ECON 323- Midterm Exam Guide - Comprehensive Notes for the exam ( 66 pages long!)
13 Oct 2017
School
Department
Course
Professor
Document Summary
When the dependent variable is discrete and takes only a few values, it doesn"t make sense to treat it as normally distributed. When tting a linear probability model (lpm), the tted probabilities can be less than zero or greater than one. The partial e ect of a regressor is also assumed to be constant. These limitations can be overcome by using di erent models than the lpm. In this chapter, we will cover logit, probit, poisson and tobit models. athese notes are based on wooldridge, 2016 and stock watson, 2015. Let p(y=1 | x)=g( + x + z + ) = g( + x ) where g() is a function that will take values strictly between 0 and. In the logit model, g is the logistic function: In the probit model, g is the standard normal cdf. The standard normal cdf has a similar shape to that of the logistic cdf.
