Students in a college introductory statistics class were evaluated with conceptual and computational tests and the relationship between their levels of knowledge on the two forms of testing was assessed. There were significant correlations between their abilities to perform computations and to answer more conceptual questions on individual tests and across separate tests, for the final examination, and for the total number of points earned throughout the semester. The correlations indicate that the two styles of testing provide partially redundant, but not totally overlapping, information about student knowledge. Further, although student averages did not differ in the two types of tests, they seldom preferred only conceptual tests of their knowledge, judging the computational tests a better means of evaluation. Applications of these results extend to prediction of missing test scores and to the testing of students for whom English is a second language. (Author)