IntroductionMixed‐effects models are a popular approach for analysis of longitudinal and/or clustered binary outcomes, with many texts either devoted to or with substantial treatment on the topic (Raudenbush and Bryk, ; Demidenko, ; Skrondal and Rabe‐Hesketh, ; Molenberghs and Verbeke, ; Brown and Prescott, ; Hedeker and Gibbons, ; McCulloch et al., ; Fitzmaurice et al., ; Goldstein, ). Such models are often referred to as belonging to the class of models known as generalized linear mixed models (GLMMs). It is well known that the regression parameters in such models have the “subject‐specific” (SS) or conditional interpretation (Neuhaus et al., ), representing effects of the covariates controlling for, or holding constant, the random effects. This is in contrast to the marginal or “population‐averaged” (PA) estimates of, most notably, the GEE approach (Zeger and Liang, ) and the marginalized multilevel model (Heagerty and Zeger, ). The SS and PA regression parameters generally differ in meaning and value, and the SS regression parameters depend on how many random effects are included in the model (e.g., conditional on random intercepts only or conditional on random intercepts and trends). Thus, Zeger et al. () recommend that SS models are useful when the focus
Biometrics – Wiley
Published: Jan 1, 2018
Keywords: ; ; ; ;
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud