In a two-sample t-test, the grouping variable should be which type?

• A. Categorical with only two categories (dichotomous)
• B. Continuous
• C. Categorical with three or more categories
• D. Ordinal

Answer: (A)

The test compares the average outcome between two separate groups, so the grouping variable must define exactly two categories. A binary, dichotomous categorical variable creates two independent samples whose means you can contrast. If the grouping variable has more than two categories, you’d need a method that handles multiple groups, like ANOVA, to compare all means. If the grouping variable were continuous, you wouldn’t form two groups to compare; instead you’d analyze the relationship between the continuous factor and the outcome via regression or correlation. So the standard setup uses a categorical variable with two categories, such as treatment vs. control, to split the data into two groups for comparing their means.