apexog
How large would a violation of the exclusion restriction have to be before the average partial effect from your probit changes sign, or stops being significant? apexog carries the plausibly exogenous framework of Conley, Hansen & Rossi (2012) from linear 2SLS coefficients to average partial effects and average interaction effects on the probability scale, estimated by a control-function probit (or, for one binary treatment, a recursive bivariate probit). Results are reported as a union of confidence intervals over a range of assumed violations, or as a local-to-zero interval under a prior.
The R package and the Stata command give the same results and share one documentation site.
# R
# install.packages("remotes")
remotes::install_github("jespernwulff/apexog")
* Stata (18 or later)
net install apexog, from("https://raw.githubusercontent.com/jespernwulff/stata-apexog/main/")
See also
- [Software]powerape
- [Software]bootmakr
- [Paper]From Improvised to Standardized: Testing Interaction Effects in Logit, Probit, and Other Nonlinear Models
- [Paper]Why You Should Not Use the ITCV With Robust Standard Errors (and What to Do Instead)
- [Book]Binary Regression Models: An Average Partial Effects Approach
- [Paper]Diagnostics Without Microdata: Assessing Credibility of Event Studies and Difference-in-Differences Estimates After Publication
- [Paper]Hic Sunt Dracones: On the Risks of Comparing the ITCV With Control Variable Correlations
- [Paper]A Critical Evaluation of the Impact Threshold of a Confounding Variable in Management Research