To explore the suggestion that subjects modulate their reading strategies in accordance with how they expect to be tested, several test expectancies (multiple-choice, true/false, essay, and cloze) were implemented in addition to a non-specific test expectancy as a control. Subjects were 124 students at Purdue University (Indiana). After reading three practice texts for which subjects were administered a test that was in line with their expectancies, subjects read and attempted to free-recall a fairy tale or an expository text. The only expectancy to significantly enhance recall relative to the intentional learning control was the essay expectancy, and this enhancement was limited to the expository text. Results were congruent with the hypothesis that subjects expecting an essay test perform more organizational processing of the text than do those expecting other test formats. The authors suggest that there was no increase in recall of the fairy tale due to the essay expectancy because subjects routinely perform organizational processing while reading a fairy tale. Regression analyses and recall data indicated that recognition test expectancies produced little change in reading strategy. Four tables present study data. A 28-item list of references is included. (Author/SLD)