regression models for ordinal responses a review of methods

These resulted in the following log ratios 2

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: , 3 4 = 7. 4°) 1330 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY TABLE 6 Results of fit of polytomous logistic (PL model) and adjacent category (AC model) logistic modelsa Variable Intercept1 (α1) Intercept2 (α2) Intercept3 (α3) Intercept4 (α4) Midline episiotomy 1° 2° 3° 4° a PL model ˆ β ± se (β ) ˆˆ P-value –4.5479 ± 0.0771 –5.0987 ± 0.1013 –3.3328 ± 0.0425 –5.1404 ± 0.1013 –0.4661 ± 0.3628 0.0847 ± 0.3687 0.7280 ± 0.1178 1.6845 ± 0.1945 AC model ˆ β ± se (β ) ˆˆ P-value –5.0987 ± 0.1013 –0.5508 ± 0.1268 1.7659 ± 0.1093 –1.8076 ± 0.1113 0.1988 0.8182 0.0001 0.0001 –0.4661 ± 0.3630 0.5508 ± 0.5158 0.6433 ± 0.3849 0.9565 ± 0.2235 0.1991 0.2856 0.0947 0.0001 Response variable is degree of laceration: none, 1°, 2°, 3°, and 4°. likelihood ratio test of H0: β = γ j = 0 resulted in a χ2 = 89.8 (4 d.f.), a significant improvement over the proportional odds model. We also fit a constrained partial proportional odds model (Table 5b) to the data, with the following c...
View Full Document

This document was uploaded on 02/25/2014.

Ask a homework question - tutors are online