Regression Models with Categorical Variables 141
■ (^) A linear probability model assumes that the function to be estimated
is linear and, as a result, it is possible to obtain negative probabilities.
■ (^) Unlike the linear probability model, the predicted probability of the
probit regression and logit regression models is forced to be between 0
and 1.
■ (^) The probit regression model and logit regression model are nonlinear
regression models where the dependent variable is a binary variable and
the predicted value is a cumulative probability distribution.
■ (^) The logit regression model differs from the probit model because rather
than the predicted value being a standard normal cumulative probabil-
ity distribution, it is a standard cumulative probability distribution of a
distribution called the logistic distribution.