It is widely assumed that witnessing physical aggression between parents is detrimental to children’s well-being, that parent involvement in school increases academic success in offspring, that playing violent Internet games fosters aggression, and that victimization from bullying may result in emotional problems. Many environmental factors known to predict—and presumed to influence--children’s psychosocial development cannot be subject to causal—meaning experimental--analysis for ethical reasons (e.g., a research design where children are randomly assigned to exposure to bullying, harsh parenting, single parenthood/poverty) and/or practical ones (e.g., lack of peer intimacy, insecure attachment, poor self-esteem). Data presumptively illuminating the causes of individual variation in children’s emotions, cognition and behavior, then, mostly derive from observational/correlational studies, studies whose results are being combined ever more frequently for purposes of meta-analysis on the supposition that such efforts can improve causal inference. Because results of both individual studies and of meta-analyses can and are used to inform practice and policy, there is a need to evaluate how solid the empirical foundation is. By applying an econometric method, novel to psychology and medicine, we seek to determine whether prevailing understanding about diverse causes of variation in children’s functioning and mental health is sound.