Dummy variables are used to...?

• A. A test for serial correlations between errors.
• B. A test for recoding a categorical variable into multiple dichotomous variables.
• C. A measure of effect size.
• D. A graphical representation of the mean with the 95% confidence interval.

Answer: (B)

Dummy variables are a coding method used to bring categorical predictors into regression analysis by turning categories into binary indicators. This lets the model estimate the effect of belonging to each category relative to a reference category. For a categorical variable with k categories, you create k−1 dummy variables, each coded 1 if the observation is in that category and 0 otherwise, with the remaining category serving as the baseline. For example, if a variable has three categories like North, South, and East, you might create two dummies: North (1 if North, 0 otherwise) and South (1 if South, 0 otherwise). The coefficient for North then represents the expected difference from the baseline category (East). This approach is specifically about coding the variable for use in models, not about testing serial correlations, measuring effect size directly, or graphically displaying means with confidence intervals.