The difficulties that many students, particularly those in the social and behavioral sciences, encounter while taking an introductory statistics course have been widely reported in many parts of the world. Factors that have been purported as relating to performance in introductory statistics include a variety of cognitive and affective variables (Feinberg & Halperin, 1978). Cognitive factors, such as mathematics ability and background, certainly play a major role in performance in an introductory statistics course; however, affective variables are also important. Gal and Ginsburg (1994) reported that "The body of research on students' attitudes, beliefs, and affect related directly to statistics education is very small and problematic." They further state that concerns for studying non-cognitive aspects of statistics education should not only be motivated by outcome (performance) but also by process considerations.<br><br>A major issue concerns the influence of attitudes on achievement. McLeod (1992) suggested that neither attitude nor achievement is dependent on the other, but they "interact with each other in complex and unpredictable ways. (p. 582)" The prospect of the reciprocal relationship of attitude and achievement has been proposed by others (Kulm, 1980) and negates the possibility of isolating the cognitive and affective domains. A number of studies have investigated the relationship between attitudes toward statistics and performance in an introductory course using a variety of correlational and regression techniques. Results generally indicate a small to moderate positive relationship. This relationship appears to be fairly consistent regardless of the instrument used, the time of administration of either the attitude or performance measure, or the level of the student.<br><br>Longitudinal studies of math attitudes and performance (Pajares & Miller, 1994; Meece, Wigfield, & Eccles, 1990; Eccles & Jacobs, 1986) have provided path analyses of the relationship of these variables. In a study of undergraduates, Pajares and Miller (1994) determined that mathematics self-efficacy was highly related to mathematics performance with mathematics self-concept and high school mathematics experience making a small, but significant contribution. Perceived usefulness was not a contributor to mathematics performance. Eccles and Jacobs' (1986) path analysis indicated that mathematics grades were influenced by the students' self-concept of math ability and math anxiety, but were not influenced by the student's perception of the task difficulty or the perceived value of mathematics. Meece, Wigfield, and Eccles' (1990) path analysis found that students' prior grades and expectancies were predictors of grades, while importance and anxiety were not. In several other papers, the first two of the current authors have developed a model suggesting that some aspects of statistics attitude at the beginning of a course affect test performance during the course and that end of course attitudes were both directly and indirectly influenced by performance during the course. The research reported here extends that line of inquiry looking at data from English and Arabic speaking samples available in the U.S. and Israel.