This simulation study investigated how residual non-normality affects cross-classified multiple membership random effects modeling estimates. The conditions manipulated in the study were residual distribution type (normal, uniform, and chi-squared), intra-unit correlation coefficients (0.1, 0.2, and 0.3), number of groups (20, 50, and 100), average group size (10 and 20), cross-classification rate (20% and 40%), and multiple membership rate (10% and 20%). The relative bias and the root mean square of the parameter estimates were evaluated. The results indicated that non-normal residuals had a larger impact on the cross-classified multiple membership random effect variance component estimates, especially when the number of groups was relatively smaller (i.e., 50 groups or less) and that the intra-unit correlation coefficient increased. Specifically, the degrees of relative bias were larger when the level-two residuals followed chi-squared distribution (i.e., a severely skewed distribution) compared to uniform distribution (i.e., a non-normal but symmetrical distribution). The coverage rates of the level-two variance component were lower than the nominal 95% when residuals followed chi-squared distribution, while they were higher when residuals followed uniform distribution. These findings can be useful in applying cross-classified multiple membership random effects modeling to account for the contextual effects of cross-classified and multiple higher-level units.