The following STATA programs developed by our research group are used in several papers posted on this website. These programs, which have been optimized to run efficiently on large datasets, may be useful in other applications as well.
Binned Scatterplots
binscatter.ado is a tool to non-parametrically visualize conditional expectation functions in large datasets. The program also produces regression-discontinuity plots in a computationally efficient manner. For examples of binned scatter plots, see Figure 2 and the discussion in Section 4 in Chetty, Friedman, and Rockoff (2014). Binscatter can be installed directly from the SSC (ssc install binscatter) and a set of examples are also available (help binscatter). binscatter.ado was written by Michael Stepner, with input from Jessica Laird and Laszlo Sandor.
Estimation of Bunching in Distributions
bunch_count.ado calculates the degree of bunching at exogenously specified points in a distribution. This ado file was used to calculate the statistics in Chetty, Friedman, Olsen, and Pistaferri (2011) to quantify the degree of bunching in income at kink points in the tax system and to plot histograms in a computationally efficient manner. bunch_count.ado was written by Tore Olsen.
Estimation of Bunching in Distributions
Estimation of Teacher Value-Added
vam.ado calculates drift-adjusted value-added measures of teacher quality from student-level test score data following the method developed by Chetty, Friedman, and Rockoff (2014). vam.ado was written by Michael Stepner, building on work by Jessica Laird and Heather Sarsons.
Estimation of Teacher Value-Added