In circumstances where the proportion of false findings is high, as we suspect is the case for candidate gene main effect and cGxE studies, the issue of publication bias should be considered by both authors and editors. Authors, journal editors, and reviewers should understand that novel positive findings may often be false, and so should set higher standards of evidence, including stricter standards of significance testing, full accounting for all sources of multiple testing, and direct independent replication prior to publication. Equally, investigators should strive to submit, and editors to publish, adequately powered, independent direct replication attempts, irrespective of the outcome -- positive, negative, or null. Meta- or mega-analyses that combine data across multiple studies while accounting for across-study heterogeneity, allow smaller sized studies to contribute to hypotheses necessitating larger samples, such as cGxE. Such collaborative research should be encouraged. These kinds of recommendations, aimed at reducing the role of publication and other biases, are being adopted by journals that have an interest in cGxE; some recent examples are Behavior Genetics (Hewitt, 2012), the Journal of Abnormal Child Psychology (Johnston,
