Relationship separation is prevalent in Western society and ongoing problems for adults and children are common, resulting in high personal and social costs. A literature review revealed that no existing measure of separation adjustment assessed multiple domains of separation adjustment, and was psychometrically sound. Additionally, a literature review revealed that few studies have investigated change in separation adjustment over time using cognitive predictors. This thesis is a program of research investigating separation adjustment in recently separated adults. Five studies were conducted with three main aims: (1) to develop a multidimensional, psychometrically reliable and valid measure of separation adjustment; (2) to assess the trajectory of separation adjustment over time; and (3) to test the hypothesis that certain cognitive variables predict the trajectory of separation adjustment. A conceptualisation of separation adjustment was developed consisting of the following five domains: Connection to the former partner, loneliness and emotional negativity (referred to as lonely negativity), parenting negotiation, parent-child relationship and general psychological adjustment. The Problems After Separation Test (PAST) was developed to reflect the first four of the five domains. In Study 1, 268 participants separated for up to 18 months, were recruited to assess the factor structure and internal consistency of the PAST. In Study 2, 209 participants were recruited to assess the stability of the factor structure, and temporal stability. In Study 3, participants from Study 2 were used to obtain convergent and divergent validity. The results of the first three studies showed that separation adjustment is a multidimensional construct, and that the PAST is a reliable and valid measure of separation adjustment. Study 4 aimed to assess the trajectory of separation adjustment over a 6 month period using the PAST and another measure of general psychological adjustment. Participants from both Study 1 and 2 were used, and a longitudinal cohort sequential design was employed. The sample consisted of three cohorts: those separated for up to six months, those separated between 6 and 12 months, and those separated between 12 and 18 months. The results showed that connection to the former partner, lonely negativity, and general psychological adjustment improved over time, but parent-child relationships and parenting negotiation were stable, and chronic parenting problems were common. Men showed greater parent-child relationship problems than women, possibly because men were most likely to be the non-resident parent. Participants separating from a de facto relationship reported only slightly more problems on lonely negativity, general psychological distress and parent-child relationships than participants separating from a marriage. Study 4 also provides information on clinical and reliable change, suggested cut-offs that might be used to define normal, moderate and severe adjustment problems, and attrition analyses. Using participants from Study 2, Study 5 aimed to assess cognitive predictors of change on connection to the former partner, lonely negativity and psychological distress. The cognitions assessed at each data collection were dysfunctional attitudes, attachment style, causal attributions, threat appraisal, and self-efficacy. Causal attributions were obtained by asking participants "For three minutes tell my about the problems you had in your relationship and what lead to the separation". The responses were transcribed, a coding manual was devised, and inter-rater reliability of coding was good. Cross-sectionally, the majority of cognitions were associated with one or more domains of separation adjustment, however longitudinally, cognitions did not predict change in separation adjustment over a 6-month period. Cognitions themselves were found to be somewhat variable, which is not surprising given that stressful life events, such as relationship separation would alter cognitions. Other variables that might be responsible for changes in separation adjustment trajectory are discussed. The results of the combined 5 studies have both theoretical and practical implications. Theoretically, separation adjustment is a construct made up of distinct domains which have different trajectories. Connection, lonely negativity and psychological adjustment generally represent transient problems. Although the current research found that a non-trivial minority of participants continued to show distress on these domains two years post separation. On the other hand, parenting negotiation and the parent-child relationship was a chronic problem for many separated people, particularly men. Practically, the results of the current research suggests that most therapeutic attention should be directed towards improving parenting adjustment post separation. Limitations of the current research and suggestions for future research are discussed.