Abstract
Research on the psychology of adolescent self-injury is relatively nascent. For some time, self-injury has predominantly been considered a behaviour pathognomonic of borderline personality disorder and, to a somewhat lesser extent, of psychosis or intellectual disability (Hayward, 2007). However, recent empirical research has suggested that self-injury – particularly that of a superficial nature, or causing “mild to moderate tissue damage” (Skegg, 2005, p. 2) – is a behaviour present to a significant degree in nonclinical populations, most notably amongst adolescents (e.g., Muehlenkamp, Williams, Gutierrez, & Claes, 2009). Indeed, amongst this demographic, the behaviour has been described as occurring at a rate of ‘epidemic’ proportions (Miller & Smith, 2008), with females being the most likely to self-injure (e.g., 64% of adolescents who self-injure; Ross & Heath, 2002). Understandably, many parents, teachers, school counsellors, and other health professionals have expressed deep concerns over the trend of non-suicidal self-injury (NSSI). The reasons for the “swift proliferation” (D'Onofrio, 2007, p. 6) of NSSI are, as yet, undetermined; however, one prominent theory is that of social contagion, which states that emotions, attitudes and behaviours are communicable across members of a social network.
The overall aim of the studies presented in this thesis was the investigation of social contagion theory as it pertained to the NSSI attitudes and behaviours of NZ adolescent females. There were two primary objectives. The first objective was to establish an independent but informed perspective, using reports obtained from school counsellors, on the prevalence of female adolescent NSSI behaviour in NZ secondary schools, its evident pattern amongst friendship groups, and beliefs held by counsellors about the reasons for and likely communicability of such behaviours.
The second objective was to investigate whether a social contagion effect was discernible in the NSSI attitudes and NSSI behaviours of NZ adolescent female students, based on direct empirical evidence. Attitude-related variables investigated included positive endorsement of NSSI and beliefs about reasons for engaging in the behaviour. Sub-elements of this second objective were as follows: i) to investigate whether a social contagion effect was discernible in the depressive symptomatology present amongst the same population, given that such pathology is commonly co-morbid with NSSI and is also considered to be transmissible via social contagion; and, in an effort to transcend the confound between peer selection and socialisation effects (a limitation present in much of the extant research on the social contagion of various behaviours), ii) to trial a longitudinal social network methodological design, towards identifying its potential value in differentiating said effects in NSSI research.
The research involved three distinct studies. In Study 1, an online self-report questionnaire was administered to NZ secondary school guidance counsellors to ascertain the estimated prevalence rate of NSSI by adolescent females, the nature of NSSI encountered, its pattern amongst friendship groups, and counsellors’ beliefs about NSSI. Results indicated that NSSI is indeed a pervasive problem within NZ secondary schools, similar to schools internationally, and that there was sufficient report of NSSI occurring within friendship cliques to warrant further, more direct research into the phenomenon of peer contagion.
In Study 2, adolescent girls from two NZ secondary schools were asked about their NSSI attitudes, behaviours, depressive symptoms, and friendship constellations. Friendship cliques and dyads were identified by way of social network analyses. Comparisons within identified friendship cliques showed that girls’ depressive symptoms and NSSI attitudes significantly corresponded with those of their clique and dyad friends, and that being in a clique seemed to be a protective factor for depression and for pathological attitudes towards NSSI. Hierarchical regression analyses showed a significant effect of peer-related variables when predicting individual depression and NSSI attitude scores after personal variables had been taken into account, offering preliminary support for contagion theory. However, the selection versus socialisation confound remained.
With a view to potentially addressing this confound, Study 3 involved a longitudinal approach whereby the original questionnaire was re-administered six months later to a subset of girls from one of the schools who had completed the questionnaire in Study 2. This constituted a trial-run of a methodological design involving both longitudinal and social network analyses for the purpose of identifying the potential strengths and weaknesses of using such an approach to differentiate selection from socialisation effects. Results showed no evidence for socialisation effects in the depression scores or NSSI attitudes of adolescent females; however, this may have been due to certain extraneous factors. On the whole, the methodological design has clear potential for differentiating selection from socialisation effects in NSSI research.
The overall finding of the thesis, and the distinct contribution to knowledge, was that NSSI attitudes and depressive symptomatology are homophilous within friendship cliques and dyads, and that the use of a longitudinal social network design has potential for differentiating selection from socialisation effects in future research investigating the social contagion of depression and NSSI attitudes and behaviour. It is hoped that further empirical research might combine similar methodology with greater longitudinal time separation and a broader sample group to provide incontrovertible empirical evidence to support a contagion theory of NSSI.