Obsessive Compulsive Disorder (OCD) is classified as an anxiety disorder characterized by distressing persistent unwanted ideas or impulses (obsessions) and urges and/or compulsion to do something to relieve the associated anxiety caused by the obsession. The thematic content of the obsessions are highly variable, ranging from symmetry, contamination to aggressive concerns. Compulsions tend to be linked to the obsessions, but can also be idiosyncratic to the intrusive thought. According to the cognitive model, Obsessive-compulsive disorder (OCD) is maintained by various belief factors such as an inflated sense of responsibility, overestimation of threat and the over-control of thoughts. Despite much support for this hypothesis, there is a lack of specificity. This series of studies sought to determine the relationship between a number of cognitive beliefs and appraisal processes and obsessive-compulsive symptoms. This thesis presents the results of three studies. The first study was designed to investigate the hypothesis that certain beliefs are more prevalent in OCD, compared with other anxiety disorders. The second study expands on earlier findings by examining whether the six metacognitive beliefs proposed by the Obsessive Compulsive Cognitions Working Group, (OCCWG; 1997, 2001, & 2003) correlate with specific symptom-based OCD subtypes. The final study addresses some of the methodological weaknesses inherent in retrospective self-report measures by replicating the study using experimental techniques. Most importantly, this research was conducted from within the theoretical framework of Rachman (1993) and Salkovskis (1989) models which emphasise the misinterpretation of significance of the intrusive thoughts. The first study explored the relationship between thought-action fusion (TAF) and inflated responsibility beliefs across individuals diagnosed with obsessive compulsive disorder (OCD), an anxiety disorder other than OCD (anxious controls; AC), and a non-anxious control group (NAC). It was hypothesized that the OCD group would evidence significantly higher inflated responsibility and TAF scores, compared to the AC and NAC groups. In this study, non-clinical and clinical participants were recruited for research. The non-clinical group was comprised of undergraduate students (n = 22: mean age = 26.8; SD = 9.2). The clinical groups included 20 participants with OCD as their primary diagnosis (mean age = 32.1; SD = 11.9) and 21 individuals diagnosed with another anxiety disorder (mean age = 32.2; SD = 10.9). To measure inflated responsibility beliefs and thought action fusion, self-report questionnaires were administered to the participants. The results of this study demonstrated that inflated responsibility beliefs, while present in other anxiety disorders, were significantly higher in participants with OCD, even after controlling for depressed mood and TAF levels. No group differences emerged between the OCD and anxious groups on measures of TAF. Thus, it can be tentatively concluded that inflated responsibility beliefs may have a more robust relationship with OCD than TAF beliefs, which appear to act as a general vulnerability factor occurring along a continuum of anxiety disorders. The second study examined the associations between the six OCD-related beliefs: control of thoughts, importance of thoughts, responsibility, intolerance of uncertainty, overestimation of threat and perfectionism and five empirically derived OCD subgroups. Clinical participants with a primary diagnosis with OCD (n = 67: mean age = 38.0; SD = 11.7) were recruited over a period of two years from the Anxiety Disorders Unit. Participant responses were cluster analysed to form five stable groups: aggressive obsessions-checking compulsions (n = 22: mean age = 26.8; SD = 9.2); contamination obsessions-cleaning compulsions (n = 22: mean age = 26.8; SD = 9.2); symmetry concerns-ordering/arranging compulsions (n = 22: mean age = 26.8; SD = 9.2); hoarding obsessions-hoarding compulsions (n = 22: mean age = 26.8; SD = 9.2); and miscellaneous obsessions -miscellaneous compulsions (n = 22: mean age = 26.8; SD = 9.2). The second found that intolerance of uncertainty was significantly related to the contamination subgroup. While responsibility and threat estimation beliefs were higher in the aggressive-checking subgroup, these differences did not reach statistical significance. No other significant results were found, however, there was a non-significant trend for perfectionism beliefs to be higher in symmetry-ordering and hoarding subgroup. Following the results of this study, questions remained about whether the lack of significant findings reflected the generality of these beliefs or were due to methodological differences. This led to the development of the final study presented in this thesis. The purpose of the final study was to investigate whether the second study was limited by the method of assessment (e.g. self-report questionnaires). This study was unique, as it was the first of its kind to experimentally manipulate all six beliefs in empirically derived OCD subtypes. Twenty participants (mean age = 45.0; SD = 11.0) were chosen from the second study to form the following priori groups: contamination (n = 4: mean age = 44.5; SD = 9.5); aggressive (n = 6: mean age = 46.5; SD = 7.2); hoarding (n = 4: mean age = 47.2; SD = 6.9); and symmetry (n = 6: mean age = 41.8; SD = 17.4). Six behavioural experiments designed to reflect one of the six OCCWG beliefs were specifically developed and administered to the groups. Baseline scores were obtained using self-report questionnaires. The study found strong support for the use of experimental paradigms over self-report measures, as several significant interactions between cognitive beliefs and OCD symptom-based subtypes were found. Specifically, the hoarding subgroup evidenced significantly higher overall thought action fusion scores compared to those in the contamination group. The symmetry subgroup exhibited significantly higher anxiety than the aggressive group during the perfectionism task and demonstrated significantly higher scores on several items measuring perfectionism compared to the contamination group. Finally, over-estimation of threat beliefs was significantly higher in the contamination thoughts. No statistically significant group differences were found for controllability of thoughts, responsibility and intolerance of uncertainty. In conclusion, these studies collectively showed that in some cases of OCD certain beliefs appear highly applicable, whereas in others they are not. This finding may explain why some OCD patients have poor treatment outcomes as the beliefs and appraisals were highly variable across groups. These findings are of both theoretical and clinical significance because they add to the growing understanding that OCD may consist of distinct clusters of symptoms with different underlying motivations and beliefs. This finding is of clinical significance because treatment guidelines for OCD can become more specific, factoring into the therapy situation these underlying beliefs and appraisal processes. Lastly, the findings regarding inflated responsibility deserve special mention, given the significance of this construct in contemporary cognitive models. The results of the present studies were mixed with regard to responsibility as only the first study found a significant result. It appears that, like the other belief domains proposed by the OCCWG, responsibility may not be specific to all types of OCD and current cognitive models may benefit was shifting the emphasis to other belief domains.