ABSTRACT
A statistical model can be seen as a simplified type of “theory”, describing the process through which observed data were generated. In meta-analysis, we can choose between two alternative models: the fixed-effect model, and the random-effects model. While the fixed-effect model assumes that there is only one true effect size, the random-effects model states that the true effects also vary due to between-study heterogeneity.
This chapter describes the idea behind the fixed- and random-effects model, and how they are implemented in practice. A discussion of various methods to estimate the between-study heterogeneity variance τ 2 https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781003107347/de07b140-3dcb-4e86-8547-42a1b8d32683/content/math0_1.tif"/> in random-effects models is provided.
Using hands-on examples with “real-life” data sets, the chapter illustrates how effect sizes can be pooled in R, depending on the chosen summary measure. Guidelines on how to interpret various results are also provided.
