A path analytic model of state test anxiety was tested in 169 college students who were enrolled in statistics courses. Variables in the model included gender, mathematics ability, trait test anxiety (trait worry and trait emotionality as separate variables), statistics course anxiety, statistics achievement (scores on midterm examinations), perceived test difficulty, and state test anxiety (state worry and state emotionality) measured at the time the final examination was taking place. Female students reported higher trait test anxiety and statistics course anxiety than did males. Mathematics ability was negatively related to statistics course anxiety and positively related to statistics achievement. Students with low mathematics ability perceived statistics courses as difficult; and their perception, in turn, strongly influences their level of statistics course anxiety. Trait worry had a significant direct effect on state worry, but not on state emotionality, and the same pattern occurred with trait-state emotionality. Trait worry, but not trait emotionality, was positively related to statistics course anxiety. Students reporting high statistics course anxiety perceived the final examination as difficult; however, statistics course anxiety had a significant effect on neither statistics achievement or state test anxiety. Students' perception of statistics course difficulty influenced their perception of test difficulty, which in turn had significant direct effects on both state worry and emotionality. As expected, achievement on statistics midterm examinations had a direct effect on state worry but not state emotionality. (Contains 2 figures, 2 tables, and 42 references.) (Author)