Logit is used when predicting limited dependent variables. By virtue of the binary dependent variable, these models do not meet the key assumptions of OLS. Logit uses maximum likelihood estimation (ML...Logit is used when predicting limited dependent variables. By virtue of the binary dependent variable, these models do not meet the key assumptions of OLS. Logit uses maximum likelihood estimation (MLE), which is a counterpart to minimizing least squares. MLE identifies the probability of obtaining the sample as a function of the model parameters. It answers the question, what are the values for BB’s that make the sample most likely?
