From the course: Introduction to Stata 15

Unlock the full course today

Join today to access over 22,600 courses taught by industry experts or purchase this course individually.

The linear probability, logit, and probit models

The linear probability, logit, and probit models - Stata Tutorial

From the course: Introduction to Stata 15

Start my 1-month free trial

The linear probability, logit, and probit models

“

- When a dependent variable is measured as a binary variable, we'll need to make a choice on how to model this. We could estimate such a model, like we've done previously, through ordinary least squares. Such a model is called a linear probability model. However, it has some disadvantages, and this is why we often prefer to specify logit or probit models. One major disadvantage of the linear probability model is that it causes out of bound predictions. Here is a classical example of out of bound predictions from a linear probability model. There are two variables, one continuous x variable, and one binary y variable. The red line represents the predicted values of the linear probability model. Hopefully, you can quickly identify what the problem is. The linear probability model predicts values below zero and above one. However, it's not possible to have a probability that is lower than zero or higher than one. These results are non-nonsensical. The solution is to transform the…

Contents