The presence of GxE correlation complicates the interpretation of GxE interactions because differences in genetic effects in different environments may reflect differences in gene frequency if individuals are selecting themselves into different environments (rather than differences in the effect of genes as a function of the environment, as GxE would indicate). Accordingly, tests of GxE interaction usually control for GxE correlation, or explicitly test for GxE correlation (to ensure there is none for the variables being studied) before testing for GxE interaction. This is an inherently unsatisfactory solution, but it reflects the fact that our analytic methods are limited for capturing the complexity of developmental transactions.
