#Marginsplot stata full#

#Marginsplot stata full#

At the top, the option “Factor variable” should be selected.įrom the list of drop-down “Specification” options, select “2-way full factorial.” In the “Variable 1” box, select voter. This opens the “Create varlist with factor or time-series variables” dialog box as shown in Figure 2. The “regress - Linear regression” dialog box should still be open.Ĭlear the independent variables from the first model from the “Independent variables” drop-down box and click on the button with three dots to the right.

However, the model can also be estimated by using the menu options. (As before, if we were including categorical variables with more than two categories we would need to instruct Stata to treat them accordingly, by inserting “i.” in front of both of them, i.voter#i.gndr.)Īs before, entering the command into the Stata Command window is the simplest way to carry out this estimation. Inserting “#” between the variables instructs Stata to include both variables and an interaction between them when estimating the model. It is possible to have three-way interactions or more, but we focus on the two-way case for ease of explanation. This example will focus on interactions between one pair of variables that are both categorical in nature. An interaction can occur between independent variables that are categorical or continuous and across multiple independent variables. In a linear regression model, the dependent variables should be continuous. Focus is given instead to the difference in slopes which is described by the interaction coefficient. In a model including an interaction term, the slope estimates cannot be interpreted in the same way, as they are now conditional on other values. Most attention is focused on the slope estimates because they capture the relationship between the dependent and independent variables. This requires estimating an intercept (often called a constant) and a slope for each independent variable that describes the change in the dependent variable for a one-unit increase in the independent variable. In a “main effects” multiple regression model, a dependent (or response) variable is expressed as a linear function of two or more independent (or explanatory) variables.

Interaction describes a particular type of non-linear relationship, where the “effect” of an independent variable on the dependent variable differs at different values of another independent variable in the model.