Ordinal Regression Analysis: Using Generalized Ordinal Logistic Regression Models to Estimate Educational Data

摘要: The proportional odds (PO) assumption for ordinal regression analysis is often violated because it strongly affected by sample size and the number of covariate patterns. To address this issue, partial (PPO) model generalized logit were developed. However, these models are not typically used in research. One likely reason restriction current statistical software packages: SPSS cannot perform SAS requires data restructuring. This article illustrates use logistic to predict mathematics proficiency levels using Stata compares results from fitting PO models.