The first step in fitting any linear models should be exploratory graphing with the raw data. In many cases, this step will provide preliminary answers to the hypotheses and will provide a general idea of the effect sizes for the predictors. Sometimes, descriptive graphs will not provide the whole picture, in particular when the response is binary or with multiple continuous predictors. In these cases, it useful to plot the model predictions in order to fully understand the model.
This workshop will cover both how to create descriptive graphs and model-prediction graphs for a range of linear and logistic models. This workshop assumes familiarity with linear models (regression and anova) as well as logistic regression.
Methods will be demonstrated in a hands-on fashion using Stata.
If you intend follow along with the Stata coding during the workshop, please make sure that you have the latest version of Stata installed. The latest version of Stata can be purchased on their website, https://www.stata.com/order. Researchers can also request a research computing account through the Cornell Center for Social Sciences by visiting https://accounts.computing.socialsciences.cornell.edu.