This research sought to examine the link between different situational demands, in this case three different types of science achievement tests (multiple-choice, constructed response, and performance assessments), and perceptions, beliefs, and performance of 491 high school students. The students were characterized by three well-established motivational patterns: intrinsic-mastery, ego-success, and academically helpless. Students were compared across these three patterns with respect to their efficacy for working on the tests, their beliefs about whether these tests were valid measures of their science knowledge, and their observed performance on the multiple-choice and constructed response tests. It was found that students varying in motivational pattern could be distinguished reliably by: (1) efficacy for working on multiple-choice and constructed response tests; (2) beliefs about the validity of multiple-choice and constructed response tests in revealing their science knowledge; and (3) performance on the multiple-choice and constructed response tests when quantitative and verbal ability are controlled for. However, motivational pattern did not relate to efficacy for or validity beliefs about performance assessments. Perhaps the flexible nature of performance assessments facilitates all students goal pursuits and performancewhereas multiple-choice and constructed response tests do not. An appendix contains some items used in the study. (Contains 7 tables and 27 references.) (Author/SLD)