Home

Menu

Loading wiki pages...

View
Wiki Version:
<p>This page provides replication do files for "Heterogeneous Treatment Effects in the Low Track: Revisiting the Kenyan Primary School Experiment"</p> <p>Most code will be provided for Stata (currently all). The Wild Gradient Bootstraps (not yet available) will be provided in R code, but the Stata code here employs an alternative standard error estimator that produces very similar confidence intervals.</p> <p>The code is generally written so as to be easily manipulable by the replicating researcher, facilitating robustness checks and exploration of alternative specification.</p> <p>To begin:</p> <p>1 - Download the study data from here:</p> <p><a href="https://www.aeaweb.org/articles?id=10.1257/aer.101.5.1739" rel="nofollow">https://www.aeaweb.org/articles?id=10.1257/aer.101.5.1739</a></p> <p>2 - Download the "DDK_Setup_OSF" do file from the "Replication Do Files" folder below, and open in Stata editor</p> <p>3 - Enter the appropriate folder names into the global Macros at the top. i. I suggest you set the bootstrap reps to 3 initially if you want to just replicate the point estimates. Bootstraps of local non-parametric regressions can be computationally intensive. These take approximately 20-30 minutes to run on my computer (Macbook Dual Core).</p> <p>4 - Run "DDK_Setup". i. This will generate a clean, analysis-ready data set with sub-groups defined similarly for all further analysis. ii. it will point any files you save to the folder you name in "DDKsave" global, as well as saving the clean dataset there.</p> <p>5 - You can now run any of the analysis do files. In most, at the top of the do file you can define your sub-group of interest using any of the definition generated in Setup or any definition you like. i. a number of files also allow you to adjust parameters of the model, such as bandwidth, regression specifications, etc.</p> <p>6 - More extensive documentation, along with the missing do files, will be posted as soon as possible.<br> i. Missing do files include the Wild Gradient Bootstrap code and the test of rank similarity. </p>
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.
Accept
This website relies on cookies to help provide a better user experience. By clicking Accept or continuing to use the site, you agree. For more information, see our Privacy Policy and information on cookie use.
Accept
×

Start managing your projects on the OSF today.

Free and easy to use, the Open Science Framework supports the entire research lifecycle: planning, execution, reporting, archiving, and discovery.