![](img018.GIF) |
Random
effects model is based on the assumption that the studies are part of a large universe of
similar studies. The relative risk estimate can be used to answer the question that will
the intervention or the exposure be associated with the outcome studied (compare this with
fixed effects model where the question was: did the studies show on average that
the exposure or treatment associated with the outcome of interest?). However, a random
effects model is computationally more intensive than a fixed effects model. |