The correctional experience of adult offenders has largely been ignored in comparison to youth and adolescent offenders in the correctional literature and even more so in terms of repeated offending behaviour. The perspective that self-esteem is also considered an important behavioural motivator and outcome of life experiences is reflected in the high amount of inclusion it enjoys as a treatment outcome in numerous correctional rehabilitation programmes. However, the dearth of research on the level of self-esteem of repeat offenders internationally and in the South African context does not reflect the importance of this relationship and with a 95% recidivism rate in South African prisons one can understand the need for investigation into this phenomenon. Descriptive, comparative and explanatory analyses were conducted using the self-concept theory of Carl Rogers to guide the research. A purposive sample of 73 male repeat offenders from correctional centres in the Zululand area of the KwaZulu-Natal province were included in the study. Self-esteem was assessed through the use of the Coopersmith Self-Esteem Inventory (Adult Form) and relative theoretical and empirical perspectives were utilised in the development of a questionnaire used to quantitatively assess multiple variables related to self-esteem and repeat offending behaviour. The level of self-esteem was therefore assessed in terms of the type of offence/s committed by the participants and in relation to a number of demographic (age, education, marital status, number of children and childhood socioeconomic status), psychosocial (parental closeness, parental availability, need for acceptance, familial abuse and treatment by community upon release) and correctional (exposure to correctional environment, experience of victimisation inside the correctional facility, visitations, conditions of the correctional centres and programme participation) variables. Inferential (t-tests, ANOVA and correlations) and descriptive (means, standard deviations and frequency distributions) statistical analyses were utilised to compare the self-esteem scores of the participants and provide a general description of the characteristics of the sample respectively. The reliability of the Coopersmith Self-Esteem Inventory (Adult Form) was also assessed producing a Cronbach alpha of .69 indicating that it was a reliable measure of self-esteem in the South African correctional context. The results indicated that the self-esteem scores of South African repeat offenders varied significantly in terms of offence type with aggressive offenders showing significantly higher levels of self-esteem than sexual, economic, narcotic and “other” offenders. Participants from the remaining offending categories did not show significant differences in self-esteem between one another. Demographic variables (age, education, marital status, number of children and childhood socioeconomic status) were not found to have a significant effect on self-esteem scores. Psychosocial variables related to familial relations (parental closeness and parental availability) were not found to be significantly different despite clear patterns emerging indicating an increase in self-esteem when parental bonds were strong and parents were perceived as available during childhood. The analysis of data related to the need for acceptance by friends and family members indicated that despite the repeated offending nature of the sample that there was still a degree of acceptance with conventional norms and values. Self-esteem was also seen to improve with perceptions of community support and fairness of treatment by community upon release. Experiences of abuse prior to incarceration and victimisation within the correctional facility were seen to negatively affect self-esteem however time and context related variables were determined to play a mediating role due to the increased effect witnessed for experiences that were more recent. Further correctional variables analysed showed increases in self-esteem in terms of prolonged exposure to the correctional environment, and increase in the frequency of visitations and positive perceptions of satisfaction with the conditions in the correctional centres. Programme participation was was however found to have no significant effect on self-esteem. Recommendations formulated for future research included the need for a longitudinal research design and the incorporation of contextual data in the form of non-offending control groups for the establishment of cultural norms and values through the use of qualitative interviews. The generally negative life experiences of repeat offenders were concluded to increase the need for the use of psychological defences and the creation of an increasingly false reality in an attempt to maintain some form of psychological consistency. / Thesis (M.Soc.Sc.)-University of KwaZulu-Natal, Durban, 2012.

MA ( Psychology) / Dpartment of Psychology