The tendency to make critical versus uncritical responses was examined, using data from two of the following three report writing tests: The Alternative Expressions Test, Recognizing Ambiguities Test, and Evaluation of Revisions Test. Subjects included 208 undergraduate students, 320 graduate business students, and 400 first-year law students who reached the last item in the two tests. Four variables were obtained for each item: item location, response style validity index, content validity index, and modified difficulty index. The relationship between item location and criticalness response style seemed to be causal; the later the item appeared in the test, the more it measured response style. This finding is striking because it demonstrated changes within the same test, even a very short one. When considered with the lack of relationship between item location and content, it suggests that many observed changes in test scores may occur because of changes in response style tendencies, rather than content. There are also implications for the measurement of response style. The assumption that items in different locations are equivalent is likely to be incorrect. Paper presented at APA, Los Angeles, September, 1964.