Inferential confusion, obsessive beliefs, and obsessive-compulsive symptoms: A replication and extension

https://doi.org/10.1016/j.janxdis.2009.02.017Get rights and content

Abstract

This study replicated and extended previous research regarding utility of an inference-based approach (IBA) to the study of Obsessive-Compulsive Disorder (OCD). The IBA is a model for the development of OCD symptoms through false reasoning. One of its key features is inferential confusion—a form of processing information in which an individual accepts a remote possibility based only on subjective evidence. In a nonclinical sample, this study examined the specificity of relations between the expanded Inferential Confusion Questionnaire (ICQ-EV) and OC symptoms. Results were that the ICQ-EV significantly predicted OC symptoms after controlling for general distress, anxiety, and depression. This finding supports the unique association between inferential confusion and OCD. Further, the ICQ-EV was a stronger predictor of certain OC symptoms than scales from the Obsessive Beliefs Questionnaire, which itself has shown strong relations with OC symptoms. Thus, both inference-based and cognitive appraisal models appear useful for understanding OCD.

Introduction

Cognitive models of Obsessive-Compulsive Disorder (OCD) have garnered substantial attention during the past 25 years. These models derive from Beck's (1976) cognitive specificity hypothesis in which different psychological disorders are characterized by specific dysfunctional beliefs. Salkovskis, 1985, Salkovskis, 1989 and Rachman (1997) provided early conceptualizations of OCD as stemming from an individual's appraisal of intrusive thoughts rather than from the intrusive thoughts themselves. In this way, intrusive thoughts—which are experienced nearly universally across clinical and nonclinical samples alike (e.g., Burns, Formea, Keortge, & Sternberger, 1995; Rachman & deSilva, 1978)—are stimuli for subsequent negative automatic thoughts. By holding certain beliefs, typically those involving responsibility for feared harm to oneself or others, one places weight on these otherwise benign cognitive experiences which in turn increases susceptibility to experiencing distress. The obsessive-compulsive cognitions working group (OCCWG, 1997), in particular, has focused on identifying and measuring these beliefs. An example is the Obsessive Beliefs Questionnaire (OBQ; OCCWG, 2005) which covers three domains: responsibility/threat estimation, perfectionism/certainty, and importance/control of thoughts. These beliefs correlate moderately to strongly with OCD symptoms, even after controlling for depression and general anxiety (OCCWG, 2005) and have been shown to predict OCD symptoms prospectively (Abramowitz, Khandker, Nelson, Deacon, & Rygwall, 2006). It has been suggested that some of the OBQ beliefs may not be specific to OCD (Tolin, Woods, & Abramowitz, 2003) and research continues to refine the content and measurement of belief domains, but overall this work has the potential to be highly useful for understanding OCD symptom development and for informing cognitive intervention.

Another line of research has targeted a prior step in the process of OCD symptom development. That is, whereas contemporary cognitive models focus at the point of feared negative consequences of potential events (e.g., if the stove is left on, people may be hurt and it will be my fault), also of interest is how people come initially to accept the premise of the intrusive thought (e.g., why they question whether the stove is on after checking that it is not). One possibility is that people with OCD are impaired with respect to certain kinds of memory. Studies have provided mixed conclusions, however (see Muller & Roberts, 2005), with some finding that individuals with OCD perform poorly on one or more forms of memory (Savage et al., 2000; Woods, Vevea, Chambless, & Bayen, 2002), others that only individuals with certain OCD symptoms show impaired memory (Sher, Frost, & Otto, 1983), and yet others that find no impairment for OCD versus other groups (Tolin et al., 2001). One possibility for the inconsistent findings relates to differences in comparison groups used: Against nonclinical controls, differences are more likely to be found than when non-OCD clinical participants are involved. Thus, general distress or other nonspecific psychopathology variables rather than OCD-specific variables may account for group differences. Conversely, a more consistent finding has been that individuals with OCD tend to show lower confidence in their memory, even when they perform as well as or better than comparison groups (Foa, Amir, Gershuny, Molnar, & Kozak, 1997; McNally and Kohlbeck, 1993, Sher et al., 1983, Tolin et al., 2001, Woods et al., 2002). The basis for this reduced confidence, however, is not clear, but one possibility offered by Constans, Foa, Franklin, and Mathews (1995) is that individuals with OCD experience a disparity between their actual and preferred memory vividness. That is, in order to feel confident about their memory, individuals with OCD require a higher degree of vividness than they—or other people—actually experience, and this may foster pathological doubt.

A second approach to understanding the transition from relatively benign thoughts to obsessions is provided by an inference-based approach (IBA; O’Connor, Aardema, & Pelissier, 2005). According to the IBA, a person with OCD does not fail to correctly perceive or sense reality, but rather overestimates the likelihood of a given possibility based on purely subjective data. Aardema, O’Connor, and Emmelkamp (2006) describe the IBA as:

“a characteristic reasoning process associated with the occurrence of obsessions and, as such, is more concerned with the form and context of the obsession rather than its (ab)normal content [and] holds that the obsessional doubt finds its justification in a wide variety of idiosyncratic narratives that contain inductive reasoning processes peculiar to OCD. As such, the “intrusions” in OCD… inherit their persistence and reality value from reasoning processes associated with their occurrence rather than being the result of a specific belief” (p. 139)

A key feature of such processing is the phenomenon of inverse inference in which an individual accepts a remote, subjective possibility either in the absence of supportive evidence or even despite the presence of contradictory evidence. For example, instead of reaching a typical inference through reality-based information (e.g., my hands are discolored so they must be dirty), inverse inference relies on subjective information to arrive at a feared conclusion (e.g., I was outside where dangerous germs are so my hands must be dirty). One problem with such reasoning is that in the absence of objective criteria (e.g., visible stains), how does the individual know when his or her hands are clean and therefore when to stop washing (e.g., how can I be certain my hands are clean enough?). The answer to this question also is subjective and may contribute to one's search for an elusive feeling of completion and eventually compulsions of increasing complexity. This description is consistent with clinical observation that it is extremely difficult to “convince” individuals with OCD to cease their irrational behavior using only psychoeducation or presentation of objective evidence.

To assess for presence/strength of such distorted inference, Aardema, O’Connor, Emmelkamp, Marchand, and Todorov (2005) developed the Inferential Confusion Questionnaire (ICQ). The original ICQ contained 15 items such as I am sometimes more convinced by what might be there than by what I actually see, Just the thought that there could be danger is proof enough for me that there is, and I often react to a scenario that might happen as if it is actually happening. The ICQ distinguished an OCD group from an anxious control group and showed strong relations to delusional symptoms (Aardema et al., 2005). Further, the ICQ showed unique associations with OCD symptoms after controlling for OBQ scores. Conceptually, the IBA is not incompatible with contemporary appraisal models, rather these two models appear to target different stages within the obsessional sequence (Clark & O’Connor, 2004). Therefore, research need not focus on whether one versus the other domain should be retained or rejected, but instead whether each has unique explanatory value for the experience of OCD symptomatology.

This study will replicate but then extend previous findings in four ways. First, the IBA has been supported via associations with OCD symptoms beyond the variance contributed by the OBQ. However, the original ICQ mainly contained items that reflect inverse inference and a distrust of the senses that lead to an invalid doubting inference (e.g., Even if I have all sorts of visible evidence against the existence of a certain danger, I still feel it will occur). The expanded version (ICQ-EV; Aardema, Wu, Careau, O’Connor, & Dennie, submitted for publication) includes additional reasoning processes hypothesized to give rise to inferential confusion, including (a) over-reliance on possibility during reasoning (e.g., Sometimes every far-fetched possibility my mind comes up with feels real to me); (b) absorption into imaginary sequences at the expense of reality (e.g., My imagination is sometimes so strong that I feel stuck and unable to see things differently); (c) irrelevant associations (e.g., I often connect ideas or events in my mind that would seem far-fetched to others or even to myself); and (d) category errors (e.g., I often confuse different events as if they were the same). The ICQ-EV replaces the original version and therefore it is necessary to examine if it performs similarly to its predecessor in terms of its relevance to OCD. Second, whereas previous research has used only one marker of OC symptoms, this study will use three measures that have shown clear convergent and discriminant correlations (see Section 2). In particular, the three measures have psychometrically strong scales assessing the core domains of checking, washing, and rituals pertaining to symmetry, ordering, and grooming (Wu & Carter, 2008b). Use of multiple symptom scales provides broader coverage of each domain and also reduces the likelihood of spurious findings resulting from idiosyncrasies of any one instrument. Third, whereas Aardema et al. (2006) examined ICQ-OCD relations after accounting for general anxiety, a recognized limitation was that it did not control for depressed mood. There is a well-known pattern of comorbidity between OCD and depression (Denys, Tenney, van Megen, de Geus, & Westenberg, 2004; Rasmussen & Eisen, 1992)—in fact, they are so consistently correlated that Richter, Cox, and Direnfeld (1994) advised that depression routinely should be assessed within OCD studies. Aardema, Radomsky, O’Connor, and Julien (2008) recently addressed this issue by replicating the Aardema et al. (2006) results after controlling for the Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979). However, despite its common use and utility for certain applications, the BDI has been shown to contain a substantial general distress component and not simply depression (Watson, Weber, et al., 1995). This study will use a depression-specific measure as a covariate in order to examine specific associations between inferential confusion and OC symptoms. Fourth, whereas Aardema et al. (2006) included 85 individuals with OCD and Aardema et al. (2008) reported data for 130 English-speaking students, this study will use a much larger sample to allow for examination of specific patterns of correlations among target constructs.

Section snippets

Participants

Participants were 317 undergraduate psychology students at Northern Illinois University who participated in the study as partial fulfillment of a research exposure requirement. Sample characteristics included 57% women and mean age of 19.1 years (S.D. = 1.30; range = 18–27). Whereas racial identification was not assessed, previous studies using the broader undergraduate body from which this sample was taken shows substantial racial diversity (approximately 65% White, 19% Black, 7% Asian, 5%

Descriptive data and internal consistency

Means and standard deviations are reported in Table 1. Overall, the current group scored similarly to previous groups on all measures.2

Discussion

This study replicated Aardema et al. (2006) and Aardema et al. (2008) in finding that inferential confusion has unique relations with OC symptoms. Strengths of the current design included use of an expanded item pool to assess inferential confusion (i.e., the ICQ-EV), administration of multiple OC symptom measures, controlling for nonspecific variance owing to anxiety and depression, and a large sample. Results supported that the ICQ-EV is a predictor of OC symptoms even after accounting for

References (44)

  • S. Rachman

    A cognitive theory of obsessions

    Behaviour Research and Therapy

    (1997)
  • S. Rachman et al.

    Abnormal and normal obsessions

    Behavior Research and Therapy

    (1978)
  • S.A. Rasmussen et al.

    The epidemiology and clinical features of obsessive compulsive disorder

    Psychiatric Clinics of North America

    (1992)
  • M.A. Richter et al.

    A comparison of three assessment instruments for obsessive-compulsive symptoms

    Journal of Behavior Therapy and Experimental Psychiatry

    (1994)
  • P.M. Salkovskis

    Obsessional-compulsive problems: a cognitive-behavioural analysis

    Behaviour Research and Therapy

    (1985)
  • P.M. Salkovskis

    Cognitive-behavioural factors and the persistence of intrusive thoughts in obsessional problems

    Behaviour Research and Therapy

    (1989)
  • K.J. Sher et al.

    Cognitive deficits in compulsive checkers: an exploratory study

    Behaviour Research and Therapy

    (1983)
  • L.J. Summerfeldt et al.

    Symptom structure in Obsessive-Compulsive Disorder: a confirmatory factor-analytic study

    Behaviour Research and Therapy

    (1999)
  • D.F. Tolin et al.

    Memory and memory confidence in obsessive-compulsive disorder

    Behaviour Research and Therapy

    (2001)
  • K.D. Wu et al.

    Further investigation of the obsessive beliefs questionnaire: factor structure and specificity of relations with OCD symptoms

    Journal of Anxiety Disorders

    (2008)
  • K.D. Wu et al.

    Further investigation of the Obsessive-Compulsive Inventory: psychometric analysis in two nonclinical samples

    Journal of Anxiety Disorders

    (2003)
  • K.D. Wu et al.

    Relations between hoarding and obsessive-compulsive disorder

    Behaviour Research and Therapy

    (2005)
  • Cited by (32)

    • Dr. Jack Rachman's contributions to our understanding and treatment of obsessive-compulsive disorder

      2023, Journal of Behavior Therapy and Experimental Psychiatry
      Citation Excerpt :

      O’Connor and Robillard (1995) argued that OCD is a state of mind characterized by inferential confusion, in which people confuse imagined possibilities with actual probabilities (e.g., “Another table like this was dirty, therefore this table is dirty until I can prove with certainty it is not”). This model has empirical support (e.g., Aadrdema, O’Connor, & Emmelkamp, 2006; Wu, Aardema, & O'Connor, 2009). Szechtman and Woody (2004) viewed OCD as a disturbance of security motivation.

    • Inverse reasoning processes in obsessive-compulsive disorder: Replication in a clinical sample

      2019, Journal of Anxiety Disorders
      Citation Excerpt :

      Higher scores on the ICQ-EV indicate a greater degree of reliance on these reasoning processes and consequently an increased tendency to confuse imagined possibilities with reality. There is general support for the IBA, with researchers across multiple studies having reported positive and significant associations between the ICQ-EV and OCD symptoms, even when controlling for negative mood states and the maladaptive belief domains proposed by the cognitive appraisal model (Aardema et al., 2005, 2013; Aardema, Radomsky, O’Connor, & Julien, 2008; Aardema & Wu, 2011; Aardema et al., 2010; Paradisis, Aardema, & Wu, 2015; Wong & Grisham, 2017b; Wu, Aardema, & O’Connor, 2009). Most recently, in a large clinical sample of individuals diagnosed with OCD, Aardema, Wu, Moulding, Jean-Sebastien, and Louis-Philippe (2018) found that the ICQ-EV uniquely predicted OCD symptoms related to repugnant obsessions, checking, and just right experiences.

    • The relationship of inferential confusion and obsessive beliefs with specific obsessive-compulsive symptoms

      2018, Journal of Obsessive-Compulsive and Related Disorders
      Citation Excerpt :

      Yet, experimental studies generally have supported a link between belief domains and OC symptoms (e.g., Lopatka & Rachman, 1995; Moulding, Kyrios, & Doron, 2007; Moulding, Doron, Kyrios, & Nedeljkovic, 2008; Rowa, Purdon, Summerfeldt, & Antony, 2005; Tolin, Abramowitz, Hamlin, Foa, & Synodi, 2002), and several researchers have noted that the relationship between cognitive beliefs and OC symptoms is complex, with some beliefs relevant only to certain presentations of OCD (e.g., Del Borrello & O’Connor, 2014; Julien, O’Connor, Aardema, & Todorov, 2006; Taylor et al., 2006). Indeed, most studies have supported the notion that specific OC domains are associated with specific beliefs, with relatively consistent findings supporting a specific link between the importance of and control of thoughts and repugnant obsessions and other unwanted thoughts (Tolin, Brady, & Hannon, 2008; Tolin, Woods, & Abramowitz, 2003; Wheaton et al., 2010); between perfectionism and certainty and checking, symmetry, and ritualizing (Summerfeldt, 2008; Wheaton et al., 2010; Wu et al., 2009); and between overestimation of threat and/or responsibility and contamination (Julien et al., 2006; OCCWG, 2005; Tolin, Brady, & Hannan, 2008; Wheaton et al., 2010; Wu et al., 2009). The OBQ remains the most frequently used instrument to investigate cognitions in OCD, as research continues to refine and expand the measurement of cognitive domains and processes relevant to specific presentations of OCD (e.g., Doron, Derby, Szepsenwol, Nahaloni, & Moulding, 2016; Melli, Gelli, Moulding, Stopani, & Pinto, 2018; Radomsky et al., 2014a, 2014b).

    View all citing articles on Scopus
    View full text