An exploratory factor analysis and several confirmatory analyses were performed to evaluate the factorial structure of the Revised Mathematics Anxiety Rating Scale (RMARS) through the responses of 805 college students. On 559 students' scores, the instrument's construct validity was tested through a confirmatory factor analysis (CFA) and was found to be inadequate. An exploratory factor analysis suggested a modification by dropping five items. After the modification, a second CFA showed that the modified model fit the theoretical model well. Cross-validation of the modified model was tested on a different sample of 246 students and was found to be satisfactory. In addition, concurrent validity of the instrument was found from significant relationships between the modified RMARS scores and students' self-perceptions of their general and current mathematics anxiety levels. The modified RMARS is valid and reliable and could be used as a screening tool, a placement tool, or a research tool.