The aim of this study was to examine the relationship between adult attachment representations, OC related cognitive structures and OC symptoms. The degree of fit of four models was examined, including full and partial mediations. The results indicated the superiority of a fully mediated model where IWMs of other and IWMs of self affected OC symptoms via their influence on cognitive-affective structures and OC related cognitions. Particularly, IWMs of self directly predicted OC related cognitions, sensitive self and world view assumptions. Similarly, IWMs of other directly predicted sensitive self and world-view assumptions. These cognitive-affective structures then predicted OC symptoms directly or indirectly. Specifically, sensitive self predicted OC symptoms via its effect on OC related cognitions, while world assumptions predicted OC symptoms directly and indirectly via OC related cognitions.
These findings are consistent with previous suggestions implicating attachment experiences in the later development of more general cognitive affective structure such as perceptions of self, other and the world (e.g., Bretherton & Munholland, 1999; Doron & Kyrios, 2005; Guidano & Liotti, 1983) and linking such structures with OC symptoms and beliefs (e.g., Guidano & Liotti, 1983; Newth & Rachman, 2001; Rowa, Purdon, Summerfeldt, & Antony, 2005; also see chapter 4 and 5 of this thesis). In particular, these results are consistent with Doron and Kyrios' (2005) recent model, linking vulnerability to OCD with dysfunctional perceptions of self and others stemming from attachment relationships.
Previous studies have linked world-view and sensitive self structures with OC phenomena (see chapter 4 and 5 of this thesis). The results of this study replicate these findings using two independent samples and a more sophisticated from of analysis (i.e., SEM) that enables to distinguish between indirect and direct relationships between variables and to analyze relationships between latent variables without random error. Recent research has also implicated attachment in the development of OC related beliefs such as perfectionism, thought suppression and overestimation of threat (see Mikulciner et al., 2003). Indeed, cognitive theory implicates early experiences in the development of the underlying beliefs associated with OCD (Salkovskis, Shafran, Rachman, & Freeston, 1999). Particular types of intrusive thoughts then interact with such beliefs (e.g., inflated sense of responsibility) leading to the development of a resistance cycle (i.e., avoidance and neutralizing) resulting in the exacerbation of intrusions into obsessions and compulsions (Salkovskis, 1989; Rachman, 1997, 1998).
Consistent with these cognitive models, Doron and Kyrios (2005; also see chapter 2 of this thesis) argued that unfavorable attachment experiences increase the likelihood for the development of OC related cognitions and other cognitive-affective structures associate with OCD such as 'sensitive' self structures (i.e., perceptions of incompetence in valued domains) and world assumptions (e.g., negative perceptions of the benevolence of world). Individuals possessing negative attachment internal representation chronically overactivate their attachment-system, leading to anxiety and readiness for action (see Shaver and Mikulciner, 2002). The individual constantly scans for internal/external stimuli that may threaten their sensitive self domains (e.g., an intrusion about hurting someone may endanger the individuals' perceptions of morality), and activate their attachment system. This, in turn, may trigger negative interpretations of such intrusions associated with dysfunctional OC related beliefs. Assumptions about the world may further increase anxiety associated with intrusions and to some extent the response to the intrusions (see chapter 4 of this thesis).
The present study also has implications for therapy. In particular, it suggests that attachment representations may partially underlie the beliefs currently considered within cognitive theory of OCD as well as other OC related cognitive structures. Thus, therapy aimed towards attachment relationships (e.g., Bowlby, 1988; Young, 1999) may improve the efficacy of treatment, particularly for those individuals who are refractory to current treatments. For instance, addressing individuals' fears of abandonment associated with negative IWMs of self may reduce the tendency to overestimate social threat, focus on danger cues and negatively evaluate oneself. Addressing negative attachment representation may also help in reducing sensitive self structures associated with OC symptoms and cognitions by facilitating the re-evaluation of importance and/or feelings of competence in specific self domains.
The results of the present study also suggest that it may be important to address the negative IWMs of others in individuals vulnerable to OCD. Mikulciner and colleagues suggested that work with such individuals should address the tendency for negative expectations from others and attempt to enhance clients´â”¢ insight into their emotional life. This, in turn, may reduce the need to avoid painful experiences (i.e., emotions and thoughts) resulting in more functional way of dealing with such experiences. Finally, it is important to consider attachment relationships in terms of in the effects that these have on the therapy process itself. For example, differences in attachment have also been linked with differences in therapy outcome (e.g., Kanninen, Salo, Punamaeki, & Raija-Leena, 2000), engagement in treatment (e.g., Eames & Roth, 2000), help seeking behaviors (e.g., Dozier, 1990) and evaluation of smoothness and depth of therapeutic sessions (e.g., Mohr, Gelso, & Hill, 2005). The therapist should consider such tendencies during the engagement or socialization phase of treatment, evaluate their effects on the therapeutic relationship, and gently address such perceptions (see Mikulciner et al., 2003).
Future research may benefit from assessing the relationship between OC dimensions and attachment representations as these may have different developmental pathways and require different treatment strategies (McKay et al., 2004). Further, while attachment representations, cognitive affective structures of self and world wer found to relate to OC symptoms in these two analogue samples, research should study the relationship in clinical cohorts with OCD and other anxiety disorders. As individuals presenting with OCD report more often than not past or present depressive symptoms (e.g., Angst et al., 2004; Rasmussen & Eisen, 1992) an examination of attachment representations, OCD and the identified cognitive affective structures would be important to consider.
The present study was cross-sectional and although it was based on structural equation modeling techniques, it was still based on correlational data. Research findings support the relative stability of attachment representations across the life span (see Farley 2002 for meta analysis; also see Chapter 2 of this thesis). However, a variety of environmental changes may modify the attachment bond and consequently the attachment related representations (Bowlby, 1973; Guidano & Liotti. 1983; Sroufe et al., 1999). Several examples for such changes are the death or illnesses of the main caregiver, changes in family structure (e.g., divorce, remarriage) or diminishing social support network (Bowlby, 1973; Guidano & Liotti. 1983). Thus, the link found between adult IWMs and OC related phenomena may partly reflect the influence of negative life events on the attachment system, rather than solely the influence of early attachment experiences. Future research using longitudinal and experimental approaches would provide stronger evidence of causality. Also, as most participants included in this study were women and of Australian ethnic background, sample representativeness was limited. Therefore, future studies could benefit from including more representative samples, as well as replicating across different cultures. Indeed, a number of researchers have identified the influence of cultural factors on OCD (e.g., Fontenelle, Mendlowicz, Marques, & Versiani, 2004; Sica, Novara, Sanavio, Dorz, & Coradeschi, 2002). Finally, as the present study is based on self reports measures, replication with other methods of data collection such as experimental manipulations of the attachment system or self related constructs would be beneficial in future research (see Riskind, Williams, & Kyrios, 2002).
This study has found that attachment representations are linked to OC symptoms and OC-related cognitive-affective structures. This, it is hoped, will lead to the expansion of therapeutic models of OCD. Chapter 7 will examine the proposed relationships with a clinical sample.