Interpreting and Visualizing Regression Models Using Stata
Tuesday, April 24, 2012 at 1:01PM
Michael Mitchell

I am very excited to announce that my newest book, Interpreting and Visualizing Regression Models Using Stata, is shipping. Taking advantage of the contrast, margins, and marginsplot commands, the book shows how you can use the features of Stata to interpret and visualize your results.

The heart of the book illustrates how to interpret interactions through the liberal use of graphs created using the marginsplot command (the book contains nearly 200 figures). These visual interpretive tools are tied to the regression coefficients, helping you visualize (and visually interpret) your results. The book covers nearly every conceivable kind of interaction... continuous by continuous interactions, categorical by continuous, and categorical by categorical, as well as three way interactions.

The book engages many seemingly complex kinds of interactions, showing how the margins, marginsplot and contrast commands can make sense of such interactions. Categorical by continuous interactions are illustrated where the continuous variable is fit non-linearly (such as using a quadratic term, or using a piecewise regression). The book also includes extensive exploration of the contrast command to interpret interactions among categorical predictors (including interactions of three categorical variables).

The book is not limited to just ordinary regression, also covering models with repeated time points (e.g., repeated measures designs, time series), multilevel models, non-linear models (e.g., logit, mlogit), as well as data collected using complex surveys.

You can find more details, including the table of contents, at the StataPress web site

Article originally appeared on Michael Norman Mitchell (
See website for complete article licensing information.