A family of ordered logistic regression models fit by data expansion

Combined analysis of retrospective and prospective occurrences in cohort studies: HIV-1 serostatus and incident pneumonia

“…We propose expanding [2][3][4] the observed data set from N 5 2610 to 2N 5 5220, where each subject contributes two records, one retrospective and one prospective. The retrospective record contains data on the occurrence time Extending the work of Allison, 5 the hazard or instantaneous risk of a bidirectional occurrence takes its standard form, conditional on occurrence type R, as…”

Markov Models for Repeated Ordinal Data

“…Therefore,t he factorization (2)a llows to compute the continuation ratiom odelb y meanso f" data expansion" [11] ands oftware for ordinary binaryr egression. Each observation is copied up to D timestoobtain one copy for each of the auxiliaryvariables z s as indicated above.…”

“…Theo rdinary proportionalo dds model can be computedb yt he SAS procedure LOG-ISTIC. Forthe continuation ratiomodeland the generalized proportionalodds model the SAS code showninFigure 2can be used to preparet he data [11]. Thed atah avet ob e sorted by au nique variablet hati dentifies each "unit" ("id"say)and thenumber of the consecutive observations per unit ("visit" in the example of the supplementation study).…”