Mothers tend to receive lower wages than comparable women who remain childless. This persisting finding has been reported by the number of studies and is generally referred to as a ‘motherhood wage penalty’ or the ‘family wage gap’. Numerous studies have been conducted in order to measure this gap and understand its sources. Nonetheless, we still lack comprehensive knowledge which factors are responsible for the persistence of motherhood wage penalty. The reported family wage gap varies also heavily across studies. One of the reasons for this state of affairs is that existing studies use different data sources and analytical methods and refer to different samples, countries and periods. In this paper we summarize existing empirical evidence on this topic and analyze sources of the variation among the reported estimates using meta-analysis. The method allows to estimate the true unknown parameter possibly best using the available empirical evidence and to identify reasons for across-study differences in this estimated effect size. Based on 214 wage effects associated with having exactly one child and 252 wage effects associated with the total number of children, we find an average motherhood wage gap of around 3%. The meta-regression reveals that the size of the wage gap is mostly explained by mothers-non-mothers’ differences in human capital measured by age, education, work experience and tenure, as well as occupational and sectoral segregation. Contrary to the Beckerian specialization hypothesis, we find the wage gap to be larger for the single than married mothers. Our results also show that the gap is smallest in Nordic countries, where public policies actively support gender equality and reconciliation of work and family, and largest in the post-socialist Central and Eastern European countries and Anglo-Saxon countries.