Individual's beliefs regarding control have been linked to many domains of psychopathology; and, in particular, have been integrated to contemporary cognitive conceptualisations of anxiety, anxiety disorders and depression (e.g., Bandura, Reference Bandura1997; Clark, Reference Clark2004; Purdon & Clark, Reference Purdon, Clark, Frost and Steketee2002). For example, the learned helplessness model of depression (Miller & Seligman, Reference Miller and Seligman1973, Reference Miller and Seligman1975) proposes that a cognitive distortion of one's ability to control one's world is a major component of a reactive affective disorder. It has been proposed that a depressed individual may have a tendency to perceive oneself as relatively ineffective in exerting control over significant life events and their outcomes. In a similar way, Barlow, Chorpita, and Turovsky (Reference Barlow, Chorpita and Turovsky1996) have defined anxiety as a coherent cognitive-affective structure that is characterised by a ‘sense of uncontrollability focused on possible future threat, danger, or other upcoming, potentially negative events’ (p. 253). Consistently, many frequent symptoms in obsessive–compulsive disorder (OCD) may be understood as an individual's concern regarding the potential loss of control over his/her own thoughts and actions, and a high desire to maintain control over thoughts, actions, and the world through rituals (Clark, Reference Clark2004; Purdon & Clark, Reference Purdon, Clark, Frost and Steketee2002).
Early cognitive models of OCD (Carr, Reference Carr1974; McFall & Wollersheim, Reference McFall and Wollersheim1979) conceptualised compulsive behaviour as an attempt to achieve a sense of control over undesired outcomes; however, in modern cognitive theories of OCD, mental control is incorporated as a product of misunderstanding normal intrusions as unacceptable/threatening. Clark's (Reference Clark2004) comprehensive cognitive model of OCD emphasises the role of faulty appraisals of mental control efforts over unwanted intrusive thoughts and the perceived negative consequences of failed mental control in the persistence of obsessions. Although the adverse effects of faulty appraisal of thought control and its consequences have been successfully explained and documented in OCD literature, the role of individuals’ beliefs associated with controllability of life events has been largely ignored in recent cognitive formulations of OCD (Moulding & Kyrios, Reference Moulding and Kyrios2006).
Moulding and Kyrios (Reference Moulding and Kyrios2006) have suggested that two control-related beliefs, in particular, may be important to understand why individuals with OCD cannot ignore their intrusive thoughts, and feel an excessive need to perform neutralisation behaviours in response to their intrusive thoughts. The first belief is referred to as sense of control (SC), which refers to the degree to which an individual believes that a situation is controllable and that one has the skills necessary to bring about a desired outcome or prevent an undesirable outcome (Skinner, Reference Skinner1996). The second construct is desire for control (DC), which refers to an individual's motivation to gain control over life events (Burger & Cooper, Reference Burger and Cooper1979; Deci & Ryan, Reference Deci and Ryan2000; Skinner, Reference Skinner1996). The concept of SC is discussed within the social and behavioural science in a variety of forms, such as personal control, locus of control orientation, self-efficacy, mastery, self-directedness, personal autonomy, and helplessness (Ross & Sastry, Reference Ross, Sastry, Aneshensel and Phelan1999). Locus of control (LOC) is one of the most influential concepts widely examined in the framework of sense of control. It refers to an individual's attributions of control as being either internal or external (Rotter, Reference Rotter1966). Individuals with a predominantly internal locus of control believe that events in their lives are controlled by themselves; in contrast, those with predominantly external locus of control believe that such events are controlled by external sources, such as chance, fate, God, powerful others, and they have little or no control on the outcomes of events in which they are involved. Although LOC and SC terms are often used interchangeably to refer to one's perceived control over life events, Skinner (Reference Skinner1996) points out the basic conceptual differences between these two control dimensions. LOC would actually refer to the first part of sense of control that refers to an individual's belief about the controllability of situations, but not the perceived skill to control the outcomes of an event. While an external locus of control is usually accompanied by a low sense of control, it does not necessarily follow that individuals with an internal LOC have a high sense of control. For example, one may believe that events in general are dependent an skill level (i.e., internal locus of control) and at the same time believing that one does not have the skill necessary to control that event (i.e., low sense of control, for further discussion; see Skinner, Reference Skinner1996, Reference Skinner2007). In this manner, the present study mainly focused on examining the role of beliefs about controllability of life events in depression and obsessive–compulsive (OC) symptoms.
Since the first conceptualisation of LOC (Rotter, Reference Rotter1966), elaborated models of the relation between cognitions and depression have proposed that the attribution of causality that an individual makes for a traumatic event would determine the type and the extent of his/her depression. Consistently, subsequent experimental and theoretical investigations (e.g., Arnkoff & Mahoney, Reference Arnkoff, Mahoney, Perlmuter and Monty1979; Burger, Reference Burger1984, Reference Burger1991, Reference Burger1992; Dağ, Reference Dağ1992, Reference Dağ2002; Layton, Reference Layton1985; Liu et al., Reference Liu, Kurita, Uchiyama, Okawa, Liu and Ma2000), as well as clinical studies (e.g., Alloy & Clements, Reference Alloy and Clements1992; Martin, Abramson, & Alloy, Reference Martin, Abramson and Alloy1984) have consistently indicated that depressed individuals have a lower internal locus of control when compared to nondepressed ones. However, different from depression studies, the role of a person's control beliefs over life events in OCD has been largely ignored. This is quite surprising, considering the salient role of the motivation of gaining control over particular emotions, thoughts, and life events, as well as fears of losing control over thoughts and actions in the maintenance and persistence of OCD symptoms (Moulding & Kyrios, Reference Moulding and Kyrios2007).
Although few studies have investigated more general levels of DC and SC in individuals with OC-symptoms (for a review, see Moulding & Kyrios, Reference Moulding and Kyrios2006), they have provided consistent evidence that higher levels of DC and lower SC are associated with higher levels of OCD-relevant beliefs and symptoms (e.g., Doron, Kyrios, & Moulding, Reference Doron, Kyrios and Moulding2007; Moulding & Kyrios, Reference Moulding and Kyrios2007; Moulding, Kyrios, Doron, & Nedeljkovic, Reference Moulding, Kyrios, Doron and Nedeljkovic2009). Based on their a series of studies (Doron et al., Reference Doron, Kyrios and Moulding2007; Moulding, Doron, Kyrios, & Nedeljkovic, Reference Moulding, Doron, Kyrios and Nedeljkovic2008; Moulding & Kyrios, Reference Moulding and Kyrios2007; Moulding, Kyrios, & Doron, Reference Moulding, Kyrios and Doron2007; Moulding et al., Reference Moulding, Kyrios, Doron and Nedeljkovic2009), Moulding and his colleagues suggested that when an individual's sense of control does not reach the desired level, the discrepancy between SC and DC may motivate a person to perform neutralisation behaviours to regain sense of control. However, to date, only two studies have examined the relationship between LOC and OCD symptoms (Altin & Karanci, Reference Altin and Karanci2008; Kennedy, Lynch, & Schwab, Reference Kennedy, Lynch and Schwab1998). Kennedy et al. (Reference Kennedy, Lynch and Schwab1998) examined LOC in various clinical groups presenting depression or anxiety disorders, and a normal comparison group. They found that while the internal locus of control scale scores did not differ among patient and control groups, the OCD group's externality scores were the lowest of any of the anxiety and depressive disorder patients studied. The authors suggest that OCD patients are aware of the need for internal control, and their obsessions and rituals may serve a maladaptive function to increase their internal SC and reduce their external SC. Altin and Karanci (Reference Altin and Karanci2008) examined the relationship of LOC with responsibility appraisal in a Turkish nonclinical adolescent sample, and found that while LOC did not directly relate to OC symptoms, it moderated the effect of responsibility appraisals on OC symptoms, specifically on obsessive symptoms. They reported that a combination of high responsibility and low sense of control (i.e., external locus of control) produced the highest level of OC symptoms. As a result, their results underline the importance of understanding the relationship between general control beliefs and obsessive relevant appraisals in OC symptomatology.
Cultural variables also play a role between OCD and expression of symptoms (Siev & Cohen, Reference Siev and Cohen2007). Consistently, several researcher have alluded to the role of cultural context in OCD-specific cognitions and symptoms. To illustrate, Yorulmaz, Gençöz, and Woody (Reference Yorulmaz, Gençöz and Woody2010) found that there were cross-cultural differences in some specific appraisals such as importance and control of thoughts among Canadian and Turkish participants. Yet, the relationship between control cognitions in OC symptoms severity and depression in a non-Western culture has not been thoroughly examined and thus, it is not well understood. Therefore, we aimed to examine the associations between general control beliefs and OCD-relevant control beliefs in OC and depression symptoms using a non-Western culture. In light of theoretical assertions and previous research findings, in the current study we hypothesised that external locus of control would directly relate to symptoms of depression. However, based on the findings about impact of responsibility appraisals combined with low sense of control in OC symptoms (Altin & Karanci, Reference Altin and Karanci2008), we expected that an increased external LOC would increase the influence of obsessive-relevant beliefs on the severity of OC symptoms. In addition, as different OC symptoms may have different etiologies and related cognitions, we hypothesised that the association of LOC with OCD-relevant beliefs would differ in terms of different OC symptoms. However, as beliefs regarding the need to control thoughts are conceptually more similar to control variables, we expected that a low sense of control would interact with beliefs about control of intrusive thoughts, and would produce more persistent and severe OC symptoms.
Method
Participants
The sample for the study comprised 584 volunteer undergraduate university students from various departments of Uludag University in Bursa and Gazi University in Ankara, Turkey. Fifty-four (9.2%) participants were excluded from analyses because they had received a psychiatric diagnosis at the time of the study, thus the final sample comprised 530 undergraduates. The sample comprised 376 women (71.3%) and 154 men (28.7%), with a mean age of 20.78 (SD = 1.94).
Instruments
Locus of Control Scale (LCS; Dag, Reference Dağ2002)
LCS is a 47-item scale assessing whether individuals attribute the consequences of their behaviours to external or internal sources. It is a 5-point Likert-type scale, where 1 is totally inappropriate, and 5 is totally appropriate. Possible score range is from 47 to 235. Higher scores in this measure indicate a higher tendency to believe that life events are uncontrollable (i.e., external locus of control). The psychometric properties of the LCS for the Turkish university sample have been found satisfactory. An internal consistency analysis showed that the alpha coefficient was .92, while a 4-week test–retest reliability was .88. The results of a convergent validity analyses showed that LCS correlated significantly with Internal-External Locus of Control Scale (r = .67; Rotter, Reference Rotter1966), Rosenbaum's Learned Resourcefulness Schedule (r = -.39; Rosenbaum, Reference Rosenbaum1980), the Symptom Check List-90 (r = .25; SCL-90-R, Derogatis, Reference Derogatis1977) and the Paranormal Beliefs Scale (r = .46; Tobacyk, Reference Tobacyk1988). In the present study, the internal consistency of the LCS was high, α = . 87 (95% CI = .84–.88).
Obsessive Beliefs Questionnaire-Revised (OBQ; Obsessive Compulsive Cognitions Working Group [OCCWG], 2003, 2005)
The OBQ aims to investigate faulty belief domains which are assumed to be relatively specific to OCD. The current version of the scale has 44 items, with three subscales such as Responsibility/Threat Estimation (RT), Perfectionism/Certainty (PC) and Importance/Control of Thoughts (ICT) (OCCWG, 2005). A 7-point response format is used. The OBQ has been found to have satisfactory reliability and validity in various samples from different cultures (e.g., OCCWG, 2003, 2005; Yorulmaz & Gençöz, Reference Yorulmaz and Gençöz2008). Like the original version, the Turkish version of the OBQ was found to have good psychometric properties in both nonclinical and clinical samples (Yorulmaz & Gençöz, Reference Yorulmaz and Gençöz2008; Yorulmaz, Baştuğ, Tüzer, & Göka, Reference Yorulmaz, Baştuğ, Tüzer and Göka2011). A proportionality agreement coefficient or Tucker phi indicated that there was a high degree of similarity between the factors of ICT (Tucker phi = 0.88), RT (Tucker phi = 0.92), and PC (Tucker phi = 0.93). In addition, the Turkish OBQ-44 had satisfactory internal consistency (αs = .92 for the Total Score and .80, .86, .85 for ICT, PC, and RT, respectively. Similarly, in the present study, the internal consistency was very high for ICT, α = .81 (95% CI = .78–.83); PC, α = .86 (95% CI = .84–.88); RT, α = .85 (95% CI = .83–.87), and the Total Score, α = .93 (95% CI = .92–.94).
Padua Inventory-Washington State University Revision (PI-WSUR; Burns, Keortge, Formea, & Sternberger, Reference Burns, Keortge, Formea and Sternberger1996)
The PI-WSUR is a 39-item scale on a 5-point response option that evaluates OCD symptoms in five dimensions: checking compulsions, contamination obsessions/cleaning compulsions, dressing/grooming compulsions, obsessive thoughts, and impulses involving harm to oneself or others. The total scale scores ranges from 0 to 156, with high scores referring to higher severity of the OCD symptoms. The original factor structure was confirmed in clinical and nonclinical samples Turkey (Yorulmaz et al., Reference Yorulmaz, Karancı, Dirik, Baştuğ, Kısa, Göka and Burns2007), and internal consistency of the Turkish PI-WSUR were also found to be satisfactory (e.g., α = .76–.95). Moreover, the Turkish version was positively related with the Maudsley Obsessive Compulsive Inventory (Rachman & Hodgson, Reference Rachman and Hodgson1980) and Thought-Action Fusion Scale (Shafran, Thordarson, & Rachman, Reference Shafran, Thordarson and Rachman1996) among OCD patients and a nonclinical sample (r's ranging from 0.56 to 0.84). In the present study, alpha coefficients were satisfactory for the total score, α = .92 (95% CI = .91–.93), and five subscales, αs = .92 (95% CI = .83–.86) for checking, .85 (95% CI = .83–.86) for contamination/washing, .78 (95% CI = .74–.79) for obsessive thoughts of harm, .74 (95% CI = .71–.76) for obsessive impulsive of harm, 79 (95% CI = .76–.81) for grooming/dressing.
Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, Reference Beck, Epstein, Brown and Steer1988)
The BAI has 21 items on a 4-point response option, assessing cognitive and somatic symptoms of anxiety. The possible range of scores is 0 to 84, where higher scores correspond to higher anxiety. Ulusoy, Şahin, and Erkmen (Reference Ulusoy, Şahin and Erkmen1998) examined the Turkish version and found it to have reasonable psychometric properties with a satisfactory internal consistency and satisfactory construct validity. Furthermore, it was moderately correlated with the measures of depression, hopelessness and anxiety. In the present study, alpha coefficient of the BAI was very high, α = .92 (95% CI = .91–.92).
Beck Depression Inventory (BDI: Beck, Steer, & Garbin, Reference Beck, Steer and Garbin1988)
The 21-item BDI evaluates depressive symptoms in emotional, somatic, cognitive and motivation domains. The scores range from 0 to 21 and higher scores represent higher severity of depressive tendency also increases. The Turkish version also revealed acceptable psychometric characteristics (Hisli, Reference Hisli1988). The split half reliability of the Turkish BDI-R was between .74 and .78 for university students, and .61 for depressive patients. A 4-week test–retest reliability was .65 and .73 for students and patients, respectively (Tegin, Reference Tegin1980; Hisli, Reference Hisli1988, Reference Hisli1989). The concurrent validity of scale, based on the correlation with the Minnesota Multiphasic personality Inventory Depression Scale, was found to be .63, for the psychiatric sample (Hisli, Reference Hisli1988), and .50 for a university sample (Hisli, Reference Hisli1989). In the present study, the alpha coefficient was high, α = .83 (95% CI = .79 – .87).
Procedure
The administration of the research instrument to undergraduate students was carried out during regular class hours. The rationale of the study and the criterion of voluntary participation were emphasised. Only those who signed the informed consent form were included in the study. Aside from the first part of the questionnaire set, which included an explanation and rationale of the research, the informed consent form, and socio-demographic information, the scales were presented in a randomised sequence in order to control for order effects. Each administration took approximately 30–35 minutes.
Results
Correlation Coefficients Among Variables
The means, standard deviations and correlations between the measures are presented in TABLE 1. There was a weak but significant correlation between the LCS and depression; while the LCS had no significant correlation with OCD symptoms and beliefs. Additionally, the LCS showed only a significant relationship with obsessive thoughts of harm. As expected, OCD-relevant beliefs showed a moderate-to-strong significant positive association with all symptom dimensions of OCD.
TABLE 1 Mean Scores, Standard Deviations, Intercorrelations, and Reliability Coefficients for the Variables
Note: Means, and standard deviations (in parentheses) are given on diagonal. LOC = Locus of Control Scale; BDI = Beck Depression Inventory; BAI = Beck Anxiety Inventory; PI-R Total = Padua Inventory- Revised; PI-R Cleaning = Contamination/Washing; PI-Grooming = Grooming/Dressing; PI-R Obs.Harm = Obsessive Thoughts of Harm; PI-R Imp.Harm = Obsessive Impulsive of Harm; OBQ-RT = OBQ Responsibility/Overestimation of Threat; OBQ-PC = OBQ Perfectionism/Intolerance of Uncertainty; OBQ-ICT = OBQ Importance of Thoughts/Need to Control Thoughts.
*p < .05, **p < .001.
The Role of LOC and OCD-Relevant Beliefs in Depression and OC Symptoms
Following the tests of assumptions (Tabachnick & Fidell, Reference Tabachnick and Fidell2007), separate hierarchical regression analyses were performed to explore the relative role of the LOC and OCD-relevant cognitive constructs in depression and OC symptoms. In predicting OC symptoms, the first step consisted of control variables (i.e., age, gender, depression and anxiety scores). The LCS Total Score and three subscales of OBQ were entered into the analyses in the second step. In the last step, the interaction terms between LCS and obsessive beliefs were entered into the analysis. Aside from the exclusion of depressive scores among the control variables, the same steps were also followed in predicting depression.
Predictors of Depression Symptoms
The first regression analysis was conducted to examine the roles of LOC and OCD-relevant beliefs and their interaction in the severity of depression symptoms. TABLE 2 shows the results of this analysis. As expected, after controlling the significant variance explained by the BAI Total score, R2Change = .13, β = .34, t(500) = 8.28, p < .001, in step 2 the LCS Total score was the only significant predictor of a depressive tendency. A higher level of externality was significantly associated with higher levels of symptoms of depression, R2Change = .07, β = 20, t(500) = 5.08, p < .001. On the other hand, OBQ subscales and the interaction of LCS with obsessive beliefs did not contribute significantly to the explanatory power of the model. The predictors accounted for 20% of the variance in BDI.
TABLE 2 Summary of the Six Separate Hierarchical Multiple Regression Analyses for Predicting OC-Symptoms and Depression
Note: LOC = Locus of Control Scale, BDI = Beck Depression Inventory; BAI = Beck Anxiety Inventory; PI-R = Padua Inventory Revised; OBQ-RT = OBQ Responsibility/Overestimation of Threat; OBQ-PC = OBQ Perfectionism/Intolerance of Uncertainty; OBQ-ICT = OBQ Importance of Thoughts/Need to Control Thoughts.
*p < .05, ** p < .01, *** p < .001.
For ease of understanding, values for the nonsignificant variables are excluded. Each value is taken from the represented steps.
Predictors of OC Symptoms
The results of the regression analyses are also presented in TABLE 2, and it can be seen that among the control variables, which explained 16% of the variance, the depression, β = .22, t(500) = 5.25, p < .001), anxiety, β = .24, t(500) = 5.71, p < .001, and age scores, β = -.11 t(500) = -2.79, p < .001, had a significant association with the PI-WSUR Total Score. After excluding this variance, the second step, which comprised the OBQ subscales and LCS scores, explained 22% of the total variance, but only the subscales of RT, β = .28, t(496) = 5.39, p < .001, and PC, β = .23, t(496) = 5.09, p < .001, significantly predicted the PI-WSUR Total Score. In the last step, the interaction of LCS with ICT subscale weakly but significantly improved the explained variance in the PI- WSUR Total Score, β = 0.8, t(490) = 2.13, p < .05. All these variables combined accounted for 39% of the variance.
To better understand the nature of the significant interaction between LCS and ICT beliefs, whether the slopes of these two regression lines significantly differed from zero was tested following the procedure outlined by Aiken and West (Reference Aiken and West1991). According to this procedure, simple regression lines for moderated variables are plotted for significant interaction effects using centered data. FIGURE 1 suggests that types of locus of control orientation affected individuals differently, depending on their level of beliefs about the importance and control of thoughts in determining OC symptoms. The results revealed that the simple slope test was only significant for high levels of ICT; simple slope β = .29, t(543) = 4.71, p < .001. That is, it would appear that the degree of an individual's beliefs about the controllability of life events is only important for defining the severity of OC symptoms when an individual's belief in the importance and the control of thoughts is high. For the subjects with high ICT beliefs, OC symptom severity was lower among those with an internal locus of control than those with an external locus of control. However, for participants with low ICT beliefs, the simple slope was not significant, indicating that their levels of OC symptoms were low regardless of their level of locus of control.
FIGURE 1 Interaction between importance/control of thoughts and locus of control in the prediction of OCD symptoms.
Predictors of OC Symptom Dimensions
After examining the predictors of general OC symptoms, the predictors of the symptom subtypes were examined, and as can be seen from the TABLE 2, the results of the regression analyses revealed some OC symptom differentiations. To illustrate, among the control variables, the participants’ age showed a significant negative association with cleaning, dressing/grooming, and obsessive thoughts of harm. Moreover, while being male predicted the symptoms of checking and obsessive impulses of harm, being female was more related to cleaning/washing symptoms. After controlling the influence of the control variables in the first step, LCS did not predict any OC symptoms. RT and PC beliefs were significantly associated with all OC symptoms, except in the PI-WSUR Obsessive Impulses of Harm subscale. Finally, the interaction between LCS and ICT subscale was only weakly but significantly related to the PI-WSUR Checking subscale.
As seen in FIGURE 2, the simple slope coefficient for the checking of LCS at the two levels of ICT beliefs (i.e., low and high) was significant only for high ICT belief levels; simple slope β = .11, t(543) = 2.05, p < .05. The combination of externality and the over-importance of thought and control beliefs produced the highest degree of checking symptoms, whereas the simple slope was not significant for participants with low ICT beliefs, suggesting that their levels of OC symptoms were low, regardless of their levels of locus of control.
FIGURE 2 Interaction between importance/control of thoughts and locus of control in the prediction of checking symptoms.
Discussion
In this study, the relationship between the LOC and obsessive-related beliefs to OC and depression symptoms was examined comparatively in a nonclinical undergraduate sample. Based on previous findings, it was hypothesised that while LOC would show a direct relationship with the severity of symptoms of depression, it would interact with obsessive-related beliefs, in particular, with the importance of thought and control beliefs, to increase OC symptom severity. Furthermore, consistent with the suggestion that a discrepancy in control beliefs may be a trigger for engaging in OC symptoms we predicted that the combination of high desire for control of thoughts and external locus of control (i.e., a low sense of control) would relate more strongly to symptom dimensions with overt compulsive features, such as checking and cleaning, as opposed to obsessional impulses of harm and thoughts of harm. Consistent with the hypotheses, it was found that while greater external locus of control was significantly associated with greater levels of depressive symptoms, LOC did not show any significant relationship with either general OC symptom severity or any OC symptom dimensions. On the other hand, the interaction of LOC with importance and control of beliefs predicted weakly but significantly the severity of general OC symptoms as well as symptoms of checking.
The finding that LOC is a predictor of depressive symptoms is consistent with the theoretical importance of control beliefs in depressive symptoms. Since Rotter's (Reference Rotter1966) original work, most studies conducted on samples of students and other groups have revealed that LOC externality level has a positive relationship with depressive symptoms (Burger, Reference Burger1984; Dyal, Reference Dyal and Lefcourt1984; Ganellen & Blaney, Reference Ganellen and Blaney1984; Holder & Levi, Reference Holder and Levi1988; Lester, Reference Lester1989; Whitman, Desmond, & Price, Reference Whitman, Desmond and Price1987). Furthermore, Martin et al. (Reference Martin, Abramson and Alloy1984) found that clinically depression was related to low internality and high externality, supporting similar research findings among student populations. More importantly, the present study revealed that, consistent with the cognitive model of OCD, general control beliefs did not interact with any OCD-relevant beliefs to predict more severe depressive symptoms. These findings suggest that in spite of a high commorbity between depression and OCD (e.g., Fireman, Koran, Leventhal, & Jacabson, Reference Fireman, Koran, Leventhal and Jacabson2001; Samuels & Nesdadt, Reference Samuels and Nestadt1997), OCD-relevant cognitions still seem to be a unique feature of OCD.
The studies conducted by Keinan (Reference Keinan1994, Reference Keinan2002) showed that when the desire for control is increased, or the individual's sense of control is endangered, the individual is motivated to increase his/her subjective sense of control, even when the beliefs and actions used for this enhancement are implausible. Consistently, the results of the present study showed that the interaction between LOC and beliefs about the importance and control of thought was significantly associated with general OC symptoms and specifically checking symptoms. These results suggest that when external locus of control, which indicates the presence of decreased sense of control, is combined with a high desire to control thoughts, this increased discrepancy may motivate a person to perform actions to decrease the discrepancy between the perceived and desired control by performing compulsive checking. Therefore, checking symptoms may be seen as an attempt to regain control over external events. Once anxiety reduction and sense of control has been achieved through compulsive behaviours, such as in the case of checking, people with OCD may experience relief and may stop the ritual activity. Unfortunately, the relief is short-lived, anxiety returns, and the ritualised activity must be re-performed (Magril, Dar, & Liberman, Reference Magril, Dar and Liberman2008; Salkovkis, Reference Salkovskis1985).
The specific relationship between control beliefs and checking symptoms is also consistent with the hypothesis of this study. As suggested by Moulding and Kyrios (Reference Moulding and Kyrios2007), a higher DC and lower SC would relate more strongly to the symptom dimensions with overt compulsive features (e.g., checking and contamination/washing obsessions), but not covert compulsive features, such as obsessional impulses of harm or thoughts of harm. This is consistent with the suggestion that discrepancies in control beliefs may be a factor in motivating actions within OCD. However, the interaction between LOC and beliefs about the importance/control of thoughts was not significantly related to cleaning/washing symptoms. While this finding was not predicted, as a past study has linked control cognitions to contamination obsessions/washing compulsions rather than checking (Doron et al., Reference Doron, Kyrios and Moulding2007), it is possible that contamination and checking symptoms may be related to different control beliefs. As suggested by different researchers (Rachman, Reference Rachman2004; Steketee, Grayson, & Foa, Reference Steketee, Grayson and Foa1985), individuals with contamination obsessions are characterised by attempts to control their immediate environments, while checking symptoms are more related to ensuring the prevention of future harm. Therefore, this nonsignificant association of LOC with desire for thought control in terms of cleaning symptoms may be a result of item characteristics of LCS, which are mostly related to general control beliefs and control over future events, rather than an individual's immediate environment. Hence, this result should be replicated in future studies using different control measures.
The present study presents some limitations. First, the sample used in this study comprised a nonclinical university student from within a limited age range. In particular, as previous literature has indicated that the relationship between LOC and psychopathology may vary according to demographic variables such as age (Lachman, Reference Lachman1990), sex (Weisz, Rothbaum, & Blackburn, Reference Weisz, Rothbaum and Blackburn1984), and ethnicity (Lefcourt, Reference Lefcourt and Maher1972), it is important to examine the relationship between LOC and OC symptoms in different age and education groups. Although most of the previous studies reported that almost 80% of nonclinical subjects experienced unwanted intrusive thoughts that have very similar content as seen in clinical obsessions (e.g., Rachman & de Silva, Reference Rachman and de Silva1978; Salkovskis & Harrison, Reference Salkovskis and Harrison1984), the concepts examined in the present study are closely related to psychopathology and thus it is also strongly encouraged to examine these constructs in clinical samples. Additionally, the cross-sectional nature and the inclusion of general locus of control measures may also contribute to the limitations of the study. Accordingly, a replication of the present findings in clinical groups having received psychiatric diagnoses in studies with different methodologies (e.g., longitudinal) and alternative instruments will provide a clearer understanding of the role of control beliefs in OCD.
To conclude, the results of this study support the specificity of obsessive-related beliefs with OC symptoms, indicating these beliefs not showing any association with symptoms of depression in a young adult sample from a non-Western country, Turkey. The results seem to support the theoretical consistency of OC symptomatology in the international context. More importantly, the results point out the facilitating role of external locus of control when coupled with a high level of desire to control thoughts in OC symptoms, particularly, in checking symptoms. These novel findings suggest that understanding possible interaction between different control constructs would improve our understanding of phenomenology of OCD, and could help clinicians and researchers to develop more comprehensive cognitive formulations of OCD.
