Findings: Mathematics Achievement
The evidence related to the association between math achievement and violence has thus far been ambiguous. Two of four studies in this review report a PoC in the predicted direction (Mutschler, 1997; Zentall & Smith, 1993), one reports that the relationship is as predicted for males but not females (Koda, 1999) and one and one reports null findings (Tsamis et al., 2009). Only one study of nonviolent-only offending was reported, and it suggests that math ability is negatively associated with nonviolent offending (Mutschler, 1997). This evidence suggests an association between math ability and violence, at least for males, but provides no information about the differential etiology of violence.
Among studies comparing violent to nonviolent offenders, the evidence leans in the direction of a differential etiology. Four of six studies reported a PoC in the predicted direction (Bryant et al., 1984; Lewis et al., 1979; Mutschler, 1997; Zentall & Smith, 1993), with two of those reporting a preponderance of statistically significant coefficients. Two studies report findings that do not support math ability as a differential predictor of violence (Marcus & Gray, 1998; Tarter et al., 1983). The ratio is higher than chance would predict, though not as strong as we would need to make confident conclusions, given the methodological weaknesses of these designs. Nevertheless, a strong majority of comparisons (7 of 9) were negative, as predicated, and findings from one of the remaining were just ambiguous, not contradictory.