Method: In a systematic literature search 26,000 titles were scanned for eligibility: 1) evaluation of a psychosocial family-based prevention programs 2) applying a randomized-controlled design 3) within a low-income and/or migration sample. The search delivered a sample of 106 studies to be included in the meta-analysis.
Results: Within the sample of studies 66 programs were conducted in a low-income sample, 11 in a migration sample, and 29 in a sample showing both of these characteristics. A random-effects model delivered medium average effects (d = .270, SE = .0283, p = .000). Effects were robust up to a one-year follow-up. Programs aiming to buffer the effects of poverty were most successful in improving family-related outcomes. In migration samples effects on cognitive development were the strongest, while for mixed groups educational outcomes improved the most. Programs aiming to prevent substance abuse (k = 11) and problematic sexual behavior (k= 16) showed mostly non-significant effects, but will be explored more deeply. Different moderators of effectiveness were tested. It can be shown, that programs were significantly more effective when they were conducted in a disadvantaged neighborhood rather than in a setting where participants were recruited from diverse locations. A significant negative linear relationship between attrition in the control group and effect sizes indicates that implementation quality is essential to program effectiveness.