This study shows how separate and combined contextual and conceptual problem rewording can positively influence student performance in solving mathematical word problems. Participants included 80 seventh-grade Iranian students randomly assigned in groups of 20 to three experimental groups involving three types of rewording and a control group. All participants completed a pretest. The three experimental groups completed posttests involving problem rewording, and the control group received a posttest identical to the pretest. Rewording in general was shown to decrease student errors in solving word problems. There was no significant difference among contextual, conceptual, and combined contextual/conceptual rewording in reducing errors. In terms of Knifong and Holtan's (1976) classification of word problem errors, the rewordings had no significant effect on decreasing "clerical and computational" errors, but they significantly reduced "other errors" (such as use of wrong procedures or operations or providing no response).