6.5 Take-Home Points

• We use regression analysis if our dependent variable is numeric and we have at least one numeric independent (predictor) variable.

• We use dummy variables to include a categorical variable as a predictor in a regression model. We need a dummy (1/0) variable for each category on the categorical variable except for one category, which represents the reference group.

• We use an F test to test the null hypothesis that the regression model does not help to predict the dependent variable in the population. We use a t test to test the null hypothesis that a regression coefficient is zero in the population.

• In a regression model, moderation means that there are different slopes (effects of the predictor) for different groups or contexts (moderator).

• Interaction variables represent moderation in a regression model.

• An interaction variable is the product of the predictor and moderator. If a categorical moderator is represented by one or more dummy variables, we need an interaction variable for each of the moderator’s dummy variables.

• Statistical inference for an interaction variable is exactly the same as for “ordinary” regression predictors.

• The effect of the predictor in a model with an interaction variable does not represent a main or average effect. It is a conditional effect: The effect for cases that score zero on the moderator.

• To interpret moderation, describe the effects (slopes, unstandardized regression coefficients) and visualize the regression lines for different groups.

• Warn the reader if the predictor scores are not nicely distributed for all groups or levels (no common support).

• Don’t use the standardized regression coefficients (Beta) for interaction variables, variables included in interactions, or for dummy variables in SPSS.