This study investigated the propositions that (1) the expressed preference for formal-deductive style of thinking and mathematics anxiety are negatively correlated; (2) the preferred cognitive style is positively correlated with the respondent's self-assessment of mathematical ability; and (3) those expressing a preference for the analyst style of thinking would perform better on a test of statistical concepts than those expressing a disregard for the analyst style of thinking. The sample included 70 male and female graduate students. Results indicated that the propositions were supported by the data. Implications from this research are discussed. (CW)