Grooming is a fundamental component of sociality in many gregarious animal species, and elucidating the costs and benefits of this behaviour is crucial for understanding its function. There is evidence that animals giving grooming pay a cost in terms of the time and energy they invest, while recipients benefit not just from the removal of dirt and parasites, but also from the relaxing effects of being groomed. Recently, however, studies of primates have indicated that giving grooming may also provide such hedonic benefits, reducing levels of stress or anxiety in the groomer. In this study of free-ranging adult female Barbary macaques at Trentham Monkey Forest (Stoke-on-Trent, UK), we tested the hypothesis that grooming reduces anxiety in the donor and/or the recipient. During focal follows, we quantified females' rates of self-scratching as a behavioural index of their anxiety levels. Self-scratching rates in the 2-min periods after bouts of grooming (given, received and reciprocated) were compared to overall mean self-scratching rates; we predicted that if grooming reduces anxiety, self-scratching rates would be significantly lower after grooming bouts than mean levels. We first analysed all grooming bouts and then analysed separately grooming bouts with adult males, with all adult females, with subordinate adult females and with dominant adult females. Contrary to our prediction, self-scratching rates were never seen to be lower after grooming than mean levels. In fact, for the majority of grooming partner–direction combinations, we found significantly higher rates of self-scratching after grooming compared to mean levels. The hypothesis that grooming reduces anxiety was therefore not supported. Grooming seems in some cases to increase, not alleviate, anxiety. We explore possible explanations for these unexpected results.