Differences in the performance of handicapped and nonhandicapped students on the Graduate Record Examinations General Test as a function of sex group composition were investigated. Hierarchical multiple regression analyses were conducted, regressing General Test scores on sex and handicapped group membership, and adjustments to these scores were computed. Data came from 186 visually impaired students taking the regular test in a standard national administration, 151 visually impaired students taking a large type version, and 105 physically handicapped students taking the regular test type in a national administration. The resulting adjustments had only minimal effects on handicapped/nonhandicapped mean score differences, suggesting that these disparities cannot be ascribed to differences in sex group distributions. Other plausible explanations for observed score disparities between handicapped and nonhandicapped groups are discussed. (Contains 4 tables and 11 references.) (Author/SLD)