The Mathematics Anxiety Rating Scale (MARS) (F. Richardson and R. Suinn, 1972) was submitted to a reliability generalization analysis to characterize the variability of measurement error in MARS scores across administrations and to identify possible study characteristics that are predictive of reliability variation. The meta-analysis was performed with 67 studies that met study criteria. In general, the MARS and its variants yielded scores with strong internal consistency and test-retest reliability estimates, although variation was observed. Adult samples were related to lower score reliability compared to other age groupings. Inclusion of total score standard deviation in the regression models resulted in roughly 25% increases in R squared effects. (Contains 3 tables and 44 references.) (Author/SLD)