Previous theoretical models of bulimia nervosa (BN) and binge eating disorder (BED) have implicated cross-domain risk-taking behaviour as a significant maintenance factor in both disorders. The present study aimed to test this hypothesis by administering the Balloon Analogue Risk Task (BART) to 25 women with BN or BED and 27 healthy comparison women without a history of an eating disorder. Furthermore, we tested the effect of a divided dose of 64 IU of oxytocin on risk-taking behaviour in the BART. Contrary to our hypothesis, women with BN or BED did not exhibit baseline differences in performance on the BART in the placebo condition (t = 1.42, df = 50, P = 0.161, d = 0.39). Oxytocin did not have a main effect on performance in the BART (F = 0.01, df = 1, P = .907, η  < 0.001); however, there was an interaction, such that participants in the BN/BED participant group, compared to the healthy comparison group, demonstrated safer behaviour on the BART in the oxytocin condition, but not in the placebo condition (F = 4.29, df = 1, P = 0.044, η  = 0.082). These findings cast doubt on the common assumption that individuals with BN and BED exhibit greater risk-taking behaviour in all domains and add to the evidence that oxytocin plays a functional role in modulating behaviours that entail trade-offs between reward approach and risk in humans. We recommend that future dose-response studies investigate the effect of oxytocin on reward approach behaviour further in women with recurrent binge eating behaviour, as well as the clinical significance of this effect.