An investigation of the relationship between self-ambivalence, self-discrepancy, and autogenous and reactive obsessions
Copyright © Dr Jonathan Pointer 2011. All rights reserved.
Objective: Recent theorising highlights a role for self-concept discrepancies in the onset and maintenance of obsessions. This theorising suggests that self-concept discrepancies might be more influential in the occurrence of autogenous as opposed to reactive obsessions. The aim of the current study was to investigate the relationship between self-ambivalence, self-discrepancy, and autogenous and reactive obsessions whilst controlling for dysphoria. Design: This study employed a cross-sectional correlational design in a non-clinical general population sample. Three hundred and seventy-six(22.6% male, mean age: 26.60, SD: 9.27; 77.4% female, mean age: 26.13, SD: 8.63) participants completed measures of self-ambivalence, self-discrepancy, obsessionality (both autogenous and reactive obsessions) and dysphoria. Findings: Self-ambivalence was positively associated with obsessionality, after controlling for dysphoria. There was no difference in the strength of the positive association between autogenous obsessions and self-ambivalence, and between reactive obsessions, after controlling for dysphoria. There was no association between self-discrepancies and obsessionality, after controlling for dysphoria. Conclusions: Self-ambivalence plays a small part in obsessionality, but subtyping obsessions into autogenous and reactive may not be a helpful distinction. Although self-discrepancy was not associated with obsessionality, there was a strong association between self-discrepancy and dysphoria, which may have clinical implications for alleviating negative mood. Other theoretical and clinical implications are discussed, as are suggestions for further research.
Phenomenology of obsessions
Obsessions are frequent, repetitive unwanted thoughts, images, and impulses (Julien, O'Connor & Aardema (2007)with an intrusive quality that are associated with negative affect (Stein & Finberg, 2007). Both Clark (2004) and Purdon et al. (2007) argue that ego-dystonicity is a defining characteristic of obsessions. Whereasego-dystonic defines thoughts that are unacceptable, unwanted, and ‘alien’ that contradict valued self-perceptions,ego-syntonic refers to thoughts which are acceptable to, or in harmony with valued self-perceptions (Bhar & Kyrios, 2007).
However, according to Purdon et al. (2007), considering a thought as either ego-syntonic or ego-dystonic is simplistic. Purdon et al. propose that the same thought can be both syntonic and dystonic in relation to valued self-perceptions. They give the example of contamination fear as an obsessional thought that is syntonic in terms of the wish not to harm others being in harmony with valued self-perceptions, but the experience of excessive feelings of responsibility and fear of the consequence of not preventing contamination are experienced by the individual as aliento his or her sense of rational thought. Purdon et al. (2007) also consider that although a thought may contradict an individual’s valued self-perceptions, it can still be consistent with his or her preferences and past experiences, thus not making it an alien thought. Additionally, dystonic thoughts can become increasingly experienced as syntonic the more an individual experiences them and shifts self-perceptions accordingly.
The distinction between intrusive thoughts and obsessions
The experience of having discrete intrusive thoughts, images, and impulses/urges is a normal, universal human phenomenon (Veale & Wilson, 2005), and in the literature these phenomena are sometimes referred to as ‘normal obsessions’ (e.g. Barrett & Healy, 2003). Intrusive thoughts are experienced by 90% of the general population (Wells & Morrison, 1994; Purdon & Clark, 1993; Freeston et al., 1992; Clark & de Silva, 1985; Salkovskis & Harrison, 1984; Rachman & de Silva, 1978) and refer to fleeting unbidden, unwanted thoughts, images, and impulses/urges which are experienced as uncontrollable and attributed to internal origin (Rachman, 1981; Wells & Morrison, 1984). However, according to cognitive models, as a result of the influence of appraisal processes (Salkovskis, 1985, 1989; Clark & Purdon, 1993; Rachman, 1993, 1997, 2002, 2004), intrusive thoughts can develop into obsessions; the difference between them being in terms of frequency, duration, intensity, and consequences (Rachman & de Silva, 1978). The more ego-dystonic the intrusive thoughts are, the more they are experienced as threatening to an individual’s sense of self (Clark, 2004), thus these intrusive thoughts become more frequent and intense (Clark, Purdon & Wang, 2003).
The classification of obsessions
Although obsessions may share key features (e.g. intrusiveness and ego-dystonicity), attempts have been made to sub-classify them. This has included grouping them into themes relating to ‘contamination fears, pathological doubt, a need for symmetry or order, body-related worries, and sexual or aggressive obsessions’(Doron et al., 2009, p.3). Lee and Kwon’s (2003) research indicated that obsession might be more broadly subtyped into autogenous and reactive obsessions, based on the analysis of responses to the Revised Obsessive Intrusions Inventory (ROII, Clark & Purdon, 1993; Purdon & Clark, 1994), which is an inventory that measures the frequency of commonly experienced intrusive thoughts. These subtypes are categorized by three key differences: (1) differences in the contentof intrusions, (2) different cognitive appraisalsof these intrusions, and (3) different control strategiesused to reduce psychological distress. Regarding reactive obsessions, the content concerns asymmetry, accidental mistakes, contamination, and loss; the cognitive appraisal of these intrusions is linked to identifiable triggers, which are normally external and relate to the feared threat (e.g. contact with dirt will trigger the fear of contamination). The psychological distress caused by the feared threat results in the use of control strategies, which for reactive obsessions are neutralising behaviours, such as checking, or in the case of contamination fear, compulsive washing/cleaning behaviour (Doron & Kyrios, 2005). Regardingautogenous obsessions, the content is proposed to be more ego-dystonic and irrational thanreactive obsessions, such as unwanted immoral, aggressive and sexual impulses and thoughts that the experiencer finds offensive, disturbing and sometimes horrific. In terms of appraisal, unlike reactive obsessions, the origins of these triggers are difficult to trace (Doron & Kyrios, 2005) and it is argued that it is the contentof thoughts, images, and impulses featuring sex, blasphemy, aggression (i.e. autogenous obsessions) which is the source of the triggers which are then misinterpreted (Clark, 2004; Doron & Kyrios, 2005; Wells, 1997).These triggers are often tenuously or symbolically related with thoughts (Lee & Kwon, 2005), such as the letter b triggering the intrusive thought of wanting to harm one’s brother (ibid.). Because these intrusive thoughts appear to be self-generated they are therefore appraised as signifying the true feared nature of the self. The control strategies used to reduce psychological distress related to autogenous obsessions are suppression and attempts to banish these intrusive thoughtsthrough the use of thought-control and/or compulsive behaviours (Lee & Kwon, 2003).
Purdon and Clark (1993) performed an exploratory factor analysis of the ROII, and found a similar two-factor structure for women (the content and form primarily divided between dirt/contamination and sex/aggression), whilst men seemed to have a one-factor structure (the content and form primarily being that of sex/aggression). It is argued that a possible reason for this study finding that men have a one-factor structure that excludes content and form relating to dirt/contamination, is because of low respondent levels (Belloch, Morillo et al., 2004, cited in Moulding et al., 2007). When analysing the responses to the Spanish version of the ROII, Belloch et al. (2004, cited in Moulding et al., 2007) found a similar two-factor model to that of Lee and Kwon (2003), divided between “(1) aggression, sexually and socially inappropriate behaviours; and (2) doubt, checking and cleanliness,” (Moulding et al., 2007, p.679).Further research evidence has supported the subtyping of obsession into autogenous and reactive subtypes (Lee & Kwon, 2003; Lee, Kwon, Kwon, & Telch, 2005; Lee, Lee, Kim, Kwon, & Telch, 2005; Lee & Telch, 2005; Lee, Zoung-Soul, & Kwon, 2005; Moulding, Kyrios, & Doron, 2007). However, although obsessions may be put into one of these two categories, it is still possible to assess on a continuum, as suggested by Lee & Telch (2005). Therefore, for example, a person may score high on autogenous obsessions and lower on reactive obsessions rather than being assigned to a single category of autogenous or reactive obsessions.
The dimensional perspective on both obsessionality and OCD
In tandem with the view that autogenous and reactive subtypes of OCD may best be conceptualized on a continuum (Lee & Telch, 2005), taxometric studies (Haslam et al., 2005; Olatunji et al., 2008, cited in Doronet al., 2009) also provide evidence that supports the dimensional rather than categorical conceptualization of OCD cognitions and symptomology. Also, as aforementioned, the content of intrusive thoughts is similar in around 90% of both clinical and nonclinical samples (Rachman and de Silva, 1978; Salkovskis & Harrison, 1994; Burns, Formea, Keortge, & Sternberger, 1995), and that the difference between normal and abnormal obsessions is also on a continuum. Additionally, both clinical and non-clinical populations reduce anxiety through using similar cognitive and behavioural strategies (Murris, Merckelbach, & Clavan, 1997). More recently, research by Garćia-Soriano et al. (2011) has also confirmed that normal obsessive intrusive thoughts and clinical obsessions are on a continuum. To assess this, the researchers used the Obsessional Intrusive Thoughts Inventory (INPIOS), a 48 item inventory which measures the frequency and content of obsessive intrusive thoughts. Their sample population was made up of 734 non-clinical participants and 55 clinical participants meeting the criteria for OCD. Garćia-Soriano et al. (2011) used confirmatory factor analysis to confirm that both the clinical and non-clinical participants experienced obsessive intrusive thoughts across the whole range of obsessional content. Therefore, if one accepts that obsessional phenomena are continuous then the study of non-clinical expressions is legitimate in theorising the nature of obsessions. This has provided the rationale for using a non-clinical population in previous studies (e.g. Doron et al., 2009), and the rationale for drawing on a community-based population in this current study; i.e. that obsessionality is experienced both in non-clinical and clinical populations and that an understanding of obsessions in non-clinical samples might further the understanding of clinical obsessions. However, although it is possible for findings from a non-clinical population to be extrapolated to a clinical population, such as to individuals with OCD where obsessionality is a key feature of the disorder, this is not automatic.For example, non-clinical samples might produce a restricted range on obsessional measures.
Phenomenology of OCD
OCD is a psychiatric diagnosis defined by the experience of obsessions that produce anxiety, and/or compulsive acts (compulsions) aimed at reducing anxiety (Stein & Fineberg, 2007) to a degree whereby these experiences cause the experiencer distress and can negatively affect the person’s ability to function in his or her daily life (Roth & Fonagy, 2005; Wells, 1997). Compulsions are repetitive, ritualistic, inflexible behaviours that can be covert (such as mental acts) or overt (such as repetitive handing washing) that the individual is compelled to carry out in response to their obsessional thoughts. The purpose of performing these compulsions is to experience temporary relief from the distress caused by the obsessive thoughts, which may include attempting to prevent a feared event from happening (Doron et al., 2009). However, 25% of people with OCD do not present with overt compulsive rituals, and may instead use covert compulsive rituals, such as counting or reciting ‘good’ thoughts to try and replace ‘bad’ ones (Obsessive-Compulsive Cognitions Working Group [OCCWG], 2001). According to the Epidemiological Catchment Area survey, OCD is indicated has having a lifetime prevalence rate of 2.5% (Roth & Fonagy, 2005).Depression is commonly comorbid with OCD (Stein & Fineberg, 2007). Debates currently exist in the literature regarding the characterisation and classification of OCD as a single disorder.
As noted, classifications have been made on the basis of symptom content (e.g. ‘checkers,’ ‘washers’) or more broadly (e.g. autogenous versus reactive obsessions). The diagnostic status of hoarding as separate disorder associated to, but distinct from, OCD has also been debated and it is possible that it will be categorised as a separate disorder in the new DSM. The multitude of various presentations of OCD has led to differing attempts to categorize what is nowadays considered to be a heterogeneous disorder.
Theories of OCD
Both cognitive-behavioural and psychodynamic theories consider that a conflicted sense of self (i.e. self-ambivalence regarding whether self is essentially ‘good’ or ‘bad’) is a key feature in OCD psychopathology (Kempke & Luyten, 2007). Rachman (1997, 1998) proposed that obsessions developed from individuals appraising repugnant thoughts and images as revealing their true personality (i.e. “dangerous, evil, unreliable, and potentially uncontrollable”; Bhar & Kyrios, 2006, p. 1846), which then leads to obsessive rumination. Individuals with OCD are considered to have an ambivalent sense of self, and have schemas that relate to perfectionism, control, overemphasis of thinking, and responsibility (Kempke & Luyten, 2007). Rachman (1997, 1998, 2002, 2004) developed different cognitive-behavioural models for different subtypes of OCD, using the idea of appraisal being the significant factor in the development and maintenance of this disorder. Psychodynamic theories have considered self-ambivalence in term of unconscious conflicts (e.g. between the id and superego, which is mediated by the ego; and between internalised ‘good’ and ‘bad’ objects), and have argued that OCD occurs in individuals where the differing aspects of the self are not adequately integrated (Gomez, 1997; McWilliams, 1994). The growing convergence between cognitive-behavioural and psychodynamic conceptualizations of the role of self-ambivalence may help to explain the formation and development of OCD symptomology and lead to more effective psychological treatments for people with this psychological disorder (ibid.). These theories will now be discussed.
Cognitive models of OCD
Cognitive theories of OCD (e.g. Salkovskis, 1985, 1989; Clark & Purdon, 1993; Rachman, 1993, 1997, 2002, 2004) suggest that OCD symptoms arise from, and are maintained by, the misinterpretation of intrusive thoughts, images, and impulses, and that this leads to these cognitive processes being experienced as highly significant and threatening. That is, it is not the occurrence of intrusions per se that is problematic, but that the way in which they are focussed upon once they occur.
Researchers have focused on differing conceptualizations of the role that appraisal might play in OCD symptomology, but have agreed that these misinterpretations occur from specific cognitions and dysfunctional beliefs, categorized within the following domains: (1) over-emphasizing the importance of thoughts, which refers to misinterpreting ordinary intrusive thoughts as showing the ‘true’ nature of the self rather than an ordinary insignificant experience; (2) an inflated sense of responsibility (related to harm to self/others), which refers to the belief that one holds a more powerful position concerning provoking or preventing oneself and/or other people experiencing negative outcomes; (3) intolerance of uncertainty, which refers to “beliefs about the necessity of being certain, about the capacity to cope with unpredictable change, and about adequate functioning in situations which are inherently ambiguous” (OCCWG, 1997, p. 168);(4) thought-action fusion, which refers to a type of metacognitive beliefs (beliefs about having particular types of thoughts) that conflates thought with action; (5) overestimating the probability and severity of threat, which refers to perceiving the world as a more dangerous place than it is, and leads to a belief that constant vigilance is required to protect oneself and others from harm; and (6) perfectionism, which refers to a belief that it is possible to achieve perfection (i.e. having no faults) if one tries hard enough (OCCWG, 1997, 2001). Further, these appraisal processes are linked. For example, Rachman (1993) argued that elevated levels of responsibility develop from thought-action fusion beliefs. TAF beliefs have been divided into likelihood TAF beliefs (Rachman & Shafran, 1999), which focus on the belief that having specific intrusive thoughts increases the likelihood of them occurring; and morality TAF beliefs (ibid.), which focus on intrusive thoughts being thought of as morally equivalent to having put these thoughts into action and/or failing to do something that could prevent harm is as morally wrong as deliberately causing harm(Wroe & Salkovskis, 2000). However, often both these subtypes of thought-action beliefs are experienced related to a single intrusive thought. For example, the experiencer both believes that the thought may come true and that they would be morally responsible if this were to occur.
Salkovskis’s (1985, 1989) theoryof OCD is presently the most comprehensive cognitive-behavioural theory of this disorder, and emphasises both how people with OCD tend to appraise their intrusive thoughts, images, and impulses as a sign that there are significant risks to self and others, and that they may be responsible for either the harm they fear has occurred or its prevention (Salkovskis & McGuire, 2003). This appraisal of their intrusive thoughts leads people with OCD to feel extreme discomfort and, in this theoretical model, is linked to their urge to engage in neutralizing behaviour (for example, compulsive washing or checking rituals) as a way of diminishing the experience of discomfort, which thus becomes a negative reinforcer of compulsive behaviour. This theory suggests that there are also other consequences of interpreting normal intrusive thoughts as a sign of responsibility, based on the idea that people with OCD try too hard to control their thoughts and actions. These include an increased attention to (and thus access to) their intrusive thoughts and the external triggers that lead to intrusive thoughts, as well as behaviours that attempt to reduce responsibility – neutralizing, reassurance-seeking, compulsions, avoidance, and thought-suppression. Rather than experiencing the normal extinction of anxiety, the result of engaging in these behaviours is counterproductive because these behaviours focus individuals’ attention on these distressing thoughts, thereby leading to them being experienced with increased frequency. Additionally, engaging in these behaviours prevents distressing intrusive thoughts from being tested, and the feared consequences of not engaging in these behaviours being disconfirmed (ibid).
Salkovskis’ model has been criticized for not focusing upon motivational factors, for not questioning why appraisals of intrusive thoughts are distressing or why there is an urge to engage in repetitive behaviour, and for focusing more upon cognitive factors to the detriment of considering emotional factors (Jakes, 1996; O’Kearney, 1998). However, Salkovskis has responded by reiterating how people with OCD are motivated by their need to attempt to diminish the experience of discomfort caused by their appraisal of their intrusive thoughts, and thus engage in neutralising and compulsive behaviour (Salkovskis & Freeston, 2001). Salkovskis et al. (1999) proposed that the origins of why some people may experience a distorted appraisal of intrusive thoughts that leads them to experience an inflated sense of responsibility may be related to them having experienced particular learning experiences. These include the pairing of an early developed sense of responsibility along with implicit and/ or explicit encouragement during the person’s formative years; having the experience of a thought and/or action that incorrectly is attributed to the occurrence of a negative event; an experience whereby an act of commission or omission leads to a negative event; being exposed to extreme and rigid codes of conduct and duty; or, not experiencing one’s own sensitivity to thoughts regarding responsibility being challenged by others (ibid.). Salkovskis et al.’s (1999) proposal that there are 5 primary pathways to the development of inflated responsibility was later given support by Coles and Schofield (2008) who were able to test this through developing the Pathways to Inflated Responsibility Beliefs Scale (PIRBS) which they correlated with other measures of OCD symptomology, parental styles, OCD-related beliefs, and levels of childhood responsibility (ibid). However, a critique of Cole and Schofield’s (2008) study is that because the data was derived from participants with OCD and hoarding behaviour, it is possible that the experience of living with these disorders may have distorted recall of early experiences. According to Abramowitz, Schwartz, and Moore (2003) experiences of blame and criticism, as well as situational/actual increases in responsibility, may interact with early experiences, thus predisposing individuals to appraise negatively their normal intrusive thoughts.
Psychodynamic theories of OCD
OCD as a concept existed in the psychoanalytic tradition and as such can be found in the early literature, where it is referred to as ‘obsessional neurosis’ (Freud, 1926). Freud initially speculated that this neurosis resulted from individuals experiencing something, primarily sexual, that was incompatible with his or her ideational lives. This caused such strong emotional distress, that thoughts related to this experience were supressed. Freud proposed that unconscious thoughts resurface into conscious awareness in the form of symptoms, and that supressed thoughts related to sexual acts that occurred as a child would return as obsessional thoughts (Freud, 1824, 1896). Freud also proposed that obsessive behaviours were not senseless, but instead had meaning which could be interpreted (Freud, 1907). However, Freud began to question his initial belief that people with obsessional neurosis had experienced sexual acts as children, and began to postulate that these experiences may have been fantasies rather than reality for most people. Freud’s focus on repressed sexuality lead him to suggest that children need to successfully negotiate psychosexual stages (oral, anal, and phallic), and that fixation at, or regression to, the anal stage results in the development of OCD symptomology in adults (Jakes, 1996). Freud associated the themes of toilet training with the conflictual themes of “aggression and submissiveness, cruelty and gentleness, dirtiness and cleanliness, order and disorder” (Fenichel, 1977, p.273, cited in Jakes, 1996) with those that are seen in OCD symptomology (Jakes, 1996). He argued that regression to this stage may occur as a result of an unresolved Oedipus conflict (ibid.), at a stage when issues of control, sexuality and aggression were being worked through in relation to the child’s primary caregivers.
However, other psychodynamic theorists have argued that OCD is not connected to psychosexual stages of development, and instead results from an individual’s incapacity to feel in control of life (e.g. Alder, 1964, cited in Jakes, 1996), and thus attempt to overly control aspects of themselves (Malan, 1979, cited in Jakes, 1996). More recent views of OCD have focused on Object Relations (e.g. Fairbairn, 1943, cited in Gomez, 1997) and attachment theories (e.g. Bowlby, 1979). Object relations theorists moved away focusing on the idea that humans need to find ways of reducing tension caused by instinctual drives (pleasure-seeking) and instead proposed that people’s overriding need is for intimacy from other humans (object-seeking) (Gomez, 1997; Howard, 2006).
Freud (1917), in his work on melancholia, refers to unconscious ambivalence towards objects, thus allowing them to be kept good. Klein (1946, cited in Gomez, 1997) also developed a theory of how, within the first few months of life, infants’ minds are dominated by the schizoid-position. This refers to the unconscious way they ‘split’ their objects into ‘good’ and ‘bad’ ‘part-objects,’ as a way of attempting to both develop love and trust towards their primary caregivers whilst also managing their overwhelming feelings of hate towards them (borne out of feelings of persecution when their infantile needs are not met).
Drawing on both Freud’s and Klein’s work on ‘splitting,’ Fairbairn (1943) (who was a founder member of the Object Relations movement) developed the idea of ambivalenceby postulating that infants internalize aspects of their primary caregivers that they find frightening and bad because they both need the objects and want to control them (so they are less threatening). By internalizing these objects the infant is “taking upon himself the burden of badness which appears to reside in his objects’ (Fairbairn, 1943, p.65), thereby experiencing the self as bad as opposed to the caregiver. Fairbairn refers to this as ‘moral defence’ (1943, p.66) which the infant unconsciously chooses because ‘it is better to be a sinner in a world ruled by God than to live in a world ruled by the Devil’ (Fairbairn, 1943, p.66); i.e. that it better to be bad but to live in a good, safe, secure, world, even if one is bad, than to escape being a sinner but live in a world that is ‘bad,’ unsafe and insecure wherein the destruction of the self is inevitable. Therefore, it is possible that Fairbairn’s (1943) theory helps to explain the observation that people with OCD adopt the most negative appraisal of themselves, a motivation that is contrary to research and theory that proposes that people are motivated to appraise information such that it maintains their positive self-image, such as recent research exploring the fundamental attribution error (Maruna & Mann, 2006) and the self-serving bias (Shepperd, Malone & Sweeny, 2008). Instead, it would appear that people with OCD have self-concepts (self-representations) that result in negative self-appraisals when experiencing unwanted intrusive thoughts (Sherman & Cohen, 2002). It might be speculated therefore that people with OCD feel safer and more secure when they appraise their intrusive thoughts as an indication that they are ‘bad’ rather than the world they occupy. This conceptualisation of OCD suggests that self-representations may be an important factor to be considered when considering obsessionality across both non-clinical and clinical populations.
Drawing on both cognitive-behavioural and psychodynamic ways of conceptualising conflicting parts of the self (both conscious and unconscious), this paper now considers two conceptualisations of self-representation: self-ambivalence and self-discrepancy.
Self-ambivalencerefers to the distressing discomfort and uncertainty caused by holding conflicting self-representations/beliefs about the self; for example, that one is both worthy and unworthy, moral and immoral, loveable and unlovable (Kempke & Luyten, 2007). In formulating their theory of the relevance of self-ambivalence to obsessional thinking, Guidano and Liotti (1983) drew on Bowlby’s (1970) attachment theory to help explain how opposing self-representations/beliefs about the self (self-ambivalence) may develop within an individual. Bowlby (1970) proposed that primary caregiver attachment relationships are internalized in infancy to create what he describes as ‘internal working models’ of the self, others, and relationships. Accordingly, if infants experience their primary caregiver as responsive and supportive, then Bowlby’s (1970) attachment model suggests that this will enable them to develop an internalized self-representation of being lovable, worthy, and competent. Similarly, if infants experience their primary caregiver as unresponsive and unsupportive then they will develop an internalized self-representation of being unlovable, unworthy, and incompetent.
Guidano and Liotti (1983) proposed that if infants experience ambivalent attachments with their primary caregivers then they would develop an ambivalent self-representation. They proposed that ambivalent attachments could form either when infants experience their primary caregiver as oscillating between being responsive and supportive, and unresponsive and unsupportive/critical, or when infants experience one of their parents as responsive and supportive, and the other one as unresponsive and unsupportive/critical. Guidano and Liotti (1983) proposed that both of these experiences would lead to the development of internalized conflicting self-representations of being loveable, worthy, and competent andunlovable, unworthy, and incompetent.
In considering the relevance of self-ambivalence to obsessions, Guidano and Liotti (1983) argue that people with OCD are ambivalent about whether they are intrinsically moral, lovable, and worthy. Consequently, intrusive thoughts become more significant and distressing for individuals with OCD because they activate the negative aspect of conflicting beliefs about the self (e.g. ‘I am immoral’), and are appraised as evidence of their ‘true’ revealed self. The more ego-dystonic the thoughts, the more these thoughts are experienced as a threat to the valued aspects of the self (i.e. moral, worthy, competent). Similarly to Rachman’s theory (1997,1998), Guidano and Liotti considered that compulsions could be viewed as attempts to recommit to moral and social ideals (Bhar & Kyrios, 2007) thus restoring the positive aspect of conflicting beliefs about the self (e.g. ‘I am moral’) of the ambivalent self-concept, thereby resolving the ambivalence. Frost et al. (2007) propose that individuals with conflicting self-representations (i.e. self-ambivalence) become preoccupied (i.e. obsessional) with seeking environmental evidence to validate each of their polarized views of self in an effort to seek certainty about their self.
Drawing on Guidano and Liotti’s (1983) work, Bhar (Bhar, 2004; Bhar & Kyrios, 2000; Bhar and Kyrios, 2007) researched the associations between self-ambivalence and OCD, the results of which suggested that people with OCD are more self-ambivalent than healthy controls (Bhar and Kyrios, 2007). To conduct this research, Bhar and colleagues developed the Self-Ambivalence Measure (SAM; Bhar & Kyrios, 2007) which measures conflicting beliefs about morality, self-worth and social acceptance.Bhar and Kyrios (2007) found thatself-ambivalence positively correlated with severity of OCD-related dysfunctional beliefs (i.e. responsibility, perfectionism, appraisal of thoughts as important) and symptoms. Additionally, participants with OCD scored significantly higher on the SAM compared to a non-clinical control group (the results remained significant after controlling for mood and self-esteem). Conversely, no significant difference was found between those individuals with OCD and those with general anxiety disorder (GAD), suggesting that self-ambivalence is a general feature of anxiety disorders rather than a distinct feature of OCD. However, a confounding methodological issue was that the participants in both the OCD and anxiety group had high levels of OCD dysfunctional beliefs, thus the study did not actually differentiate between these two clinical groups. George (2010) found that self-ambivalence (measured by the SAM) was significantly higher in an OCD group compared with both a mixed non-OCD group and a non-clinical group. This may be due to actively ensuring that only the OCD group, and not the anxiety group, contained participants who met the DSM-IV diagnostic criteria for OCD. This was achieved by a screening process that used the Structured Clinical Interview for DSM-IV AXIS I Disorders - Clinical Version (SCID-CV, First et al., 1997). Following the idea that people with OCD are of a “tender conscience” (Rachman & Hodgson, 1980), Doron and Kyrios (2005) argue that although overall self-ambivalence may not be a distinct factor of OCD to that of other anxiety disorders, it may contain specific domains that are more sensitive in OCD, such as the morality domain. This conceptualization led to further research that showed that people with OCD had sensitive domains within their self-conceptions (social acceptability, job/school competence, and morality) (Doron, Kyrios and Moulding, 2007; Doron, Moulding, Kyrios, and Nedeljkovic (2008), which in turn enabled for a broader conceptual understanding of OCD than had previously been achieved.
Bhar and Kyrios (2007) argue that self-ambivalence may be a meta-vulnerability for OCD that underpins the dysfunctional schemas identified by the Obsessive-Compulsive Cognitions Working Group (1997). For example, this could result in the following: a perfectionist schema that requires everything to be faultless otherwise it results in a negative appraisal of the self; an intolerance of uncertainty; wishing to control thoughts so as to reduce the activation of distressing negative beliefs; focusing on personal responsibility so as to ward thoughts that their ‘real’ self is irresponsible’; over-emphasising the importance of thoughts compared to emotions, because thoughts are easier to control than emotions. Currently, little research has looked at the relationship between self-ambivalence and obsessionality and even less has looked at the differential relationships with obsessional subtypes. Given the earlier comments about the heterogeneity of obsessional content, theory could be advanced by looking at self-ambivalence as a predictor of different subtypings.
Ascertaining the relationship between autogenous and reactive obsessions and self-ambivalence
Autogenous obsessions are predicted to be more ego-dystonic than reactive obsessions, and to have fewer identifiable triggers (Lee & Kwon, 2003). Moulding et al. (2007) have proposed, based on Lee and Kwon’s differentiation between autogenous and reactive obsessions presented above, that autogenous obsessions have a more negative impact on the person’s sense of self. Additionally, Purdon and Clark (1999) suggest that autogenous obsessions which are experienced as alien to the self (Purdon et al., 2007) are likely to induce more doubt about a person’s sense of self than reactive obsessions. The researcher in the current study postulates that this may be because of the lack of perceived situational context (i.e. no trace of an external trigger that generates these intrusive thoughts) induces fear that these unwanted thoughts can only be attributed to personal traits, and thus are appraised as more significant than reactive obsessions. It is also possible that the significance given to these normal autogenous thoughts is further compounded in individuals with high levels of self-ambivalence. This is because the activity of attributing unwanted ego-dystonic normal intrusive thoughts to a personal trait, rather than understanding it as a response to a situational context, is likely to activate the negative aspects of self-ambivalence. This causes intrusive thoughts to be evaluated as a sign of the self being dangerous and immoral, thus evoking the fear that these unwanted ego-dystonic thoughts and impulses may be acted upon.
Consequently, the current study predicts that autogenous obsessions will be more highly related to self-ambivalence than reactive obsessions. In the only published study looking at the relation between autogenous and reactive obsessions and self-ambivalence, Moulding et al. (2007) found a higher correlation between autogenous obsessions and self-ambivalence (r= .30) than between reactive obsessions and self-ambivalence (r= .25) using a non-clinical sample, although the difference between the correlations is very small. Given that this is the only study, the first aim of the current study is to ascertain the relationship between autogenous and reactive obsessions and self-ambivalence using a sample of non-clinical participants.
There will be a positive correlation between self-ambivalence and both autogenous and reactive obsession scores, but the effect size for the correlation between self-ambivalence and autogenous obsessions will be greater.
The theories of self-ambivalence (e.g. Guidano & Liotti, 1983; Bhar & Kyrios, 2007; Frost et al. (2007) and self-discrepancy (e.g. Higgins, 1987; Carver et al., 1999; Ferrier & Brewin, 2005) are similar in that they both conceptualize individuals as having multiples selves. However, whilst self-ambivalence theory refers to conflict between differing self-representations (e.g. moral versus immoral), self-discrepancy theory is more explicit in describing this feature in terms of differing incongruent selves.
Higgins (1987) self-discrepancy theory offers a different approach to evaluating conflicting self-representations than self-ambivalence theory (i.e. that people have different self-state representations). These are the ‘actual self,’ being the representation of who they currently are; the ‘ideal self,’ being the representation of who they wish to be; and the ‘ought self,’ being the representation of who they ought to be. Higgins’s (1987) hypothesis that experiencing discrepancies between ‘actual’ and ‘ideal self’ would result in depressive type symptomology, whereas experiencing discrepancies between ‘actual’ and ‘ought self’ would result in anxiety-related symptomology, were later supported by empirical research (e.g. Higgins et al., 1986; Higgins, 1987; Higgins, 1996). Additionally, other researchers (e.g. Oglivie, 1987; Carver et al., 1999) have proposed and studied the ‘feared self’; that is the representation of who a person fears they might be or become. The feared self may be particularly pertinent to OCD given the threat content of obsessions. Drawing on Rachman’s (1997, 1998) theory, Ferrier and Brewin (2005) proposed that people with OCD have an uncertain sense of self which results in them attempting to protect others from their ‘feared self’ which they fear will otherwise cause people harm, thus proving that they are intrinsically dangerous and immoral. Also, based on Rachman’ theory that people with OCD are distressed by unwanted ego-dystonic intrusive thoughts because they believe that they reveal who they really are (i.e. dangerous, immoral and bad), Ferrier and Brewin (2005) hypothesized that a smalldiscrepancy between ‘actual’ and ‘feared self’ may result in OCD symptomology. Ferrier and Brewin (2005) used the Selves Questionnaire (Carver et al., 1999) to research this hypothesis.A smaller discrepancy between ‘actual self’ and ‘feared self’ was found in the OCD sample compared to healthy controls, but no significant difference in discrepancy size was found between OCD and the non-OCD anxiety group. However, content analysis revealed that the ‘feared self’ in OCD was that of being dangerous (immoral, bad, or insane), whereas the ‘feared self’ within the non-OCD anxiety groups was related to that of being hopeless and fearful. The feared self of the healthy controls was characterised by a broad spectrum of negative traits (e.g. pride, selfishness).These findings seem to support Doron et al.’s (2005) findings that suggest that people with OCD have a sensitive morality domain.
More broadly, Rowa and colleagues (Rowa & Purdon, 2003; Rowa et al., 2005) researched obsessional content and distress levels and found that ego-dystonic thoughts (i.e. thoughts which contradict valued self-perceptions) were found to be more distressing than ego-syntonic thoughts (i.e. thoughts which match valued self-perceptions) in people with OCD and non-clinical participants. Rowa and Purdon’s (2003) research into the role of ‘self’ as a vulnerability factor in OCD found that students who were allocated the task of reporting upsetting intrusive thoughts reported that these thoughts were more contradictory to their sense of self than students who were allocated the task of reporting their least upsetting intrusive thoughts. A replication of this study by Rowa, Summerfeldt and Antony (2005) using a clinical cohort gave similar results, indicating that ego-dystonic intrusive thoughts were rated by people with OCD as more meaningful and distressing than ego-syntonic thoughts (Rowa, Purdon, Summerfeldt, & Antony, 2005). These studies indicate that intrusive thoughts are appraised in terms of valence and significance based on self-representations (Doron et al., 2009).
Self-discrepancy and autogenous and reactive obsessions
The Selves Questionnaire (Carver et al., 1999) allows empirical assessment of the discrepancy between actual self and ideal, ought, and feared self. In this way it becomes possible to employ another method to assess differing self-representations and how these might be related to autogenous and reactive obsessions.No current published studies exist that explore the associations between self-discrepancy and the autogenous/reactive subtyping. Such research might increase theoretical development and also indicate the relevance of looking at self-discrepancies in clinical samples of people with obsessions and compulsions. Therefore, the second aim of the current study is to look at the relationship between discrepancies in self-positions and autogenous and reactive obsessions.
Is there a relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and total obsession scores.
Is there a differential relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and autogenous versus reactive obsession scores?
This study employed a cross-sectional correlational design in a non-clinical sample. Participants completed all questionnaire-based measures in a survey at one time-point via a website on the internet (created using SSIWeb7, Version 7.0.22., developed by Sawtooth Software).Therefore participants could access the online questionnaire from the setting of their choosing and were not geographically bound to the local area of the researcher.
Sample size calculation
A priori power analysis was completed for both the hypothesis and exploratory research questions, with the aim being to estimate the size of the correlation coefficient (r). According to Cohen (1992), a large effect is r ≥ 0.5 and small effect size is r = 0.1. The sample size affects the precision of the estimates of correlation coefficient. Stata (Statacorps, 2009) was used to estimate the width of 95% confidence intervals for different effect sizes. With a sample size of 150 participants the 95% confidence interval (95% CI) would be of width 0.11 for a large effect and 0.10 for a small effect. With a smaller sample size of 100 the 95% CI would increase to 0.20 for a large effect and 0.13 for a small effect. For a sample size of 50 the 95% CI would be 0.29 for a large effect and 0.19 for a small effect. A sample size of between 100 and 150 was therefore aimed for. A total of 506 participants began the survey, and out of these 374 participants provided useable data. Therefore, the intended sample size was surpassed.
The sample criterion was any individual aged 18 or above. Regarding the rationale for using a non-clinical population, Doron et al. (2009) propose the following evidential support: first, that content of intrusive thoughts is similar, though less frequent in both clinical and nonclinical samples (Rachman & de Silva, 1978; Salkovskis & Harrison, 1984); second, that clinical and non-clinical populations reduce anxiety through using similar cognitive and behavioural strategies (Muris, Harald & Clavan, 1997). Additionally, there is a long tradition of utilising non-clinical samples to understand clinical OCD due to dominant cognitive-based theories conceptualising obsessions and compulsions as phenomena that are continuous in the general population.
The study was advertised on the following four websites:
1) The ‘Volunteer and Charity Work’ section (across UK geographical areas with a high population so as to reach as many potential participants as possible) on Gumtree (www.gumtree.com), which is a website specialising in online advertising.
2) The University of Hanover (http://psych.hanover.edu/research/exponnet.html),
3) Online psychological research UK (www.onlinepsychresearch.co.uk)
4) Social psychology network (www.socialpsychology.org/expts.htm).
All these sites hosted a one line advert (Appendix 1), within which a link to the study Information Webpage was embedded. The latter three websites have been specifically designed for researchers to advertise their studies online. Administrative staff and students within the Faculty of Arts and Human Sciences (FAHS) at Surrey University were also invited to participate via an email in which they were sent a full-length advert (Appendix 2).
The full-length advert, the one line advert, and the Information Webpage (Appendix 3) offered a financial incentive for participants who completed the survey. The full-length advert and the Information Webpage explained that this offer could be taken up by the participant adding their email address to the Demographic Questionnaire (Appendix 4), on the understanding that it would be placed in a raffle, and would only be used, and the corresponding participant contacted, if it was one of the first four pulled out. First prize was £40, and the 2nd, 3rdand 4thprices were £30, £20, £10, respectively.
Due incomplete responses, there were 130 (25.69%) unusable surveys out of the 506 who began the surveys. Therefore there were 376 useable surveys, of which there were 85 (22.6%) males and 291 (77.4%) females. The age range of participant was between 18 and 64, with mean average of 26.24 (standard deviation = 8.77). For males, the mean average was 26.60 (standard deviation = 9.27), and for females, the mean average was 26.13 (standard deviation = 8.63). See table 1 for additional demographics.
Table 1. Summary of additional demographics has been removed for this copy of the research
In terms of the implications for the external validity of study, it is important to observe that the participants were mainly female (77.4%), young (mean average of 26.24), and from the UK. Although the literature does not predict a difference in the strength of associations between self-ambivalence, self-discrepancy, and autogenous and reactive obsessions, based on gender, age or ethnic origin, this may still be possible. Therefore, if this study was to be replicated, it might be useful to consider using a sampling procedure that ensures equanimity in terms of male and females participants, and that actively attempts to attract participants from across all ages. Difficulties that can occur when using self-reports across cultural groups are addressed in the section entitled, ‘using an online survey.’
Reactive and autogenous obsessions
The Revised Obsessive Intrusions Inventory, Part 1 (ROII; Purdon & Clark, 1993; Purdon & Clark, 1994)is a widely used measure of obsessions in clinical and non-clinical groups (Appendix 5). It comprises of a 52-item self-report questionnaire measuring the frequency of intrusive thoughts, images and impulses analogous to obsessions on a 7-point Likert scale (from ‘I have never had this thought’ (0), to ‘I have this thought frequently during the day’ (6)). Examples of items from this inventory include: ‘I have had unacceptable intrusive thoughts that I left the heat, stove or lights on which may cause a fire,” “While driving I have had unacceptable intrusive thoughts of swerving into oncoming traffic.” The reliability and validity of this measure has been demonstrated within clinical and non-clinical populations, where it demonstrated high internal consistency, and high discriminant and concurrent validity with depression and anxiety(Clark & Purdon 1993; Purdon & Clark, 1994).The researcher of current study was granted permission from the authors of the ROII to both use this measure online and for the instructions in this measure to be slightly modified so as to work in an online environment. Lee and Kwon (2003) carried out a factor analysis on the ROII (Part I), and from this derived that questions 1-21 and 25-44 relate to autogenous obsessions, and questions 22-24 and 45- 52 relate to reactive obsessions. These findings have been replicated by Lee, Kwon and Telch (2005); Lee, Kim, Kwon and Telch (2005); Lee & Telch (2005); Lee, Zoung-Soul and Kwon (2005); and Moulding, Kyrios and Doron (2007). Therefore, each completed ROII provides both an autogenous obsession frequency score and reactive obsession frequency score. Given that autogenous obsession frequency scores are measured with 41 items and reactive obsession frequency scores with 11 items, in the current study total and average scores were calculated to make the scales comparable.
The Self-Ambivalence Measure (SAM; Bhar, 2004; Bhar & Kyrios, 2007) draws on Guidano and Liotti’s (1993) idea of conflicting beliefs about the self, and is a 21 item self-report questionnaire that measures conflicting beliefs about self-worth, morality, and social acceptance (Appendix 6). Respondents rate each item on a five-point Likert scale (‘Not at all’ 0, to ‘Agree Totally’ 4). The SAM consists of two subscales: self-worth ambivalence (SA) comprising 13 items, and moral ambivalence (MA) comprising 6 items, with higher scores indicating increased self-ambivalence. Both subscales and the full scale (total ambivalence scored, calculated by summing the two subscales) demonstrated good internal reliability in clinical and non-clinical samples (Cronbach’s α ranging from .85 to .88), and acceptable test-retest reliability (r= .77, p= .000) (Bhar, 2004; Bhar & Kyrios, 2007). The SAM is readily available online.
The Selves Questionnaire (Carver et al., 1999) is a self-report questionnaire based on Higgins’s (1985) self-discrepancy theory. Participants are given descriptions of self-concepts (‘ought self,’ ‘ideal self,’ ‘feared’ self) and are then asked to identify and list seven of their own traits for each of these self-concepts (see Appendix 7 for definitions of these self-concepts). Participants then rate the discrepancy on a 7-point scale between each of their listed traits and their perceived ‘actual self’: the scale ranges from 1 (‘I am just like this trait’) to 7 (‘I am the opposite of this trait’), with higher scores indicating a greater discrepancy between the perceived actual self and the other self-position (i.e. ought, ideal, feared). These ratings are summed to produce total discrepancy scores for each pair. Regarding psychometric properties of the Selves Questionnaire, reliability data is not available for this questionnaire, but it does have good face validity (Ferrier & Brewin, 2005). Reliability data may not be available because the construct of self-discrepancy is too fluid because it requires participants to operationalize their self traits in similar ways. However, this does not invalidate the questionnaire, because the questionnaire is measuring the discrepancies between differing self traits. The researcher of the current study was granted permission from the author of the Selves Questionnaire to both use this measure online and for the instructions in this measure to be slightly modified so as to work in an online environment.
The Depression, Anxiety and Stress scale (DASS; Lovibond & Lovibond, 1995)is included in this study so that relationships between key variables can be assessed whilst controlling for dysphoria (Appendix 8). It is a widely used 42-item self-report questionnaire, based on a dimensionalconceptualisation of psychological disorder. It comprises three self-report scales, measuring depression, anxiety, and stress. Each subscale consists of 14 items which participants score on a 3 point Likert scale (where 0 = ‘did not apply to me at all’, and 3 = ‘applied to me much, or most of the time’) to rate their experiences of depression, anxiety, and stress over the past week. The depression scales assesses inertia, anhedonia, lack of interest/involvement, hopelessness, devaluation of life, self-depreciation, and dysphoria. The Anxiety scale assesses situational anxiety, subjective experience of anxious affect, skeletal muscle effects, and autonomic arousal. The Stress scale assesses nervous arousal, difficulty relaxing, being irritable/over-reactive, upset/agitated, and impatient.The DASS demonstrated excellent internal consistency for each of the subscales (Depression = .96; Anxiety = .92; Stress = .95), and for the total scale (Cronbach α = .97) (Page, Hooke, & Morrison, 2007), and suitability for clinical and non-clinical sample populations (Lovibond & Lovibond, 1993, cited in Lovibond & Lovibond, 1995; Antony, Beling, Cox, Enns, & Swinson, 1998; Crawford & Henry, 2003). All three subscales demonstrate adequate discriminant and convergent validity. The DASS demonstrated construct validity through the capability of the Depression and Anxiety Scales to differentiate between depressed and anxious clinical groups (Page, Hooke, & Morrison, 2007). The DASS is readily available online.
All measures described above were formatted into a single online survey. The first screen participants saw was the Information Webpage. Participants were then asked to read the Consent Webpage (Appendix 9) carefully and then click ‘yes’ if they wished to consent to participation in the study, and ‘no’ if they did not wish to consent. Participants then completed the Demographics Questionnaire (Appendix 4).In order to maintain anonymity, participants were not asked to give their name, but were asked to give their email address if they wished to be included in the raffle. Participants then completed the measures in the following sequence: SAM (Bhar, 2004; Bhar & Kyrios, 2007), ROII (Part I)(Purdon & Clark, 1993; Purdon & Clark, 1994), DASS (Lovibond & Lovibond, 1995), and then Selves Questionnaire (Carver et al., 1999). At the end of the questionnaire stage, a Debriefing Webpage gave a fuller explanation of the purpose of the study, stated that this study was not designed to diagnose mental health conditions, and then recommended that participants see their GP/physician if they felt they need professional support. This screen directed participants to a list of organizations from which they could access support if they found any aspect of the questionnaire process distressing.
Once the recruitment phase has been completed, the numerical data was exported from the online data collection website into a SSPS datafile in order to analyse the data. During this process participants were automatically allocated a unique identification number. The email addresses entered into the demographic questionnaire from participants who requested to take part in the raffle were noted down and were put in a box. The emails were then deleted from the data file as they potentially identified participants’ data. The first four email addresses pulled out of the box were used to contact the corresponding participants to say they had won. These participants provided the researcher with further contact details in order to receive the aforementioned financial prizes. The researcher’s supervisor acted as a judicator throughout this process. Participants who ticked a box on the demographic questionnaire indicating that they would like to receive a summary of the study findings once they are available were sent a copy via email.
This study received a favourable ethical opinion from Surrey University’s Faculty of Arts and Human Sciences Research Ethics Committee.
An information sheet was the first webpage of the online questionnaire, and clearly briefed participants of the study’s purpose. It stated that all confidential material would be anonymised, and that participants did not have to include their email address in the demographic questionnaire unless they wished to be entered in to a raffle. The information webpage also stated that participants could withdraw from the study during survey completion. Consent was given by participants clicking on the ‘yes’ button on the website. At the end of the questionnaire there was a debrief webpage which explained the study in more detail. Although the study used a community-based sample and the measures have been used in previous studies, the researcher could not rule out that some participants might experience a degree of anxiety when completing some of the questionnaires. Therefore, the debriefing webpage also directed participants to a list of organizations from which they could access support if they found any aspect of the questionnaire process distressing. It also stated that this study was not designed to diagnose mental health conditions, and recommended that participants see their GP if they felt they needed professional support subsequent to taking part.
Statistical analyses were performed using PASW (SPSS) version 18.0 for Microsoft Windows.
All data was screened for errors, outliers and missing values. Due to the large sample size, normality of distribution of each variable was explored by both using visual inspection of histograms and both the Shapiro-Wilk and one–sample Kolmogorov- Smirnov tests. Due to some variables being positively skewed, log transformations were used to achieve normality of distribution for these variables.
Zero-order Pearson’s correlations and partial correlations were used to explore relationships between the variables in order to address the primary research questions.
Data screening confirmed that out 376 useable surveys, 310 participants completed the Selves Questionnaire. However, 56 of these respondents did not input all seven traits for each of the ought self, ideal self, and feared self sections of the questionnaire. Although respondents were asked to insert seven traits for each section, not all participants inserted the full 7 traits. Therefore, to standardise the scores an average discrepancy score was calculated by dividing the total discrepancy by the number of traits identified. Data screening confirmed that the Selves Questionnaire total average discrepancy and subscale average discrepancy (e.g. ought self, ideal self, feared self) scores were approximately normally distributed. The SAM total scores also approximate a normal distribution. However, data was positively skewed for the DASS total and subscale totals (depression, anxiety, stress), and the ROII total and subscale totals (e.g. autogenous obsessions and reactive obsessions). This was expected because the study used a nonclinical population. However, after a log transformation, further screening confirmed that these transformed variables were not significantly different from a normal distribution.
Descriptive statistics, prior to log transformations, are shown in Table 2, which has been removed from this internet-published research document.
The overall mean scores for autogenous and reactive obsession in the current study are slightly higher than those from Lee and Telch’s (2005) sample of college students, (autogenous obsessions mean score = 21.47, SD = 20.81, and reactive obsessions mean score = 8.32, SD = 7.70). Even when acknowledging that comparisons are not exact, it suggests that the participants in the current study were more obsessional across both obsessional subtypes. Also, similarly to Lee and Telch’s (2005) results, in the current study there is a higher mean score for autogenous obsessions compared with reactive obsession. However, the reason for this is because Lee and Kwon’s (2003) analysis (and further replicative analyses) of the ROII found that 41 of the questions were related to autogenous obsessions whilst only 11 were related to reactive obsessions. However, when looking at the item means for the current study, it becomes clear that the reactive item mean is higher than the autogenous item mean, indicating that this type of obsessions were more common in the current sample.
It is interesting to note that the mean score for stress was higher than both anxiety and depression, and that anxiety was rated as lowest out of the three dysphoria states. This may because non-clinical participant samples may be more used to experiencing stress rather than anxiety.
Regarding self-discrepancies, both the overall average mean and the item mean (taking account of the fact that not all participants completed 7 traits for each self-representation) show that there is a greater discrepancy between actual and feared self than the other self-discrepancies. However, this needs to be understood within the context that larger discrepancies between actual and ought self, and also between actual and ideal self, have negative consequences, whilst the opposite is true for discrepancies between actual and feared self where smaller discrepancies would have negative consequences for the self.
In the analyses reported, sample sizes vary due to participants not completing all of the measures. In analyses involving self-discrepancy, average discrepancy scores are used.
Relationship between self-ambivalence and obsessive subtypes
Hypothesis:There will be a positive correlation between self-ambivalence and both autogenous and reactive obsession scores, but the effect size for the correlation between self-ambivalence and autogenous obsessions will be greater.
Zero-order Pearson correlations indicated significant positive associations with medium effect sizes between self-ambivalence and both autogenous and reactive obsessions. However, contrary to the hypothesis, even though both effect sizes were within the medium range, there was a greater sizefor the correlation between self-ambivalence and reactive obsessions (r= .412, p <.001, df = 374) compared with autogenous obsessions (r= .403, p< .001, df= 374). However, there were also significant positive associations with large effect sizes between self-ambivalence and the measure of dysphoria (DASS) (r= .650, p < .001, df= 352), between autogenous obsessions and dysphoria (r= .510, p < .001, df= 352), and between reactive obsessions and DASS (r= .521, p < .001, df= 352). Therefore, it was considered important to examine the relationship between self-ambivalence and autogenous and reactive obsessions after partialling out dysphoria.
Partial correlations between self-ambivalence and obsessions controlling for dysphoria yielded significant positive associations between self-ambivalence and both autogenous and reactive obsessions. However, the large effect sizes were reduced to small effect sizes. Similarly to the zero-order analysis, there was a correlation between self-ambivalence and reactive obsession (pr= .113, p = .033, df= 351) than self-ambivalence and autogenous obsessions (pr = .109, p = .041, df= 351).
Relationship between self-ambivalence and total obsessionality
Zero-order Pearson correlations indicated significant positive associations with medium effect sizes between self-ambivalence and total obsessionality scores (ROII) (r= .475, p< .001, df= 374). However, there were also significant positive associations with large effect sizes between total obsessionality scores and dysphoria (r= .593, p < .001, df= 352). Therefore, it was considered important to examine the relationship between self-ambivalence and total obsessionality after partialling out dysphoria. After partialling out dysphoria, the large effect size was reduced to a small effect size but was still significant (pr= .147, p= .006, df= 351).
Relationship between actual/ought, actual/ideal actual/feared, total discrepancy and total obsession scores
Exploratory question:Is there a relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and total obsession scores.
There was a significant positive association of a small effect size between actual/ought self-discrepancy and total obsession score (r= .282, p< .001, df= 307). There was a significant positive association of a small effect size between actual/ideal self-discrepancy and total obsessions score (r= .175, p= .002, df= 307). There was a significant negative association of a small size between actual/feared self-discrepancy and total obsessions (r= -.186, p= .001, df= 307). There was a positive association of small effect size between total self-discrepancy score (average total of actual/ought, actual/ideal, and actual/feared self-discrepancy) and total obsession score (r= .132, p< .001, df= 308).
However, there were also significant positive associations with medium effect sizes between actual/ought self-discrepancy and dysphoria (r= .370, p< .001, df= 307), and actual/ideal discrepancy and dysphoria (r= .370, p< .001, df= 307). There was also a significant negative association with a medium effect size between actual feared discrepancy and dysphoria (r= -.355, p< .001, df= 307). Therefore, it was considered important to examine the relationship between actual/ought, actual/ideal, actual feared self, and Selves total self-discrepancy and total obsession scores after controlling for dysphoria.
When controlling for dysphoria, the effect size for the positive association between actual/ought self-discrepancy and total obsession scores (pr= .084, p = .144, df= 306) reduced and was no longer significant. The effect size between actual/ideal self-discrepancy and obsession total scores almost reduced to zero (pr= -.011, p = .850, df= 306) and was no longer significant. The same was true for the correlation between actual/feared self-discrepancy and total obsession scores (pr= .017, p =.770, df= 306).
Relationship between actual/ought, actual/ideal, actual/feared, total self-discrepancy and autogenous and reactive obsession scores
Exploratory question:Is there a relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and autogenous versus reactive obsession scores.
There were significant positive associations with a small effect size between actual/ought self-discrepancy and autogenous obsessions score (r= .268, p < .001, df= 307), between actual/ought self-discrepancy and reactive obsessions score (r = .205, p < .001, df= 307), between actual/ideal self-discrepancy and autogenous obsessions score (r=.168, p = .003, df= 307), and between actual/ideal self-discrepancy and reactive obsessions score (r=.175, p= .002, df= 307). There were significant negative associations with a small effect size between actual/feared self-discrepancy and autogenous obsessions score (r = -.155,p= .006, df= 307), and betweenactual/feared self-discrepancy and reactive obsessions score (r= -.140, p= .014, df= 307). There were significant positive associations with a small effect size between Selves total discrepancy and autogenous obsessions score (r=.143, p= .011, df= 308), and between total discrepancy and reactive obsessions score (r= .131, p= .021, df= 308). As found previously, discrepancy scores were correlated with dysphoria so partial correlations were conducted.
When controlling for dysphoria, the positive associations were no longer significant and the effect size reduced between actual/ought self-discrepancy and autogenous obsessions score (pr= .099, p= .084, df= 306), between actual/ought self-discrepancy and reactive obsessions score (pr = .015, p = .795, df= 306), between actual/ideal self-discrepancy and autogenous obsessions score (pr= .013, p = .827, df= 306), and between actual/ideal self-discrepancy and reactive obsessions score (pr= .017, p = .761, df= 306). There was no significant associations and a reduced effect size between actual/feared self-discrepancy and autogenous obsessions score (pr = .019, p = .733, df= 306), and between actual/feared self-discrepancy and reactive obsessions score (pr= .044, p = .446, df= 306). The positive associations were no longer significant and the effect size reduced between Selves total discrepancy and autogenous obsessions score (pr=.064, p = .259, df= 307), and between Selves total discrepancy and reactive obsessions score (pr= .048, p = .401, df= 307).
Supplementary Analysis: the relationship between self-discrepancies and depression, anxiety, and stress
The current study’s previous finding that there was no relationship between self-discrepancies and obsessionality once dysphoria had been partialled out suggested that dysphoria was an important factor in the relationship between these variables. This was not surprising as Higgins’ (1986) self-discrepancy theory proposed that there would be a strong association between actual/ought self-discrepancy and anxiety, and between actual/ideal self-discrepancy and depression. Supplementary analysis found the following results
There was a significant positive association with a medium effect size between actual/ought self-discrepancy and stress (r= .312, p< .001, df= 309), and between actual/ought self-discrepancy and depression (r= .369, p< .011, df= 309). There was a significant positive associations with a small effect size between actual/ought self-discrepancy and anxiety (r= .249, p< .001, df= 309). There was a significant positive associations with small effect sizes between actual/ideal self-discrepancy and stress (r= .258, p< .001, df= 309), and between actual/ideal self-discrepancy and anxiety (r= .243, p< .001, df= 309). There was a significant positive association with a medium effect size between actual/ideal self-discrepancy and depression (r= .320, p< .001, df= 309). There was significant negative associations with medium effect sizes between actual/feared self-discrepancies and stress (r= -.318, p< .001, df= 309), actual/feared self-discrepancies and anxiety (r= -.303, p< .001), and actual/feared self-discrepancy and depression (r= -.311, p< .001, df= 309). There was a significant negative association with a small effect size between total self-discrepancy and stress (r= -.136, p= .017, df= 310), a non-significant positive associations with a small effect size between total self-discrepancy and anxiety (r= .095, p= .095, df= 310), and significant positive association with a small effect size between total self-discrepancy and depression (r= .202, p< .001, df= 310).
This study investigated the relationship between self-ambivalence, self-discrepancy and autogenous and reactive obsessions. It was hypothesised that there would be a positive correlation between self-ambivalence and both autogenous and reactive obsessions, but the effect size for the correlation between self-ambivalence and autogenous obsessions would be greater. Also, due to the lack of evidence base on self-discrepancy (which is similar to self-ambivalence in terms of them both measuring conflicting self-representations), the study also addressed the following: is there a relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and total obsessions scores; and is there a relationship between actual/ought, actual/ideal and actual/feared self-discrepancy and autogenous /reactive obsessions scores?
The findings of the study were that (prior to partialling out dysphoria) there were significant positive associations between self-ambivalence and obsessionality, but that the difference in effect size for the correlation between self-ambivalence and both autogenous and reactive obsessions was very small. Once dysphoria was partialled out, the relationship between self-ambivalence and obsessionality still existed although the effect size was reduced, and there was no difference in the relationship between self-ambivalence and autogenous, and self-ambivalence and reactive obsessions. These findings confirm the first part of the hypothesis; i.e. that there will be a positive correlation between self-ambivalence and both autogenous and reactive obsession scores. However, these findings disconfirm the second part of the hypothesis: i.e. that the effect size for the correlation between self-ambivalence and autogenous obsessions will be greater than between self-ambivalence and reactive obsessions.
Regarding the research questions, the study found that prior to partialling out dysphoria, there was a positive relationship between actual/ought self-discrepancy and obsessionality, and between actual/ideal self-discrepancy and obsessionality, and a negative relationship between actual/feared self-discrepancy and obsessionality. Also, prior to partialling out dysphoria, there was a larger effect size correlation between actual/ought and autogenous compared with reactive obsessions, a larger effect size between actual/ideal and reactive compared with autogenous obsessions, and non-significant associations between actual/feared self-discrepancies and autogenous obsessions, and between actual/feared self-discrepancies and reactive obsessions. There was a slightly larger effect size for the correlation between total self-discrepancy scores and autogenous obsession compared with reactive obsessions. The largest effect sizes between self-discrepancy and obsessions was for actual/ought. However, once dysphoria had been partialled out, all of the relationships between self-discrepancies and obsessionality (and thus also both autogenous and reactive obsessions) reduced to almost zero. Limitations of the study, clinical implications, and future research recommendations can be found at the end of this discussion section, following a more detailed discussion of the findings.
The relationship between self-ambivalence and obsessionality
Prior to partialling out dysphoria, self-ambivalence was positively associated with obsessionality and these variables shared 22% of their variance. This finding seemed to give support to Bhar and Kyrios’ (2007) theory, given that it is based on clinical OCD, that the higher the obsessionality scores in a non-clinical sample, the greater the self-ambivalence. This proposal was also given support by George’s (2010) findings that self-ambivalence was significantly higher in an OCD group compared with both a general anxiety group and a non-clinical group. Interestingly, although taking account that the current study’s design uses only one sample group from a non-clinical population and is therefore not directly comparable to George’s study, the zero-order correlations in the current study indicated that self-ambivalence was positively correlated with obsessionality. After dysphoria was partialled out self-ambivalence and obsessions shared only 2% of their variance.
Therefore, the current study founda statistically significant relationship between self-ambivalence and obsessionality once partialling out the effects of dysphoria. As such, there is a unique relationship, between self-ambivalence and obsessions, thereby suggesting that self-ambivalence may have some part to play in obsessionality. However, the relationship was weak and, given the correlational nature of the data, causality cannot be inferred. It may be that the experience of obsessions results in self-ambivalence or vice versa.Therefore, although recognising that these results are based on a non-clinical sample, they do concur with Bhar and Kyrios (2007), who found that the relationship between self-ambivalence and obsessions was independent of self-esteem, depression and anxiety. However, Bhar and Kyrios’s sample included people with OCD and other anxieties. Therefore the different results in terms of strength of the relationship between self-ambivalence and obsessionality may be due to using a non-clinical sample in the current study.
Guidano and Liotti (1983) proposed, based on clinical cases, that individuals with OCD experience doubt related to whether they are intrinsically moral, lovable, and worthy, and see egodystonic intrusive thoughts as revealing evidence that their true self is immoral, unlovable and unworthy. Therefore, all the findings from the current study seem to support this theory, even though they suggest that this relationship is weak, but it is still important to recognise that these findings cannot simply be extrapolated from a non-clinical sample to clinical population. However, the finding that there is only a weak relationship between self-ambivalence and obsessions raises the question of whether the SAM (Bhar & Kyrios, 2007) is adequate as an index of self-ambivalence; i.e. that the SAM is not a robust enough measure, or does not operationalize self-ambivalence sufficiently well. Alternatively, it may not focus enough on the domains of self-ambivalence that are considered most sensitive to OCD according to Doron, Kyrios and Moulding (2007) and Doron, Moulding, Kyrios, and Nedeljkovic (2008), such as social acceptability, job/school competence, and morality. Although the SAM does have a morality subscale, it only has 6 items and may not measure moral self-ambivalence sufficiently.
Dysphoria was strongly correlated with both self-ambivalence and obsessions meaning that any assessment of the relationship between the two necessarily needed to be partial out the effect of dysphoria. This means that controlling for dysphoria gives a truer picture of the relationship between self-ambivalence and obsessionality in this non-clinical sample and, as such, there is little variance shared between these variables after controlling for dysphoria. Therefore, self-ambivalence does not explain much variance in obsessionality and instead the data reveals the nature of the relationship between self-ambivalence, obsessions, and dysphoria. However, it is not surprising that self-ambivalence would be associated with dysphoria, given Higgins’ (1987) assertion that contradictory self-states have consequences for positive mood has been supported in a number of studies (e.g. Strauman, 1989; Moretti & Higgins,1990; Scott & O’Hara, 1993; Higgins, Shah & Friedman,1997). Additionally, obsessionality is commonly comorbid with depression (Stein & Fineberg, 2007).
The relationship between self-ambivalence and autogenous obsessions and between self-ambivalence and reactive obsessions
The hypothesis that the correlation between self-ambivalence and autogenous obsessions would have a greater effect size than the correlation between self-ambivalence and reactive obsessions was not confirmed by the analysis. In fact, the shared variance was the same for both correlations at 16%.Therefore, there was no evidence of a difference between the strength of association been autogenous and reactive obsessions and self-ambivalence.This replicates Moulding et al.’s (2007) findings to some extent, because even though they did observe a greater effect size between self-ambivalence and autogenous compared to reactive obsessions, the difference was very small at r =.30 versus r =.25, respectively.The effect size for the two correlations differed by only .05 and, as such, Moulding et al.’s (2007) study does not provide compelling evidence for a difference in the relationship. In the current study,after partialling out dysphoria, variance shared between self-ambivalence and reactive obsessions, and between self-ambivalence and autogenous obsessions reduced to 1%, thus disconfirming the hypothesis that there is more of an association between self-ambivalence and autogenous obsessions than between self-ambivalence and reactive obsessions. Therefore,to the extent that SAM is a measure of sense of self and it is only one way of conceptualising sense of self, the findings of the current study disconfirm Moulding et al.’s (2007) hypothesis that autogenous obsessions have a more negative impact on the person’s sense of self, where sense of self is conceptualised within a self-ambivalence framework. Additionally, these findings fit with Frost et al.’s (2007) proposal that individuals with conflicting self-representations (e.g. that they are simultaneously moral and immoral) become preoccupied with seeking environmental evidence to establish the truth about themselves.
Dysphoria was strongly correlated with both autogenous and reactive obsessions, and with self-ambivalence, meaning that any assessment of the relationship between the two necessarily needed to be partial out of the effect of dysphoria. This replicates Moulding et al.’s (2007) finding that the relationship between and autogenous obsession and self-ambivalence, and between reactive obsessions and self-ambivalence, reduces when controlling for mood. Therefore, self-ambivalence does not explain much variance in both autogenous and reactive obsessions independent of dysphoria. This is perhaps not surprising because distress is a core part of OCD (Roth & Fonagy, 2005; Well, 1997) and is commonly comorbid with depression (Stein & Fineberg, 2007).However, what the current analysis indicates is that, in a non-clinical sample, self-ambivalence is associated strongly with negative affect.
How robust is the theoretical subtyping of obsessions into autogenous and reactive obsessions?
As noted, there was a strong association between dysphoria and obsessions. Further, it is also interesting to note that the differences in effect size between autogenous obsessions and dysphoria and between reactive obsessions and dysphoria is small. This suggests that the concept of subtyping obsessions in this way may not be valid or useful. Lee and Kwon’s (2003) assertion that one of the distinguishing characteristics between autogenous and reactive obsessions is that the former type are more ego-dystonic is not supported by the current analysis. This might be explained by considering Purdon et al.’s (2007) argument that the same thought can both be syntonic and dystonic in relation to valued self-perceptions; that although a thought may contradict an individual’s valued self-perceptions, it can still be consistent with his or her preferences and past experiences, thus not making it an alien thought. Further, that dystonic thoughts can become increasingly experienced as syntonic the more an individual experiences them and shifts his or her self-perceptions accordingly. With this in mind, the researcher suggests that perhaps it is overly simplistic to consider that autogenous obsessions are more ego-dystonic than reactive obsessions. This is important as it would help to explain why there was not a larger effect size between self-ambivalence and autogenous obsessions than between self-ambivalence and reactive obsessions as was predicted by the hypothesis based on the theory that autogenous obsessions are more ego-dystonic. Therefore, if ego-dystonicity is not a clear distinguishing feature between the obsessional subtypes, then the subdivision itself starts to become questionable. Additionally, it might be the case that some contents are more ego-dystonic than others but that does not imply then that self-ambivalence should be differentially related. Perhaps what is needed is for research to establish the association between with self-ambivalence and ego-dystonicity regardless of thought type.
The relationship between self-discrepancies and obsessionality
Prior to partialling out dysphoria, actual/ought self-discrepancy was positively associated with obsessionality (these variables shared 7% of their variance), actual/ideal self-discrepancy was positively associated with obsessionality (these variables shared 3% of their variance), total self-discrepancy was positively associated with obsessionality (these variables shared 1% of their variance), and actual/feared self-discrepancy was negatively associated with total obsessionality (i.e. the higher the obsessionality the less actual/fear self-discrepancies) and these variables shared 3% of their variance. As such then, all of these associations had small effect sizes.
Although the current study sampled a non-clinical population, findings from the zero-order Pearson correlations seemed to have face validity because in a similar way to how intrusive thoughts are theorised to interact with self-ambivalence (Guidano and Liotti, 1983), intrusive thoughts may activate/highlight the discrepancies between actual/ought self-discrepancies and/or actual/ideal self-discrepancies. Also, these finding appear to support Ferrier and Brewin’s (2005) hypothesis that small discrepancies between actual and feared self would result in an increase in obsessionality because they would be appraised as evidence of a ‘true’ revealed self. However, the data is correlational and so it may be that the experience of obsessions creates discrepancies in self-perception equally as a discrepant self-image may motivate obsessions.
After partialling out dysphoria, variance shared between all discrepancy variables and obsessionality reduced to zero. Therefore, the researcher argues that Frost et al.’s (2007) proposal that individuals with conflicting self-representations become preoccupied with seeking environmental evidence to establish the truth about themselves, also does not fit with findings from the current study regarding the relationship between all discrepancy variables and obsessionality, if it assumed that the conceptualisation of conflicting self-representations is extended to include the existence of self-discrepancies. This is because in the current study self-discrepancies were not associated with increased obsessionality.
The relationship between self-discrepancies and autogenous obsessions, and between self-discrepancies and reactive obsessions
Prior to partialling out dysphoria, actual/ought discrepancy was positively associated with both autogenous (these variables shared 7% of their variance) and reactive obsessions (these variables shared 4% of their variance. Actual/ideal self-discrepancy was positively associated with both autogenous (these variables shared 2% of their variance)and reactive obsessions (these variables shared 3% of the their variance). Also, total selves discrepancy was positively associated with both autogenous (these variables shared 2% of their variance) and reactive obsessions (these variables shared 1% of their variance). These initial results were not surprising because self-discrepancies might be predicted to relate to obsessionality in a similar way to how self-ambivalence and obsessionality might be related.
Prior to partialling out dysphoria, actual/feared self-discrepancy was negatively associated with both autogenous (2% shared variance) and reactive obsessions (1% shared their variance). These findings were notsurprising if it assumed that feared self traits are likely to be ego-dystonic, as are autogenous obsessions comparedwith reactive obsessions. A smaller discrepancy score means that individuals perceive their actual self is like their feared self. Theoretically, such a fear of being bad produces the occurrence of obsessions because the individual attends to negative intrusions more because these activate the possibility of being bad. But equally, it is possible that individuals who have lots of unpleasant obsessions are likely to come to feel that they are a bad person (i.e. that they are the person they fear).However, after partialling out dysphoria, variance shared between all discrepancy variables and both autogenous and reactive obsessions reduced to zero.
The relationship between self-discrepancies and depression, anxiety, and stress
A supplementary analysis, suggested by the results above, was conducted to test Higgins’ (1985) self-discrepancy theory that there would be a strong relationship between actual/ought self-discrepancy and anxiety, and between actual/ideal self-discrepancy and depression. The findings showed that there were significant positive associations with medium effect sizes between actual/ought self-discrepancy and stress and depression, and a significant positive association with a small effect size between actual/ought self-discrepancy and anxiety. There were significant positive associations with small effect sizes between actual/ideal self-discrepancy and both stress and anxiety, and positive association with a medium effect sizes between actual/ideal self-discrepancy and depression. There were significant negative associations with medium effect sizes between actual/feared self-discrepancy and stress, anxiety, and depression. There was a significant negative association with a small effect size between total self-discrepancy and stress, a non-significant positive association with a small size between total self-discrepancy and anxiety, and a significant positive association with a small effect size between self-discrepancy and depression. Therefore, these findings also confirm Higgins (1987) self-discrepancy theory that proposes that there would be a strong positive association between actual/ought self-discrepancy and anxiety. However, these findings indicate that the effect size between actual/ought self-discrepancy and stress is higher. Thismay because participants may be used to experiencing stress rather than anxiety in a non-clinical population.Additionally, what is interesting is that the effect size between actual/ought and depression is even larger, suggesting that Higgins’ (1987) self-discrepancy theory is not as predictive of what self-discrepancies have the strongest associations with depression, anxiety, and stress.The finding that there was as strong positive association between actual/ideal self-discrepancy and depression confirms Higgins’s (1987) self-discrepancy theory in terms of the relationship between these two variables. However, there were also strong positive associations with only slightly smaller effect sizes between actual/ideal self-discrepancy and anxiety, and between actual/ideal self-discrepancy and stress. Again, these findings confirm Higgins’s (1987) self-discrepancy theory, but additionally suggest that other types of dysphoria other than depression are almost equally as strongly associated with actual/ideal self-discrepancy.
The finding that actual/feared self-discrepancy is negatively associated with stress, anxiety, and depression is not surprising since the further away one perceives one’s actual self to be from one’s feared self, the lower the level of stress, anxiety, and depression.However, what is worthy of noting is that the differences between these effect sizes is very small, suggesting that actual/feared self-discrepancies have a similar strength of association with stress, anxiety, and depression.
Using a non-clinical population
The main disadvantages of using a non-clinical populationconcerns the question of the extent to which findings in a non-clinical sample can be extrapolated to clinical populations. For example, Purdon and Clark (1994) suggest that some intrusive thoughts are more common within clinical populations and thus clinical populations would have scored higher on obsessionality (ROII). Thus, in terms of content differences, non-clinical samples may not endorse such a range of responses on obsessionality measures as clinical samples. Indeed, the item means for the autogenous and reactive subscales suggest infrequent occurrence of these experiences. This will have implications for the effects sizes achieved for correlations. Further, clinical samples may endorse extreme content more frequently which might call into question whether non-clinical samples allow a consideration of the full range of obsessionality. Additionally, the problem with the idea that constructs, such as self-ambivalence, obsessionality, and dysphoria are on a continuum between non-clinical and clinical populations (as suggested in cognitive-behavioural models), is that it does not address the possibility that although the surface manifestations of obsessionality appear similar, the underlying processes may be different.
However, there are mixed findings regarding the validity of analogue research. For example, although Coyne’s (1994) findings cast doubt on the validity of this type of research, and suggest that there are distinct differences between depression in non-clinical populations compared with clinical populations, these findings are in contrast to research by Cox et al. (1999) who found that the nature of depressive experience is similar in both analogue and clinically depressed samples in terms of the covariance matrices of depressive symptoms.
A disadvantage of using a variety of recruitment pathways was that the researcher was unable to specify exactly how participants became aware about the survey (even though the demographic questionnaire asked for some information regarding this) and thus hypotheses about what drew participants to take part in the research (other than a potential financial incentive) and how this may have affected the results are not possible. However, a major advantage of using a variety of recruitment pathways was that the study was able to generate a large sample size that enabled the statistical analysis to find small effect sizes. This would have not been possible with a smaller sample.
Using an online survey
Although a possible disadvantage of using an online survey was an ethical concern that it might have been more difficult for participants to seek support than if they could speak to the researcher directly after completion of the survey, the researcher did supply several organisation’s contact details from where potentially distressed participants could seek support. An advantage of using an online survey was that the participants crossed demographic boundaries such as between students and nonstudents, and were not bound to a local region. However, it is possible that because the population sample in the study included participants from across different cultural groups, some of these groups may have interpreted some of the questions differently which were used to assess the constructs measured. However, even if some of the cultural groups did interpret some of the questions differently, the majority of the participants were from western cultures, thus limiting the effect of non-western perceptions of the constructs being studied. Using an online survey also meant that if participants wished, they could remain anonymous unless they wished to leave their email so as to take advantage of a potential financial incentive of taking part in a prize draw.
Demand characteristics, and other possible problems related to participants’ responses
A weakness of self-report questionnaires is that they are subject to demand characteristics – i.e. that participants wish to show their ‘better self’ rather than give an accurate reflection which they may perceive as less socially acceptable. Although this effect may be diminished for those participants that chose to remain anonymous, most participants chose to leave their email address so as to take part in the prize draw because of the financial incentive. However, the data suggests that participants were attempting to be truthful about themselves – this is particularly apparent in the self-discrepancy part of the questionnaire in which participants listed their actual, ought, ideal, and feared self. The demand characteristic effect may also have been reduced because the information sheet and questionnaire emphasise how normal it is to have intrusive thoughts.
Regarding reporting intrusive obsessions in the ROII, it is possible that non-clinical populations may feel more comfortable with thinking about intrusive thoughts than clinical populations because individuals with higher levels of self-ambivalence may be more worried about what others may think of them. However, this does not mean that non-clinical samples are not self-ambivalent, but instead that the experience of feeling comfortable with thinking about intrusive thoughts may be on a continuum related to level of self-ambivalence. Also, individuals with higher levels of obsessionality (i.e. moving towards the clinical end of the continuum) may have thought-action-fusion beliefs such as thinking that if they indicate what intrusive thoughts they have in the survey then these might come true. This belief would impede these individuals from wanting to indicate their level of experience of intrusive thoughts in the survey (i.e. they would indicate that they do not have intrusive thought which in fact worries them), thereby lowering their obsessionality score.
Order of measures within the survey
It is possible that by placing Carver et al.’s (1999) Selves questionnaire last in the survey, some participants may have been too tired to focus on the task which requires self-reflection. However, the reason for placing this questionnaire at the end of the survey was because it needed a lot of concentration and self-reflection, and therefore it was thought that some respondents may be less inclined to complete the survey if their first experience of the survey was to do to a particularly difficult and tiring task. However, the researcher noted that participants’ responses had face validity in that they fitted with the self-representations of actual, ought, ideal and feared self. Therefore, the order of the questionnaires within the survey did not seem to have an impact on the responses that participants gave.
Out of the 376 participants who provided useable data, 309 participants completed the part of the survey that included Carver et al.’s (1999) Selves’ Questionnaire. Out of this number, 254 participants completed 7 traits for each self-representation, but 55 gave less than 7 traits for at least one of the self-representations. This was accounted for by analysing the average score for each completed self-presentation. However, the fact that 67 participants chose not to start the Selves Questionnaire and a further 55 were unable to manage to give seven traits for each section suggests that the task of being asked to complete seven traits for each self-representation was experienced as too arduous, which therefore has implications for the Selves Questionnaire measure if used again. As there is no particular reason why Carver et al.’s (1999) Selves Questionnaire needs to specifically ask for seven traits, future use of this questionnaire could be adapted to ask for fewer traits.
Alternative methods of data analysis
The current study used zero-order and partial correlation analysis to explore the strength of associations between the differing constructs being measured (i.e. self-ambivalence, self-discrepancy, and autogenous and reactive obsessions). Arguably these were the appropriate statistical tests for the research hypotheses and questions. However, there are potential advantages of using alternative methods of data analysis. For example, whilst correlation analysis gives the strength of relationships between two variables but not the direction of causation, multiple regression analysis predicts which independent variables influence the dependent variable. Also, correlation analysis is limited to involving two variables at a time whereas multiple regression analysis can include all of the variables in a study (assuming the sample size is large enough).Therefore, had the study used multiple regression analysis, a model could have been generated that included self-ambivalence, self-discrepancy, and dysphoria as independent variables and obsessionality as a dependent variable. Running multiple regression analysis on this model would have benefited the study because it would have shown which of these independent variables is the most important predictor of autogenous and reactive obsessions.
Is self-discrepancy a problematic construct to measure?
It is possible that self-discrepancy could be critiqued as a difficult construct to study because it is questionable how stable are self-representations/ self-positions (i.e. they may change over the life course, as and when people enter into relationships, marriage/civil partnerships, have children, et cetera). However, the important question that has been explored in this study is not howpeople define themselves, but the sizeof their self-discrepancies. So, whilst individuals’ self-perceptions may change over time, what it measured here is the discrepancy not the content, and so this study does not assume that self-characterisation does not change over time.
Summary of Theoretical Implications
Within the literature there is a debate about whether OCD is a unitary or heterogeneous disorder. Theoretical contributions to understanding obsessions therefore can be made if specific factors are found to have a differential relationship with different obsessional subtypes. This study tested the prediction that self-ambivalence and self-discrepancy would be more strongly associated with autogenous than reactive obsessions, based on the convergence of both cognitive-behavioural and psychodynamic theories regarding that self-concept is an important factor across both models in OCD (Kempke & Luyten, 2007). The findings do not support this idea.Theoretically, the reason for this might be that sensitive domains of self-concept might predict frequency of particular types of obsessions. It is also possible that given self-ambivalence has been found across anxiety disorders and self-discrepancy relates to negative affect then this may be why we do not observe a differential relationship with obsession subtypes. Another reason for the current study’s finding is that the theoretical distinction between autogenous and reactive obsessions is weak. However, the current research findings have added to the psychological understanding of the strength of the relationship between conflicting self-representations and dysphoria. The importance of this finding will be discussed further within clinical implications.
The current study’s therapeutic focus on self-concept discrepancies is driven by the work of Guidano and Liotti (1983), and the convergence between cognitive-behavioural and psychodynamic models in relation to the importance of self-ambivalence in the development and maintenance of obsessions, as argued by Kempke and Luyten (2007).
From a cognitive-behavioural perspective, if compulsions are a way to resolve ambivalence and compulsions maintain obsessions, then targeting the problem of ambivalence might ameliorate obsessions. From the psychodynamic perspective, the therapeutic relevance is that core intrapychic conflicts trigger sensitivity to obsessional thoughts. Hence, a therapeutic rationale in both models for dealing with core ambivalence is that it might help to reduce obsessions. The current study suggests a small role for self-ambivalence but this does not necessarily undermine therapeutic potential as it is based on new measures (SAM) or previously untested model (self-discrepancy) and comes from a non-clinical sample that, generally, scored lower onobsessionality. Therefore, the findings of the current study suggest that clinical work with clients suffering from OCD and other psychopathologies that feature intrusive thoughts, such as post-traumatic stress disorder (PTSD), and/or negative moods (such as anxiety, depression, and stress), could benefit from including a focus on reducing self-ambivalence. However, because the direction of the causation between these variables is unclear, it could also be the case that focusing on lowering negative mood could be used to reduce obsessionality in clinical disorders such as OCD. If this is the case, then focusing on psychological treatments that support clients lifting their mood (e.g. CBT) could also benefit them in terms of experiencing a reduction in obsessional thinking.
An interesting finding of the study is the strong relationships between self-ambivalence, self-discrepancy and dysphoria. The clinical implications of finding that there is strong relationship between self-discrepancies and dysphoria may be that mood disorders could be alleviated by work on clients’ self-discrepancies. This could take the form of developing acceptance of the discrepancies between actual and ought self, between actual and ideal self, and between actual and feared self. Therefore, it might be possible that both Acceptance and Commitment Therapy (ACT; Hayes, Strosahl, and Wilson, 1999) and Compassion-Focused Therapy (CFT; Gilbert, 2010), which are third-wave CBT therapies, could be used to work with client’s acceptance of their self-discrepancies as way of working indirectly with reducing negative mood. Similarly, this therapeutic approach could be used to help decrease self-ambivalence, also with the aim of indirectly reducing negative mood.
The findings of the current study need to be replicated in a clinical population, such as participants who meet the diagnostic criteria for OCD as well as other disorders that manifest intrusive cognitions (e.g. PTSD). This is because although the current study used research evidence that suggests that the constructs of self ambivalence, self-discrepancy, obsessions, and dysphoria are dimensional, and thus can be researched on a non-clinical sample, there is still a debate regarding whether data from analogue samples can be extrapolated to clinical contexts. This is because non-clinical samples might produce a restricted range on obsessional measures. Therefore, replication of the current study would enable the hypotheses to be tested in clinical samples so as to see if conflicting self-representations has greater predictive power.
As aforementioned, contrary to the literature that suggests that useful distinctions can be made between autogenous and reactive obsessions, the current research suggest that this may not be useful concept as no significant differences were found between these two subtypes and their associations with both self-ambivalence and self-discrepancy. Building on these findings, it is important for future research to establish whether there is any association with self-ambivalence and ego-dystonicity regardless of thought type. To support this, more focus could be placed on developing differing ways of operationalizing self-concept discrepancies. Additionally, future research could also be extended to more conceptual work on different obsessional subtypes.
Abramowitz, J.S., Schwartz, S. A., and Moore, K.M. (2003). Obsessional thoughts in postpartum females and their partners: Content, severity and relationship with depression. Journal of Clinical Psychology in Medical Settings, 10, 157- 164.
Abramowitz, J., McKay, D. and Taylor, S. (2005).Subtypes of obsessive-compulsive disorder. Behavior Therapy, 36, 367-369.
Antony, M. M., Bie1ing, P. J., Cox, B. J., Enns, M. W., and Swinson, R. P. (1998). Psychometric properties of the 42-item and 21-item versions of the Depression Anxiety Stress Scales in clinical groups and a community sample. Psychological
Assessment, 10, 176-181.
Barrett, P. M. and Healy, L. J. (2003). An examination of the cognitive processes
involved in childhood obsessive-compulsive disorder. Behaviour Research and Therapy, 41, 285-299.
Bhar, S. (2004). Self-ambivalence on obsessive-compulsive disorder. Unpublished PhD thesis: University of Melbourne.
Bhar, S. and Kyrios, M. (2007). An investigation of self-ambivalence in obsessive-compulsive disorder. Behaviour Research and Therapy, 45, 1845-1857.
Bowlby, J. (1970). Attachment and Loss (Vol. 1 Attachment). London: The Hogarth Press.
Burns, G. L., Formea, G. M., Keortge, S. G., & Sternberger, L. G. (1995). The utilization of nonpatient samples in the study of obsessive-compulsive disorder. Behaviour Research and Therapy, 33, 133–144.
Carver, C. S., Lawrence, J. W., and Scheier, M. F. (1999). Self-discrepancies and affect: incorporating the role of feared selves.Personality and Social Psychology Bulletin, 25, 783–792.
Clark, D.A., Purdon, C., and Wang, A. (2003). The meta-cognitive beliefs questionnaire: Development of a measure of obsessional beliefs. Behaviour Research and Therapy, 41, 655-669.
Clark, D. A. (2004). Cognitive-behavioral therapy for OCD. New York: Guilford Press.
Clark, D. A. and Purdon, C. (1993). New perspectives for a cognitive theory of obsessions. Australian Psychologist,28(3), 161-167.
Cohen, J. (1992). Quantitative methods in psychology.The American Psychological Association, 112(1), 155-159.
Coles, M.E. and Schofield, C.A. (2008). Assessing the development of inflated responsibility beliefs: the pathways to inflated responsibility beliefs scale. Behavior Therapy, 39, 332- 335.
Cox, B.J., Enns, M.W., Borger, S.C. and Parker, J.D.A. (1999). The nature of the depressive experience in analogue and clinically depressed samples. Behaviour Research and Therapy, 37(1), 15-24.
Coyne, J. C. (1994). Self-reported distress: analog or ersatz depression? Psychological Bulletin, 116(1), 29-45.
Crawford, J.R. and Henry, J.D. (2003). The Depression Anxiety Stress Scales (DASS): normative data and latent structure in a large non-clinical sample. The British Journal of Clinical Psychology, 42, 111-131.
Cuttler, C. and Graf, P. (2009).Sub-clinical compulsive checkers show impaired performance on habitual, event- and time-cued episodic prospective memory tasks. Journal of Anxiety Disorders, 23(6), 813–823.
de Silva, P. and Rachman, S. (2004). Obsessive-compulsive disorder(3rd edn). Oxford: Oxford University Press.
Doron, G., and Kyrios, M. (2005). Obsessive compulsive disorder: a review of possible internal representations within a broader cognitive theory. Clinical Psychology Review, 25, 415–432.
Doron, G., Kyrios, M., and Moulding, R. (2007). Sensitive domains of self-concept in obsessive compulsive disorder (OCD): further evidence for a multidimensional model of OCD. Journal of Anxiety Disorders, 21, 433–444.
Doron, G., Moulding, R., Kyrios, M., and Nedeljkovic, M. (2008). Sensitivity of self-beliefs in Obsessive Compulsive Disorder (OCD). Depression and Anxiety, 25, 874–884.
Doron, G., Moulding, R., Kyrios, M., Nedeljkovic, M., & Mikulincer (2009). Adult attachment insecurities are related to obsessive compulsive phenomena. Journal of Social and Clinical Psychology, 28, 1025-1052.
Fairbairn, R. (1943) The repression and return of bad objects (with special reference to the ‘war neuroses’). In R. Fairbairn (Ed.) Psycho-Analytic Studies of the Personality (Rev. edn. pp. 59-81). London: Routledge.
Faull, M., Joseph, S., Meaden, A., and Lawrence, T. (2004). Obsessive beliefs and their relation to obsessive-compulsive symptoms. Clinical Psychology and Psychotherapy, 11, 158-167.
Ferrier, S., and Brewin, C. (2005). Feared identity and obsessive-compulsive disorder. Behaviour Research and Therapy, 43, 1363-1374.
Freud, S. (1824). The defence neuro-psychoses.InJ. Strachey (Ed.) The Standard Edition of the Complete Psychological Works of Sigmund Freud, 3, (Rev. edn., pp. 45-61). London: Vintage Press.
Freud, S. (1896). Further remarks on the neuro-psychoses of defence.In J. Strachey (Ed.)The Standard Edition of the Complete Psychological Works of Sigmund Freud, 3, (Rev. edn., pp. 162-189) London: Vintage Press.
Freud, S. (1907). Obsessive actions and religious practice.In J. Strachey (Ed.) The Standard Edition of the Complete Psychological Works of Sigmund Freud, 9, (Rev edn., pp. 115-128) London: Vintage Press.
Freud, S. (1917) Mourning and Melancholia,In J. Strachey (Ed.) The Standard Edition of the Complete Psychological Works of Sigmund Freud, 14, (Rev edn., pp. 243-252) London: Vintage Press.
Frost, R., Kyrios, M., McCarthy, K., and Mathews, Y. (2007). Self-ambivalence and attachment to possessions. Journal of Cognitive Psychotherapy: An International Quarterly , 21 (3), 232-242.
Garćia-Soriano, G., Belloch, A., Morillo, C. and Clark, D.A. (2011). Symptom dimensions in obsessive-compulsive disorder: from normal cognitive intrusions to clinical obsessions. Journal of Anxiety Disorders, 25, 474-482.
Ghisi, M., Chiri, L.R., Marchetti, I., Sanavio, E., Sica, C. (2010). In search of specificity: “Not just right experiences” and obsessive–compulsive symptoms in non-clinical and clinical Italian individuals.Journal of Anxiety Disorders, 24(8) 879–886.
George, C. (2010). An investigation of the relationship between perceived parental rearing style, mentalization and self-ambivalence in obsessive compulsive disorder. Unpublished Clinical Doctorate Thesis, University of Surrey.
Gilbert, P. (2010). Compassion focused therapy. London: Routledge.
Gomez, L. (1997). An introduction to object relations. London: Free Association Books.
Grisham, J.R., Brown, T.A., Liverant, G.I. and Campbell-Sills, L. (2005). The distinctiveness of compulsive hoarding from obsessive-compulsive disorder. Journal of Anxiety Disorders, 19(17), 767-779.
Guidano, V. F., and Liotti, G. (1983). Cognitive processes and emotional disorders. New York: Guilford.
Haslam, N., Williams, B.J., Kyrios, M., McKay, D. and Taylor, S. (2005). Subtyping obsessive-compulsive disorder: a taxometric analysis. Behavior Therapy, 36(4), 381-391.
Hayes, S.C., Strosahl, K.D. and Wilson, K.G. (1999). Acceptance and commitment therapy: an experimental approach to behavior change. New York: Guildford Press.
Henderson, G., & Pollard, A. (1988). Three types of obsessive compulsive disorder in a community sample. Journal of Clinical Psychology, 44(5), 747–752.
Higgins, E. T., Bond, R. N., Klein, R., and Strauman, T. (1986). Self-discrepancies and emotional vulnerability: How magnitude, accessibility, and type of discrepancy influence affect. Journal of Personality and Social Psychology, 51, 5–15.
Higgins, E. T. (1987). Self-discrepancy: a theory relating self and affect. Psychological Review, 94, 319–340.
Higgins, E. T. (1996). The “self-digest”: self-knowledge serving self-regulatory functions. Journal of Personality and Social Psychology, 71(6), 1062-1083.
Higgins, E. T., Shah, J. and Friedman, R. (1997). Emotional responses to goal attainment: Strength of regulatory focus as moderator. Journal of Personality and Social Psychology, 72, 515–525.
Howard, S. (2006).Psychodynamic Counselling in a Nutshell. London: Sage
Jakes, I. (1996). Theoretical approaches to obsessive-compulsive disorder. New York: Cambridge University Press.
Julien, D., O'Connor, K. P. and Aardema, F. (2007). Intrusive thoughts, obsessions, and appraisals in obsessive-compulsive disorder: a critical review. Clinical Psychology Review, 27(3), 366-83.
Kempke, S. and Luyten, P. (2007). Psychodynamic and cognitive-behavioural approaches of obsessive-compulsive disorder: Is it time to work through our ambivalence? Bulletin of the Menninger Clinic, 15(4) 291-311.
Lee, H. J. and Kwon, S. M. (2003). Two different types of obsession: autogenous obsessions and reactive obsessions. Behaviour Research and Therapy, 41, 11-29.
Lovibond, S. H., and Lovibond, P. F. (1995a). Manual for the Depression Anxiety Stress Scales. Sydney: Psychology Foundation.
Lovibond, S.H. and Lovibond, P.F. (1995b). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335-343.
Maruna, S. and Mann, R.E. (2006). A fundamental attribution error? Rethinking cognitive distortions. Legal and Criminal Psychology, 11, 155-177.
Mataix-Colis, D., Rosario-Campos, M. C., & Leckman, J. F. (2005). A multidimensional model of obsessive-compulsive disorder. American Journal of Psychiatry, 162(2), 228–238.
McKay, D., Abramowitz, J. S., Calamari, J. E., Kyrios, M., Radomsky, A., Sookman, D., Taylor, S., and Wilhelm, S. (2004). A critical evaluation of obsessive-compulsive disorder subtypes: Symptoms versus mechanisms. Clinical Psychology Review, 24, 283–313.
McWilliams, N. (1994). Psychoanalytic diagnosis: understanding personality structure in the clinical process. London: The Guildford Press.
Moretti, M., and Higgins, E. T. (1990). Relating self-discrepancy to self-esteem: The contribution of discrepancy beyond actual-self ratings. Journal of Experimental and Social Psychology, 26, 108–123.
Moulding, R., Kyrios, M., Doron, G. and Nedeljkovic., M. (2007). Autogenous and reactive obsessions: Further evidence for a two-factor model of obsessions. Journal of Anxiety Disorders, 21(5) 677-690.
Murris, P. Merckelbach, H. and Clavan, M. (1997). Abnormal and normal obsessions. Behaviour Research and Therapy, 35(3), 249-252.
Obsessive-Compulsive Cognitions Working Group. (1997). Cognitive assessment of obsessive-compulsive disorder. Behaviour Research and Therapy, 35(7), 667–681.
Obsessive-Compulsive Cognitions Working Group. (2001). Development and initial validation of the obsessive beliefs questionnaire and the interpretation of intrusions inventory. Behaviour Research and Therapy, 39, 987–1006.
O’Kearney, R. (1998). Responsibility appraisals and obsessive-compulsive disorder. A critique of Salkovskis’s cognitive theory. Australian Journal of Psychology, 50, 43-47.
Oglivie, D.M. (1987). The undesired self: a neglected variable in personality research. Journal of Personality and Social Psychology, 52, 379-385.
Page, A.C., Hooke, G.R. and Morrison, D.L. (2007). Psychometric properties of the Depression Anxiety Stress Scales (DASS) in depressed clinical samples. British Journal Clinical Psychology, 46, 283-297.
Petusa, A., Fullana, M.A., Singh, S., Alonso, P., Menchon, J.M. and Mataix-Cols, D. (2008). Compulsive hoarding: OCD symptoms, distinct syndrome, or both? American Journal of Psychiatry, 165, 1289-1298.
Purdon, C. and Clark, D. A. (1994). Obsessive intrusive thoughts in nonclinical subjects II: cognitive appraisal, emotional response and thought control strategies. Behaviour Research and Therapy, 32, 403–410.
Purdon, C., Cripps, E., Faull, M., Joseph, S. and Rowa, K. (2007). Development of a measure of egodystonicity. Journal of Cognitive Psychotherapy, 21(3), 198-216.
Rachman, S. and de Silva, P. (1978). Abnormal and normal obsessions. Behaviour Research and Therapy, 31, 1449-1454.
Rachman, S., and Hodgson, R. (1980). Obsessions and compulsions. Englewood Cliffs: Prentice-Hall, Inc.
Rachman, S. (1993). Obsessions, responsibility and guilt. Behavior Research and Therapy, 31, 149–154.
Rachman, S. (1997). A cognitive theory of obsessions.Behavior Research and Therapy, 35, 793–802.
Rachman, S. (1998). A cognitive theory of obsessions: elaborations. Behavior Research and Therapy, 36, 385–401.
Rachman, S. and Shafran, R. (1999). Cognitive distortions: thought-action fusion. Clinical Psychology and Psychotherapy, 6, 80-85.
Rachman, S. (2002). A cognitive theory of compulsive checking. Behavior Research and Therapy, 40, 625–639.
Rachman, S. (2004). Fear of contamination. Behavior Research and Therapy, 42, 1227–1255.
Rowa, K., and Purdon, C. (2003). Why are certain intrusive thoughts more upsetting than others? Behavioural and Cognitive Psychotherapy, 31, 1-11.
Rowa, K., Purdon, C., Summerfeldt, L., and Antony, M. (2005). Why are some obsessions more upsetting than others? Behaviour Research and Therapy, 43, 1453-1465.
Roth, A. & Fonagy, P. (Eds) (2005). What works for whom? A critical review of psychotherapy research (2ndedn). The Guilford Press: New York, London.
Ruscio, A. M., Stein, D. J., Chiu, W. T., & Kessler, R. C. (2008). The epidemiology of obsessive–compulsive disorder in the national comorbidity survey replication.
Molecular Psychiatry 15(1), 53-63.
Salkovskis, P.M. and Freeston, M.H. (2001). Obsessions, compulsions, motivation and responsibility for harm. Australian Journal of Psychology, 53, 1-6.
Salkovskis, P.M. and Harrison, J. (1984). Abnormal and normal obsessions – a replication. Behaviour Research and Therapy, 22, 1-4.
Salkovskis, P. M. (1985). Obsessional-compulsive problems: a cognitive-behavioral analysis. Behavior Research and Therapy, 23, 571–583.
Salkovskis, P. M., Forrester, E., and Richards, C. (1998). Cognitive-behavioral approach to understanding obsessional thinking. British Journal of Psychiatry, 173, 53–63.
Salkovskis, P.M. (1989). Cognitive-behavioural factors and the persistence of intrusive thoughts in obsessional problems. Behaviour Research and Therapy, 27, 677-682.
Salkovskis, P. M. and Harrison, J. (1994). Abnormal and normal obsessions: a replication. Behaviour Research and Therapy, 22, 549-552.
Salkovskis, P.M., Shafran, R., Rachman, S., and Freeston, M.H. (1999). Multiple pathways to inflated responsibility beliefs in obsessional problems: Possible origins and implications for therapy and research. Behaviour Research and Therapy, 37, 1055-1072.
Salkovskis, P.M. and McGuire, J. (2003). Cognitive-behavioural theory of OCD. In R. Menzies & P. de Silva (Eds.) Obsessive compulsive disorder: Theory, research and treatment(pp. 39-58). Chichester, England: Wiley.
Scott, L. and O’Hara, M.W (1993). Self-discrepancies in clinically anxious and depressed university students. Journal of Abnormal Psychology 102(2), 282-287.
Shepperd, J., Malone, W. and Sweeny, K. (2008). Exploring causes of the self-serving bias. Social and Personality Compass, 2(2), 895-908.
Sherman, D.K., & Cohen, G.L. (2002). Accepting threatening information: Self affirmation and the reduction of defensive biases. Current Directions inPsychological Science, 11(4), 119-123.
Smith, G. (2010). An exploratory study looking at the way people with ObsessiveCompulsive Disorder conceptualise self and significant others. Unpublished Clinical Doctorate Thesis, University of Surrey.
StataCorp. (2009). Stata Statistical Software: Release 11. College Station, TX: StataCorp LP.
Stein, D.J. and Fineberg, N.A. ( 2007). Obsessive-compulsive disorder. Oxford: Oxford University Press.
Strauman, T. J. (1989). Self-discrepancies in clinical depression and social phobia: Cognitive structures that underlie emotional disorders? Journal of Abnormal Psychology, 98, 14-22.
Taylor, S. Afifi, T.O. Stein, M.B., Asmundson, G.J.G. and Jang, K.L. (2010). Etiology of obsessive beliefs: a behavioral-genetic analysis. Journal of Psychotherapy, 24(3), 177-186.
Veale, D., and Wilson, R. (2005). Overcoming obsessive compulsive disorder: a self-help guide using cognitive behavioural techniques. London: Constable & Robinson Ltd.
Wells, A. (2000). Emotional disorders and metacognition: innovative cognitive therapy. Chichester: Wiley.
Wells, A. (1997). Cognitive therapy of anxiety disorders.John Wiley & Sons.
Wroe, A. and Salkovskis, P. M. (2000). Causing harm and allowing harm: A study of beliefs in obsessional problems. Behaviour Research and Therapy, 38, 1141-1162.
Wu, K.D. and Watson, D. (2003). Further investigation of the obsessive –compulsive inventory: Psychometric analysis in two non-clinical samples. Journal of Anxiety Disorders, 17, 305-319.
Please note: the appendices have been removed since the research was submitted, so as to ensure its format remains appropriate for its internet readership, and so that it does not to breach inventory copyrights.
According to Cohen (1992), r=.10 (small effect size), the effect size accounts for 1% of the total variance; r=.30 (medium effect size), the effect size accounts for 9% of the total variance; and r=.50 (large effect size), the effect accounts for 25% of the variance.
The researcher has used the American Psychological Society’s symbol for partial correlation, pr, because the British Psychology Society does not have a symbol for this.