Sampling scale can cause bias in positive assortative mating estimates: evidence from two intertidal snails

Assortative mating in the wild is commonly estimated by correlating between traits in mating pairs (e.g. the size of males and females). Unfortunately, such an approach may suffer from considerable sampling bias when the distribution of different expressions of a trait in the wild is nonrandom (e.g. when segregation of different size classes of individuals occurs in different microhabitats or areas). Consequently, any observed trait correlation in the wild can be an artefact of pooling heterogeneous samples of mating pairs from different microhabitats or areas rather than true nonrandom matings. This bias in estimating trait correlations as a result of sampling scale is termed the scale‐of‐choice effect (SCE). In the present study, we use two intertidal littorinid species from Hong Kong to show how the SCE can bias size‐assortative mating estimates from mating pairs captured in the wild, empirically demonstrating the influence of this effect on measures of positive assortative mating. This finding cautions that studies overlooking the SCE may have misinterpreted the magnitude and the cause of assortative mating, and we provide a new analytical approach for protecting against this potential bias in future studies.