Regression with a Binary Dependent Variable Chapter 11 Regression with a Binary Dependent Variable
Regression with a Binary Dependent Variable (SW Chapter 11)
Example: Mortgage denial and race The Boston Fed HMDA data set
The Linear Probability Model (SW Section 11.1)
The linear probability model, ctd.
The linear probability model, ctd.
Example: linear probability model, HMDA data
Linear probability model: HMDA data, ctd.
Linear probability model: HMDA data, ctd
The linear probability model: Summary
Probit and Logit Regression (SW Section 11.2)
Probit regression, ctd.
STATA Example: HMDA data
STATA Example: HMDA data, ctd.
Probit regression with multiple regressors
STATA Example: HMDA data
STATA Example, ctd.: predicted probit probabilities
STATA Example, ctd.
Logit Regression
Logit regression, ctd.
STATA Example: HMDA data
Predicted probabilities from estimated probit and logit models usually are (as usual) very close in this application.
Example for class discussion:
Hezbollah militants example, ctd.
Predicted change in probabilities:
Estimation and Inference in Probit (and Logit) Models (SW Section 11
Probit estimation by nonlinear least squares
Probit estimation by maximum likelihood
Special case: the probit MLE with no X
The MLE in the “no-X” case (Bernoulli distribution), ctd.:
The MLE in the “no-X” case (Bernoulli distribution), ctd:
The probit likelihood with one X
The probit likelihood function:
The Probit MLE, ctd.
The logit likelihood with one X
Measures of fit for logit and probit
Application to the Boston HMDA Data (SW Section 11.4)
The HMDA Data Set
The loan officer’s decision
Regression specifications
Table 11.2, ctd.
Table 11.2, ctd.
Summary of Empirical Results
Remaining threats to internal, external validity
Summary (SW Section 11.5)