A re-conceptualisation of Rotter's Locus of Control construct - renamed as Personal Causality in this study - is presented and the development of a scale for its measurement is described. The role of causal beliefs in psychopathology and research into the modification of causal beliefs is reviewed. The conclusions from these reviews are that a weak sense of personal causality is frequently associated with a variety of forms of psychopathology and that causal beliefs can be modified by a variety of procedures, including psychotherapy. A longitudinal study of psychotherapy (defined herein as the non-physical treatment of psychological disturbance) with 24 psychiatric out-patients is reported. The focus of the study is on changes in patients' causal beliefs, assessed using the Personal Causality Scale, and on the identification of the elements of psychotherapy to which these changes are related. The results indicate that changes in patients' causal beliefs in relation to most life areas may be construed as outcomes in the same sense as are more commonly considered outcomes such as symptom alleviation and more general personality change - that changes in causal beliefs are related to these outcome variables and show a similar pattern of relationships to pre-therapy and in-therapy variables. Therapeutic variables (such as participant expectations and constructions of the therapeutic process) having implications for changes in causal beliefs, for symptom alleviation and for more global personality change are identified. In addition, the utility of distinguishing causal beliefs according to different areas of life, of assessing the therapeutic experience from both the therapist's and the patient's viewpoints and of taking in-therapy as well as pre- and post-therapy measures is illustrated. Implications for the practice of psychotherapy are discussed.