Investigations of memory and associated phenomena in obsessive-compulsive disorder (OCD) can advance our understanding of this often debilitating problem. Theoretical models predict both the presence and absence of memory biases in favour of threat-relevant information in association with anxiety disorders generally, and with OCD specifically. Two previous experiments (one involving compulsive washing and another involving compulsive checking) that demonstrated such a memory bias are reviewed in the context of the existing literature. Additionally, a new experiment failing to demonstrate such a bias (in association with compulsive ordering and arranging) is presented. The results are discussed in terms of cognitive-behavioural and information processing approaches to understanding OCD. It is argued that experiments which utilize stimuli that are low in ecological validity are unlikely to detect explicit memory biases in OCD. As such, experimental paradigms that are perceived as particularly significant, relevant and important to participants with OCD are encouraged.