Covariance structure modeling was used to test a model of test anxiety
with a sample of 204 male and 412 female college students. Variables
in the model included number of prior math courses, student course lev
el, math self-concept, self-efficacy, achievement, and general and sta
tistical test anxiety. Results indicated that whereas some paths diffe
red across gender, these differences were not statistically significan
t according to tests of invariance. General and statistical test anxie
ty were found to be distinct, but related constructs, as seen in the d
ifferential relationships between these variables and other variables
in the model. Lack of gender differences was discussed in light of pre
vious research.