“…Despite this, the considerable heterogeneity in effect sizes found in these meta-analytic studies raises some concerns about the aggregation of diverse effect sizes across studies. Furthermore, it suggests that there may be sample characteristics (e.g., age, racial, intelligence, socioeconomic status, or cultural background; Heilbrun, 1979;Kosson et al, 1990;Walsh and Kosson, 2007) In line with this suggestion, a prospective study of 199 U.S. offenders found that for European Americans, psychopathy predicted recidivism at lower levels of socioeconomic status (SES) but was unrelated at higher levels of SES. However, for African Americans the predictive power of psychopathy was relatively stable across SES (Walsh and Kosson, 2007).…”