Abstract: After we fit a model, our analysis does not stop. We want to use our results to construct counterfactual scenarios. We want to study the effects of changes in variables over the population or for a specific subpopulation. Answering such questions is more challenging for nonlinear models and, in particular, for models in which we make no assumptions about functional forms—nonparametric models. In this course, we will illustrate how to answer these and other relevant empirical questions for nonlinear cross-sectional and panel-data models and for nonparametric models. We do this within a unified framework using Stata.

Abstract: After reviewing the several user-written packages that exist to study networks with Stata, I will focus on the nwcommands package developed by Thomas Grund and especially detail my extensions to this package. I will stress how to declare data to be a network with the help of some commands I developed (nw_fromlist, nw_fromneighbor, nw_fromstem, etc.). Then, I will show how to compute standard econometrics regressions (OLS, etc.) using network metrics, as well as how to use network-designed econometrics methods (QAP). I will rely on data on international trade, known as the world trade web, as well as data on the international investments of multinational firms and show how to measure the resemblance between those two networks.

Abstract: The user-written command oaxaca (Jann 2008) is often used for analyzing gender wage gaps in Stata, and extensions for nonlinear decomposition have been proposed by Yun (2005), Fairlie (fairlie: Jann 2006), and Bartus (gdecomp). I review these commands and I present tools for extending the analysis to structural equation models to account for indirect effects and for automating the presentation of decomposition results (relying on esttab).

Abstract: Econometric modelling for causal inference and program evaluation have witnessed a tremendous development in the last decade, with new approaches and methods addressing an expanding set of challenging problems, both in medical and the social sciences. This workshop covers some recent developments in causal inference and program evaluation using Stata and, in particular, causal inference with continuous treatment (namely, dose–response models). I will discuss the logic of dose–response models, the generalized propensity score (GPS) approach, and regression-adjustment-based dose–response models (RADR). I will present applications of the GPS approach via the Stata commands gpscore and doseresponse, and applications of the RADR models via the Stata command ctreatreg, which I recently developed.

Abstract: I present two new tools that can be used for data management. reindex is a quick way to convert between levels, growth rates, and indices with different scales and base periods. indielabels imports and exports labels from or to plain format text or Excel to be shared independently from the statistical software package. Both programs will be illustrated using practical working examples.

Abstract: Unit-level data refer to a detailed dataset for the sampled units (a village, a district, a household, or any other unit), along with their sampling weights. This presentation sheds light on how Stata could be used for handling unit-level data, especially when the data are recorded in text format. I focus on extracting of unit-level data in text format, analyzing unit-level data, using multipliers and sampling weights, and visualizing unit-level data.

Abstract: Stata developers present will carefully and cautiously consider wishes and grumbles from Stata users in the audience. Questions, and possibly answers, may concern reports of present bugs and limitations or requests for new features in future releases of the software.