Generalized estimating equations: A pragmatic and flexible approach to the marginalGLMmodelling of correlated data in the behavioural sciences

• DOI: 10.1111/eth.12713
• Publication Date: 2017-12-11T07:15:07Z
• Authors: Stano Pekár; Marek Brabec
• Keywords: Gee; Marginal model; Estimating equations; Statistical Inference

Abstract Within behavioural research, non‐normally distributed data with a complicated structure are common. For instance, can represent repeated observations of quantities on the same individual. The regression analysis such is both by interdependency (response variables) and their non‐normal distribution. Over last decade, have been more frequently analysed using generalized mixed‐effect models. Some researchers invoke heavy machinery modelling to obtain desired population‐level (marginal) inference, which be achieved simpler tools—namely marginal This paper highlights (using estimating equations [ GEE ]) as an alternative method. In various situations, based fewer assumptions directly generate estimates (population‐level parameters) immediate interest researcher (such population means). Using four examples from we demonstrate use, advantages, limits approach implemented within functions ‘geepack’ package in R.