This study explored the variance in mathematical problem-solving achievement accounted for by sex and ethnicity. Each of 316 adolescents in a rural school district in southeast Texas was administered the California Achievement Test and a set of five affective scales. Reading scores accounted for 49.5% of the variance; computation, 14.6%; affective scales, 1%; sex and ethnicity, 0.58%. Thus, no significant sex or ethnic differences were found in problem-solving ability when the effects of reading and computation were removed. The researchers concluded that this could indicate that males and females, blacks, whites, and chicanos are equally good (or equally poor) problem-solvers when they have similar reading and computational skills. (MNS)