1 Computational replication

Do the authors’ results replicate?

Using the authors own code as provided?
When we re-check their data cleaning and feature construction?

1.1 Data sourcing and code structure

The authors provided the data, codebooks, and replication code, linked in their publication and hosted on Harvard Dataverse here. We downloaded these files into the folder OriginalFiles

1.2 Cleaning and reconstruction

We recode the construction of the data used in the analysis in chapter $???$ .

1.3 ‘Stata’ part of code

Re-run Stata analysis code and see if it generates the numbers reported in the text [DONE] (looks good)

Source and documentation:

Florian re-ran the do files as found in ‘original files’

“From_Survey_To_Replication_Dataset.do”: contains the Stata commands that allow to create Replication_Dataset.dta, starting from Survey.dta.

“Replication_Code_Stata.do”: contains the Stata commands that allow to replicate all the results presented in the paper. It is organized by tables.

(?All results or just regressions?)

Their .do file doesn’t create formatted tables, it just generates a log file.

Tables 2-5. Florian scanned it, saw the results looked similar; same coefficients and standard errors.

See the log file: stata_replication_log_20240419.log

1.4 R part of the code

See Replication_Code_R_minor_fixes.R recreates their results using their code with only slight modifications (file paths, package etc.)

# Computational replication {#computational}

Do the authors' results replicate?

1. Using the authors own code as provided?

2. When we re-check their data cleaning and feature construction?

## Data sourcing and code structure

The authors provided the data, codebooks, and replication code, linked in their publication and hosted on Harvard Dataverse [here](https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/NSC8TJ). We downloaded these files into the folder `OriginalFiles`


## Cleaning and reconstruction

We recode the construction of the data used in the analysis in chapter \ref{recode}.

## 'Stata' part of code

> Re-run Stata analysis code and see if it generates the numbers reported in the text [DONE] (looks good)

Source and documentation: 

Florian re-ran the do files as found in ['original files'](https://github.com/ashlab11/ReplicationGames/tree/main/OriginalFiles)

> "From_Survey_To_Replication_Dataset.do": contains the Stata commands that allow to create Replication_Dataset.dta, starting from Survey.dta.

> "Replication_Code_Stata.do": contains the Stata commands that allow to replicate all the results presented in the paper. It is organized by tables.

(?All results or just regressions?)

Their `.do` file doesn't create formatted tables, it just generates a log file.

Tables 2-5. Florian scanned it, saw the results looked similar; same coefficients and standard errors. 

<!-- Note: this could also be checked in chatgpt. -->

See the log file: [`stata_replication_log_20240419.log`](https://github.com/ashlab11/ReplicationGames/commit/c1c70a0f037e0fdfda7405dc0a089fee74f281cb#diff-0caf70398d5ec3d3683924f8a14046b5d176dc8b4a95f3aaabdadd96b9243390)

## R part of the code

See `Replication_Code_R_minor_fixes.R` recreates their results using their code with only slight modifications (file paths, package etc.)