The data type of the dependent variable in logistic regression must be binomial.

• A. Numeric
• B. Binomial
• C. Categorical
• D. Ordinal

Answer: (B)

Logistic regression is used when you want to model a binary outcome, a yes/no type result. The dependent variable is recorded as 0 or 1, representing failure or success, which means the data follow a binomial distribution (a Bernoulli trial is the special case with a single observation). The model estimates the probability of the event occurring and links it to the predictors through the logistic function. If you truly have a binary outcome for each observation, binomial is the right way to describe the data type the model expects. When there are more than two outcome classes, you’d move to multinomial (or ordinal) logistic models instead, rather than the standard binary form.