The dependent variable in logistic regression is binary.

• A. True
• B. False
• C. Not necessarily
• D. Depends on data

Answer: (A)

Logistic regression is built to model the probability of a yes/no outcome. The dependent variable is encoded as a binary indicator (0 or 1), representing one of two possible states. Because the response is a Bernoulli random variable, the model uses a binomial likelihood and a logistic (sigmoid) link to map any linear combination of predictors to a probability between 0 and 1. In this standard form, the outcome is inherently binary, which is why the statement is true. There are extensions, like multinomial or ordinal logistic regression, that handle more than two categories, but those are not the standard binary logistic regression. Therefore, the best answer is that the dependent variable is binary in logistic regression.