Measures of assortative mating may be biased by environmentally-driven resemblance of the partners
​Environmental factor affecting both members of a pair may drive resemblance in labile traits such as behavior expressed by the individuals that form a pair
​Assortative mating is traditionally measured as the correlation between the phenotypes of males and females across mated pairs. Although this approach is valid for fixed traits, different issues arise when applied to labile traits. Importantly, this approach assumes that environmental sources of resemblance between mated individuals are non-existent which means that assortative mating is likely to be overestimated when these environmental effects occur and are not taken into account.
In this methodological paper, we explored different statistical approaches which allow estimating assortative mating while dealing with environmentally-driven source of resemblance. Using simulations of two bird populations, we show that the performance of the different approaches depends on the population’s characteristics, the trait repeatability and the experimental design used to collect data.  These approaches can potentially be extended to other study systems than bird populations. The take-home message is that mated individuals can look similar not because they are assortatively mated but because they share the same environment. Although environmentally-driven resemblance of partners is very likely to occur in nature, it has yet to be empirically studied in wild populations.

Barbara Class