The purpose of this review was to conduct a meta-analysis of 25 years of mathematics interventions for students with mathematics difficulty or disability in Grades 4 through 12. A search of the literature yielded 25 studies that met the inclusion criteria. Studies were coded to extract important study information (e.g., participant information, research design, description of treatment, and comparison groups) and data needed to calculate Hedge's "g." We used robust variance estimation (RVE) to address dependence resulting from multiple outcomes per study. The RVE random-effects model estimated a treatment effect of 0.85. After adjusting for small-study effects, the final model estimated an underlying, moderate effect of 0.49 with a large amount of unexplained heterogeneity between studies. Studies with more than 15 hr of treatment and those focused on fraction content significantly moderated mathematics outcomes. Findings are limited by extreme variability across study estimates, the lack of standardized mathematics measures, and a limited number of studies across 25 years of research.