distcomp: Comparing distributions

Abstract

In this article, I introduce the distcomp command, which assesses whether two distributions differ at each possible value while controlling the probability of any false positive, even in finite samples. I discuss syntax and the underlying methodology (from Goldman and Kaplan [2018, Journal of Econometrics 206: 143–166]). Multiple examples illustrate the distcomp command, including revisiting the experimental data of Gneezy and List (2006, Econometrica 74: 1365–1384) and the regression discontinuity design of Cattaneo, Frandsen, and Titiunik (2015, Journal of Causal Inference 3: 1–24).

Technology within and across Firms

“…(2) 22 We use the STATA discomp package developed in Kaplan (2019). Because of "multiple testing problem" that increases the Type I error (α), the KS-based multiple test uses "familywise error rate" (FWER) that provides the probability of rejecting at least one true null hypothesis.…”

“…Because of "multiple testing problem" that increases the Type I error (α), the KS-based multiple test uses "familywise error rate" (FWER) that provides the probability of rejecting at least one true null hypothesis. For the details of this test, please see Goldman and Kaplan (2018) and Kaplan (2019).…”

Talent, Career Choice and Competition: The Gender Wage Gap at the Top

“…The data on individuals originate from Sweden's official wage statistics and contain detailed information on a representative sample of the labor force, including full-time equivalent wages, education, occupation, and gender. 23 All data sets are matched by unique identification codes. To make the sample of firms consistent across the time periods, we restrict our analysis to firms with at least 20 employees in the nonagricultural private sector, which are available throughout the period.…”

“…24 This wage difference declines somewhat during the later period, 2003-2009, where the wage advantage for men is approximately 15 percent. 23 The worker data originate from the Swedish annual salary survey (Lönestrukturstatistiken). The survey's sampling units consist of firms included in Statistics Sweden's firm data base (FS).…”

Field experiments and public policy: festina lente

“…10There have also been advances in flexible parametric quantile regression methods, due to Frumento and Bottai (2016, 2017) and implemented by Bottai and Orsini (2019). Similarly, there have been advances in the econometric methods for comparing unconditional distributions, due to Goldman and Kaplan (2018) and implemented by Kaplan (2019). Carter et al (2019) offer a great illustration of how one can use information ‘in the tails’ to drive better development policy.…”