In 2003, Caspi et al. published a landmark paper in Science reporting an interaction, between a polymorphism (5HTTLPR) in the promoter region of the serotonin transporter gene (5HTT renamed SLC6A4) and the number of life events experienced, in predicting risk of depression. This paper has been widely cited and has become the dominant research paradigm in psychiatric genetics because it appears to reconcile the competing claims of genetic and environmental etiology. There have been numerous replication studies, with highly inconsistent results, some supporting the interaction as reported by Caspi, some finding no interaction and others reporting the interaction in the opposite direction. Further, it has recently become apparent that genotyping of the 5HTTLPR polymorphism is somewhat more complicated than previously thought, casting doubts on the veracity of earlier studies. Most importantly, it has been suggested that the GxE detected may be an artifact of the unsatisfactory psychometric properties of the psychiatric constructs being used. We are analyzing data from our large twin family collection (5200 individuals) for whom depression diagnoses, risk factors and DNA samples are available, to bring considerable power to the question of GxE interactions for depression. Further, we are investigating whether the inconsistently replicated interaction of 5HTTLPR (and, independently, any single nucleotide polymorphism haplotype) with stressful life events on risk of depression is moderated by level of social support and whether this could explain the inconsistency. We are also exploring the extent to which IRT modeling will help resolve the effects of GxE from artifacts of measurement.