نقش منبع کنترل و عزت نفس در احتمال ابتلا به هذیان در یک نمونه غیربالینی

Possessing either an external locus of control (LOC) or low self-esteem (SE) has been demonstrated prospectively to predict schizophrenia onset. This may be due to the consequences of these qualities for stress perception and resultant cortisol release. However, research with non-clinical samples has shown that only individuals with a combination of low SE and an external LOC show a significant cortisol response in response to a stressor. The current study hypothesized that low SE and an external LOC would be associated with greater proneness to hallucination-like experiences and delusion-like beliefs in a non-clinical sample (N = 493) than any other combination of these variables. A multiple linear regression found that the interaction between SE and LOC was not a significant predictor of either hallucination-like experiences or delusion-like beliefs. In line with previous research, LOC was found to be a significant, albeit weak, predictor of hallucination-like experiences and delusion-like beliefs. Implications for future research are examined, in addition to possible interventions in prodromal psychosis.

Schizophrenia is a disabling mental condition with a relatively low incidence of around 15 per 100,000 people (McGrath, 2005). Many of the psychotic experiences associated with schizophrenia, such as persecutory delusions and auditory verbal hallucinations, have been found to exist on a continuum reaching into the general population (Johns and van Os, 2001 and Verdoux and van Os, 2002). The annual incidence of hallucinatory experiences in the healthy population has been found to be 4% (Johns, Nazroo, Bebbington, & Kuipers, 1998) and typically more than 25% of healthy samples endorse items on the Peters et al. Delusions Inventory (Peters, Joseph, Day, & Garety, 2004). It is therefore plausible that research into individual differences in proneness to psychotic symptoms in non-clinical populations will aid our understanding of the cognitive mechanisms behind such symptoms in pathological conditions. One of the best ways to unravel the conundrum of psychotic symptoms is to examine what factors prospectively predict onset. One such predictive factor is locus of control (LOC). LOC is a personality trait that reflects the degree to which a person experiences events as due to their own actions (Rotter, 1966). Possession of an internal LOC signifies that the person expects reinforcement to be contingent on their own behaviour. An external LOC indicates that the individual expects reinforcement to be a function of chance or fate, under the control of powerful others, or simply unpredictable (Rotter, 1990). In one longitudinal study (Frenkel, Kugelmass, Nathan, & Ingraham, 1995) possessing an external LOC in adolescence was found to be a strong predictor of schizophrenia in adulthood. However, this study was unable to examine the relation between LOC and specific psychotic symptoms. Whilst subsequent research has demonstrated an association between an external LOC and the experience of persecutory delusions (Kaney and Bentall, 1989 and Garety and Freeman, 1999), there is as yet only preliminary evidence that LOC is related to hallucination-proneness. In a non-clinical sample, Levine, Jonas, and Serper (2004) found that an external LOC (assessed using Rotter’s (1966) scale) was associated with hallucination-proneness (assessed using the revised Launay–Slade Hallucination Scale [LSHS-R; Launay & Slade, 1981 modified by Bentall & Slade, 1985]). However, due to their small sample size (N = 42), Levine et al. (2004) noted that their study should be regarded as an “exploratory investigation between the locus-of-control and a predisposition towards hallucinations” (p. 27). In addition to LOC, low self-esteem (SE) has been found to predict first onset of psychotic symptoms (Krabbendam et al., 2002). Numerous mechanisms have been proposed as to how SE may play a causal role in the development of psychosis. For example, it has been suggested that low SE may result in the deployment of psychosis as a defence mechanism (Bentall, 2003 and Bentall and Kaney, 1996) and that low SE may strengthen the conviction with which some psychotic beliefs are held due to its consistency with the negative content of many such beliefs (Garety, Kuipers, Fowler, Freeman, & Bebbington, 2001). However, the role of SE is complicated when its relation to individual symptoms of psychosis are examined. For example, much work has investigated SE in relation to persecutory delusions. Whilst some studies have found explicit SE to be lower in individuals with persecutory delusions (e.g., Freeman et al., 1998) others have found such individuals’ explicit SE to be the same as other patients with remitted persecutory delusions as well as healthy controls (Lyon et al., 1994 and McKay et al., 2007). Similarly, in a non-clinical sample, explicit SE (assessed using the Rosenberg self-esteem scale) has been found not to predict levels of paranoid- and grandiose-like beliefs (Fowler, Freeman, & Smith, 2006). Although low SE has been linked to hallucinations in clinical samples (Close & Garety, 1998), to our knowledge the relation of SE to hallucination-proneness has not been investigated in non-clinical samples. In designing the present study we were interested in using existing findings from the stress literature to generate predictions about the interaction between LOC and SE in the genesis of psychotic experiences. Walker and Diforio’s (1997) neural diathesis–stress model proposes that cortisol release in response to a stressor interacts with an underlying diathesis (dopamine abnormality) to cause the symptoms of schizophrenia, including psychotic experiences. LOC and SE may play a role in the aetiology of psychotic experiences through their effects on stress perception and subsequent cortisol release. Healthy individuals with an external LOC have been found to report having less control over a given stressor, and to find such a stressor more subjectively stressful than those with internal LOCs (Bollini, Walker, Hamann, & Kestler, 2004). However, there are conflicting findings as to whether there is a direct relation between the amount of cortisol released in response to a stressor and an individual’s LOC. Whilst some studies have found an external LOC in healthy participants to be associated with increased cortisol in response to a stressor (Pruessner et al., 1997), others have failed to find such a relation (Bollini et al., 2004 and Kirschbaum et al., 1995). This appears to be firstly because the effects of an individual’s LOC on cortisol response are task-dependent. When individuals perceive that they have control over a stressor, those with an internal LOC will have a lower cortisol response than those with an external LOC (Bollini et al., 2004). However, when there is no perceived control, the cortisol responses are equivalent for those with internal and external LOCs (Bollini et al., 2004). A second reason why previous studies have failed to find a direct link between cortisol release in response to a stressor and an individual’s LOC may be due to the confounding effects of failing to control for participants’ levels of SE. Cortisol release in response to a stressor has been found to be jointly dependent on both participants’ LOC and their SE. In tasks involving performing mental arithmetic under time pressure with a social-evaluative component (individuals had to report their results to other participants), it was found that only those with a combination of low (self-reported) SE and an external LOC showed a significant salivary cortisol response (Pruessner, Hellhammer, & Kirschbaum, 1999). Furthermore, individuals with low (self-reported) SE and an external LOC have been found not to habituate to such psychosocial stressors (with social-evaluative components) as effectively as individuals with other combinations of LOC and SE (Kirschbaum et al., 1995). This results in such individuals having continuously high cortisol responses on exposure to such stressors (Kirschbaum et al., 1995). In summary, the findings reviewed here provide grounds for hypothesizing that individuals with an external LOC and low (self-reported) SE would be more prone to experiences lying at the extreme end of the continuum of psychosis-like experiences in the non-psychiatric population. The choice of a non-psychiatric population was deemed to be most appropriate, as the findings on the effects of LOC and SE on cortisol release in response to stressors have been drawn from research on such samples. In conclusion the current study set out to examine the following hypotheses: 1. LOC should predict levels of hallucination-like experiences and delusion-like beliefs, with a more external LOC being associated with higher levels of these variables. 2. SE should not be a significant predictor of either hallucination-like experiences or delusion-like beliefs. 3. Descriptive statistics should indicate that individuals with low SE and an external LOC would be more prone to hallucination-like experiences and delusion-like beliefs than individuals with either low SE and an internal LOC, high SE and an internal LOC, or high SE and an external LOC. If these differences are significant this should result in the interaction term between LOC and SE being a significant predictor of hallucination-like experiences and delusion-like beliefs in a multiple regression.

Descriptive statistics for the present study, along with comparative data from other studies, are presented in Table 1. Bivariate correlations are presented in Table 2. Due to the large number of correlations, only those correlations significant at p < 0.001 were taken as having practical significance for the purposes of this investigation. Age and gender did not significantly correlate with any of the variables. For illustrative purposes relating to our third hypothesis, psychopathology scores after a tertile split based on SE score and LOC score are presented in Table 3. The process of splitting continuous data into categorical groups and consequent analysis with MANOVA/ANOVA remains a controversial procedure ( Royston, Altman, & Sauerbrei, 2006). In this dataset the existence of a correlation (r = −0.31, p < 0.001) between SE and LOC made the use of MANOVA/ANOVA inappropriate due to the violation of the assumption of orthogonal relations between the independent variables. Furthermore, use of ANOVA with categorical variables derived from continuous variables has also been argued to increase the probability of a Type I error occurring ( Maxwell & Delaney, 1993). Thus, the complete, continuous dataset was analysed using a multiple linear regression, rather than these derived categorical groups.