In self-adapted testing, examinees are allowed to choose the difficulty of each item to be presented immediately before attempting it. Previous research has demonstrated that self-adapted testing leads to better performance than do fixed-order tests and is preferred by examinees. The present study examined the strategies that 29 college students used in selecting items during a self-adapted test. After completing the Test Anxiety Inventory, subjects took the self-adapted test. The test contained 40 items sorted into 8 categories of difficulty based on Rasch model estimates. Three test-taking strategies were identified. Most subjects adopted a flexible strategy in which they generally selected easier items following failure and harder items following success. Some subjects adopted a "failure intolerant" strategy in which they generally selected easier items following failure and items of the same difficulty after success. Finally, some subjects adopted a "failure tolerant" strategy in which they chose items of the same difficulty level after failure, but harder items after success. The failure-tolerant strategy was associated with lower estimated ability than were the other two strategies. This finding may reflect the attributions examinees adopting that strategy make and the effort they expend following failures. The results provide general support for the value of continued development of self-adapted testing. (Author/TJH)