Getting Started in Logit and Ordered Logit Regression - Data and ...
Getting Started in Logit and Ordered Logit Regression - Data and ...
Getting Started in Logit and Ordered Logit Regression - Data and ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Ord<strong>in</strong>al logit: predicted probabilities<br />
Follow<strong>in</strong>g Hamilton, 2006, p.279, ologit estimates a score, S, as a l<strong>in</strong>ear function of the<br />
X’s:<br />
S = 0.22X 1<br />
-0.05X 2<br />
Predicted probabilities are estimated as:<br />
+ 0.11X3 + 0.22X4 -0.29X 5<br />
+ 0.003X6 + 1.57X7 P(y_ord<strong>in</strong>al=“disagree”) = P(S + u ≤ _cut1) = P(S + u ≤ -0.5528058)<br />
P(y_ord<strong>in</strong>al=“neutral”) = P(_cut1 < S + u ≤ _cut2) = P(-0.5528058 < S + u ≤ 0.5389237)<br />
P(y_ord<strong>in</strong>al=“agree”) = P(_cut2 < S + u ) = P(0.5389237 < S + u)<br />
To estimate predicted probabilities type predict right after ologit model. Unlike<br />
logit, this time you need to specify the predictions for all categories <strong>in</strong> the ord<strong>in</strong>al<br />
variable (y_ord<strong>in</strong>al), type:<br />
predict disagree neutral agree<br />
PU/DSS/OTR
Change language