This thesis presents empirical evidence supporting the relevance of adult attachment theory in understanding the chronic pain experience. The thesis arose from the recognition that, although pain-related psychosocial variables have been used to differentiate subgroups of individuals in chronic pain, little is known about the developmental origins or cognitive, behavioural, or affective correlates of these subgroups. Attachment theory offers a development framework, which is of potential value in guiding early pain intervention and prevention efforts, as well as tailoring interventions to suit specific patient needs. In order to investigate the value of applying adult attachment theory to the chronic pain field, five central research questions were identified: 1) Does insecure attachment contribute to a more negative experience of acute pain for those not diagnosed with a chronic pain condition? That is, does the association between insecure attachment and maladaptive pain-related variables precede the onset of chronic pain)? 2) How do the subgroups of adult attachment, identified in a chronic pain sample, compare to the clusters identified by pain-related psychosocial measures in the same sample? 3) Are the documented associations between attachment and psychosocial variables (i.e., cognitive appraisals, self-efficacy, coping, social support) evident in the context of chronic pain? 4) Do attachment variables moderate the associations between these psychosocial variables and painrelated variables?, and 5) What is the association between adult attachment variables and outcomes from chronic pain rehabilitation programs? To address the first question, an experimental study was designed to explore the perceptions, emotions and coping patterns of 58 pain-free individuals exposed to an acute (coldpressor) pain experience. Findings revealed that higher levels of anxiety over relationships were associated with lower pain thresholds, more stress, depression and anxiety, diminished perceptions of control over pain, and diminished ability to decrease pain. More secure attachment, on the other hand, was associated with greater perceptions of control over pain, and lower levels of depression and catastrophizing. In order to explore research questions 2-5, data from two clinical samples were collected on the first and last days of participation in one of two chronic pain rehabilitation programs (program length varied from two to three weeks). A total of 152 participants completed questionnaires before treatment, with 100 of these participants also completing questionnaires after treatment. Participants were aged between 18 and 82 years (M = 38.8, SD = 12.4), and had been in pain for an average of 3.7 years (SD = 6.9). Approximately 41% of participants were women, and 58% were men (gender for one participant was unknown). In order to address question two regarding the similarity between adult attachment subgroups and pain clusters based on a psychosocial pain measure (Multidimensional Pain Inventory, MPI), pre-treatment data were analyzed. The clustering approach using the MPI, known as Multiaxial Assessment of Pain (MAP), identifies three subgroups of individuals, labeled interpersonally distressed, dysfunctional, and adaptive copers. These clusters were compared with the four subgroups and two dimensions obtained with adult attachment measures. Results revealed that the interpersonally distressed MAP subgroup was associated with fearful attachment, while the dysfunctional MAP subgroup was linked with avoidant (low comfort) attachment. The comfort attachment dimension was associated with more adaptive MPI variables, while the anxious attachment dimension was linked with more maladaptive MPI variables. Consideration of research questions three to five prompted development of an integrative model, the Attachment-Diathesis Model of Chronic Pain (ADMCP), which provided a framework for exploring the psychosocial mechanisms linking attachment security with the outcome variables of perceived pain intensity, disability, and treatment outcomes. Based on predictions emerging from this model, a set of five studies explored pre-treatment associations between adult attachment variables, six psychosocial mechanisms (pain appraisal, depression, pain self-efficacy, anxiety, perceptions of social support, and coping), and outcome variables pre-treatment. One further paper was devoted to implications of adult attachment for treatment outcome. In general, findings from these studies revealed that variables representing attachment insecurity were significantly associated with more maladaptive coping, cognitive, affective and interpersonal characteristics, both before and after treatment, compared to variables reflecting secure attachment. Overall, both securely and insecurely attached individuals made progress in response to pain rehabilitation. Finally, a summary of the evidence of associations between attachment variables, psychosocial mechanisms, and treatment outcomes revealed by this research is presented, permitting discussion regarding the explanatory value of the ADMCP and clinical implications. Limitations to the present research, including family-wise error, the use of self-report measures, and the self-selecting nature of the sample, are discussed, and directions for further research are proposed. The evidence presented in this thesis advances our understanding of the developmental origins of chronic pain conditions, and provides a lens through which to view the heterogeneity of people with chronic pain. Overall, this project highlights the value of applying adult attachment principles to this field. It is anticipated that findings will assist in the development of a more differentiated range of treatment approaches for the individuals who experience chronic pain. In addition, the findings of this project may have important implications for prevention and early intervention of these conditions.