Anorexia Nervosa (AN) is a complex psychiatric disorder of unknown etiology. AN has strong neurobiological underpinnings that are reflected in observable differences in neural circuitry and cognitive functioning. One pathway to improving our understanding of the disease is through the study of underlying neuropsychological impairments. Previous studies of neurocognitive function in eating disorders have indicated impaired cognitive flexibility associated with AN. Specifically, this has been examined using a range of setshifting tasks. Coupling our present understanding of the neural circuitry of set-shifting with current research into the neural correlates of AN will allow us to better understand and explain aberrant behaviors and cognitions typically associated with AN. The current study assessed set-shifting ability across multiple tasks in a cohort of 22 women recovered from restricting-type/purging-type AN (RAN) as compared with 22 healthy control women (CW). The Delis-Kaplan Executive Function System (D-KEFS) Trail Making, Verbal Fluency (VF), and Color-Word Interference (CWI) tests were used along with the Wisconsin Card Sorting Task (WCST). For each of the measures, the shift component was isolated from other task processes by computing a ratio score. It was hypothesized that the RAN group would have impaired set-shifting performance on each of these tasks relative to the CW group. There were no significant differences between groups on the ratio scores. The RAN group had slower completion times on all trials of the D-KEFS Trail Making and CWI tests as well as greater error totals on the WCST compared with CW. The fact that these did not translate to significant impairments when the switching components were isolated suggests that other component processes of these tasks, such as inhibition, may be contributing to between group differences. The RAN group performed better than the CW on the D-KEFS VF task. It was also hypothesized that within the RAN group there would be across-task correlation on the set-shifting measures, and that impaired set-shifting performance would be associated with increased perfectionism and obsessionality. There was a significant negative correlation within the RAN group between lifetime obsessionality and the WCST ratio score. Taken together, the findings of this study suggest that while the CW generally out-performs the RAN on most of the tasks, this cannot be attributed to a clear-cut group difference in setshifting. Rather, it highlights the need for a nuanced approach to examining neuropsychological performance in AN with attention to task design, analysis, and the diagnostic characteristics of the sample in order to clarify the nature of any impairments.