Extra-pair paternity (EPP), where offspring are sired by a male other than the social male, varies enormously both within and among species. Trying to explain this variation has proved difficult because the majority of the interspecific variation is phylogenetically based. Ideally, variation in EPP should be investigated in closely related species, but clades with sufficient variation are rare. We present a comprehensive multifactorial test to explain variation in EPP among individuals in 20 populations of nine species over 89 years from a single bird family (Maluridae). Females had higher EPP in the presence of more helpers, more neighbours or if paired incestuously. Furthermore, higher EPP occurred in years with many incestuous pairs, populations with many helpers and species with high male density or in which males provide less care. Altogether, these variables accounted for 48% of the total and 89% of the interspecific and interpopulation variation in EPP. These findings indicate why consistent patterns in EPP have been so challenging to detect and suggest that a single predictor is unlikely to account for the enormous variation in EPP across levels of analysis. Nevertheless, it also shows that existing hypotheses can explain the variation in EPP well and that the density of males in particular is a good predictor to explain variation in EPP among species when a large part of the confounding effect of phylogeny is excluded.