Aggressiveness in patients with obsessive-compulsive disorder as assessed by the Implicit Relational Assessment Procedure

Highlights

• Aggressiveness can be assessed using the Implicit Relational Assessment Procedure.

• Patients with OCD appear more ambivalent about their aggressiveness than controls.

• Patients with checking compulsions are not especially affected by aggressiveness.

Abstract

Objective

Psychodynamic and cognitive theories postulate a prominent role of aggressiveness in patients with obsessive-compulsive disorder (OCD). Explicit assessment of aggressiveness in OCD has yielded diverging results. The present study aimed to investigate aggressiveness in OCD using the Implicit Relational Assessment Procedure (IRAP).

Method

Patients with OCD (n = 59) were compared to non-clinical controls (NCs; n = 31) on an IRAP using self-referential statements and the explicit State-Trait Anger Expression Inventory-II (STAXI-II). During the computer-based IRAP, participants were required to respond as quickly and accurately (“correct” or “incorrect”) to the relation of two presented stimuli (e.g., “I am” + “aggressive”).

Results

D_IRAP-Scores for the I am aggressive trial type were significantly higher in NCs compared to the OCD sample (d = 0.73). Patients with OCD scored significantly higher on the Trait Anger scales and the Anger Expression-Out scale of the STAXI-II. The I am aggressive D_IRAP-Score correlated with the overall Trait Anger scale (r = −.33, p = .001) and with the Anger Expression-In scale (r = −0.31, p = .003).

Conclusions

Patients with OCD were more ambivalent about their own aggressiveness than NCs. These findings were in line with patients’ explicit aggressiveness.

Introduction

Obsessive-compulsive disorder (OCD) puts considerable strain on affected individuals’ daily lives (Moritz et al., 2005; Ruscio, Stein, Chiu, & Kessler, 2010). A recent meta-analysis showed that patients with OCD have a severely impaired quality of life in the domains of work, social life, emotions and family (Coluccia et al., 2016). A likely contributor thereto are aggressive feelings, which are increased in patients with OCD compared to the general population (e.g., Moritz, Kempke, Luyten, Randjbar, & Jelinek, 2011; Radomsky, Ashbaugh, & Gelfand, 2007).

Two theories underlie much of the research conducted on aggressiveness in patients with OCD. For one, Freud's psychodynamic theory (1976) construes the cause of OCD as the oedipal conflict between ambivalent sexual and aggressive impulses towards a subject's parents. According to the theory, these aggressive impulses are unconscious, yet they clash with the subject's hypermoral superego. As a dysfunctional coping mechanism for aggressive impulses, which may in turn surface as sexual or hostile obsessions, patients may develop compulsions characterized by perfectionism and conscientiousness (Fenichel, 1945; Kempke & Luyten, 2007). In current research, Freud's conceptualization of aggression is often called ‘latent aggression’, highlighting its inaccessibility to the conscious mind (Moritz et al., 2011).

In contrast to Freud, Rachman (1993) identified cognitive factors at the root of OCD. Particularly, an inflated sense of responsibility (Mitchell, Hanna, & Dyer, 2019; Obsessive Compulsive Cognitions Working Group, 1997) may cause patients with OCD to hold themselves accountable for preventing the manifestation of their obsessive thoughts. For example, a person who obsessively fears a fire in their apartment building might feel overly responsible for this eventuality, causing them to compulsively check their own electrical appliances. According to Rachman (1993), the ways in which this pathological mechanism may lead to aggressiveness are twofold: Firstly, when a person with OCD becomes angry, an inflated sense of responsibility for the cause of that anger may cause them to direct their anger inwardly, (i.e., inward rather than outward aggressiveness). Secondly, the attempt to decrease anxiety resulting from obsessions by performing compulsions is typically futile, leading patients with OCD to become frustrated and angry¹. Rachman supposes this pattern is particularly pronounced in persons with checking compulsions, as they are often confronted with unchanged feelings of insecurity despite prolonged efforts. In summary, both theories complement each other: While Freud (1976) sees unconscious aggressiveness as a cause and promoter of OCD, Rachman (1993) emphasizes how obsessive compulsive (OC) cognitions may trigger aggressive feelings.

The body of research on aggressiveness in OCD is heterogeneous, applying numerous questionnaires and yielding diverging results. Moritz et al. (2011) found elevated levels of latent aggression in patients with OCD and indications that these were related to OC symptom severity. Similarly, hidden aggression, construed as a composite score of the State-Trait Anger Expression Inventory-II's (STAXI-II; Spielberger, 1999) subscales Trait Anger and Anger Expression-In, was correlated with OC symptomology in a population of outpatients (Moosavi, Naziri, & Mohammadi, 2014). Moreover, Krug et al. (2009) found that inpatients diagnosed with OCD scored higher on Trait Anger and overall Anger Expression than non-clinical controls (NCs). Additionally, a study by Cludius, Mannsfeld, Schmidt, and Jelinek (2020) showed higher STAXI-II Trait anger, Anger Expression-In and Anger Expression-Out scores in patients with OCD compared to NCs (Cludius et al., 2020). Two studies showed that OC symptoms were associated with anger and aggressiveness even in a community sample (Liu, Liu, & Zhao, 2017; Whiteside & Abramowitz, 2004). Whiteside and Abramowitz (2004), however, found this relationship disappeared after controlling for depressive symptoms. Several other studies observed a confounding influence of depression. When compared to controls, students formally diagnosed with OCD were found to have elevated levels of aggressiveness which were covaried by general distress (Whiteside & Abramowitz, 2005). Elevated levels of anger (Moscovitch, McCabe, Antony, Rocca, & Swinson, 2008) and anger rumination (Jessup, Knowles, Berg, & Olatunji, 2018) in patients with OCD disappeared after controlling for depression and anxiety, respectively. Conversely, two studies found the correlation between OC symptomology and anger/aggressiveness prevailed even after controlling for depression and anxiety (Liu et al., 2017; Tallis, Rosen, & Shafran, 1996).

While Rachman (1993) suggests that aggressiveness may be particularly pronounced in patients with checking compulsions, thus far the evidence is limited. A sample of patients with checking compulsions showed higher levels of anger and aggressiveness than NCs (Radomsky et al., 2007), however no comparison was drawn to a non-checking OCD sample, limiting the conclusiveness of these findings. In another study, checking symptoms in a student population were moderately correlated with aggressiveness and internalization of anger, yet so were washing and doubting (Whiteside & Abramowitz, 2004). Finally, two studies found that neither anger rumination nor aggressiveness were related to checking (Jessup et al., 2018; Whiteside & Abramowitz, 2005).

So far, research on aggressiveness in OCD has mainly relied on self-report questionnaires. However, sensitive target concepts, such as aggressiveness, are susceptible to social desirability bias or imprecise reporting (Barnes-Holmes et al., 2006; Greenwald, Farnham, Greenwald, McGhee, & Schwartz, 2000). Implicit measures can circumvent such barriers by capturing more automatic components of individual attitudes (Barnes-Holmes, Barnes-Holmes, Stewart, & Boles, 2010; Greenwald & Banaji, 1995). Notably, implicit measures were formerly construed as gateways into unconscious thought (Greenwald & Banaji, 1995), yet there is mounting evidence to the contrary (Gawronski, 2009). They do indeed produce responses that are more automatic, spontaneous, and immediate than explicit measures (e.g., questionnaires; Hofmann, Gwaronski, Gschwendner, Le, & Schmitt, 2005; Houwer, 2005).

A study by Cludius, Schmidt, Moritz, Banse, and Jelinek (2017), upon which the current research builds, examined the self-concept of aggressiveness in 58 patients with OCD and 25 NCs using an implicit measure, the Aggressiveness Implicit Association Task (Agg-IAT). In the task, participants categorized themselves (“me”) as opposed to others (expressed by vocations, e.g. “architect”, “dentist”, “carpenter”) in conjunction with peaceful and aggressive descriptors. No significant group difference emerged, with participants displaying a bias towards a peaceful self-image. Surprisingly, participants with checking compulsions showed a significantly stronger bias towards self – peaceful and others – aggressive than NCs, contradicting prior theory and findings. However, the Agg-IAT findings were not compared with an explicit measure, rendering inconsistence with the study's hypotheses unclear. A subsequent study by Cludius et al. (2020) investigating the Agg-IAT again found no difference between patients with OCD and NCs, disconfirming the hypotheses of a more aggressive implicit self-concept in patients with OCD. Additionally, no group difference on the Agg-IAT was found between NCs and patients with checking compulsions in subsequent exploratory analyses.

The IAT has several structural shortcomings (for an overview, see Fiedler, Messner, & Bluemke, 2006). Most importantly, it merely shows that a certain association is relatively stronger than another; e.g. in the aforementioned study, the association self – peaceful/others – aggressive was more readily reproduced than the self – aggressive/others – peaceful association. It does not, however, provide information about how independently peaceful/aggressive the self or others are perceived, nor whether aggressiveness is perceived as a trait amongst many in a certain vocation, or if the entire vocation is per se considered aggressive.

The Implicit Relational Assessment Procedure (IRAP; Barnes-Holmes et al., 2006) disentangles these underlying relations. Conceptually rooted in relational frame theory (RFT; Hayes, Barnes-Holmes, & Roche, 2001), the IRAP applies latency and accuracy pressure to produce brief and immediate relational responses (BIRRs). Well-learned verbal relations (i.e. self – peaceful – correct) are taken to imply the subject's implicit attitude, therefore producing faster responses than unfamiliar relations (i.e., self – peaceful – incorrect; Barnes-Holmes, Barnes-Holmes, et al., 2010). Their counterparts, extended and elaborated relational responses (EERRs), are the types of answers produced by questionnaires (Barnes-Holmes, Barnes-Holmes, et al., 2010). They are not subject to latency pressure and incorporate verbal relations beyond the most well-learned – they are reflective, explicit, and therefore subject to response bias (Greenwald et al., 2000).

Several studies have successfully applied the IRAP to socially sensitive subjects, such as normative and paraphilic sexual interests (Dawson, Barnes-Holmes, Gresswell, Hart, & Gore, 2009; Rönspies et al., 2015) and gender stereotypes (Drake, Primeaux, & Thomas, 2018). Moreover, the IRAP has previously been used in patients with OCD (Nicholson, McCourt, & Barnes-Holmes, 2013; Vella, 2017). A meta-analysis of IRAP studies found relatively strong evidence for validity pertaining to contrasted groups, as examined in our study, whereas discriminant validity was found to vary according to whether implicit-explicit dissociations (between the IRAP and an explicit measure) were to be expected (Golijani-Moghaddam, Hart, & Dawson, 2013).

Contrary to most prior research on aggressiveness in OCD in general, the current research focuses specifically on participants' self-concept of their own aggressiveness. That is, the implicit measure (IRAP) employs self-referential statements, whereas the explicit measure (STAXI-II) requires participants to rate their agreement with statements from the first-person perspective. The patients’ individual beliefs about their propensity for aggression, particularly considering that patients with OCD rarely indeed act violent, appears most relevant for the treatment of OCD. Regarded in conjunction with cognitive biases such as over-responsibility, these cognitions may have significant behavioral implications.

Based on the body of prior research (e.g., Moritz et al., 2011; Whiteside & Abramowitz, 2005) as well as theories by Rachman (1993) and Freud (1976), we assumed that patients with OCD differ from NCs in their self-concept of aggressiveness. However, prior research differs in the directionality thereof: the majority of explicit findings showed elevated aggressiveness in patients with OCD (e.g., Liu et al., 2017; Moritz et al., 2011; Whiteside & Abramowitz, 2005), whereas the two prior studies using an implicit measure found no difference in self-image with regard to aggressiveness (Cludius et al., 2017, 2020). Therefore, our hypotheses were non-directional. We refrained from including a clinical control group in this first study due to the novelty of our IRAP design. We expected a group difference in aggressiveness using (1) an implicit measure, the IRAP, as well as (2) an explicit measure, the STAXI-II questionnaire. Moreover, as both instruments assess aggressiveness, albeit in structurally different ways, we expected (3) a positive correlation between scores. Finally, based on Rachman's (1993) theory, corroborated by findings by Radomsky et al. (2007) and Cludius et al. (2017), we expected the same pattern (1, 2, and 3) to emerge in a subsample of participants with checking compulsions.