Ethnicity remains one of the most salient layers of individuals' social identities, and information about the distribution of ethnic identities turns out to be crucial for many studies that investigate political or social processes in divided societies. In the aftermath of civil wars, however, censuses providing such data are controversial and often delayed. Where census data are lacking, researchers can make use of the multiplier method to infer the distribution of ethnic identities based on indirect observations from sub-samples of the population. However, due to the selective nature of their data, the sub-samples might not be representative of the population. This paper proposes a method, which corrects for such selection effects. Based on the ethnic identity of birth-giving parents, the authors estimate the distribution of ethnic identities in the municipalities of the Federation of Bosnia and Herzegovina for 2008–2010. They correct for possible selection biases by including economic, demographic and war-related variables. Multiple tests of validity show that this estimation appears to be the most accurate procedure currently available for the distribution of ethnic identities in municipalities in Bosnia and Herzegovina.