README

The files in this replication package replicate the results of Image Concerns in Pledges to Give Blood: Evidence from a Field Experiment by Christian Johannes Meyer and Egon Tripodi. The master file code/replication.do runs all of the code contained in tables_and_figure/do_files/ and will recreate all figures and tables from the paper. The markdown file replication.md contains the dynamic output of all analyses conducted (i.e. the whole code along with the output generated through Stata/R).

Data Availability and Provenance Statements

The data used in this project was collected through a survey and a field experiment conducted in Bonn, Germany. The data is generated by means of a short survey and of the observed behavior of people involved (i.e. whether they pledge to donate and whether they donate with our partner organizations). Survey data is complemented with administrative data on donation behavior obtained from two blood banks in Bonn. The data in this replication package was de-identified to ensure anonymity of respondents.

Dataset List

The analysis in the paper is based on four different datasets:

Data file	Source	Notes
`data/main.dta`	Survey and field experiment	Used to replicate all figures and tables except for Table 4, Appendix Table A1 and Appendix Table D1.
`data/data_table_4.dta`	Administrative data	Used to replicate Table 4.
`data/data_table_a1.dta`	Survey and municipal statistics	Used to replicate Table A1.
`data/data_table_d1.dta`	Survey and field experiment	Used to replicate Table D1.

Computational Requirements

The analysis was conducted using Stata 16.0 and R 4.0.3 and, to the best of our knowledge, the code could be run on any computer with appropriate software installed. Replicating all the results should take a few minutes, although the exact computation time may vary according the specific computer machine used.

Software Requirements

We used the following software:

Stata (code was last run with version 16.0)
- cibar (as of 2021-04-26)
- coefplot (as of 2021-04-26)
- dm79 (as of 2021-04-26)
- estout (as of 2021-04-26)
- kdens (as of 2021-04-26)
- lassopack (as of 2021-04-26)
- mixlogit (as of 2021-04-26)
- moremata (as of 2021-04-26)
- ralpha (as of 2021-04-26)
- rsource (as of 2021-04-26)
- semean (as of 2021-04-26)
- seq (as of 2021-04-26)

R (code was last run with version 4.0.3)
- cowplot (as of 2021-04-26)
- DiagrammeR (as of 2021-04-26)
- dplyr (as of 2021-04-26)
- foreign (as of 2021-04-26)
- grf (as of 2021-04-26)
- haven (as of 2021-04-26)
- magrittr (as of 2021-04-26)
- rlang (as of 2021-04-26)
- tidyverse (as of 2021-04-26)
- xtable (as of 2021-04-26)

Description of Files in Replication Package

This section provides a description of all the files in this replication package.

code/do_files/ is the directory containing 10 Stata do-files, one to replicate each figure and table of the paper.
code/replication.do is the master file that runs all the do-file contained in code/do_files/ and produce all the figures and tables. As already mentioned, you are not expected to run this file as the data are not publicly available.
data/ is the directory containing the datasets used for the analysis.
logs/replication.smcl is the log file produced by running replication.do. The log is already compiled, so that you can look at the commands execution without running the script.
results/ is the directory where all the figures and tables produced by the analysis are stored.
README.md is this readme file.
replication.md is the markdown containing the dynamic output. It contains the to code, together with the relevant output and the figures generated by the code, allowing you to observe code and output without the need of running the related do-file. Please notice this is a markdown file and some of its content are taken from other directories of the replication package, therefore if you make some changes (e.g. modifying or erasing some figures in results/, or moving replication.md to another directory) you may not visualize properly figures within replication.md.

Instructions to Replicators

All results can be replicated by running replication.do. You may need to make the following changes to the file:

Edit line 38 in replication.do to adjust the default path. You should enter the full path leading to ~/replication_package (e.g. "C:/Users/nameoftheuser/Documents/replication_package").
In case you do not want to reproduce the whole results but just specific figures and tables, you only need to comment out lines related to undesired do-files (by means of // or *).

List of Tables and Programs

The following is a list of all the figures and tables of the paper together with the specific program generating each of them:

Figure/Table	Program	Output	Note
Figure 1	`Figure 1.do`	`Figure_1.png`
Figure A1	no data
Figure B1	`Figure B1.do`	`Figure_B1.png`	This figure is generated using R, Stata is only used to run the R commands.
Figure C1	no data
Figure C2	no data
Figure C3	no data
Table 1	`Table 1.do`	`Table_1.xlsx`, and console output
Table 2	`Table 2.do`	`Table_2.tex`, and console output
Table 3	`Table 3.do`	`Table_3.tex`, and console output
Table 4	`Table 4.do`	console output
Table A1	`Table A1.do`	`Table_A1.xlsx`, and console output	Data in column 1 have been taken directly from Bonn City Government Statistical Office 2017 population statistics (https://www2.bonn.de/statistik/dl/ews/Bevoelkerungsstatistik2017.pdf)
Table A2	`Table A2.do`	`Table_A2.tex`, and console output
Table A3	`Table A3.do`	`Table_A3.tex`, and console output
Table B1 Panel A	`Table B1a.do`	`Table_B1a.tex`, and console output
Table B1 Panel B	`Table B1b.do`	`Table_B1b.tex`, and console output
Table D1	`Table D1.do`	console ouptut

chrjmeyer / ImageConcernsPledgestoGiveBlood