This is an ongoing list of resources for understanding the 'margins' and 'marginsplot' commands in Stata. If you know of links I have omitted, please let me know.
The 'margins' and 'marginsplot' commands were introduced in Stata 11. They are very useful as a way of estimating and graphing predicted probabilities and are potentially much more informative than regression tables for certain kinds of analysis (e.g. when examining interaction effects, probit or logit regressions or any model which returns non-intuitive coefficients, like a negative binomial).
Despite their advantages, these commands are still somewhat underutilized, at least in terms of the papers I read. It may be the case that since margins is only a few years young, many researchers simply don't know about it or don't fully understand how useful it can be at clarifying regression results. Hopefully one side-benefit of this post will be to generate discussion from people who are using margins in their own research and/or flush out people who have valid objections to it that I may not be aware of.
1. Margins overviews:
- From Stata.com: Overview of margins & marginsplot.
- From UCLA: A brief overview of what margins can do & how to use it to examine interactions.
- From SSCC: Exploring Regression Results using Margins.
- The help file for margins in Stata 13.
- The help file for marginsplot in Stata 13.
2. I highly recommend Richard Williams's slides on "Using Stata’s Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects" (2011). If you want to cite Williams, use his Stata Journal article on the same topic (2012).
3. Bill Rising's slides on "How to get an edge with margins and marginsplot" (2012).
4. Ben Jann's slides on "Predictive Margins and Marginal Effects in Stata" (2013). Benn Jann also has a paper and ppt on his excellent graphing command 'coefplot'.
5. The definitive text on margins is J. Scott Long & Jeremy Freese's newly released book Regression Models for Categorical Dependent Variables using Stata by J. Scott Long & Jeremy Freese (3rd edition) (2014).