IX. Risk, causation, mediation, and moderation

Throughout this monograph, there has been frequent reference to levels of risk, inference of causation, testing for mediating variables, and the need to consider possible moderating influences. In this chapter, we review what is meant by these concepts, and then seek to pull together the findings from the English and Romanian Adoptee (ERA) studies that were relevant for these issues. When the findings have been presented in detail in earlier chapters, we simply summarize the main salient points. However, with respect to possible genetic moderation of the effects of institutional deprivation, we present new data because these were not considered in earlier chapters. There was a time when most developmental research, particularly that dealing with social development, moved blithely ahead using cross-sectional studies to investigate developmental processes without consideration of the multiple complex ways in which these processes may work together or separately. That is no longer acceptable (Kraemer et al., 1997; Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001; Murray, Farrington, & Eisner, 2009; Rutter, 1988, 2009). Not only must the various processes, and their interplay, be clearly conceptualized, but also it will be essential to pit different refutable causal hypotheses against each other (Lahey, D'Onofrio, & Waldman, 2009; Rutter, 2003, 2006b).