode data
+    encode measure, generate(measure_encode) 
+    encode group, generate(group_encode) 
+	 
+    recode measure_encode ///
+        (2 = 1 "Education") ///
+        (1 = 2 "Age") ///
+        (3 = 3 "Female") ///
+        (7 = 4 "Baseline usage") ///
+        (5 = 5 "Restriction index") ///
+        (6 = 6 "Addiction index") ///
+        (4 = 7 "Income less than $50,000"), ///
+        gen(measure_recode)
+        
+    recode group_encode ///
+        (1 = 2 "Below median") ///
+        (2 = 1 "Above median"), ///
+        gen(group_recode)
+		
+	* Plot data (app categories together)
+    cispike Motivation if measure_recode != 7, ///
+        over1(group_recode) over2(measure_recode) ///
+        horizontal reverse ///
+        spikecolor(maroon gray) ///
+        cicolor(maroon gray) ///
+        graphopts($CISPIKE_HORIZONTAL_GRAPHOPTS ///
+                  xtitle(" " "Behavior change premium") ///
+		    	  $SMALL_LABELS)
+        
+    graph export "output/cispike_motivation_by_group.pdf", replace
+	
+	* Restore data
+	restore
+end
+
+program plot_limit_wtp
+    * Preserve data
+    preserve
+    
+    * Specify groups
+    include "input/lib/stata/define_heterogeneity.do"
+    
+    foreach mod in I E A G U R L {
+        foreach group in 0 1 {
+            gen WTP_`mod'_`group' = S3_MPLLimit ``mod'`group''
+        }
+    }
+    
+    * Reshape data
+    keep UserID WTP_*
+    reshape long WTP, i(UserID) j(measure) string
+    split measure, p("_")
+    drop measure measure1
+    rename measure2 measure
+    rename measure3 group
+    
+    * Recode data
+    encode measure, generate(measure_encode) 
+    encode group, generate(group_encode) 
+     
+    recode measure_encode ///
+        (2 = 1 "Education") ///
+        (1 = 2 "Age") ///
+        (3 = 3 "Female") ///
+        (7 = 4 "Baseline usage") ///
+        (5 = 5 "Restriction index") ///
+        (6 = 6 "Addiction index") ///
+        (4 = 7 "Income less than $50,000"), ///
+        gen(measure_recode)
+        
+    recode group_encode ///
+        (1 = 2 "Below median") ///
+        (2 = 1 "Above median"), ///
+        gen(group_recode)
+        
+    * Plot data (app categories together)
+    cispike WTP if measure_recode != 7, ///
+        over1(group_recode) over2(measure_recode) ///
+        horizontal reverse ///
+        spikecolor(maroon gray) ///
+        cicolor(maroon gray) ///
+        graphopts($CISPIKE_HORIZONTAL_GRAPHOPTS ///
+                  xtitle(" " "Willingness to pay for limit ($)") ///
+                  $SMALL_LABELS)
+        
+    graph export "output/cispike_limit_motivation_by_group.pdf", replace
+    
+    * Restore data
+    restore
+end
+
+
+***********
+* Execute *
+***********
+
+main
+

17/replication_package/code/analysis/treatment_effects/code/HeterogeneityInstrumental.do

@@ -0,0 +1,477 @@
+// Response to commitment, moderated by demand for flexibility
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_constants
+    yaml read YAML using "input/config.yaml"
+    yaml global STRATA = YAML.metadata.strata
+end
+
+program define_plot_settings       
+    global COEFPLOT_HORIZONTAL_SETTINGS ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(cols(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(6.5)
+
+    global COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(6.5)
+        
+    global SMALL_LABELS ///
+        xlabel(, labsize(small)) ///
+        xtitle(, size(small)) ///
+        ylabel(, labsize(small)) ///
+        ytitle(, size(small)) ///
+        legend(size(small))
+        
+    global COEF_SMALL_LABELS ///
+        coeflabels(, labsize(small)) ///
+        $SMALL_LABELS
+
+    global COEFPLOT_SETTINGS_STD ///
+        xline(0, lwidth(thin) lcolor(black)) ///
+        bgcolor(white) graphregion(color(white)) grid(w) ///
+        legend(rows(1) region(lcolor(white))) ///
+        xsize(6.5) ysize(4.5) ///
+        xtitle(" " "Treatment effect (standard deviations per hour/day of use)") 
+
+    global COLOR_MAROON ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+
+    global COLOR_LIGHT_RED ///
+        mcolor(maroon*0.7) ciopts(recast(rcap) lcolor(maroon*0.7))
+
+    global COLOR_DARK_RED ///
+        mcolor(maroon*1.3) ciopts(recast(rcap) lcolor(maroon*1.3))
+
+    global COLOR_LIGHT_GREY ///
+        mcolor(gray*0.8) ciopts(recast(rcap) lcolor(gray*0.8))
+
+    global COLOR_DARK_GREY ///
+        mcolor(gray*1.3) ciopts(recast(rcap) lcolor(gray*1.3))
+
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    define_constants
+    define_plot_settings
+    import_data
+
+    reg_survey_heterogeneity
+    reg_iv_self_control_no_B3
+    reg_welfare_itt
+    reg_welfare_late
+end
+
+program import_data
+    use "input/final_data_sample.dta", clear
+end
+
+program gen_coefficient
+    syntax, var(str) suffix(str) label_var(str)
+    
+    cap drop C`suffix'
+    gen C`suffix' = `var'   
+    
+    local vlabel: variable label `label_var'
+    label var C`suffix' "`vlabel'"
+end
+
+program reshape_self_control_outcomes_h
+    * Reshape wide to long
+    gen S4_Usage_FITSBY = PD_P3_UsageFITSBY
+    gen S3_Usage_FITSBY = PD_P2_UsageFITSBY
+    
+    keep UserID S3_Bonus S2_LimitType Stratifier ///
+        S1_Income S1_Education S0_Age S0_Gender ///
+        StratWantRestrictionIndex StratAddictionLifeIndex PD_P1_UsageFITSBY ///
+        S*_Usage_FITSBY ///
+        S*_PhoneUseChange_N ///
+        S*_AddictionIndex_N ///
+        S*_SMSIndex_N ///
+        S*_SWBIndex_N ///
+        S*_LifeBetter_N ///
+        S*_index_well_N
+    
+    local indep UserID S3_Bonus S2_LimitType Stratifier S1_* S0_* Strat* PD_*
+    rename_but, varlist(`indep') prefix(outcome)
+    reshape long outcome, i(UserID) j(measure) string
+    
+    split measure, p(_)
+    replace measure = measure2 + "_" + measure3 + "_" + measure4 if measure4 != ""
+    replace measure = measure2 + "_" + measure3 if measure4 == ""
+    rename measure1 survey
+    drop measure2 measure3 measure4
+    
+    * Reshape long to wide
+    reshape wide outcome, i(UserID survey) j(measure) string
+    rename outcome* *
+    
+    * Recode data
+    encode survey, gen(S)
+    
+    * Label data
+    label var PhoneUseChange "Ideal use change"
+    label var AddictionIndex "Addiction scale x (-1)"
+    label var SMSIndex "SMS addiction scale x (-1)"
+    label var LifeBetter "Phone makes life better"
+    label var SWBIndex "Subjective well-being"
+    label var index_well "Survey index"
+end
+
+program reshape_self_control_outcomes
+    * Reshape wide to long
+    gen S4_Usage_FITSBY = PD_P3_UsageFITSBY
+    gen S3_Usage_FITSBY = PD_P2_UsageFITSBY
+    
+    keep UserID S3_Bonus S2_LimitType Stratifier ///
+        S*_Usage_FITSBY ///
+        S*_PhoneUseChange_N ///
+        S*_AddictionIndex_N ///
+        S*_SMSIndex_N ///
+        S*_SWBIndex_N ///
+        S*_LifeBetter_N ///
+        S*_index_well_N
+    
+    local indep UserID S3_Bonus S2_LimitType Stratifier S1_*
+    rename_but, varlist(`indep') prefix(outcome)
+    reshape long outcome, i(`indep') j(measure) string
+    
+    split measure, p(_)
+    replace measure = measure2 + "_" + measure3 + "_" + measure4 if measure4 != ""
+    replace measure = measure2 + "_" + measure3 if measure4 == ""
+    rename measure1 survey
+    drop measure2 measure3 measure4
+    
+    * Reshape long to wide
+    reshape wide outcome, i(UserID survey) j(measure) string
+    rename outcome* *
+    
+    * Recode data
+    encode survey, gen(S)
+    
+    * Label data
+    label var PhoneUseChange "Ideal use change"
+    label var AddictionIndex "Addiction scale x (-1)"
+    label var SMSIndex "SMS addiction scale x (-1)"
+    label var LifeBetter "Phone makes life better"
+    label var SWBIndex "Subjective well-being"
+    label var index_well "Survey index"
+end
+
+program reg_usage_stacked_no_B3
+    syntax, yvar(str) [suffix(str) indep(str) if(str)]
+
+    cap drop i_S4
+    gen i_S4 = S - 1
+    gen B_`yvar'4 = i_S4 * B_`yvar'
+    
+    ivregress 2sls `yvar' (U`suffix' = B_`yvar'4 i.S#L_`yvar') `indep' `if', robust
+end
+
+program reg_survey_heterogeneity
+    syntax
+
+    est clear
+
+    preserve
+    * Reshape data
+    reshape_self_control_outcomes_h
+
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+        AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+        SWBIndex_N ///
+        index_well_N 
+
+    include "input/lib/stata/define_heterogeneity.do"
+
+
+    * Run regressions
+    foreach if in R0 R1 L0 L1 U0 U1 {
+        foreach yvar in `yvarset' {
+            local baseline = "S1_`yvar'"
+
+            * Treatment indicators
+            gen_treatment, suffix(_`yvar') simple
+            cap drop B3_`yvar'
+            cap drop B4_`yvar'
+            gen B3_`yvar' = B_`yvar' * (S == 1)
+            gen B4_`yvar' = B_`yvar' * (S == 2)
+            
+            * Specify regression
+            local indep i.S i.S#$STRATA i.S#c.`baseline'
+            
+            * Limit
+            gen_coefficient, var(L_`yvar') suffix(_`yvar') label_var(`yvar')
+            reg `yvar' C_`yvar' B3_`yvar' B4_`yvar' `indep' ``if'', robust cluster(UserID)
+            est store L_`yvar'_`if'
+            
+            * Bonus
+            gen_coefficient, var(B4_`yvar') suffix(_`yvar') label_var(`yvar')
+            reg `yvar' L_`yvar' B3_`yvar' C_`yvar'  `indep' ``if'', robust cluster(UserID)
+            est store B_`yvar'_`if'
+        }
+    }
+    
+    * Plot regressions
+    foreach mod in R L U {
+        local coef_plot0 ///
+            label("`label_`mod'0'") ///
+            mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+            
+        local coef_plot1 ///
+            label("`label_`mod'1'") ///
+            mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+        
+        coefplot (L_*_`mod'1, `coef_plot1') ///
+                 (L_*_`mod'0, `coef_plot0'), ///
+            keep(C_*) ///
+            $COEFPLOT_HORIZONTAL_SETTINGS ///
+            xtitle(" " "Treatment effect" "(standard deviations)") ///
+            $COEF_SMALL_LABELS
+        
+        graph export "output/coef_limit_itt_by_`suffix_`mod''.pdf", replace
+
+        coefplot (B_*_`mod'1, `coef_plot1') ///
+                 (B_*_`mod'0, `coef_plot0'), ///
+            keep(C_*) ///
+            $COEFPLOT_HORIZONTAL_SETTINGS ///
+            xtitle(" " "Treatment effect" "(standard deviations)") ///
+            $COEF_SMALL_LABELS
+        
+        graph export "output/coef_bonus_itt_by_`suffix_`mod''.pdf", replace
+    }
+    
+    * Restore data
+    restore
+
+end
+
+program reg_iv_self_control_no_B3 
+    est clear
+     
+    * Preserve data
+    preserve
+
+    * Reshape data
+    reshape_self_control_outcomes
+    
+    * Specify regression 
+    local yvarset ///
+        PhoneUseChange_N ///
+        AddictionIndex_N ///
+        SMSIndex_N ///
+        LifeBetter_N ///
+        SWBIndex_N ///
+        index_well_N 
+        
+    * Run regressions
+    foreach yvar in `yvarset' {
+        local baseline = "S1_`yvar'"
+        
+        * Treatment indicators
+        gen_treatment, suffix(_`yvar') simple
+
+        * Specify regression
+        local indep i.S i.S#$STRATA i.S#c.`baseline'
+        
+        * Run regression
+        gen_usage_stacked, yvar(`yvar') suffix(_`yvar') var(`yvar') 
+        reg_usage_stacked_no_B3, yvar(`yvar') suffix(_`yvar') indep(`indep') 
+        est store U_`yvar'
+    }
+        
+    * Plot regressions
+    coefplot (U_*, $COLOR_MAROON), ///
+             keep(U_*) ///
+             $COEFPLOT_SETTINGS_STD ///
+             legend(off)
+             
+    graph export "output/coef_iv_self_control_no_B3.pdf", replace 
+    
+    * Restore data
+    restore
+end
+
+program reg_welfare_itt
+    est clear
+
+    preserve
+    * Reshape data
+    reshape_self_control_outcomes_h
+
+    * Specify regression 
+    local yvar index_well_N 
+
+    include "input/lib/stata/define_heterogeneity.do"
+
+    local label_E "Education"
+    local label_A "Age"
+    local label_G "Female"
+    local label_U "Baseline usage"
+    local label_R "Restriction index"
+    local label_L "Addiction index"
+
+    gen MG_Indicator =  0 if S0_Gender == 1
+    replace MG_Indicator = 1 if S0_Gender == 2
+
+    local baseline = "S1_`yvar'"
+
+    * Treatment indicators
+    gen_treatment, simple
+    cap drop B3_`yvar'
+    cap drop B4_`yvar'
+    gen B3 = B * (S == 1)
+    gen B4 = B * (S == 2)
+    
+    * Specify regression
+    local indep i.S i.S#$STRATA i.S#c.`baseline'
+
+    * Run regressions
+    foreach group in E A G U R L {
+        foreach s in 0 1 {
+
+            local if "if M`group'_Indicator == `s'"
+            * Limit
+            cap drop C_`group'
+            gen C_`group' = L
+            label var C_`group' "`label_`group''"
+            reg `yvar' C_`group' B3 B4 `indep' ``if'', robust cluster(UserID)
+            est store L_`group'`s'
+            
+            * Bonus
+            cap drop C_`group'
+            gen C_`group' = B4
+            label var C_`group' "`label_`group''"
+            reg `yvar' L B3 C_`group'  `indep' ``if'', robust cluster(UserID)
+            est store B_`group'`s'
+        }
+    }
+    
+    * Plot regressions
+
+    local coef_plot0 ///
+        label("Below median") 
+        
+    local coef_plot1 ///
+        label("Above median") 
+    
+    coefplot (L*1, `coef_plot1' $COLOR_DARK_GREY) ///
+             (L*0, `coef_plot0' $COLOR_LIGHT_GREY), ///
+        keep(C_*) ///
+        $COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xtitle(" " "Treatment effect" "(standard deviations)") ///
+        $COEF_SMALL_LABELS
+    
+    graph export "output/coef_heterogenous_limit_itt_welfare.pdf", replace
+
+    coefplot (B*1, `coef_plot1' $COLOR_DARK_RED) ///
+             (B*0, `coef_plot0' $COLOR_LIGHT_RED), ///
+        keep(C_*) ///
+        $COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xtitle(" " "Treatment effect" "(standard deviations)") ///
+        $COEF_SMALL_LABELS
+    
+    graph export "output/coef_heterogenous_bonus_itt_welfare.pdf", replace
+    
+    * Restore data
+    restore
+end
+
+program reg_welfare_late
+    est clear
+
+    preserve
+    * Reshape data
+    reshape_self_control_outcomes_h
+
+    * Specify regression 
+    local yvar index_well_N 
+
+    include "input/lib/stata/define_heterogeneity.do"
+
+    local label_E "Education"
+    local label_A "Age"
+    local label_G "Female"
+    local label_U "Baseline usage"
+    local label_R "Restriction index"
+    local label_L "Addiction index"
+
+    gen MG_Indicator =  0 if S0_Gender == 1
+    replace MG_Indicator = 1 if S0_Gender == 2
+
+
+    local baseline = "S1_`yvar'"
+    
+    * Specify regression
+    local indep i.S i.S#$STRATA i.S#c.`baseline'
+
+    * Run regressions
+    foreach group in E A G U R L {
+        foreach s in 0 1 {
+
+            local if "if M`group'_Indicator == `s'"
+            
+            * Create usage variable (make negative per issue 184 comments)
+            cap drop U_`group'
+            gen U_`group' = -1 * Usage_FITSBY 
+
+            * Converts usage to hours /day from minutes/day
+            replace U_`group' =  U_`group'/60
+            label var U_`group' "`label_`group''"
+
+            * Run regression
+            gen_treatment, suffix(_`yvar') simple
+            reg_usage_stacked, yvar(`yvar') suffix(_`group') indep(`indep') if(``if'')
+
+            est store U_`group'`s'
+        }
+    }
+    
+    * Plot regressions
+
+    local coef_plot0 ///
+        label("Below median") ///
+        mcolor(gray) ciopts(recast(rcap) lcolor(gray))
+        
+    local coef_plot1 ///
+        label("Above median") ///
+        mcolor(maroon) ciopts(recast(rcap) lcolor(maroon))
+    
+    coefplot (U*1, `coef_plot1') ///
+             (U*0, `coef_plot0'), ///
+        keep(U_*) ///
+        $COEFPLOT_HORIZONTAL_MED_SETTINGS ///
+        xtitle(" " "Treatment effect" "(standard deviations per hour/day of use)") ///
+        $COEF_SMALL_LABELS
+    
+    graph export "output/coef_heterogenous_late_welfare.pdf", replace
+
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/code/ModelHeterogeneity.R

@@ -0,0 +1,1406 @@
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Setup
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+# Import plotting functions and constants from lib file
+source('input/lib/r/ModelFunctions.R')
+p_B <- (hourly_rate / num_days) / 60
+F_B <- (hourly_rate * max_hours) / num_days
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Helper Functions
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Adds what decile (or alternatively different step) the variable `x` is in
+add_deciles <- function(x, step=0.1){
+decile  <- cut(x,
+               breaks=quantile(x,
+                               probs=seq(0,1,by=step),
+                               na.rm = TRUE),
+               include.lowest=TRUE,
+               labels=FALSE)
+return(decile)
+}
+
+# Regresses for tau in bins by decile and plots
+plot_tau <- function(df, tau_data, decile_name, variable_name, xlabel, eq='usage ~ PD_P1_UsageFITSBY + B + S', filename){
+  
+  tau_data$decile <- tau_data[[decile_name]]
+  df$decile       <- df[[decile_name]]
+  df$amount_var   <- df[[variable_name]]
+
+  taus <- c()
+  deciles <- sort(unique(tau_data$decile))
+  
+  formula <- eq
+  for (dec_idx in deciles){
+    var <- paste('L', dec_idx, sep="")
+    tau_data[[var]] <- ifelse(is.na(tau_data$decile),
+                                    0,
+                                    ifelse(tau_data$decile == dec_idx, 
+                                           tau_data$L, 
+                                           0))
+    formula <- paste(paste(formula, '+'), var)
+  }
+  fit <- lm(data    = tau_data, 
+            formula = formula) 
+
+  print(formula)
+  
+  for (dec_idx in deciles){
+    var <- paste('L', dec_idx, sep="")
+    taus <- c(taus, as.numeric(fit$coefficients[var]))
+  }
+  
+  decile_amnts <- df %>%
+                    group_by(decile) %>%
+                    summarize(amount = mean(amount_var), .groups = "drop")
+                    
+  plot_tau <- data.frame(taus)
+  plot_tau$decile <- deciles
+  
+  plot_tau %<>% left_join(decile_amnts,
+                          by ="decile",
+                          how="left")
+
+  print(plot_tau)
+  
+  a <-   ggplot(plot_tau, aes(x=amount, y=taus)) + 
+          geom_point(color=maroon) + 
+          theme_classic() +
+          geom_smooth(method = "lm",
+                      formula = "y ~ x",
+                      se = FALSE,
+                      color="black",
+                      size=0.6) +
+          labs(x = xlabel,
+               y = "Tau L")
+
+  ggsave(sprintf('output/%s.pdf', filename), plot=a, width=6.5, height=4.5, units="in")
+}
+
+# Binscatter by zkashner (more or less)
+plot_value <- function(df, decile_name, variable_name, variable_amount, xlabel, ylabel, filename){
+  
+  df$decile     <- df[[decile_name]]
+  df$value_var  <- df[[variable_name]]
+  df$amount_var <- df[[variable_amount]]
+  
+  values <- c()
+  deciles <- unique(df$decile)
+  
+  for (dec_idx in deciles){
+    subset <- df$decile == dec_idx
+    value  <- mean(df[subset,]$value_var, na.rm = T) / num_days
+    values <- c(values, value)
+  }
+  
+  decile_amnts <- df %>%
+                    group_by(decile) %>%
+                    summarize(amount = mean(amount_var), .groups = "drop")
+                    
+  plot_value <- data.frame(values)
+  plot_value$decile <- deciles
+  
+  plot_value %<>% merge(decile_amnts,
+                        by ="decile",
+                        how="left")
+
+  print(plot_value)
+  
+  a <- ggplot(plot_value, aes(x=amount, y=values)) + 
+        geom_point(color=maroon) + 
+        theme_classic() +
+        geom_smooth(method = "lm",
+                    formula = "y ~ x",
+                    se = FALSE,
+                    color="black",
+                    size=0.6) +
+        labs(x = xlabel,
+             y = ylabel)
+
+  ggsave(sprintf('output/%s.pdf', filename), plot=a, width=6.5, height=4.5, units="in")
+}
+
+reshape_tau_data <- function(df){
+  tau_data <- df %>% 
+    select(
+      UserID, 
+      w,
+      L, 
+      B, 
+      S, 
+      addiction_decile,
+      restriction_decile,
+      tightness_decile,
+      PD_P1_UsageFITSBY,
+      PD_P2_UsageFITSBY,
+      PD_P3_UsageFITSBY,
+      PD_P4_UsageFITSBY,
+      PD_P5_UsageFITSBY
+    )
+
+  tau_data %<>% 
+    gather(
+      key = 'period', 
+      value = 'usage', 
+      -UserID, 
+      -w,
+      -L, 
+      -B, 
+      -S, 
+      -PD_P1_UsageFITSBY,
+      -addiction_decile,
+      -restriction_decile,
+      -tightness_decile
+    )
+
+  return(tau_data)
+}
+
+reshape_tightness <- function(df){
+
+  pt1_usage <- df %>%
+                 select(
+                     UserID, 
+                     paste('PD_DailyUsage_', 1:10, sep="")) %>%
+                 gather(
+                   key = 'period', 
+                   value = 'usage', 
+                   -UserID,
+                 ) %>%
+                 group_by(UserID) %>%
+                 summarize(PD_P1_PT1_UsageFITSBY = mean(usage, na.rm=TRUE), .groups = "drop")
+
+  pt2_usage <- df %>%
+                 select(
+                     UserID, 
+                     paste('PD_DailyUsage_', 11:20, sep="")) %>%
+                 gather(
+                   key = 'period', 
+                   value = 'usage', 
+                   -UserID, 
+                 ) %>%
+                 group_by(UserID) %>%
+                 summarize(PD_P1_PT2_UsageFITSBY = mean(usage, na.rm=TRUE), .groups = "drop")
+   
+  tightness_df <- df %>% 
+                    select(
+                           UserID, 
+                           w, L, B, S, 
+                           tightness_decile) %>%
+                    merge(pt1_usage, how="left", on="UserID") %>% 
+                    merge(pt2_usage, how="left", on="UserID")
+
+  return(tightness_df)
+}
+
+reshape_mispredict <- function(df){
+mpd_df <- df %>%
+            mutate(Mispredict_P2_S2  = PD_P2_UsageFITSBY - S2_PredictUseNext_1_W) %>%
+            mutate(Mispredict_P3_S2  = PD_P3_UsageFITSBY - S2_PredictUseNext_2_W) %>%
+            mutate(Mispredict_P4_S2  = PD_P4_UsageFITSBY - S2_PredictUseNext_3_W) %>%
+            mutate(Mispredict_S2     = (Mispredict_P2_S2 + Mispredict_P3_S2 + Mispredict_P4_S2)/3) %>%
+            mutate(mispredict_decile = add_deciles(Mispredict_P2_S2)) %>%
+            mutate(Mispredict_P3_S3  = PD_P3_UsageFITSBY - S3_PredictUseNext_1_W) %>%
+            mutate(Mispredict_P4_S3  = PD_P4_UsageFITSBY - S3_PredictUseNext_2_W) %>%
+            mutate(Mispredict_P5_S3  = PD_P5_UsageFITSBY - S3_PredictUseNext_3_W) %>%
+            mutate(Mispredict_S3     = (Mispredict_P3_S3 + Mispredict_P4_S3 + Mispredict_P5_S3)/3) %>%
+            mutate(Mispredict_P4_S4  = PD_P4_UsageFITSBY - S4_PredictUseNext_1_W) %>%
+            mutate(Mispredict_P5_S4  = PD_P5_UsageFITSBY - S4_PredictUseNext_2_W) %>%
+            mutate(Mispredict_S4     = (Mispredict_P4_S4 + Mispredict_P5_S4)/2) %>%
+            mutate(Mispredict_S34    = (3*Mispredict_S3 + 2*Mispredict_S4)/5) %>% #reweight 
+            select(UserID, w, mispredict_decile, Mispredict_P2_S2, Mispredict_S2, Mispredict_S3, Mispredict_S4, Mispredict_S34)
+
+  return(mpd_df)
+}
+
+plot_taus <- function(df, tau_data, tightness_df){
+  plot_tau(df,
+           tau_data,
+           decile_name  = 'addiction_decile', 
+           variable_name = 'StratAddictionLifeIndex', 
+           xlabel = "Addiction Index",
+           filename = "binscatter_heterogeneity_tau_addiction")
+
+  plot_tau(df,
+           tau_data,
+           decile_name  = 'restriction_decile', 
+           variable_name = 'StratWantRestrictionIndex', 
+           xlabel = "Restriction Index",
+           filename = "binscatter_heterogeneity_tau_restriction")
+
+  plot_tau(df,
+           tau_data,
+           decile_name  = 'tightness_decile', 
+           variable_name = 'PD_P2_LimitTightFITSBY', 
+           xlabel = "Limit Tightness",
+           filename = "binscatter_heterogeneity_tau_tightness")
+
+  plot_tau(df,
+           tightness_df,
+           decile_name  = 'tightness_decile', 
+           variable_name = 'PD_P2_LimitTightFITSBY', 
+           xlabel = "Limit Tightness",
+           eq     = 'PD_P1_PT2_UsageFITSBY ~ PD_P1_PT1_UsageFITSBY + B + S',
+           filename = "binscatter_heterogeneity_tau_tightness_placebo")
+}
+
+plot_valuations <- function(df){
+  vars      <- c('behavioral_change_premium', 'S3_MPLLimit')
+  names     <- c('Behavioral Change Premium', 'Limit Valuation')
+  file_exts <- c('behavioral_change_premium', 'v_L')
+
+  for (i in 1:2){
+    var_name <- vars[i]
+    ylabel   <- names[i]
+    file_ext <- file_exts[i]
+
+    plot_value(df,
+               decile_name = "addiction_decile",
+               variable_name = var_name,
+               variable_amount = "StratAddictionLifeIndex",
+               xlabel = "Addiction Index",
+               ylabel = ylabel,
+               filename = sprintf("binscatter_heterogeneity_%s_addiction", file_ext))
+
+    plot_value(df,
+               decile_name = "restriction_decile",
+               variable_name = var_name,
+               variable_amount = "StratWantRestrictionIndex",
+               xlabel = "Restriction Index",
+               ylabel = ylabel,
+               filename = sprintf("binscatter_heterogeneity_%s_restriction", file_ext))
+
+    plot_value(df,
+               decile_name = "tightness_decile",
+               variable_name = var_name,
+               variable_amount = "PD_P2_LimitTightFITSBY",
+               xlabel = "Limit Tightness",
+               ylabel = ylabel,
+               filename = sprintf("binscatter_heterogeneity_%s_tightness", file_ext))
+  }
+}
+
+plot_mispredict <- function(mpd_df){
+  plot_value(mpd_df,
+             decile_name = "mispredict_decile",
+             variable_name = "Mispredict_S34",
+             variable_amount = "Mispredict_P2_S2",
+             xlabel = "Survey 2 Misprediction (minutes/day)",
+             ylabel = "Surveys 3 and 4 Misprediction (minutes/day)",
+             filename = "binscatter_heterogeneity_misprediction")
+}
+
+find_tau_spec <- function(df){
+  
+  days_beg <- 1:10
+  days_end <- 11:20
+
+  tau_data <- df %>% 
+    mutate(tightness=ifelse(L,PD_P2_LimitTightFITSBY, 0)) %>%
+    mutate(PD_P1beg_Usage_FITSBY = 
+                                rowSums(.[paste0('PD_DailyUsageFITSBY_',days_beg)], na.rm=TRUE)/length(days_beg),
+           PD_P1end_Usage_FITSBY = 
+                                rowSums(.[paste0('PD_DailyUsageFITSBY_',days_end)], na.rm=TRUE)/length(days_end)) %>%
+    select(
+      UserID, 
+      w, L, B, S, 
+      PD_P1_UsageFITSBY,
+      PD_P1beg_Usage_FITSBY,
+      PD_P1end_Usage_FITSBY,
+      PD_P2_UsageFITSBY,
+      PD_P3_UsageFITSBY,
+      PD_P4_UsageFITSBY,
+      PD_P5_UsageFITSBY,
+      PD_P2_LimitTightFITSBY,
+      tightness
+    )  
+           
+
+fit_1 <-lm('PD_P1end_Usage_FITSBY ~ B + L + tightness + PD_P1beg_Usage_FITSBY + S',
+  data= tau_data, weights = w)
+
+cluster_se1 <- as.vector(summary(fit_1,cluster = c("UserID"))$coefficients[,"Std. Error"])
+
+# the last command prints the stargazer output (in this case as text)
+
+fit_2 <- lm('PD_P2_UsageFITSBY ~ B + L + tightness + PD_P1_UsageFITSBY+ S',
+  data= tau_data, weights = w)
+
+cluster_se2 <- as.vector(summary(fit_2,cluster = c("UserID"))$coefficients[,"Std. Error"])
+
+
+fit_3 <- lm('PD_P3_UsageFITSBY ~ B + L + tightness + PD_P1_UsageFITSBY  + S',
+  data=tau_data,weights = w)
+
+cluster_se3 <- as.vector(summary(fit_3,cluster = c("UserID"))$coefficients[,"Std. Error"])
+
+
+stargazer(fit_1, fit_2, fit_3,
+              omit.stat         = c("adj.rsq","f","ser"),
+              se = list(cluster_se1, cluster_se2, cluster_se3),
+              covariate.labels  = c("Bonus treatment", "Limit treatment", 
+                                    "Limit treatment $\\times$ period 2 limit tightness", 
+                                    "1st half of period 1 FITSBY use", "Period 1 FITSBY use"),
+              align             = TRUE, 
+              dep.var.labels.include = FALSE, 
+              column.labels    = c('\\shortstack{2nd half of period 1 \\\\ FITSBY use}',
+                                    '\\shortstack{Period 2 \\\\ FITSBY use}',
+                                    '\\shortstack{Period 3 \\\\ FITSBY use}'),
+              title             = "", 
+              omit              = c("Intercept", "S1", "S2", "S3", "S4",
+                                    "S5", "S6", "S7", "S8", "Constant"),
+              type              = "latex", 
+              omit.table.layout = "n",
+              float             = FALSE,
+              dep.var.caption   = "",
+              star.cutoffs      = NA,
+              out               = "output/heterogeneity_reg.tex"
+              )
+  
+  return() 
+}
+
+
+
+
+plot_weekly_effects <- function(df, filename1, filename2){
+  get_df <- function(df){
+  bonus_coefs <- c()
+  limit_coefs <- c()
+  bonus_lower <- c()
+  bonus_upper <- c()
+  limit_upper<- c()
+  limit_lower<- c()
+  
+  for (t in 4:15){
+    dep_var <- sprintf('PD_WeeklyUsageFITSBY_%s', t)
+    eq <- paste0(dep_var, '~ PD_WeeklyUsageFITSBY_3 + L + B + S')
+  
+  # Run regression
+    fit <- lm(data    = df, 
+              formula = eq,
+              weights = w) 
+    
+
+
+    bonus_coefs <- c(bonus_coefs, summary(fit)$coefficients[4,1])
+    limit_coefs <- c(limit_coefs, summary(fit)$coefficients[3,1])
+    bonus_lower <- c(bonus_lower, summary(fit, cluster= c("UserID"))$coefficients[4,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+    bonus_upper <- c(bonus_upper, summary(fit, cluster= c("UserID"))$coefficients[4,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+    limit_upper<- c(limit_upper, summary(fit, cluster= c("UserID"))$coefficients[3,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+    limit_lower<- c(limit_lower, summary(fit, cluster= c("UserID"))$coefficients[3,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+
+    }
+
+  weeklydataframe <- as.data.frame(cbind(bonus_coefs, limit_coefs, bonus_lower, 
+                                              bonus_upper, limit_lower, limit_upper ))
+
+
+  names(weeklydataframe) <- c("bonus_coefs", "limit_coefs", "bonus_lower", 
+                                              "bonus_upper", "limit_lower", "limit_upper")
+ 
+  
+  return(weeklydataframe)
+  }
+
+
+  df_weekly <- get_df(df)
+
+  x <- c('4', '5', '6', '7', '8', '9', '10', '11', '12', '13', '14', '15')
+  names <- factor(x, levels=x)
+    
+  weeklydf <- data.frame(names, df_weekly)
+
+  b <- ggplot(weeklydf, aes(x=names, width=.2)) +
+    geom_point(aes(y=bonus_coefs), colour=maroon, stat="identity") +
+    geom_errorbar(aes(ymin=bonus_upper, ymax=bonus_lower),  colour=maroon, stat="identity") +
+    scale_y_continuous(name="Treatment effect (minutes/day)") +
+    theme_classic() + 
+    #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+    labs(x = "Week of experiment")  +
+    theme(legend.text.align = 0,
+          legend.key.height = unit(1, "cm"),
+          legend.position="bottom") +
+    theme(legend.margin=margin(0,0,0,0),
+          legend.box.margin=margin(-10,-10,-10,-10))  +
+    theme(axis.text.x = element_text(colour="black"))  +
+    coord_cartesian(ylim = c(-70, 5)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.05, color="grey" ) 
+      )
+
+  l <- ggplot(weeklydf, aes(x=names, width=.2)) +
+    geom_point(aes(y=limit_coefs), colour=grey, stat="identity") +
+    geom_errorbar(aes(ymin=limit_upper, ymax=limit_lower),  colour=grey, stat="identity") +
+    scale_y_continuous(name="Treatment effect (minutes/day)") +
+    theme_classic() + 
+    #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+    labs(x = "Week of experiment")  +
+    theme(legend.text.align = 0,
+          legend.key.height = unit(1, "cm"),
+          legend.position="bottom") +
+    theme(legend.margin=margin(0,0,0,0),
+          legend.box.margin=margin(-10,-10,-10,-10))  +
+    theme(axis.text.x = element_text(colour="black"))  +
+    coord_cartesian(ylim = c(-70, 5)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.05, color="grey" ) 
+      )
+
+
+  ggsave(sprintf('output/%s.pdf', filename1), plot=b, width=6.5, height=4.5, units="in")
+  ggsave(sprintf('output/%s.pdf', filename2), plot=l, width=6.5, height=4.5, units="in")
+  }
+
+
+
+
+
+
+plot_treatment_effects <- function(df, filename1, filename2, filename3){
+  period_usage <- c("PD_P2_UsageFITSBY", "PD_P3_UsageFITSBY", "PD_P4_UsageFITSBY", "PD_P5_UsageFITSBY") 
+
+  bonus_coefs <- c()
+  limit_coefs <- c()
+  bonus_lower <- c()
+  bonus_upper <- c()
+  limit_upper<- c()
+  limit_lower<- c()
+
+  for (period in period_usage){
+      dep_var <- period
+      eq <- paste0(dep_var, '~ PD_P1_UsageFITSBY + L + B + S')
+      
+       fit <- lm(data    = df, 
+              formula = eq,
+              weights = w) 
+
+      bonus_coefs <- c(bonus_coefs, summary(fit)$coefficients[4,1])
+      bonus_lower <- c(bonus_lower, summary(fit, cluster= c("UserID"))$coefficients[4,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+      bonus_upper <- c(bonus_upper, summary(fit, cluster= c("UserID"))$coefficients[4,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+      
+      limit_coefs <- c(limit_coefs, summary(fit)$coefficients[3,1])
+      limit_lower <- c(limit_lower, summary(fit, cluster= c("UserID"))$coefficients[3,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      limit_upper <- c(limit_upper, summary(fit, cluster= c("UserID"))$coefficients[3,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      
+  }
+  
+  x <- c('Period 2', 'Period 3', 'Period 4', 'Period 5')
+  names <- factor(x, levels=x)
+  periodtreatments <- data.frame(names, bonus_coefs, limit_coefs, bonus_lower, 
+                                                bonus_upper, limit_lower, limit_upper )
+
+
+
+ cols <- c("Bonus"=maroon ,
+            "Limit"=grey)
+  
+  cols_shape <- c("Bonus"=16 ,
+              "Limit"=15)
+
+  a <- ggplot(periodtreatments, aes(x=names, width=.2)) +
+    geom_point(aes(y=bonus_coefs, colour="Bonus"), stat="identity", position = position_nudge(x = -.1)) +
+    geom_point(aes(y=limit_coefs, colour="Limit"), stat="identity", position = position_nudge(x = .1))+
+    geom_errorbar(aes(ymin=bonus_upper, ymax=bonus_lower, width=0.05),  stat="identity", colour=maroon, position = position_nudge(x = -.1)) +
+    geom_errorbar(aes(ymin=limit_lower, ymax=limit_upper, width=0.05), stat="identity", colour=grey, position=position_nudge(x = .1)) +
+    scale_y_continuous(name="Treatment effect (minutes/day)") +
+    theme_classic() + 
+    #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+    labs(x = "")  +
+    theme(legend.text.align = 0,
+          legend.key.height = unit(1, "cm"),
+          legend.position="bottom") +
+    theme(legend.margin=margin(0,0,0,0),
+          legend.box.margin=margin(-10,-10,-10,-10))  +
+    theme(axis.text.x = element_text(colour="black"))  +
+    coord_cartesian(ylim = c(-70, 5)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.05, color="grey" ) 
+      )+ 
+    scale_colour_manual(name = "", values=cols, 
+                        labels = c("Bonus", "Limit")) + 
+    guides(colour=guide_legend(title.position="top", 
+                                       title.hjust =0.5))
+
+
+  b <- ggplot(periodtreatments, aes(x=names, width=.2)) +
+    geom_point(aes(y=bonus_coefs), colour=maroon, stat="identity") +
+    geom_errorbar(aes(ymin=bonus_upper, ymax=bonus_lower, width=0.05),  colour=maroon, stat="identity") +
+    scale_y_continuous(name="Treatment effect (minutes/day)") +
+    theme_classic() + 
+    #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+    labs(x = "")  +
+    theme(legend.text.align = 0,
+          legend.key.height = unit(1, "cm"),
+          legend.position="bottom") +
+    theme(legend.margin=margin(0,0,0,0),
+          legend.box.margin=margin(-10,-10,-10,-10))  +
+    theme(axis.text.x = element_text(colour="black"))  +
+    coord_cartesian(ylim = c(-70, 5)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.05, color="grey" ) 
+      )
+
+
+
+  l <- ggplot(periodtreatments, aes(x=names, width=.2)) +
+    geom_point(aes(y=limit_coefs), colour=grey, stat="identity") +
+    geom_errorbar(aes(ymin=limit_upper, ymax=limit_lower, width=0.05), colour=grey,  stat="identity") +
+    scale_y_continuous(name="Treatment effect (minutes/day)") +
+    theme_classic() + 
+    #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+    labs(x = "")  +
+    theme(legend.text.align = 0,
+          legend.key.height = unit(1, "cm"),
+          legend.position="bottom") +
+    theme(legend.margin=margin(0,0,0,0),
+          legend.box.margin=margin(-10,-10,-10,-10))  +
+    theme(axis.text.x = element_text(colour="black"))  +
+    coord_cartesian(ylim = c(-70, 5)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.05, color="grey" ) 
+      )
+
+ggsave(sprintf('output/%s.pdf', filename1), plot=a, width=6.5, height=4.5, units="in")
+ggsave(sprintf('output/%s.pdf', filename2), plot=b, width=6.5, height=4.5, units="in")
+ggsave(sprintf('output/%s.pdf', filename3), plot=l, width=6.5, height=4.5, units="in")
+
+  
+}
+
+plot_treatment_effects_interaction <- function(df, filename1){
+    period_usage <- c("PD_P2_UsageFITSBY", "PD_P3_UsageFITSBY", "PD_P4_UsageFITSBY", "PD_P5_UsageFITSBY") 
+
+  bonus_coefs <- c()
+  limit_coefs <- c()
+  bonus_lower <- c()
+  bonus_upper <- c()
+  limit_upper<- c()
+  limit_lower<- c()
+  interaction_coefs <- c()
+  interaction_lower <- c() 
+  interaction_upper <- c()
+
+  for (period in period_usage){
+      dep_var <- period
+      eq <- paste0(dep_var, '~ PD_P1_UsageFITSBY + L + B + L*B + S')
+      
+       fit <- lm(data    = df, 
+              formula = eq,
+              weights = w) 
+
+      bonus_coefs <- c(bonus_coefs, summary(fit)$coefficients[4,1])
+      bonus_lower <- c(bonus_lower, summary(fit, cluster= c("UserID"))$coefficients[4,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+      bonus_upper <- c(bonus_upper, summary(fit, cluster= c("UserID"))$coefficients[4,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[4,2])
+      
+      limit_coefs <- c(limit_coefs, summary(fit)$coefficients[3,1])
+      limit_lower <- c(limit_lower, summary(fit, cluster= c("UserID"))$coefficients[3,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      limit_upper <- c(limit_upper, summary(fit, cluster= c("UserID"))$coefficients[3,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      
+      interaction_coefs <- c(interaction_coefs, summary(fit)$coefficients[12,1])
+      interaction_lower <- c(interaction_lower, summary(fit, cluster= c("UserID"))$coefficients[12,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[12,2])
+      interaction_upper <- c(interaction_upper, summary(fit, cluster= c("UserID"))$coefficients[12,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[12,2])
+      
+  }
+  
+
+
+x <- c('Period 2', 'Period 3', 'Period 4', 'Period 5')
+names <- factor(x, levels=x)
+
+periodtreatments <- data.frame(names, bonus_coefs, bonus_lower, bonus_upper, limit_coefs, limit_lower, limit_upper,interaction_coefs, interaction_lower, interaction_upper)
+
+periodtreatments$bonus <- "Bonus"
+periodtreatments$limit <- "Limit"
+periodtreatments$BL <- "Limit x Bonus"
+
+
+maroon <- '#94343c'
+grey   <- '#848484'
+skyblue <- '#87CEEB'
+black  <- '#000000'
+deepskyblue <- '#B0C4DE'
+
+ cols <- c("Bonus"=maroon ,
+           "Limit"=grey,
+           "Limit x Bonus"= deepskyblue)
+
+cols_shape <- c("Bonus"=15 ,
+           "Limit"=19,
+           "Limit x Bonus"= 17)
+
+a <- ggplot(periodtreatments, aes(x=names, width=.2)) +
+  geom_point(aes(y=bonus_coefs, colour=bonus, shape =bonus),  stat="identity", position = position_nudge(x = -.2)) +
+  geom_point(aes(y=limit_coefs, colour=limit, shape=limit), stat="identity", position = position_nudge(x = 0))+
+  geom_point(aes(y=interaction_coefs, colour=BL, shape=BL),  stat="identity", position = position_nudge(x = 0.2)) +
+  geom_errorbar(aes(ymin=bonus_upper, ymax=bonus_lower, width=0.05),  stat="identity", colour=maroon, position = position_nudge(x = -.2)) +
+  geom_errorbar(aes(ymin=limit_lower, ymax=limit_upper, width=0.05), stat="identity", colour=grey, position=position_nudge(x =0)) +
+  geom_errorbar(aes(ymin=interaction_lower, ymax=interaction_upper, width=0.05), stat="identity", colour=deepskyblue, position=position_nudge(x =0.2)) +
+  scale_y_continuous(name="Treatment effect (minutes/day)") +
+  theme_classic() + 
+  #theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
+  labs(x = "")  +
+  theme(legend.text.align = 0,
+        legend.key.height = unit(1, "cm"),
+        legend.position="bottom") +
+  theme(legend.margin=margin(0,0,0,0),
+        legend.box.margin=margin(-10,-10,-10,-10))  +
+  theme(axis.text.x = element_text(colour="black"))  +
+  coord_cartesian(ylim = c(-80, 20)) +
+  theme(legend.text=element_text(size=11))  +
+    theme( # remove the vertical grid lines
+           panel.grid.major.x = element_blank() ,
+           # explicitly set the horizontal lines (or they will disappear too)
+           panel.grid.major.y = element_line( size=.05, color="grey" ) 
+    )+ 
+  scale_colour_manual(name = "", 
+                      values=cols) + 
+  scale_shape_manual(name = "",
+                     values = cols_shape)  
+
+ggsave(sprintf('output/%s.pdf', filename1), plot=a, width=6.5, height=4.5, units="in")
+
+  
+}
+
+get_opt <- function(df) {
+  # Specify regression
+  
+    analysisUser <- read_dta("input/AnalysisUser.dta") 
+
+    limit <- analysisUser %>%
+      filter(AppCode %in% df$UserID) %>%
+      select(OptedOut) %>%
+      filter(!is.na(OptedOut))
+
+
+    estimate <- 
+        list(nrow(limit %>% filter(OptedOut==1)) ,
+            signif(nrow(limit %>% filter(OptedOut==1))/ nrow(df %>% filter(L==1))*100, digits=1))
+
+    names(estimate) <- c('numberpeopleoptedout', 'percentoptedout')
+
+    save_nrow(estimate, filename ="optingout", suffix="")
+}
+
+
+get_addiction_treatment_effect <- function(df, filename){
+  survey_outcomes <- c("index_well_N", "SWBIndex_N", "LifeBetter_N", "SMSIndex_N", "AddictionIndex_N", "PhoneUseChange_N") 
+
+  bonus_coefs <- c()
+  limit_coefs <- c()
+  bonus_lower <- c()
+  bonus_upper <- c()
+  limit_upper<- c()
+  limit_lower<- c()
+  
+  
+  df <- df %>%
+    mutate(S43_PhoneUseChange_N = (S4_PhoneUseChange_N + S3_PhoneUseChange_N)/2,
+           S43_AddictionIndex_N = (S4_AddictionIndex_N + S3_AddictionIndex_N)/2,
+           S43_SMSIndex_N = (S4_SMSIndex_N + S3_SMSIndex_N)/2,
+           S43_LifeBetter_N = (S4_LifeBetter_N + S3_LifeBetter_N)/2,
+           S43_SWBIndex_N = (S4_SWBIndex_N + S3_SWBIndex_N)/2,
+           S43_index_well_N = (S4_index_well_N + S3_index_well_N)/2)
+
+
+  for (outcome in survey_outcomes){
+      dep_var <- sprintf("S4_%s", outcome)
+      indep_var <- sprintf("S1_%s", outcome)
+      eq <- paste0(paste0(dep_var, '~ L + B + S + '), indep_var)
+      
+       fit <- lm(data    = df, 
+              formula = eq,
+              weights = w) 
+
+      bonus_coefs <- c(bonus_coefs, summary(fit)$coefficients[3,1])
+      bonus_lower <- c(bonus_lower, summary(fit, cluster= c("UserID"))$coefficients[3,1] -1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      bonus_upper <- c(bonus_upper, summary(fit, cluster= c("UserID"))$coefficients[3,1] +1.96*summary(fit, cluster= c("UserID"))$coefficients[3,2])
+      
+      
+      dep_var_limit <- sprintf("S43_%s", outcome)
+      indep_var <- sprintf("S1_%s", outcome)
+      eq_limit <- paste0(paste0(dep_var_limit, '~ L + B + S + '), indep_var)
+      
+       fit_limit <- lm(data    = df, 
+              formula = eq_limit,
+              weights = w) 
+
+      limit_coefs <- c(limit_coefs, summary(fit_limit)$coefficients[3,1])
+      limit_lower <- c(limit_lower, summary(fit_limit, cluster= c("UserID"))$coefficients[3,1] -1.96*summary(fit_limit, cluster= c("UserID"))$coefficients[3,2])
+      limit_upper <- c(limit_upper, summary(fit_limit, cluster= c("UserID"))$coefficients[3,1] +1.96*summary(fit_limit, cluster= c("UserID"))$coefficients[3,2])
+      
+  }
+    
+  weeklydataframe <- as.data.frame(cbind(bonus_coefs, limit_coefs, bonus_lower, 
+                                                bonus_upper, limit_lower, limit_upper ))
+    
+  names(weeklydataframe) <- c("bonus_coefs", "limit_coefs", "bonus_lower", 
+                                                "bonus_upper", "limit_lower", "limit_upper")
+   
+    
+
+  x <- c('Survey index', 'Subjective well-being', 'Phone makes life better', 'SMS addiction scale x (-1)', 'Addiction scale x(-1)',   'Ideal use change')
+    names <- factor(x, levels=x)
+      
+    weeklydf <- data.frame(names, weeklydataframe)
+    
+    
+    cols <- c("Bonus"=maroon,
+              "Limit"=grey)
+
+  a <- ggplot(weeklydf, aes(x=names, width=.2)) +
+    geom_point(aes(y=bonus_coefs, colour="Bonus"), stat="identity", position = position_nudge(x = -.1)) +
+    geom_point(aes(y=limit_coefs, colour="Limit"), stat="identity", position = position_nudge(x = .1))+
+    geom_errorbar(aes(ymin=bonus_upper, ymax=bonus_lower, width=0.05),  stat="identity", colour=maroon, position = position_nudge(x = -.1)) +
+    geom_errorbar(aes(ymin=limit_lower, ymax=limit_upper, width=0.05), stat="identity", colour=grey, position=position_nudge(x = .1)) +
+    scale_y_continuous(name="Treatment effect (standard deviation)") +
+    theme_classic() +
+    scale_colour_manual(name = "", values=cols, 
+                        labels = c("Bonus", "Limit")) +
+    labs(x = "") +
+    geom_hline(yintercept=0) + 
+    coord_flip(ylim = c(-0.2,0.6)) +
+    theme(legend.text=element_text(size=11))  +
+      theme( # remove the vertical grid lines
+             panel.grid.major.x = element_blank() ,
+             # explicitly set the horizontal lines (or they will disappear too)
+             panel.grid.major.y = element_line( size=.09, color="grey" ) 
+      )  +
+    theme(legend.position="bottom")
+
+  ggsave(sprintf('output/%s.pdf', filename), plot=a, width=6.5, height=4.5, units="in")  
+}
+
+
+get_addiction_scalar <- function(df){
+  addiction <- df %>%
+  select(contains("Addiction"))
+
+    df_addiction <- df 
+    for (i in 1:16){
+      df_addiction <- df_addiction %>%
+        mutate(!!as.name(paste0("High_S1_Addiction_", i)) := ifelse(!!as.name(paste0("S1_Addiction_",i))>0.5, 1, 0)) %>%
+        mutate(!!as.name(paste0("High_S3_Addiction_", i)) := ifelse(!!as.name(paste0("S3_Addiction_",i))>0.5, 1, 0))
+    }
+
+    df_means <- df_addiction
+    for (i in 1:16){
+      df_means <- df_means %>%
+        mutate(!!as.name(paste0("Mean_High_S1_Addiction_", i)) := mean(!!as.name(paste0("High_S1_Addiction_",i)), na.rm = T)) %>%
+        mutate(!!as.name(paste0("Mean_High_S3_Addiction_", i)) := mean(!!as.name(paste0("High_S3_Addiction_",i)), na.rm = T))
+    }
+
+    df_S3_addiction <- df_means %>%
+      select(contains("Mean_High_S3")) %>%
+      unique()  
+
+
+    df_S3_addiction$top_seven <- rowMeans(df_S3_addiction[1:7], na.rm=TRUE)
+    df_S3_addiction$bottom_nine <- rowMeans(df_S3_addiction[8:16], na.rm=TRUE)
+
+    mean_top_seven <- (df_S3_addiction$top_seven)*100
+    mean_bottom_nine <- (df_S3_addiction$bottom_nine)*100
+
+    df_addiction_high <- df_addiction %>%
+      select(contains("High_S3_Addiction_")) 
+
+    df_addiction_high$top_seven_any <- rowSums(df_addiction_high[1:7], na.rm=TRUE)
+    df_addiction_high$bottom_nine_any <- rowSums(df_addiction_high[8:16], na.rm=TRUE)
+
+    df_addiction_high <-df_addiction_high %>%
+      mutate(top_seven_any_indicator = ifelse(top_seven_any>0, 1, 0),
+             bottom_nine_any_indicator = ifelse(bottom_nine_any>0, 1, 0))
+
+    mean_top_seven_any <- mean(df_addiction_high$top_seven_any_indicator, na.rm=TRUE)*100 
+    mean_bottom_nine_any <- mean(df_addiction_high$bottom_nine_any_indicator, na.rm=TRUE)*100
+
+    mean_top_seven <- signif(mean_top_seven, digits=2)
+    mean_top_seven_any <- signif(mean_top_seven_any, digits=2)
+    mean_bottom_nine  <- signif(mean_bottom_nine, digits=2)
+    mean_bottom_nine_any  <- signif(mean_bottom_nine_any, digits=2)
+
+  limit_tightness_df <- df %>%
+    filter(L==1) %>%
+    select(contains("PD_P5432_LimitTight"))
+
+  limit_tightness_df_nomissing <- df %>%
+    filter(L==1) %>%
+    filter(PD_P5432_LimitTight>0)
+
+  limit_tightness_df_pfive <- df %>%
+    filter(L==1) %>%
+    select(contains("PD_P5_LimitTight"))
+
+  limit_tightness_df_nomissing_pfive <- df %>%
+    filter(L==1) %>%
+    filter(PD_P5_LimitTight>0)
+
+  percent_positive_tightness <- nrow(limit_tightness_df_nomissing) / nrow(limit_tightness_df)
+  percent_positive_tightness_pfive <- nrow(limit_tightness_df_nomissing_pfive) / nrow(limit_tightness_df_pfive)
+
+  average_limit_tightness <- mean(limit_tightness_df$PD_P5432_LimitTight, na.rm=TRUE)
+  average_limit_tightness_pfive <- mean(limit_tightness_df_pfive$PD_P5_LimitTight, na.rm=TRUE)
+
+  percentpositivetightness <- signif(percent_positive_tightness, digits=2)*100
+  percentpositivetightnesspfive <- signif(percent_positive_tightness_pfive, digits=2)*100
+  
+  averagelimittightness <- signif(average_limit_tightness, digits=2)
+  averagelimittightnesspfive <- signif(average_limit_tightness_pfive, digits=2)
+
+
+  limit_df <- df %>%
+  filter(L==1) %>%
+  select(PD_P5432_LimitTight_Facebook, PD_P5432_LimitTight_Browser, PD_P5432_LimitTight_YouTube, PD_P5432_LimitTight_Instagram)
+
+
+  limit_df[is.na(limit_df)] <- 0
+
+  mean_fb <- mean(limit_df$PD_P5432_LimitTight_Facebook)
+  mean_browser <- mean(limit_df$PD_P5432_LimitTight_Browser)
+  mean_youtube <- mean(limit_df$PD_P5432_LimitTight_YouTube)
+  mean_insta <- mean(limit_df$PD_P5432_LimitTight_Instagram)
+
+  mean_insta_nice <- signif(mean_insta, digits=1)
+  mean_youtube_nice <- signif(mean_youtube, digits=1)
+  mean_browser_nice <- signif(mean_browser, digits=1)
+  mean_fb_nice <- signif(mean_fb, digits=1)
+
+  mpl_df <- df %>%
+    filter(B==1) %>%
+    select(S2_PredictUseInitial, S2_PredictUseBonus)
+
+    mean_initial_use <- mean(mpl_df$S2_PredictUseInitial, na.rm=TRUE)/60
+    mean_use_bonus <- mean(mpl_df$S2_PredictUseBonus, na.rm=TRUE)/60
+
+
+    df <- df %>%
+        mutate(F_B_uncensored = 50*PD_P1_UsageFITSBY/20) %>%
+        mutate(F_B_min = ifelse(F_B_uncensored<150, F_B_uncensored, 150)) %>%
+        mutate(F_B     = F_B_min/num_days)
+
+    FB <- mean(df$F_B, na.rm = T)
+
+    num_days    <- 20
+    hourly_rate <- 50
+    max_hours   <- 3
+    p_B   <- (hourly_rate / num_days)  
+    abcd <- p_B*0.5*(mean_use_bonus+mean_initial_use)
+    MPL <- FB - abcd
+
+    MPLearningsmean <- MPL*20
+
+    MPLvalued <- signif(mean(df$S2_MPL, na.rm=T), digits=2)
+    MPLearningsnice <- signif(MPLearningsmean, digits=2)
+    MPLpremiumnice <- MPLvalued -  MPLearningsnice
+
+    p <-paste0(p_B, "0")
+    meanpredictuse <- signif(mean_initial_use, digits=2)
+    meanpredictbonus <- signif(mean_use_bonus, digits=2)
+    abcd <- signif(abcd, digits=3)
+    MPL <- signif(MPL, digits=3)
+    vB <- MPLvalued/20
+
+    behaviourpremium <- vB - MPL
+
+    fit_3 <- lm(data=df, PD_P5432_Usage_Other ~ PD_P1_Usage_Other + L + B + S)
+    limitotherfitsby <- fit_3$coefficients[['L']]
+    limitotherfitsbynice <- signif(limitotherfitsby, digits=2)
+
+
+    estimate <- 
+      list(mean_top_seven, mean_top_seven_any, mean_bottom_nine, mean_bottom_nine_any, 
+        percentpositivetightness, averagelimittightness, percentpositivetightnesspfive, averagelimittightnesspfive, mean_insta_nice, mean_youtube_nice,  mean_browser_nice, 
+        mean_fb_nice, MPLvalued, MPLearningsnice, MPLpremiumnice,
+        p,meanpredictuse,meanpredictbonus, abcd, MPL,  behaviourpremium, limitotherfitsbynice)
+      names(estimate) <- c('meantopsevenaddiction', 'meantopsevenanyaddiction', 'meanbottomnineaddiction', 
+        'meanbottomnineanyaddiction', 'percentpositivetightness', 'averagelimittightness', 'percentpositivetightnesspfive', 'averagelimittightnesspfive', 'instalimittight', 'youtubelimittight',
+        'browserlimittight', 'fblimittight', 
+        'MPLvalued', 'MPLearningsnice', 'MPLpremiumnice','p','meanpredictuse','meanpredictbonus', 'abcd', 'MPL',  'behaviourpremium', 'limitotherfitsbynice')
+      
+    save_nrow(estimate, filename ="addiction_scalars", suffix="")
+
+
+  
+}
+
+get_swb_effect_exported_limit <- function(df){
+  df <- df %>%
+  mutate( S43_PhoneUseChange_N = (S4_PhoneUseChange_N + S3_PhoneUseChange_N)/2,
+          S43_AddictionIndex_N = (S4_AddictionIndex_N + S3_AddictionIndex_N)/2,
+          S43_SMSIndex_N = (S4_SMSIndex_N + S3_SMSIndex_N)/2,
+          S43_LifeBetter_N = (S4_LifeBetter_N + S3_LifeBetter_N)/2,
+          S43_SWBIndex_N = (S4_SWBIndex_N + S3_SWBIndex_N)/2,
+          S43_index_well_N = (S4_index_well_N + S3_index_well_N)/2 ,
+          S43_PhoneUseChange = (S4_PhoneUseChange + S3_PhoneUseChange)/2,
+          S43_AddictionIndex = (S4_AddictionIndex + S3_AddictionIndex)/2,
+          S43_SMSIndex = (S4_SMSIndex + S3_SMSIndex)/2,
+          S43_LifeBetter = (S4_LifeBetter + S3_LifeBetter)/2,
+          S43_SWBIndex = (S4_SWBIndex + S3_SWBIndex)/2,
+          S43_index_well= (S4_index_well + S3_index_well)/2)
+
+
+  fit<- lm_robust(data=df, formula = S43_PhoneUseChange_N ~ S1_PhoneUseChange_N + B+ L+ S, cluster=UserID )
+
+  estimate   <- list (fit$coefficients[['L']])
+  names(estimate) <- c('limitidealcoefn')
+  se   <- list (summary(fit)$coefficients[4,2])
+  names(se) <- c('limitidealsen')
+  pval <- list (summary(fit)$coefficients[4,4])
+  names(pval) <- c('pvallimitideal')
+  p_value_list <- fit[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvallimit <- list(p_adj[4])
+  names(pvallimit) <- 'qadjlimitphonechange'
+
+
+  fit1 <- lm_robust(data=df, formula = S43_PhoneUseChange ~ S1_PhoneUseChange + B+ L+ S, cluster=UserID )
+  estimate1   <- list (fit1$coefficients[['L']])
+  names(estimate1) <- c('limitidealcoef')
+  se1   <- list (summary(fit1)$coefficients[4,2])
+  names(se1) <- c('limitidealse')
+
+  fit2 <- lm_robust(data=df, formula = S43_AddictionIndex_N ~ S1_AddictionIndex_N + B+ L+ S, cluster=UserID )
+  estimate2   <- list (fit2$coefficients[['L']])
+  names(estimate2) <- c('limitaddictioncoefn')
+  se2   <- list (summary(fit2)$coefficients[4,2])
+  names(se2) <- c('limitaddictionsen')
+  pval2 <- list (summary(fit2)$coefficients[4,4])
+  names(pval2) <- c('pvallimitaddict')
+  p_value_list <- fit2[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvallimit2 <- list(p_adj[4])
+  names(pvallimit2) <- 'qadjlimitaddictionindex'
+
+  fit3 <-  lm_robust(data=df, formula = S43_AddictionIndex ~ S1_AddictionIndex + B+ L+ S, cluster=UserID )
+  estimate3   <- list (fit3$coefficients[['L']])
+  names(estimate3) <- c('limitaddictioncoef')
+  se3   <- list (summary(fit3)$coefficients[4,2])
+  names(se3) <- c('limitaddictionse')
+
+  fit4 <- lm_robust(data=df, formula = S43_SMSIndex_N ~ S1_SMSIndex_N + B+ L+ S, cluster=UserID )
+  estimate4   <- list (fit4$coefficients[['L']])
+  names(estimate4) <- c('limitsmscoefn')
+  se4   <- list (summary(fit4)$coefficients[4,2])
+  names(se4) <- c('limitsmssen')
+  pval4 <- list (summary(fit4)$coefficients[4,4])
+  names(pval4) <- c('pvallimitsmsindex')
+  p_value_list <- fit4[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvallimit3 <- list(p_adj[4])
+  names(pvallimit3) <- 'qadjlimitsmsindex'
+
+
+  fit5 <-  lm_robust(data=df, formula = S43_SMSIndex ~ S1_SMSIndex + B+ L+ S, cluster=UserID )
+  estimate5   <- list (fit5$coefficients[['L']])
+  names(estimate5) <- c('limitsmscoef')
+  se5   <- list (summary(fit5)$coefficients[4,2])
+  names(se5) <- c('limitsmsse')
+
+
+  fit6 <-  lm_robust(data=df, formula = S43_LifeBetter_N ~ S1_LifeBetter_N + B+ L+ S, cluster=UserID )
+  estimate6   <- list (fit6$coefficients[['L']])
+  names(estimate6) <- c('limitlifebettercoefn')
+  se6   <- list (summary(fit6)$coefficients[4,2])
+  names(se6) <- c('limitlifebettersen')
+  pval6 <- list (summary(fit6)$coefficients[4,4])
+  names(pval6) <- c('pvallimitlifebetter')
+  p_value_list <- fit6[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+
+  pvallimit4 <- list(p_adj[4])
+  names(pvallimit4) <- 'qadjlimitlifebetter'
+
+
+
+  fit7 <-  lm_robust(data=df, formula = S43_LifeBetter ~ S1_LifeBetter_N + B+ L+ S, cluster=UserID )
+  estimate7   <- list (fit7$coefficients[['L']])
+  names(estimate7) <- c('limitlifebettercoef')
+  se7   <- list (summary(fit7)$coefficients[4,2])
+  names(se7) <- c('limitlifebetterse')
+
+
+
+
+  fit8 <-  lm_robust(data=df, formula = S43_SWBIndex_N ~ S1_SWBIndex_N + B+ L+ S, cluster=UserID )
+  estimate8   <- list (fit8$coefficients[['L']])
+  names(estimate8) <- c('limitswbindexcoefn')
+  se8   <- list (summary(fit8)$coefficients[4,2])
+  names(se8) <- c('limitswbindexsen')
+  pval8 <- list (summary(fit8)$coefficients[4,4])
+  names(pval8) <- c('pvallimitswbindex')
+  p_value_list <- fit8[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvallimit5 <- list(p_adj[4])
+  names(pvallimit5) <- 'qadjlimitswbindex'
+
+
+  fit9 <-  lm_robust(data=df, formula = S43_SWBIndex ~ S1_SWBIndex + B+ L+ S, cluster=UserID )
+  estimate9   <- list (fit9$coefficients[['L']])
+  names(estimate9) <- c('limitswbindexcoef')
+  se9   <- list (summary(fit9)$coefficients[4,2])
+  names(se9) <- c('limitswbindexse')
+
+
+  fit10 <-  lm_robust(data=df, formula = S43_index_well_N ~ S1_index_well_N + B+ L+ S, cluster=UserID )
+  estimate10   <- list (fit10$coefficients[['L']])
+  names(estimate10) <- c('limitindexwellcoefn')
+  se10   <- list (summary(fit10)$coefficients[4,2])
+  names(se10) <- c('limitindexwellsen')
+  pval10 <- list (summary(fit10)$coefficients[4,4])
+  names(pval10) <- c('pvallimitindexwell')
+  p_value_list <- fit10[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvallimit6 <- list(p_adj[4])
+  names(pvallimit6) <- 'qadjlimitindexwell'
+
+  fit11 <-  lm_robust(data=df, formula = S43_index_well ~ S1_index_well + B+ L+ S, cluster=UserID )
+  estimate11   <- list (fit11$coefficients[['L']])
+  names(estimate11) <- c('limitindexwellcoef')
+  se11   <- list (summary(fit11)$coefficients[4,2])
+  names(se11) <- c('limitindexwellse')
+
+  limit_effect <- list.merge(estimate, estimate1, estimate2, estimate3, estimate4, estimate5, estimate6, 
+    estimate7, estimate8, estimate9, estimate10, estimate11, se, se1, se2, se3, se4, se5, se6, se7, se8,
+     se9, se10, se11, pval, pval2, pval4, pval6, pval8, pval10, pvallimit, pvallimit2, pvallimit3, pvallimit4,
+     pvallimit5, pvallimit6)
+
+  return(limit_effect)
+  
+}
+
+get_swb_effect_exported_bonus <- function(df){
+  fit<- lm_robust(data=df, formula = S4_PhoneUseChange_N ~ S1_PhoneUseChange_N + B+ L+ S, cluster=UserID )
+
+  estimate   <- list (fit$coefficients[['B']])
+  names(estimate) <- c('bonusidealcoefn')
+  se   <- list (summary(fit)$coefficients[3,2])
+  names(se) <- c('bonusidealsen')
+  pvaluebonusideal <- list (summary(fit)$coefficients[3,4])
+  names(pvaluebonusideal) <- c('pvaluebonusideal')
+  p_value_list <- fit[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus <- list(p_adj[3])
+  names(pvalbonus) <- 'qadjbonusphonechange'
+
+  fit1 <- lm_robust(data=df, formula = S4_PhoneUseChange ~ S1_PhoneUseChange + B+ L+ S, cluster=UserID )
+  estimate1   <- list (fit1$coefficients[['B']])
+  names(estimate1) <- c('bonusidealcoef')
+  se1   <- list (summary(fit1)$coefficients[3,2])
+  names(se1) <- c('bonusidealse')
+
+  fit2 <- lm_robust(data=df, formula = S4_AddictionIndex_N ~ S1_AddictionIndex_N + B+ L+ S, cluster=UserID )
+  estimate2   <- list (fit2$coefficients[['B']])
+  names(estimate2) <- c('bonusaddictioncoefn')
+  se2   <- list (summary(fit2)$coefficients[3,2])
+  names(se2) <- c('bonusaddictionsen')
+  pvaluebonusaddiction <- list (summary(fit2)$coefficients[3,4])
+  names(pvaluebonusaddiction) <- c('pvaluebonusaddiction')
+  p_value_list <- fit2[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus2 <- list(p_adj[3])
+  names(pvalbonus2) <- 'qadjbonusaddictionindex'
+
+  fit3 <-  lm_robust(data=df, formula = S4_AddictionIndex ~ S1_AddictionIndex + B+ L+ S, cluster=UserID )
+  estimate3   <- list (fit3$coefficients[['B']])
+  names(estimate3) <- c('bonusaddictioncoef')
+  se3   <- list (summary(fit3)$coefficients[3,2])
+  names(se3) <- c('bonusaddictionse')
+
+
+  fit4 <- lm_robust(data=df, formula = S4_SMSIndex_N ~ S1_SMSIndex_N + B+ L+ S, cluster=UserID )
+  estimate4   <- list (fit4$coefficients[['B']])
+  names(estimate4) <- c('bonussmscoefn')
+  se4   <- list (summary(fit4)$coefficients[3,2])
+  names(se4) <- c('bonussmssen')
+  pvaluebonussms <- list (summary(fit4)$coefficients[3,4])
+  names(pvaluebonussms) <- c('pvaluebonussms')
+  p_value_list <- fit4[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus3 <- list(p_adj[3])
+  names(pvalbonus3) <- 'qadjbonussmsnindex'
+
+
+  fit5 <-  lm_robust(data=df, formula = S4_SMSIndex ~ S1_SMSIndex + B+ L+ S, cluster=UserID )
+  estimate5   <- list (fit5$coefficients[['B']])
+  names(estimate5) <- c('bonussmscoef')
+  se5   <- list (summary(fit5)$coefficients[3,2])
+  names(se5) <- c('bonussmsse')
+
+
+  fit6 <-  lm_robust(data=df, formula = S4_LifeBetter_N ~ S1_LifeBetter_N + B+ L+ S, cluster=UserID )
+  estimate6   <- list (fit6$coefficients[['B']])
+  names(estimate6) <- c('bonuslifebettercoefn')
+  se6   <- list (summary(fit6)$coefficients[3,2])
+  names(se6) <- c('bonuslifebettersen')
+  pvaluebonuslifebetter <- list (summary(fit6)$coefficients[3,4])
+  names(pvaluebonuslifebetter) <- c('pvaluebonuslifebetter')
+  p_value_list <- fit6[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus4 <- list(p_adj[3])
+  names(pvalbonus4) <- 'qadjbonuslifebetter'
+
+  fit7 <-  lm_robust(data=df, formula = S4_LifeBetter ~ S1_LifeBetter_N + B+ L+ S, cluster=UserID )
+  estimate7   <- list (fit7$coefficients[['B']])
+  names(estimate7) <- c('bonuslifebettercoef')
+  se7   <- list (summary(fit7)$coefficients[3,2])
+  names(se7) <- c('bonuslifebetterse')
+
+
+
+  fit8 <-  lm_robust(data=df, formula = S4_SWBIndex_N ~ S1_SWBIndex_N + B+ L+ S, cluster=UserID )
+  estimate8   <- list (fit8$coefficients[['B']])
+  names(estimate8) <- c('bonusswbindexcoefn')
+  se8   <- list (summary(fit8)$coefficients[3,2])
+  names(se8) <- c('bonusswbindexsen')
+  pvaluebonusswbindex <- list (summary(fit8)$coefficients[3,4])
+  names(pvaluebonusswbindex) <- c('pvaluebonusswbindex')
+  p_value_list <- fit8[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus5 <- list(p_adj[3])
+  names(pvalbonus5) <- 'qadjbonusswbindex'
+
+
+
+  fit9 <-  lm_robust(data=df, formula = S4_SWBIndex ~ S1_SWBIndex + B+ L+ S, cluster=UserID )
+  estimate9   <- list (fit9$coefficients[['B']])
+  names(estimate9) <- c('bonusswbindexcoef')
+  se9   <- list (summary(fit9)$coefficients[3,2])
+  names(se9) <- c('bonusswbindexse')
+
+
+  fit10 <-  lm_robust(data=df, formula = S4_index_well_N ~ S1_index_well_N + B+ L+ S, cluster=UserID )
+  estimate10   <- list (fit10$coefficients[['B']])
+  names(estimate10) <- c('bonusindexwellcoefn')
+  se10   <- list (summary(fit10)$coefficients[3,2])
+  names(se10) <- c('bonusindexwellsen')
+  pvaluebonusindexwell <- list (summary(fit10)$coefficients[3,4])
+  names(pvaluebonusindexwell) <- c('pvaluebonusindexwell')
+  p_value_list <- fit10[5]
+  p_value <- p_value_list[['p.value']]
+
+  p_adj <- p.adjust(p_value, method = "BH")
+  pvalbonus6 <- list(p_adj[3])
+  names(pvalbonus6) <- 'qadjbonusindexwell'
+
+
+  fit11 <-  lm_robust(data=df, formula = S4_index_well ~ S1_index_well + B+ L+ S, cluster=UserID )
+  estimate11   <- list (fit11$coefficients[['B']])
+  names(estimate11) <- c('bonusindexwellcoef')
+  se11   <- list (summary(fit11)$coefficients[3,2])
+  names(se11) <- c('bonusindexwellse')
+
+  bonus_effect <- list.merge(estimate, estimate1, estimate2, estimate3, estimate4, estimate5, estimate6, estimate7, 
+    estimate8, estimate9, estimate10, estimate11, se, se1, se2, se3, se4, se5, se6, se7, se8, se9, se10, se11, 
+    pvaluebonusideal, pvaluebonusaddiction, pvaluebonussms, pvaluebonuslifebetter, pvaluebonusswbindex,
+    pvaluebonusindexwell, pvalbonus, pvalbonus2, pvalbonus3, pvalbonus4, pvalbonus5, pvalbonus6)
+
+  return(bonus_effect)
+  
+}
+
+
+plot_histogram_predicted <- function(df, filename){
+  df_usage_predict <- df %>%
+  filter(B==0 & L==0) %>%
+  select(PD_P2_UsageFITSBY, PD_P3_UsageFITSBY, PD_P4_UsageFITSBY,
+         S2_PredictUseNext_1, S3_PredictUseNext_1, S4_PredictUseNext_1) %>%
+  mutate(diff2 =PD_P2_UsageFITSBY -S2_PredictUseNext_1,
+         diff3 = PD_P3_UsageFITSBY - S3_PredictUseNext_1,
+         diff4 = PD_P4_UsageFITSBY - S4_PredictUseNext_1) %>%
+  rowwise() %>%
+  mutate(diff = mean(c(diff2, diff3, diff4), na.rm=T))
+  
+  a<- ggplot(df_usage_predict, aes(x=diff)) + 
+      geom_histogram(aes(y = stat(count) / sum(count)), colour=maroon, fill=maroon) +
+      xlim(c(-150, 150)) +
+      theme_classic() + 
+      labs(x = "Actual minus predicted FITSBY use (minutes/day)", 
+          y="Fraction of control group")  +
+      theme(panel.grid.major.x = element_blank(),
+            panel.grid.major.y = element_line( size=.1, color="lightsteelblue"))
+  
+  ggsave(sprintf('output/%s.pdf', filename), plot=a, width=6.5, height=4.5, units="in")
+  
+}
+
+plot_individual_temptation_effects <- function(df, param_full, filename){
+  
+  tau_data <- df %>% 
+    select(
+      UserID, 
+      w, L, B, S, 
+      PD_P1_UsageFITSBY,
+      PD_P2_UsageFITSBY,
+      PD_P3_UsageFITSBY,
+      PD_P4_UsageFITSBY,
+      PD_P5_UsageFITSBY,
+      PD_P2_LimitTightFITSBY
+    )
+
+  fit_2 <- tau_data %>%
+          mutate(tightness=ifelse(L,PD_P2_LimitTightFITSBY, 0)) %>%
+          lm(formula = 'PD_P2_UsageFITSBY ~ PD_P1_UsageFITSBY + L + tightness + B + S',
+             weights = w)
+
+  const_2 <- fit_2$coefficients['L']
+  slope_2 <- fit_2$coefficients['tightness']
+  
+  
+  fit_3 <- tau_data %>%
+          mutate(tightness=ifelse(L,PD_P2_LimitTightFITSBY, 0)) %>%
+          lm(formula = 'PD_P3_UsageFITSBY ~ PD_P1_UsageFITSBY + L + tightness + B + S',
+             weights = w)
+
+  const_3 <- fit_3$coefficients['L']
+  slope_3 <- fit_3$coefficients['tightness']
+  
+  
+  
+  df <- df %>%
+      mutate(tau_tilde_L = const_3 + slope_3*PD_P3_LimitTightFITSBY,
+             tau_L_2 = const_2 + slope_2 *PD_P2_LimitTightFITSBY) %>%  
+      mutate(x_ss_i_data = PD_P1_UsageFITSBY) 
+    
+    rho                               <- param_full[['rho']]
+    alpha                             <- param_full[['alpha']]
+    lambda                            <- param_full[['lambda']]
+    delta                             <- param_full[['delta']]
+    eta                               <- param_full[['eta']]
+    zeta                              <- param_full[['zeta']]
+    omega                             <- param_full[['omega']]
+    naivete                           <- param_full[['naivete']]
+    mispredict                        <- param_full[['mispredict']]
+    
+    df <- df %>%
+    mutate(num = eta*tau_L_2/omega - (1-alpha)*delta*rho*(((eta-zeta)*tau_tilde_L/omega+zeta*rho*tau_L_2/omega) + (1+lambda)*mispredict*(-eta+(1-alpha)*delta*rho^2*((eta-zeta)*lambda+zeta))),
+           denom = 1 - (1-alpha)*delta*rho*(1+lambda),
+           gamma_spec =  num/denom,
+           gamma_tilde_spec =  gamma_spec - naivete)
+    
+    
+    df <- df %>%
+    mutate(intercept_spec = calculate_intercept_spec(x_ss_i_data, param_full, gamma_tilde_spec, gamma_spec, alpha, rho, lambda, mispredict, eta, zeta)) %>%
+    mutate(x_ss_spec = calculate_steady_state(param_full, gamma_tilde_spec, gamma_spec, alpha, rho, lambda, mispredict, eta, zeta, intercept_spec),
+             x_ss_zero_un =calculate_steady_state(param_full, 0, 0, alpha, rho, lambda, 0, eta, zeta, intercept_spec),
+             x_ss_zero =ifelse(x_ss_zero_un<0, 0, x_ss_zero_un),
+             delta_x = x_ss_spec - x_ss_zero,
+             delta_x_zero =ifelse(delta_x<0, 0, delta_x),
+             delta_x_zero_3300 = ifelse(delta_x_zero>300, 300, delta_x_zero)) 
+
+  temptation_effect_below_ten <- nrow(df %>% filter(delta_x_zero_3300<10)) / nrow(df %>% filter(!is.na(delta_x_zero_3300)))
+  temptationeffectbelowten <- signif(temptation_effect_below_ten, digits=2)*100
+  
+  temptation_effect_above_100 <- nrow(df %>% filter(delta_x_zero_3300>100)) / nrow(df %>% filter(!is.na(delta_x_zero_3300)))
+  temptationeffectabovehundred <- signif(temptation_effect_above_100, digits=2)*100
+
+
+   estimate <- 
+      list(temptationeffectbelowten, temptationeffectabovehundred)
+      names(estimate) <- c('temptationeffectbelowten', 'temptationeffectabovehundred')
+      
+    save_nrow(estimate, filename ="individual_temptation_scalars", suffix="")
+
+  
+  a<- ggplot(df, aes(x=delta_x_zero_3300)) + 
+      geom_histogram(aes(y = stat(count) / sum(count)), colour=maroon, fill=maroon) +
+      xlim(c(0, 300)) +
+      ylim(c(0,0.11)) + 
+      theme_classic() + 
+      labs(x = "Effect of temptation on FITSBY use (minutes/day)", 
+          y="Fraction of sample")  +
+      theme(panel.grid.major.x = element_blank(),
+            panel.grid.major.y = element_line( size=.1, color="lightsteelblue"))
+
+  ggsave(sprintf('output/%s.pdf', filename), plot=a, width=6.5, height=4.5, units="in")
+}
+
+
+
+
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+# Excecute
+# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+main <- function(){
+
+  df <- import_data() %>%
+          mutate(addiction_decile   = add_deciles(StratAddictionLifeIndex)) %>%
+          mutate(restriction_decile = add_deciles(StratWantRestrictionIndex, step=0.125)) %>%
+          mutate(tightness_decile   = add_deciles(PD_P2_LimitTightFITSBY, step=1/6))  %>%
+          mutate(S2_PredictUseBonus = S2_PredictUseInitial * (1 - (S2_PredictUseBonus / 100))) %>%
+          mutate(f_tilde_2_B        = (S2_PredictUseBonus + S2_PredictUseInitial) / 2) %>%
+          mutate(behavioral_change_premium = (S2_MPL/num_days) - F_B + (p_B*f_tilde_2_B))
+
+  param <- param_initial
+  param_full <- estimate_model(df, param)
+  plot_individual_temptation_effects(df, param_full, filename="hist_individual_temptation_effects")
+
+
+  bonus_effect <-get_swb_effect_exported_bonus(df)
+  limit_effect <-get_swb_effect_exported_limit(df)
+
+  swb_effects <- list.merge(bonus_effect, limit_effect)
+  save_tex2(swb_effects,     filename="swb_effects")
+  save_tex_one(swb_effects,     filename="swb_effects_onedigit", suffix="one")
+
+  tau_data     <- reshape_tau_data(df)
+  tightness_df <- reshape_tightness(df)
+  mpd_df       <- reshape_mispredict(df)
+
+  plot_taus(df, tau_data, tightness_df)
+  plot_valuations(df)
+  plot_mispredict(mpd_df)
+  print('here')
+  find_tau_spec(df)
+  print('past here')
+  plot_treatment_effects(df, filename1="treatment_effects_periods_limit_bonus", filename2="treatment_effects_periods_bonus", filename3="treatment_effects_periods_limit")
+  plot_treatment_effects_interaction(df, filename1 = "interaction_treatment_effects")
+  plot_weekly_effects(df, filename1="treatment_effects_weeks_bonus", filename2 = "treatment_effects_weeks_limit")
+  #get_opt(df)
+  get_addiction_scalar(df)
+  plot_histogram_predicted(df, filename="histogram_predicted_actual_p24")
+
+  df %<>% balance_data(magnitude=3)
+  plot_treatment_effects(df, filename1="treatment_effects_periods_limit_bonus_balanced", filename2="treatment_effects_periods_bonus_balanced", filename3="treatment_effects_periods_limit_balanced")
+  get_addiction_treatment_effect(df, filename="coef_usage_self_control_balance")
+  
+} 
+
+main()

17/replication_package/code/analysis/treatment_effects/code/SurveyValidation.do

@@ -0,0 +1,136 @@
+// Description of data
+
+***************
+* Environment *
+***************
+
+clear all
+adopath + "input/lib/ado"
+adopath + "input/lib/stata/ado"
+
+*********************
+* Utility functions *
+*********************
+
+program define_settings
+    global DESCRIPTIVE_TAB ///
+        collabels(none) nodepvars noobs replace 
+end
+
+**********************
+* Analysis functions *
+**********************
+
+program main
+    import_data
+    define_settings
+
+    correlation_motivation
+	reg_prediction_reward
+end
+
+program import_data
+	use "input/final_data_sample.dta", clear
+end
+
+program correlation_motivation
+    preserve
+
+    generate tightness=0
+    replace tightness=PD_P2_LimitTightFITSBY if  (S2_LimitType != 0)
+
+    foreach v in S1_PhoneUseChange S1_AddictionIndex S1_SMSIndex S1_LifeBetter{
+    	replace `v' = -`v'
+    }
+    correlate S2_Benchmark S3_MPLLimit tightness S1_InterestInLimits S1_PhoneUseChange S1_AddictionIndex S1_SMSIndex S1_LifeBetter
+    matrix define correlation = r(C)
+
+    drop *
+
+    svmat correlation
+    qui ds
+    foreach i of numlist 1/8 {
+       replace correlation`i' = . if _n < `i' 
+    }
+    
+    dta_to_txt, saving(output/motivation_correlation.txt) title(<tab:motivation_correlation>) nonames replace
+    dta_to_txt, saving(output/motivation_correlation_beamer.txt) title(<tab:motivation_correlation_beamer>) nonames replace
+
+    restore
+end
+
+program reg_prediction_reward
+	preserve
+
+	* make a dummy for if high reward
+	gen PredictRewardHigh = 1 if PredictReward == 5
+	replace PredictRewardHigh = 0 if PredictReward == 1
+
+
+	* Reshape data to use predictions from all three surveys
+	keep UserID PredictRewardHigh S*_PredictUseNext_1 
+	local indep UserID PredictRewardHigh
+	rename_but, varlist(`indep') prefix(outcome)
+	reshape long outcome, i(`indep') j(measure) string
+
+	gen survey = substr(measure, 2, 1)
+	drop measure
+
+	rename outcome predicted
+
+	* Save to be merged later
+	tempfile temp
+	save `temp'
+
+	restore
+
+
+	preserve
+	* Reshape data to use actual predictions from those periods
+	keep UserID PD_P2_UsageFITSBY PD_P3_UsageFITSBY PD_P4_UsageFITSBY
+	local indep UserID
+
+	rename_but, varlist(`indep') prefix(outcome)
+	reshape long outcome, i(`indep') j(measure) string
+
+	gen survey = substr(measure, 5, 1)
+	drop measure
+
+	rename outcome actual
+
+	* Re-join with the actual predictions
+	merge 1:1 UserID survey using `temp'
+
+	* Run the regressions in question
+	reg predicted PredictRewardHigh, robust 
+	est store predicted
+
+	reg actual PredictRewardHigh, robust 
+	est store actual
+
+	gen pred_min_actual = predicted - actual
+	reg pred_min_actual PredictRewardHigh, robust 
+	est store pred_min_actual
+
+	gen abs_pred_min_actual = abs(predicted - actual)
+	reg abs_pred_min_actual PredictRewardHigh, robust 
+	est store abs_pred_min_actual
+
+	* Save the regressions as a table
+	esttab predicted actual pred_min_actual abs_pred_min_actual ///
+	 	using "output/high_reward_reg.tex", ///
+		mtitle("\shortstack{Predicted\\use}" ///
+			   "\shortstack{Actual\\use}" ///
+			   "\shortstack{Predicted -\\actual use}" ///
+			   "\shortstack{Absolute value of\\predicted - actual\\use}")  ///
+		coeflabels(PredictRewardHigh "High prediction reward" ///
+				   _cons "Constant")  ///
+		$DESCRIPTIVE_TAB se nostar nonotes
+	restore
+end
+
+***********
+* Execute *
+***********
+
+main

17/replication_package/code/analysis/treatment_effects/input.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e99d601b158b198f1db1ad6c634f9c4011573ec45848fe7d4e716fc3e26cac3
+size 914

17/replication_package/code/analysis/treatment_effects/make.py

@@ -0,0 +1,75 @@
+###################
+### ENVIRONMENT ###
+###################
+import git
+import imp
+import os
+
+### SET DEFAULT PATHS
+ROOT = '../..'
+
+PATHS = {
+    'root'             : ROOT,
+    'lib'              : os.path.join(ROOT, 'lib'),
+    'config'           : os.path.join(ROOT, 'config.yaml'),
+    'config_user'      : os.path.join(ROOT, 'config_user.yaml'),
+    'input_dir'        : 'input',
+    'external_dir'     : 'external',
+    'output_dir'       : 'output',
+    'output_local_dir' : 'output_local',
+    'makelog'          : 'log/make.log',
+    'output_statslog'  : 'log/output_stats.log',
+    'source_maplog'    : 'log/source_map.log',
+    'source_statslog'  : 'log/source_stats.log',
+}
+
+### LOAD GSLAB MAKE
+f, path, desc = imp.find_module('gslab_make', [PATHS['lib']])
+gs = imp.load_module('gslab_make', f, path, desc)
+
+### LOAD CONFIG USER
+PATHS = gs.update_paths(PATHS)
+gs.update_executables(PATHS)
+
+############
+### MAKE ###
+############
+
+### START MAKE
+gs.remove_dir(['input', 'external'])
+gs.clear_dir(['output', 'log', 'temp'])
+gs.start_makelog(PATHS)
+
+### GET INPUT FILES
+inputs = gs.link_inputs(PATHS, ['input.txt'])
+# gs.write_source_logs(PATHS, inputs + externals)
+# gs.get_modified_sources(PATHS, inputs + externals)
+
+### RUN SCRIPTS
+"""
+Critical
+--------
+Many of the Stata analysis scripts recode variables using
+the `recode` command. Double-check all `recode` commands
+to confirm recoding is correct, especially when reusing
+code for a different experiment version.
+"""
+
+gs.run_stata(PATHS, program = 'code/CommitmentResponse.do')
+gs.run_stata(PATHS, program = 'code/HabitFormation.do')
+gs.run_stata(PATHS, program = 'code/Heterogeneity.do')
+gs.run_stata(PATHS, program = 'code/SurveyValidation.do')
+gs.run_stata(PATHS, program = 'code/FDRTable.do')
+gs.run_stata(PATHS, program = 'code/HeterogeneityInstrumental.do')
+gs.run_stata(PATHS, program = 'code/Beliefs.do')
+
+gs.run_r(PATHS,     program = 'code/ModelHeterogeneity.R')
+
+### LOG OUTPUTS
+gs.log_files_in_output(PATHS)
+
+### CHECK FILE SIZES
+# gs.check_module_size(PATHS)
+
+### END MAKE
+gs.end_makelog(PATHS)

17/replication_package/code/codebook.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7c741990bd21eb9bab76643df657fff0f7bfc2c8500f64325178d371127a16e
+size 63101

17/replication_package/code/config.yaml

@@ -0,0 +1,122 @@
+
+#####################################################################
+# Is git LFS required to run this repository?
+#
+# This normally remain Yes, as this prevents inadvertently
+# committing large data files
+#####################################################################
+git_lfs_required: Yes
+
+#####################################################################
+# Other required software
+#####################################################################
+gslab_make_required: Yes
+
+software_required:
+  r: No
+  stata: No
+  lyx: Yes
+  matlab: No
+  latex: No
+
+#####################################################################
+# Maximum allowed file sizes
+#####################################################################
+max_file_sizes:
+  file_MB_limit_lfs: 100         # Soft limit on file size (w/ LFS)
+  total_MB_limit_lfs: 500        # Soft limit on total size (w/ LFS)
+  file_MB_limit: 0.5             # Soft limit on file size (w/o LFS)
+  total_MB_limit: 100            # Soft limit on total size (w/o LFS)
+
+metadata:
+  payment:
+    bonus: 50
+    fixed_rate: 50
+  strata: i.Stratifier
+  
+#####################################################################
+# Repository metadata
+#####################################################################
+
+# Experiment Name (could equal, for example, to 'Pilot#', 'Temptation'. If set to 'Scratch', the pipeline will process dummy data).
+experiment_name: "Temptation"
+
+# Survey Dates (Note that the 'Phase{#}Start' surveys are just fillers for the post study phases)
+surveys:
+  Recruitment:
+    Start: !!timestamp "2020-03-22 12:00:00"
+    End:   !!timestamp "2020-04-10 10:45:00"
+  Baseline:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End:   !!timestamp "2020-04-13 16:00:00"
+  Midline:
+    Start: !!timestamp "2020-05-03 0:00:00"
+    End:   !!timestamp "2020-05-11 23:59:00"
+  Endline1:
+    Start: !!timestamp "2020-05-24 08:00:00"
+    End:   !!timestamp "2020-05-31 08:00:00"
+  Endline2:
+    Start: !!timestamp "2020-06-14 08:00:00"
+    End:   !!timestamp "2020-06-22 17:00:00"
+  Phase5Start:
+    Start: !!timestamp "2020-07-05 08:00:00"
+    End: !!timestamp "2020-07-05 23:59:00"
+  Phase6Start:
+    Start: !!timestamp "2020-07-26 08:00:00"
+    End: !!timestamp "2020-07-26 23:59:00"
+  Phase7Start:
+    Start: !!timestamp "2020-08-16 08:00:00"
+    End: !!timestamp "2020-08-16 23:59:00"
+  Phase8Start:
+    Start: !!timestamp "2020-09-06 08:00:00"
+    End: !!timestamp "2020-09-06 23:59:00"
+  Phase9Start:
+    Start: !!timestamp "2020-09-27 08:00:00"
+    End: !!timestamp "2020-09-27 23:59:00"
+  Phase10Start:
+    Start: !!timestamp "2020-10-18 08:00:00"
+    End: !!timestamp "2020-10-18 23:59:00"
+  Phase11Start:
+    Start: !!timestamp "2020-11-08 08:00:00"
+    End: !!timestamp "2020-11-08 23:59:00"
+  Enrollment:
+    Start: !!timestamp "2020-04-09 9:00:00"
+    End:   !!timestamp "2020-04-11 12:00:00"
+  WeeklyText:
+    Start: !!timestamp "2020-03-25 00:00:00"
+    End:   !!timestamp "2020-03-30 23:59:00"
+  PDBug:
+    Start: !!timestamp "2020-04-24 18:40:00"
+    End:   !!timestamp "2020-04-28 23:59:00"
+  TextSurvey1:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey2:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey3:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey4:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey5:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey6:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey7:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey8:
+    Start: !!timestamp "2020-04-12 08:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+  TextSurvey9:
+    Start: !!timestamp "2020-04-12 20:00:00"
+    End: !!timestamp "2020-06-17 00:00:00"
+
+# Date Range of Data Used in Study (range of data we pull PD data)
+date_range:
+  first_pull: !!timestamp "2020-03-21 00:00:00"
+  last_pull: !!timestamp "2020-11-15 00:00:00"

17/replication_package/code/config_user.yaml

@@ -0,0 +1,67 @@
+#####################################################################
+# Make a copy of this file called config_user.yaml and place
+# it at the root level of the repository
+#
+# This file holds local settings specific to your computing
+# environment. It should not be committed to the repository.
+#####################################################################
+
+#####################################################################
+# External dependencies
+#
+# This section defines resources used by the code that are external
+# to the repository. Code should never reference any files external
+# to the repository except via these paths.
+#
+# Each external resource is defined by a key with a value equal
+# to the local path to the resource. These
+# keys should be short descriptive names that will then be used
+# to refer to these resources in code. E.g., "raw_data",
+# "my_other_repo", etc. Defaults can optionally be placed in
+# brackets after the colon
+#
+# Replace the paths below with correct local paths on your machine
+#
+#####################################################################
+external:
+  dropbox: /project #Point to PhoneAddiction Dropbox Root
+
+#####################################################################
+# Local settings
+#
+# This section defines parameters specific to each user's local
+# environment.
+#
+# Examples include names of executables, usernames, etc. These
+# variables should NOT be used to store passwords.
+#
+# Each parameter is defined by a key with default value. These
+# keys should be short descriptive names that will then be used
+# to refer to the parameters in code.
+#
+#####################################################################
+local:
+
+  # Executable names
+  executables:
+  
+    python: python
+    r: Rscript
+    stata: stata-mp
+    matlab: matlab
+    lyx: lyx
+    latex: latex
+
+  # Data Run
+  #if true, data/run will start by reading in the latest raw master data file, instead of processing raw phone dashboard data
+  skip_building: True
+
+  # if true, will process new data in parallel. only relevant if skip_building == False
+  parallel: False
+  cores: 4
+
+ #if true, will use all data in DataTest and ConfidentialTest
+  test: False
+
+#if true, stdout will write to data/log/mb_log.log instead of to terminal
+  log: False

17/replication_package/code/data/README.md

@@ -0,0 +1,96 @@
+# data
+This module contains all code that preps for analysis and produces survey management deliverables (e.g. contact lists). The dataset needed to run this module rely on confidential data, and were thus omitted from this replication archive. 
+
+We detail how, in the presence of the raw confidential data, this module construct the main datasets. 
+
+ #### 1. Pipeline overview
+  We run the whole data pipeline by calling data/make.py, which will call run 3 main sub modules below. Note: 
+  many classes and functions required in this pipeline are located in lib/data_helpers.
+  
+     1. source/build_master
+        
+        a. Purpose: a builder object ( in builder.py) will pull all the raw data, detect gaming, clean each individual data files, merge them
+         on the user level and the user_day_app level. 
+        
+        b. Input:
+            i. Raw Survey
+            ii. Phone Dashboard
+            
+        c. Output: 
+            i. master_raw_user.pickle, a raw master file on the user level, that will contain data from surveys and PD data for each phase
+            ii. master_user_day_app.pickle, a clean master file on the user day app level that will contain use, limit, and snooze activity
+        
+     2. source/clean_master
+        
+        a. Purpose: a cleaner object (in cleaner.py) will clean the raw_master_user.pickle, and assign treatments, calculate earnings, 
+        and create outcome variables
+        
+        b. Input: raw_master_user.pickle
+        
+        c. Output: clean_master_user.pickle
+        
+     3. source/exporters
+        
+        a. Purpose: creates contact lists, tango cards, phone dashboard treatment configs, and analysis files ready for stata
+        
+        b. Input: master_clean_user.pickle and master_user_day_app.pickle
+        
+        c. Output: 
+          i. Contact Lists, Tango Cards, Phone Dashboard, Treatment Configs, and other data with identifiable info will output into /Dropbox/PhoneAddiction/Confidential
+          ii. pre_analysis_user.csv and pre_analysis_user_app_day.csv in /Dropbox/PhoneAddiction/Data/{experiment_name}/Intermediate
+    
+  ## 2. Configurations:
+     - root/config_user.yaml: configurations that alter how the pipeline is run. Read through those in the yaml file, but to highlight:
+        1. skip_building: if true, data/run will start by reading in the latest raw master data file, instead of processing raw phone dashboard data. You should not attempt run the raw PD data unless you're on Sherlock or some HPC
+        2. test: if set true, this runs nearly the full pipeline, but with dummy data. Data is saved in DataTest and ConfidentialTest. This is helpful when testing something in the build class
+     
+     - root/config.yaml: sets significant experiment dates (survey dates, and date range of PD data pull)
+     
+     - root/lib/experiment_specs: contains detailed specs for the data pipeline. Check out the README in that folder for specifics.
+     
+  ## 3. Raw Phone Dashboard Data Exports
+   - All PD Data arrives in the PhonedashboardPort dropbox folder. All these files are processed by functions 
+   in data/source/build_master/builder.py and helper functions in lib/data_helpers
+     
+  ## Snooze Events
+    1. PD receives usage ping from ForegroundApplication generator.
+
+    2. If app usage is within X minutes of budget being exhausted:
+    
+            2.a: PD does not block app, but launches warning activity with the package and metadata.
+    
+            2.b: Warning activity throws up warning dialog (event = app-block-warning).
+    
+            2.c: User closes / cancels dialog (event = closed-warning).
+    
+    3. If app usage is past budget AND has been snoozed, and snooze delay has not elapsed:
+    
+            3.a: PD blocks app (returns system to the home screen (event = blocked_app).
+    
+            3.b: PD shows dialog letting user know that delay hasn’t elapsed (event = app-blocked-no-snooze).
+    
+            3.c: User closes/cancels dialog (event = app-blocked-no-snooze-closed).
+    
+    3. If app usage is past budget AND app has not been snoozed:
+    
+            3.a: PD blocks app (returns system to the home screen (event = blocked_app).
+    
+            3.b: PD shows dialog letting user budget is exhausted (event = app-blocked-no-snooze).
+    
+            3.c: If snooze is NOT enabled for user:
+    
+                    3.c.1: PD shows dialog that user cannot use app until tomorrow (event = app-blocked-no-snooze).
+    
+                    3.c.2: User closes / cancels dialog (event = app-blocked-no-snooze-closed).
+    
+            3.d: If snooze IS enabled for user:
+    
+                    3.d.1: PD shows dialog letting user know budget is up (event = app-blocked-can-snooze).
+    
+                    3.d.2: If user closes / cancels dialog (event = skipped-snooze).
+    
+                    3.d.3: If user decides to snooze, PD shows dialog asking about snooze amount (no event generated).
+    
+                            3.d.3.a: User closes / cancels snooze dialog, without setting limit (event = cancelled-snooze).
+    
+                            3.d.3.b: User sets snooze amount (event = snoozed-app-limit).

17/replication_package/code/data/external.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8534419da84c6a8c9f902d366dd6c964c2ec49c804730cc4ad333a7d7f05a39
+size 1135

17/replication_package/code/data/input.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d958523f79a58631b695d9e13762414d22ce0055ae9ac8b9f6ad63c17026c1
+size 676

17/replication_package/code/data/make.py

@@ -0,0 +1,68 @@
+###################
+### ENVIRONMENT ###
+###################
+import git
+import imp
+import os
+import yaml
+
+### SET DEFAULT PATHS
+ROOT = git.Repo('.', search_parent_directories = True).working_tree_dir
+
+PATHS = {
+    'root'             : ROOT,
+    'lib'              : os.path.join(ROOT, 'lib'),
+    'config'           : os.path.join(ROOT, 'config.yaml'),
+    'config_user'      : os.path.join(ROOT, 'config_user.yaml'),
+    'input_dir'        : 'input', 
+    'external_dir'     : 'external',
+    'output_dir'       : 'output',
+    'output_local_dir' : 'output_local',
+    'makelog'          : 'log/make.log',         
+    'output_statslog'  : 'log/output_stats.log', 
+    'source_maplog'    : 'log/source_map.log',  
+    'source_statslog'  : 'log/source_stats.log'
+}
+
+### ADD EXPERIMENT NAME TO PATH
+with open(PATHS['config'], 'r') as stream:
+    config = yaml.safe_load(stream)
+    
+PATHS["experiment_name"] = config['experiment_name']
+
+### LOAD GSLAB MAKE
+f, path, desc = imp.find_module('gslab_make', [PATHS['lib']]) 
+gs = imp.load_module('gslab_make', f, path, desc)
+
+### LOAD CONFIG USER 
+PATHS = gs.update_paths(PATHS)
+gs.update_executables(PATHS)
+
+############
+### MAKE ###
+############
+
+### START MAKE
+gs.remove_dir(['input', 'external'])
+gs.clear_dir(['output', 'log'])
+gs.start_makelog(PATHS)
+
+### GET INPUT FILES 
+inputs = gs.link_inputs(PATHS, ['input.txt'])
+externals = gs.link_externals(PATHS, ['external.txt'])
+
+gs.write_source_logs(PATHS, inputs + externals)
+gs.get_modified_sources(PATHS, inputs + externals)
+
+### RUN SCRIPTS
+gs.run_python(PATHS, program = 'source/run.py')
+gs.run_stata(PATHS, program = 'source/prep_stata.do')
+
+### LOG OUTPUTS
+gs.log_files_in_output(PATHS)
+
+### CHECK FILE SIZES
+gs.check_module_size(PATHS)
+
+### END MAKE
+gs.end_makelog(PATHS)

17/replication_package/code/data/source/build_master/builder.py

@@ -0,0 +1,328 @@
+from datetime import datetime,timedelta
+import pandas as pd
+import os
+import sys
+import git
+from pympler.tracker import SummaryTracker
+
+#importing modules from root of data
+root = git.Repo('.', search_parent_directories = True).working_tree_dir
+sys.path.append(root)
+os.chdir(os.path.join(root))
+
+from lib.data_helpers.pull_events import PullEvents
+from lib.utilities import serialize
+from data.source.build_master.pullers.pull_events_use import PullEventsUse
+from data.source.build_master.pullers.pull_events_alt import PullEventsAlt
+
+
+from lib.data_helpers.clean_events import CleanEvents
+from data.source.build_master.cleaners.clean_surveys import CleanSurveys
+from data.source.build_master.cleaners.clean_events_use import CleanEventsUse
+from data.source.build_master.cleaners.clean_events_status import CleanEventsStatus
+from data.source.build_master.cleaners.clean_events_budget import CleanEventsBudget
+from data.source.build_master.cleaners.clean_events_snooze_delays import CleanEventsSnoozeDelays
+from data.source.build_master.cleaners.clean_events_snooze import CleanEventsSnooze
+from data.source.build_master.cleaners.clean_events_alt import CleanEventsAlt
+
+
+
+
+from lib.data_helpers.gaming import Gaming
+from data.source.build_master.master_raw_user import MasterRawUser
+from data.source.build_master.master_raw_user_day_app import MasterRawUserDayApp
+
+from lib.experiment_specs import study_config
+
+"""
+
+"""
+class Builder():
+
+    @staticmethod
+    def build_master():
+        tracker = SummaryTracker()
+
+        # print(f"\n Clean Survey Data {datetime.now()}")
+        # clean_surveys = CleanSurveys.clean_all_surveys()
+
+        # print(f"\nInitializing Master DF and add survey data {datetime.now()}")
+        # raw_user = MasterRawUser(initial_survey_df= clean_surveys[study_config.initial_master_survey])
+        # raw_user.add(clean_surveys)
+        # del clean_surveys
+
+        # print(f"\nCleaning Traditional Use and DetectGaming {datetime.now()}")
+        # trad_use_phase, trad_use_hour = Builder._build_pd_use()
+
+        # game_df = Gaming.process_gaming(error_margin=1,
+        #                                 hour_use=trad_use_hour,
+        #                                 raw_user_df=raw_user.raw_master_df)
+        # raw_user.add({"Game": game_df})
+
+        # tracker.print_diff()
+        # del [trad_use_phase, game_df]
+        # tracker.print_diff()
+
+        # if datetime.now() > study_config.surveys["Midline"]["Start"]:
+        #     print(f"\nCleaning Limit Data {datetime.now()}")
+        #     pd_snooze = Builder._build_pd_snooze()
+        #     budget_phase, pd_budget = Builder._build_pd_budget()
+        #     try:
+        #         Builder._build_pd_snooze_delay()
+        #     except:
+        #         print("couldn't process snooze delay data")
+
+        #     raw_user.add({"PDBudget": budget_phase})
+        # else:
+        #     pd_budget = pd.DataFrame()
+        #     pd_snooze = pd.DataFrame()
+
+        print(f"\nCleaning Traditional Use Individual {datetime.now()}")
+        Builder._build_pd_use_indiv()
+
+        # print(f"\n Alternative and Status Data {datetime.now()}")
+        # alt_use_hour, alt_use_phase = Builder._build_pd_alt(trad_use_hour)
+        # raw_user.add({"AltPDUse": alt_use_phase})
+
+        # clean_status, pd_latest = Builder._build_pd_status(raw_user.raw_master_df,alt_use_hour)
+        # raw_user.add({"LatestPD": pd_latest})
+        # del [alt_use_phase, pd_latest]
+
+        # print(f"\n Serialize user level data before building user-app-day data")
+        # config_user_dict = serialize.open_yaml("config_user.yaml")
+        # if config_user_dict['local']['test'] == False:
+        #     serialize.save_pickle(raw_user.raw_master_df,
+        #                           os.path.join("data", "external", "intermediate", "MasterIntermediateUser"))
+
+        # print(f"\n Create UserXAppXDate Level data {datetime.now()}")
+        # MasterRawUserDayApp.build(alt_use_hour,pd_budget,pd_snooze,clean_status)
+
+        # tracker.print_diff()
+        # del [pd_budget,pd_snooze,alt_use_hour]
+
+        # print(f"\n Recover Old Install Data")
+        # PullEventsAlt.recover_install_data()
+        # return raw_user.raw_master_df
+
+
+    @staticmethod
+    def _build_pd_use():
+        pd_use_puller = PullEvents(source="PhoneDashboard",
+                                    keyword="Use",
+                                    scratch=False,
+                                    test=False,
+                                    time_cols=["Created", "Recorded"],
+                                    raw_timezone="Local",
+                                    appcode_col='Source',
+                                    identifying_cols=["AppCode", "ForegroundApp", "ScreenActive",
+                                                      "CreatedDatetimeHour"],
+                                    sort_cols= ["CreatedDatetimeHour","RecordedDatetimeHour"],
+                                    drop_cols= ["PlayStoreCategory","UploadLag"],
+                                    cat_cols = ["ForegroundApp"],
+                                    compress_type="txt",
+                                    processing_func=PullEventsUse.process_raw_use)
+
+        raw_hour_use = pd_use_puller.update_data()
+
+        use_cleaner = CleanEvents(source="PhoneDashboard", keyword="Use")
+        use_phase, use_hour = use_cleaner.clean_events(raw_event_df=raw_hour_use,
+                                                date_col="CreatedDate",
+                                                cleaner=CleanEventsUse(use_type="Traditional"))
+
+        CleanEventsUse.get_timezones(use_hour, "CreatedDatetimeHour", "CreatedEasternDatetimeHour")
+
+
+        return use_phase, use_hour
+
+    @staticmethod
+    def _build_pd_use_indiv():
+        pd_use_puller = PullEvents(source="PhoneDashboard",
+                                    keyword="UseIndiv",
+                                    scratch=True,
+                                    test=False,
+                                    time_cols=["Created", "Recorded"],
+                                    raw_timezone="Local",
+                                    appcode_col='Source',
+                                    identifying_cols=["AppCode", "ForegroundApp", "StartTime", "UseMinutes"],
+                                    sort_cols= ["StartTime"],
+                                    drop_cols= ["PlayStoreCategory","UploadLag"],
+                                    cat_cols = ["ForegroundApp"],
+                                    compress_type="txt",
+                                    processing_func=PullEventsUse.process_raw_use_indiv)
+
+        raw_hour_use = pd_use_puller.update_data()
+
+        # use_cleaner = CleanEvents(source="PhoneDashboard", keyword="Use")
+        # use_phase, use_hour = use_cleaner.clean_events(raw_event_df=raw_hour_use,
+        #                                         date_col="CreatedDate",
+        #                                         cleaner=CleanEventsUse(use_type="Traditional"))
+
+        # CleanEventsUse.get_timezones(use_hour, "CreatedDatetimeHour", "CreatedEasternDatetimeHour")
+
+
+    @staticmethod
+    def _build_pd_status(raw_master: pd.DataFrame, alt_use_hour: pd.DataFrame):
+        pd_use_puller = PullEvents(source="PhoneDashboard",
+                                   keyword="Status",
+                                   scratch=False,
+                                   test=False,
+                                   time_cols=["LastUpload"],
+                                   raw_timezone="Local",
+                                   appcode_col='Participant',
+                                   identifying_cols=["AppCode", "Group", "Blocker",
+                                                     "LastUpload", "AppVersion","PlatformVersion","PhoneModel","OptedOut"],
+                                   sort_cols = ["LastUpload"],
+                                   drop_cols = ['PhaseUseBrowser(ms)',
+                                                'PhaseUseFB(ms)',
+                                                'PhaseUseIG(ms)',
+                                                'PhaseUseOverall(ms)',
+                                                'PhaseUseSnap(ms)',
+                                                'PhaseUseYoutube(ms)',"AsOf"],
+                                   cat_cols = [],
+                                   compress_type="txt",)
+
+        raw_status = pd_use_puller.update_data()
+        raw_status["LastUploadDate"] = raw_status["LastUpload"].apply(lambda x: x.date())
+        use_cleaner = CleanEvents(source="PhoneDashboard", keyword="Status")
+        clean_status = use_cleaner.clean_events(raw_event_df=raw_status,
+                                                date_col="LastUploadDate",
+                                                cleaner=CleanEventsStatus(),
+                                                phase_data=False)
+
+        pd_latest = CleanEventsStatus.get_latest_pd_health(clean_status, raw_master, alt_use_hour)
+        return clean_status, pd_latest
+
+    @staticmethod
+    def _build_pd_alt(clean_trad_use_hour):
+        alt_json_reader = PullEventsAlt()
+        pd_alt_puller = PullEvents(source="PhoneDashboard",
+                                       keyword="Alternative",
+                                       scratch=False,
+                                       test=False,
+                                       time_cols=["Created"],
+                                       raw_timezone="Local",
+                                       appcode_col='AppCode',
+                                       identifying_cols=["AppCode", "ForegroundApp", "CreatedDatetimeHour"],
+                                       sort_cols = ["Observed","CreatedDatetimeHour"],
+                                       drop_cols = ["Com.AudaciousSoftware.PhoneDashboard.AppTimeBudget", "Timezone",
+                                                    "CreatedDatetime","CreatedEasternDatetime","Label", "CreatedDate",
+                                                    "PlayStoreCategory","DaysObserved","Index","ZipFolder","CreatedEasternMinusLocalHours"],
+                                       cat_cols = ["ForegroundApp"],
+                                       compress_type="folder",
+                                       processing_func=alt_json_reader.process_raw_use,
+                                       file_reader=alt_json_reader.read_alt)
+
+        # This function will read in and update all types of alternative data, will only return the use data
+        # and will serialize all other data
+        raw_alt_use_hour = pd_alt_puller.update_data()
+        try:
+            combined_raw_alt_use_hour = PullEventsAlt.combine_trad_alt(raw_alt_use_hour,clean_trad_use_hour)
+        except:
+            print("could not combine trad and alt")
+            combined_raw_alt_use_hour = raw_alt_use_hour.copy()
+
+        use_cleaner = CleanEvents(source="PhoneDashboard", keyword="Alternative")
+        use_phase, use_hour = use_cleaner.clean_events(raw_event_df=combined_raw_alt_use_hour,
+                                                       date_col="CreatedDate",
+                                                       cleaner=CleanEventsUse(use_type="Alternative"))
+
+        config_user_dict = serialize.open_yaml("config_user.yaml")
+        if config_user_dict['local']['test']== False:
+            try:
+                print(f"\n Clean Alt Install data events {datetime.now()}")
+                CleanEventsAlt.process_appcode_files(
+                    input_folder = os.path.join("data", "external", "input", "PhoneDashboard", "RawAltInstall"),
+                    output_file = os.path.join("data", "external", "intermediate", "PhoneDashboard", "AltInstall"),
+                    cleaning_function= CleanEventsAlt.clean_install
+                )
+            except:
+                print("could not aggregate install data")
+        return use_hour, use_phase
+
+    @staticmethod
+    def _build_pd_budget():
+        """processes the limit setting data"""
+        pd_budget_puller = PullEvents(source="PhoneDashboard",
+                                      keyword="Budget",
+                                      scratch=False,
+                                      test=False,
+                                      time_cols=["Updated","EffectiveDate"],
+                                      raw_timezone="Local",
+                                      appcode_col="Source",
+                                      identifying_cols=["AppCode", "App", "Updated", "EffectiveDate"],
+                                      sort_cols=["Updated"],
+                                      drop_cols = [],
+                                      cat_cols = [],
+                                      compress_type="txt")
+
+        pd_budget = pd_budget_puller.update_data()
+
+        budget_cleaner = CleanEvents(source="PhoneDashboard", keyword="Budget")
+        clean_budget = budget_cleaner.clean_events(raw_event_df=pd_budget,
+                                                    date_col="EffectiveDate",
+                                                    cleaner=CleanEventsBudget(),
+                                                    phase_data = False)
+
+        budget_sum = CleanEventsBudget.get_latest_budget_data(clean_budget)
+
+        return budget_sum, clean_budget
+
+    @staticmethod
+    def _build_pd_snooze_delay():
+        """process the custom snooze data (post study functionality)"""
+        pd_snooze_delay_puller = PullEvents(source="PhoneDashboard",
+                                     keyword="Delays",
+                                     scratch = False,
+                                     test = False,
+                                     time_cols=["UpdatedDatetime", "EffectiveDatetime"],
+                                     raw_timezone = "Local",
+                                     appcode_col="App Code",
+                                     identifying_cols=["AppCode", "SnoozeDelay", "UpdatedDatetime"],
+                                     sort_cols = ["UpdatedDatetime"],
+                                     drop_cols= [],
+                                     cat_cols = [],
+                                     compress_type="txt")
+
+        raw_delayed_snooze = pd_snooze_delay_puller.update_data()
+        snooze_delay_cleaner = CleanEvents(source="PhoneDashboard", keyword="Delays")
+
+        clean_delays = snooze_delay_cleaner.clean_events(raw_event_df=raw_delayed_snooze,
+                                                   date_col= "EffectiveDate",
+                                                   cleaner= CleanEventsSnoozeDelays(),
+                                                   phase_data=False)
+
+        clean_delays.to_csv(os.path.join("data","external", "intermediate", "PhoneDashboard", "Delays.csv"))
+
+
+
+    @staticmethod
+    def _build_pd_snooze():
+        """processes the snooze event data"""
+        pd_snooze_puller = PullEvents(source="PhoneDashboard",
+                                     keyword="Snooze",
+                                     scratch = False,
+                                     test = False,
+                                     time_cols=["Recorded", "Created"],
+                                     raw_timezone = "Local",
+                                     appcode_col="Source",
+                                     identifying_cols=["AppCode", "App", "Event", "Created"],
+                                     sort_cols = ["Created"],
+                                     drop_cols= [],
+                                     cat_cols = [],
+                                     compress_type="txt")
+
+        raw_snooze = pd_snooze_puller.update_data()
+
+        snooze_cleaner = CleanEvents(source="PhoneDashboard", keyword="Snooze")
+
+        pd_snooze = snooze_cleaner.clean_events(raw_event_df=raw_snooze,
+                                                   date_col= "Date",
+                                                   cleaner= CleanEventsSnooze(),
+                                                   phase_data=False)
+
+        CleanEventsSnooze.get_premature_blocks(pd_snooze)
+
+        return pd_snooze
+
+if __name__ == "__main__":
+    pd_snooze = Builder._build_pd_snooze_delay()

17/replication_package/code/data/source/build_master/cleaners/clean_events_alt.py

@@ -0,0 +1,150 @@
+import os
+import json
+import git
+import sys
+import pandas as pd
+
+#importing modules from root of data
+root = git.Repo('.', search_parent_directories = True).working_tree_dir
+sys.path.append(root)
+os.chdir(os.path.join(root))
+
+from lib.experiment_specs import study_config
+from lib.data_helpers import data_utils
+
+from lib.data_helpers.builder_utils import BuilderUtils
+from lib.utilities import serialize
+
+class CleanEventsAlt():
+
+    @staticmethod
+    def process_appcode_files(input_folder,output_file,cleaning_function):
+        """
+        inputs:
+         - input_folder: directory where all pickle files will be read and appeneded
+         - outpul_file: the directory where the output file will be saves
+         - cleaning_function: the function used to clean the aggregated data
+        """
+        appcodes = [x for x in os.listdir(input_folder) if ".pickle" in x]
+        df_list = []
+        print(appcodes[:5])
+        for appcode in appcodes:
+            path = os.path.join(input_folder, appcode)
+            try:
+                a_df = serialize.open_pickle(path)
+                if len(a_df) == 0:
+                    continue
+                df_list.append(a_df)
+
+                #try:
+                #    d = serialize.open_hdf(path.replace(".pickle",".h5"))
+                #except:
+                #    print(f"could not open {appcode} h5 file!!!!, but pickle opened without problems")
+
+            except:
+                print(f"could not read {appcode} raw install pickle data")
+
+        if len(df_list) > 0:
+            df = pd.concat(df_list).reset_index(drop=True)
+            df = cleaning_function(df)
+
+            try:
+                serialize.save_hdf(df, output_file)
+            except:
+                print("Couldn't save hdf")
+
+            try:
+                df.to_csv(output_file + ".csv", index=False)
+            except:
+                print("couldn't save csv file")
+
+
+    @staticmethod
+    def clean_install(a_df):
+        # add column to indiciate if app is FITSBY
+        a_df = data_utils.add_A_to_appcode(a_df, "AppCode")
+        duplicate_fitsby_apps = pd.read_excel(os.path.join("lib", "experiment_specs", "FITSBY_apps.xlsx"))
+        a_df = a_df.merge(duplicate_fitsby_apps, on='App', how='left')
+        a_df = a_df.drop_duplicates(subset = ["AppCode","App","Date"])
+        return a_df
+
+    ################
+    #####OLD#########
+    ###############
+    @staticmethod
+    def process_appcode_files_OLD(input_folder,output_file,cleaning_function):
+        appcodes = [x for x in os.listdir(input_folder) if ".pickle" in x]
+        df_list = []
+        for appcode in appcodes:
+            path = os.path.join(input_folder, appcode)
+            a_dict = serialize.open_pickle(path, df_bool=False)
+            if len(a_dict) == 0:
+                continue
+
+            a_df = cleaning_function(a_dict)
+            a_df["AppCode"] = "A" + appcode.replace(".pickle", "")
+            df_list.append(a_df)
+
+        if len(df_list)>0:
+            df = pd.concat(df_list).reset_index(drop = True)
+            try:
+                serialize.save_pickle(df, output_file)
+            except:
+                print("Couldn't save Pickle")
+
+            try:
+                # DONT PUT IN TRY BECAUSE IF BACKUP FAILS, WE WANT TO RE PROCESS THE NEW FILES
+                df.to_csv(output_file+".csv", index=False, compression='gzip')
+            except:
+                print("couldn't save zip file")
+
+            return df
+
+        else:
+            print("no alt data yet!")
+            return pd.DataFrame()
+
+    @staticmethod
+    def clean_alt_block_data(a_dict):
+        a_df = pd.DataFrame.from_dict(a_dict, orient='index').reset_index().rename(columns={"index": "Created", "app": "App"})
+        a_df = data_utils.clean_iso_dates(a_df, 'Created')
+        a_df["AltLimitMinutes"] = a_df["time_budget"] / (1000 * 60)
+        a_df["AltUseMinutesAtEvent"] = a_df["time_usage"] / (1000 * 60)
+        return a_df
+
+    @staticmethod
+    def clean_warnings(a_dict):
+        df = pd.DataFrame.from_dict(a_dict, orient='index').reset_index().rename(columns={"date": "Created"}).drop(
+            columns='index')
+        df = data_utils.clean_iso_dates(df, 'Created')
+
+        df['details_dict'] = df['details'].apply(lambda x: json.loads(x))
+        chars_df = df["details_dict"].apply(pd.Series)
+        assert len(chars_df) == len(df)
+        df = pd.concat([df, chars_df], axis=1)
+        df = df.drop(columns=["details_dict", "details"])
+
+        df = df.rename(columns = {"event": "Event",
+                                  "minutes-remaining": "MinutesRemaining",
+                                  "package": "App",
+                                  "snooze-delay": "SnoozeDelay",
+                                  "snooze-minutes": "SnoozeMinutes"})
+
+        event_rename = {"app-block-warning": "App Warning Displayed",
+                        "app-blocked-can-snooze": "App Blocked - Snooze Offered",
+                        "app-blocked-delayed": "App Blocked Until Delay Elapsed",
+                        "app-blocked-no-snooze": "App Blocked - Snooze Unavailable",
+                        "app-blocked-no-snooze-closed": "User Closed App Blocked (No Snooze) Warning",
+                        "cancelled-snooze": "User Cancelled Snooze",
+                        "closed-delay-warning": "User Closed Delay Warning",
+                        "closed-warning": "User Closed Warning",
+                        "skipped-snooze": "User Declined Snooze",
+                        "snoozed-app-limit": "Snooze Enabled"}
+        df["Event"] = df["Event"].apply(lambda x: event_rename[x])
+        return df
+
+if __name__ == "__main__":
+    input_folder = os.path.join("data","external","input","PhoneDashboard","RawAltInstall")
+    output_file = os.path.join("data", "external", "intermediate", "PhoneDashboard", "AltInstall")
+    CleanEventsAlt.process_appcode_files(input_folder,output_file,CleanEventsAlt.clean_install)
+    print('donzo')

17/replication_package/code/data/source/build_master/cleaners/clean_events_budget.py

@@ -0,0 +1,58 @@
+import sys
+import os
+from datetime import datetime, timedelta
+import numpy as np
+import pandas as pd
+
+from lib.experiment_specs import study_config
+from lib.data_helpers import data_utils
+from lib.utilities import codebook
+from lib.utilities import serialize
+
+from lib.data_helpers.builder_utils import BuilderUtils
+
+""""
+The new use cleaner, which will deprecate phone_use_cleaner, and phase_use
+"""
+class CleanEventsBudget():
+
+    clean_file = os.path.join("data","external","intermediate","PhoneDashboard","CleanBudget")
+
+    def prep_clean(self, df):
+        # Update Use Data
+        df["SawLimitSettingPage"] = True
+
+        df.loc[df["App"] != "placeholder.app.does.not.exist","HasSetLimit"] = True
+        #df = df.loc[df["App"] != "placeholder.app.does.not.exist"]
+        df["EffectiveDate"] = df["EffectiveDate"].dt.date
+        df = df.loc[(df["EffectiveDate"] >= study_config.surveys["Midline"]["Start"].date())]
+
+        return df
+
+    """Called in the Event Cleaner, after the data has been subsetted to a given phase"""
+    def phase_clean(self, df, phase):
+        # prep
+        #summarize
+        ## print("hello")
+        return df
+
+    @staticmethod
+    def get_latest_budget_data(clean_budget_df):
+        df = clean_budget_df[["AppCode", "HasSetLimit", "SawLimitSettingPage"]].groupby(
+            ["AppCode"]).first().reset_index()
+
+        apps = clean_budget_df.groupby(["AppCode", "App"])["NewLimit"].last().reset_index()
+        apps = apps.loc[apps["App"].isin(study_config.fitsby)]
+        apps["LimitMinutes"] = apps["NewLimit"] / (60 * 1000)
+        apps["App"] = apps["App"].apply(lambda x: x.capitalize())
+
+        apps_p = apps.pivot_table(index=["AppCode"],
+                                  values=["LimitMinutes"],
+                                  columns=["App"],
+                                  aggfunc='first')
+        apps_p.columns = [''.join(col[::1]).strip() for col in apps_p.columns.values]
+        apps_p = apps_p.reset_index()
+
+        df = df.merge(apps_p, on = "AppCode", how = "left")
+        return df
+

17/replication_package/code/data/source/build_master/cleaners/clean_events_pc.py

@@ -0,0 +1,60 @@
+import sys
+import os
+from datetime import datetime, timedelta
+
+from lib.data_helpers.builder_utils import BuilderUtils
+from lib.utilities import serialize
+
+class CleanEventsPC():
+
+    def __init__(self):
+        self.social_hosts = []
+        self.use_subsets = {
+            "WebDesktop": {
+                "Filters": {"WebBool":[True]},
+                "DenomCol": "DaysWithWeb",
+                "NumCols": ["UseMinutes"]},
+
+            "FBDesktop" :{
+                "Filters": {"FBBool": [True]},
+                "DenomCol": "DaysWithWeb",
+                "NumCols": ["UseMinutes"]},
+
+            "IGDesktop":{
+                "Filters": {"IGBool": [True]},
+                "DenomCol": "DaysWithWeb",
+                "NumCols": ["UseMinutes"]}
+        }
+
+
+    def prep_clean(self,df):
+        # Rename some things
+        df["UseMinutes"] = df["Duration"]/60
+        df = df.drop(columns = ["Duration"])
+
+        # create date variable
+        df["StartedOnIsoDate"] = df["StartedOnIso"].apply(lambda x: x.date())
+        df["EndedOnIsoDatetime"] = df["StartedOnIso"]+df["UseMinutes"].apply(lambda x: timedelta(seconds = x*60))
+
+        # label treatment webistes
+        df.loc[df["Website"].notnull(), "WebBool"] = True
+        df.loc[df["Website"].fillna("nan").str.contains("facebook"), "FBBool"] = True
+        df.loc[df["Website"].fillna("nan").str.contains("instagram"), "IGBool"] = True
+
+        # Create List of hosts, ordered by popularity
+        top_hosts = df.groupby(['Website'])['AsOf'].agg(['count'])
+        top_hosts = top_hosts.rename(columns={'count': "WebsiteVisitCount"}).reset_index().sort_values(
+            by="WebsiteVisitCount", ascending=False)
+        top_hosts.to_csv(os.path.join("data","external","intermediate","PCDashboard", "TopSites.csv"),
+                         index=False)
+
+        # get social hosts
+        self.social_hosts = [y for y in list(top_hosts["Website"]) if any(x in y for x in study_config.social_websites)]
+        return df
+
+    """Called in the Event Cleaner, after the data has been subsetted to a given phase"""
+    def phase_clean(self, df, phase):
+        df["WebDay"] = df["StartedOnIso"].apply(lambda x: x.date())
+        df.loc[:, "DaysWithWeb"] = df.groupby(by=['AppCode'])['WebDay'].transform(lambda x: x.nunique())
+        df = BuilderUtils.get_subsets_avg_use(df, self.use_subsets)
+        return df

17/replication_package/code/data/source/build_master/cleaners/clean_events_snooze.py

@@ -0,0 +1,43 @@
+from lib.data_helpers.builder_utils import BuilderUtils
+import os
+from datetime import datetime,timedelta
+from lib.experiment_specs import study_config
+
+""""
+The new use cleaner, which will deprecate phone_use_cleaner, and phase_use
+"""
+class CleanEventsSnooze():
+
+    clean_file = os.path.join("data","external","intermediate","PhoneDashboard","CleanSnooze")
+
+    def prep_clean(self, df):
+        df["Date"] = df["Created"].dt.date
+        df = df.loc[(df["Date"] >= study_config.surveys["Midline"]["Start"].date())]
+        return df
+
+    @staticmethod
+    def get_premature_blocks(sn):
+        ud_s = sn.groupby(["AppCode", "Date", "App"]).first().reset_index()
+
+        bug = ud_s.loc[~ud_s["Event"].isin(["App Warning Displayed"])]
+        bug = bug.loc[bug["Created"] > datetime(2020, 5, 2, 0, 0), ["AppCode", "Created", "Date", "App", "Event",
+                                                                    "SnoozeExtension"]]
+        print(bug["Event"].value_counts())
+
+        # look for user-days for which last event was a display that the user didn't close
+        ud_l = sn.groupby(["AppCode", "Date", "App"]).last().reset_index()
+        b_p = ud_l.loc[ud_l["Event"].isin(["App Blocked - Snooze Offered",
+                                           "App Blocked - Snooze Unavailable",
+                                           "App Warning Displayed"])]
+        b_p["NextDate"] = b_p["Date"].apply(lambda x: x + timedelta(1))
+        b_p = b_p.rename(columns={"Event": "YesterdayEvent", "App": "YesterdayApp",
+                                  "Created": "YesterdayCreated",
+                                  "SnoozeExtension": "YesterdaySnoozeExtension"})
+
+        bug = bug.merge(b_p[["AppCode", "YesterdayApp", "NextDate", "YesterdayEvent", "YesterdayCreated",
+                             "YesterdaySnoozeExtension"]],
+                        right_on=["AppCode", "NextDate", "YesterdayApp"],
+                        left_on=["AppCode", "Date", "App"],
+                        how='left')
+
+        bug.to_csv(os.path.join("data","external","intermediate","Scratch","3b_SnoozeEvent.csv"))

17/replication_package/code/data/source/build_master/cleaners/clean_events_snooze_delays.py

@@ -0,0 +1,16 @@
+
+class CleanEventsSnoozeDelays():
+
+    def __init__(self):
+        empty = []
+
+    def prep_clean(self,df):
+        # Rename some things
+        df['EffectiveDate'] = df['EffectiveDatetime'].apply(lambda x: x.date())
+        return df
+
+    """Called in the Event Cleaner, after the data has been subsetted to a given phase"""
+    def phase_clean(self, df, phase):
+        return df
+
+

17/replication_package/code/data/source/build_master/cleaners/clean_events_status.py

@@ -0,0 +1,59 @@
+import os
+import pandas as pd
+from lib.data_helpers.builder_utils import BuilderUtils
+from lib.utilities import serialize
+from lib.experiment_specs import study_config
+from lib.data_helpers import data_utils
+from datetime import datetime, timedelta
+
+class CleanEventsStatus():
+
+    def __init__(self):
+        self.social_hosts = []
+
+    def prep_clean(self,df):
+        # Rename some things
+        return df
+
+    """Called in the Event Cleaner, after the data has been subsetted to a given phase"""
+    def phase_clean(self, df, phase):
+        #df = df.sort_values(by = ["LastUpload"])
+        #df_l = df.groupby(["AppCode"]).last().reset_index()
+        return df
+
+    @staticmethod
+    def get_latest_pd_health(st, mr, uah):
+        #OptedOut represents if the opted out of limits, and EmailEnabled suggests if Chris sends user automatic emails
+        # about missing data
+
+        keep_cols = ["AppCode","PhoneModel","AppVersion","PlatformVersion","LastUpload","Server",
+                     "OptedOut","E-MailEnabled"]
+        st_l = st.sort_values(["LastUpload"]).groupby(["AppCode"]).last().reset_index()
+        st_l["Server"] = st_l["Zipfile"].apply(lambda x: x.split("_")[1] if "_" in x else "nan")
+        st_l = st_l[keep_cols]
+
+        lt = uah.loc[uah["CreatedDate"] >= study_config.active_threshold].groupby(["AppCode"])["UseMinutes"].sum()
+        l = st_l.merge(lt, on = "AppCode", how = 'outer')
+
+
+        ############
+        #status indicates how pd data looks since study_config.active_threshold
+        ############
+        last_survey_complete = data_utils.get_last_survey()
+        code = study_config.surveys[last_survey_complete]["Code"]
+
+        #only code latest status for people that completed the last survey that ended
+        l = l.merge(mr.loc[mr[f"{code}_Complete"]=="Complete",["AppCode",f"{code}_Complete"]], on = "AppCode", how = 'right')
+
+        # has use data in past few days
+        l.loc[l["UseMinutes"]>0, "ActiveStatus"] = "Normal"
+
+        #i.e. no use data, but is status export
+        l.loc[l["ActiveStatus"].isnull(), "ActiveStatus"] = "NoUseDataLately"
+
+        #is also missing pd status data
+        l.loc[(l["ActiveStatus"]=="NoUseDataLately") & (l["PhoneModel"].isnull()),"ActiveStatus"]= "NoPDDataAtAll"
+
+        l = l[keep_cols+["ActiveStatus"]]
+
+        return l