Data cleaning code review checklist (TEMPLATE)

Data source/survey round:

Date:

List of files to be checked [Add names or links]

Master script
Clean dataset(s)
Cleaning scripts

Reproducibility

[x] All scripts run from the master after adding the correct folder path to line(s) X (and XX)
[x] The master script is organized in a way that allows you to understand the general tasks being performed in the code
[ ] The master script tracks which scripts create and use which files
[ ] The datasets created by the reviewer are exactly the same as those shared by the coder

Code organization and readability

[ ] Code file names are informative
[x] It is clear in the code why tasks are being executed
[x] The code structure facilitates understanding of the tasks
[x] Code uses white space to improve readability
[x] There is extensive use of comments to explain the code
[ ] The code is efficient (tasks are executed in the simplest way possible, loops are used when needed rather than repeating lines, pre-defined functions are used)
[ ] Common tasks are abstracted and automated (e.g. using functions or macros)

Clean dataset checks (pre-publication)

[x] The data does not include direct identifiers
[x] The dataset has a clearly labeled, uniquely and fully identifying ID variable(s)
[x] The level of observation of the dataset is clear from the dataset name, ID variables, and documentation
[x] Variables have informative labels or an accompanying dictionary
[x] Categorical variables have clear and informative value labels
[x] No modification is made from the raw to the clean data other than correcting problems
[ ] No raw variables are processed (winsorized, for example)
[ ] Variables can be easily traced back to the original questionnaire
[x] Variables are sensibly ordered and dataset is compressed and saved in appropriate format

Data cleaning tasks

[x] Are new variables being created in the cleaning do-files?
[x] Are any changes being made to observations values in the cleaning do-files?
[ ] Check merges: Are any observations dropped? If so, is there a clear justification for that? If any observations didn't match, is that explained in the comments?
[ ] Are missing values coded consistently? Are extended missing values used?

The code is excellent! Follows best practices and easy to read. Since I didn't have much to add to 01_clean_0.do, I've made a few suggestions to the rest of the code.

Suggested changes to Master:

Included section switches. Implemented in 989a81589c8373c7f9dfbbcf22809bba1b6a24c9.
Dynamically defined file paths globals depending on user. Implemented in c3d1b2586baa1f05b9dff9ebb60ca4a171701eb4.
Add a brief description to what each code does.

Suggested changes to other codes:

Changed varabbrev to off. Implemented in 3e56a7ef9579eb645e22338ac09b552bb0ae1150, and 991505b960a17daa75ab82ca16fdc718c972391f.

Using abbreviated var names is not a big deal but it is considered bad practice. It can lead to errors that are very hard to debug. In fact, there is one case in your code that could be an error, although I'm not sure if it was intentional or not. Here's the commit c8d06bf5daeef8f929a7b146bb5a19c9bdbfa35a

LeonardoViotti / code-review