Scribd
Upload
108 views

Ordered Probit and Logit Models Example

The document describes using ordered probit and logit models to analyze factors influencing health status. The models show health status is better with lower age, higher income, and fewer diseases. Marginal effects indicate an increase in income decreases the likelihood of fair or good health status while increasing the likelihood of excellent health status.

Uploaded by

tegar26
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
108 views

Ordered Probit and Logit Models Example

The document describes using ordered probit and logit models to analyze factors influencing health status. The models show health status is better with lower age, higher income, and fewer diseases. Marginal effects indicate an increase in income decreases the likelihood of fair or good health status while increasing the likelihood of excellent health status.

Uploaded by

tegar26
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Ordered Probit and Logit Models Example

Ani Katchova

2013 by Ani Katchova. All rights reserved.

Ordered Probit and Logit Models Example

Data are from the Rand Health Insurance Experiment and Deb and Trivedi (2002).
We study the factors influencing the health status (fair, good, excellent).
Health status
Fair
Good
Excellent

Codes for
alternatives
1
2
3

Percent
frequency
9%
36%
54%

The dependent variable has 3 categories/alternatives: fair, good, excellent.


There will be 1 set of coefficients with two intercepts.
There will be 3 sets of marginal effects, one for each category.

Ordered choice model coefficients


Health status

Ordered logit Ordered probit


coefficients coefficients
Age
-0.03*
-0.01*
Income
0.28*
0.16*
Number of diseases
-0.05*
-0.03*
Intercept cut1
-1.39
-0.79
Intercept cut2
0.95
0.54
Coefficient interpretation: the health status is better (from fair to good to excellent) with lower
age, higher income and lower number of diseases.
The threshold/intercept parameters are significantly different from each other so the three
categories should not be combined into one.
The logit and probit ordered model coefficients differ by a scale factor (and therefore we
cannot interpret the magnitude of the coefficients).

Ordered logit model marginal effects


Health status

Age
Income
Number of diseases

Ordered logit
marginal effects for
fair health status
0.002*
-0.02*
0.003*

Ordered logit
marginal effects for
good health status
0.005*
-0.05*
0.009*

Ordered logit
marginal effects for
excellent health status
-0.007*
0.07*
-0.01*

Marginal effects interpretation: one unit increase in income is associated with being 2% less
likely to be in the fair health status, 5% less likely to be in good health status, and 7% more
likely to be in an excellent health status.
The marginal effects sum up to zero.
The marginal effects for the probit model (not reported here) are similar to those of the logit
model.

You might also like