Determine which of the following statements is NOT true about the linear probability, logistic, and probit regression models for binary dependent variables.

• The three major drawbacks of the linear probability model are poor fitted values,heteroscedasticity, and meaningless residual analysis.
• The logistic and probit regression models aim to circumvent the drawbacks of linear probability models.
• The logit function is given by [math]\pi(z) = e^z/(1 + e^z) [/math]
• The probit function is given by [math]\pi(z) = \Phi(z) [/math] where [math]\Phi[/math] is the standard normal distribution function.
• The logit and probit functions are substantially different.

Key: E

The logit and probit models are similar (see page 307 of Frees, which also discusses items A-D).

Copyright 2023. The Society of Actuaries, Schaumburg, Illinois. Reproduced with permission.