Premium
Generalized estimating equations: A pragmatic and flexible approach to the marginal GLM modelling of correlated data in the behavioural sciences
Author(s) -
Pekár Stano,
Brabec Marek
Publication year - 2018
Publication title -
ethology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.739
H-Index - 74
eISSN - 1439-0310
pISSN - 0179-1613
DOI - 10.1111/eth.12713
Subject(s) - gee , marginal model , generalized estimating equation , generalized linear model , inference , econometrics , population , marginal distribution , statistics , mathematics , generalized linear mixed model , computer science , estimating equations , regression analysis , artificial intelligence , maximum likelihood , random variable , demography , sociology
Abstract Within behavioural research, non‐normally distributed data with a complicated structure are common. For instance, data can represent repeated observations of quantities on the same individual. The regression analysis of such data is complicated both by the interdependency of the observations (response variables) and by their non‐normal distribution. Over the last decade, such data have been more and more frequently analysed using generalized mixed‐effect models. Some researchers invoke the heavy machinery of mixed‐effect modelling to obtain the desired population‐level (marginal) inference, which can be achieved by using simpler tools—namely by marginal models. This paper highlights marginal modelling (using generalized estimating equations [ GEE ]) as an alternative method. In various situations, GEE can be based on fewer assumptions and directly generate estimates (population‐level parameters) which are of immediate interest to the behavioural researcher (such as population means). Using four examples from behavioural research, we demonstrate the use, advantages, and limits of the GEE approach as implemented within the functions of the ‘geepack’ package in R.