Insecure Attachment, Social Withdrawal and their Contribution to Paranoia
Anika Vermeulen
Student number: 10357009
University of Amsterdam
Faculty of Social and Behavioral Sciences
Department of Clinical Psychology
Supervision: dr. Lindy-Lou Boyette
External supervision: dr. Nikie Korver-Nieberg
Table of Contents
Method ... 8
Participants and procedure ... 8
Instruments ... 9
Data analysis ... 11
Results ... 13
Discussion ... 19
References ... 25
Abstract
Background. Paranoia is a common symptom of psychotic disorders, but is also present in non-clinical populations in attenuated forms. Previous studies have found that social withdrawal and insecure attachment styles are linked to (sub-)clinical paranoia. Avoidant attachment styles (fearful and dismissing) are thought to lead to more social withdrawal, while anxious styles, involving higher levels of sociability, are not. The current study was conducted in order to test whether social withdrawal mediates the relationship between avoidant attachment and (sub-) clinical paranoia.
Methods. The current study utilized data from an existing longitudinal study (Genetic Risk and Outcome of Psychosis study) on patients with psychotic disorders (N=61) and their unaffected siblings (N=59). Data from the second measurement (T2) and three-year follow-up (T3) were used for cross-sectional and prospective regressions and mediational analyses. Attachment and social withdrawal were assessed in both populations using the RQ and the SFS Social
Engagement/Withdrawal scale. The PANSS Suspiciousness item and the SIS-R Suspiciousness scale were used to assess (sub-) clinical paranoia in patients and siblings, respectively.
Results. Dismissing attachment (an avoidant style) was cross-sectionally associated with social withdrawal in both patients and siblings. Fearful attachment (high levels of both anxious and avoidant attachment) was cross-sectionally associated with sub-clinical paranoia in siblings and was prospectively related to paranoia in patients. No mediation effects were found.
Conclusion. Although limited by a small sample size and possible construct overlap, the current study was able to shed light on the theoretical relevance of attachment styles to the development of paranoia.
Insecure Attachment, Social Withdrawal and their Contribution to Paranoia
Paranoia constitutes the erroneous belief that harm is occurring or will occur, and that the persecutor intends to cause this harm (Green et al., 2008). Paranoia is a common symptom of psychotic disorders (Freeman, 2007), but can be conceptualized as lying on a continuum with normal experiences (Freeman et al., 2005). Freeman (2006) concluded that 10-15% of the general population reports paranoid ideas such as “feeling you are watched or talked about by others”. Non-clinically severe beliefs, such as “there is a possibility of a conspiracy against me”, are reported by 5-6% of the general population. Paranoid ideas of clinical severity, such as “there have been times when you felt that a group of people were plotting to cause you serious harm or injury”, are reported by 1-3% of the general population. Clinical severity is determined not only based on thought content, but also on associated levels of distress and functional impairment (Freeman, 2007). As the thought becomes rarer, distress increases (Freeman et al., 2005). The continuum model is based on the premise that non-clinical and clinical paranoia share etiological features, making it relevant to study paranoia in psychiatric and non-psychiatric populations. An advantage of studying psychosis using a single-symptom approach is the added clarity in
describing mechanisms relevant to its development and manifestation (Persons, 1986). Paranoia, intertwined with patterns of social evaluation, has been studied in relation to attachment styles. A basic premise of attachment theory is that the pattern of interaction between young children and their primary caregivers (attachment figures) shapes social and emotional functioning in later life (Bowlby, 1973). Attachment theory postulates that children crave closeness with and responsiveness from their attachment figure. When the child perceives these needs to be met, a sense of safety ensues. If this pattern of interaction is dominant, the child is said to be securely attached. However, if the dominant pattern is one in which the child perceives
that its needs are not or inconsistently being met, the child feels unsafe and may become insecurely attached (Bowlby, 1988). Attachment styles are generally trait-like and stable over time, but may change following life events, such as divorce (Bowlby, 1973; Hamilton, 2000).
One of several conceptualizations of adult attachment was suggested by Bartholomew and Horowitz (1991). These authors proposed a four-category model of adult attachment, arranged along two dimensions of anxiety and avoidance. Securely attached individuals, characterized by low anxiety and low avoidance, develop positive views of self and others, expecting them to be trustworthy and supportive. Preoccupied attachment is basically characterized by high anxiety, low avoidance, a negative view of self and a positive view of others. These individuals are interpersonally demanding and seek others’ approval. Dismissing attachment is associated with low anxiety, high avoidance, a positive view of self and a negative view of others. People with this attachment style report low levels of distress, discount close relationships and withdraw socially. Fearful attachment is characterized by high anxiety and high avoidance, involves negative views of self and others, mistrust and also social withdrawal (Bartholomew & Horowitz, 1991).
Attachment styles are differentially associated with psychopathology. For example, 73% of healthy controls compared to only 17% of schizophrenic patients reported secure attachment (Ponizovsky, Nechamkin & Rosca, 2007), implying that insecure attachment may increase susceptibility to psychopathology, such as psychosis (Read & Gumley, 2010). As paranoia involves mistrust and negative views of self and others (e.g. “I am vulnerable to threat”, “Others are dangerous”, Garety et al., 2001), attachment styles involving these views may be of particular relevance. Preoccupied, dismissing and fearful attachment have indeed been found to be related to (non-clinical) paranoia (Korver-Nieberg et al., 2013; Berry, Barrowclough & Wearden, 2008;
Korver-Nieberg et al, 2015; Berry et al., 2006), although not in MacBeth et al. (2011).
A concept relevant to both paranoia and attachment theory is social withdrawal (Garety et al., 2001). Freeman et al. (2002) theorized that vulnerable individuals become isolated when, for example mistrust, motivates them to withdraw socially (Moutoussis et al., 2007). As suspicions are no longer subject to social corrective influences, paranoia can roam freely. Indeed, Velthorst et al. (2009) found that first onset of psychosis was predicted by social withdrawal. In healthy siblings of patients with psychotic disorders, Velthorst et al. (2012) also found (cross-sectional) associations between social withdrawal and sub-clinical psychotic experiences, such as paranoia.
Attachment styles also guide social engagement (Bartholomew & Horowitz, 1991). While individuals with secure and anxious attachment styles are prone to seek social support in times of need, individuals with avoidant attachment styles tend to socially withdraw. Avoidant attachment, reported by patients with psychotic disorders and, although to a lesser extent, their non-affected siblings (Pos et al., 2015), may explain social withdrawal in both populations. Siblings of patients with psychotic disorders are an interesting population in which to study the continuum of paranoia. Siblings report more sub-clinical positive symptoms than controls (e.g. Kendler et al., 1995), yet are unaffected by illness-related variables such as medication use. Therefore, studying siblings may shed light on factors common to the etiology of both clinical and non-clinical paranoia.
Studies examining prospective relationships between insecure attachment styles and paranoia have been scarce. Also, the few studies on social withdrawal in relation to paranoia have often conceptualized social withdrawal as social anhedonia, “characterized by social disinterest, withdrawal, and a lack of pleasure from social contact, indicating a deficit in the need to belong.” (Brown et al., 2007). Although this definition encompasses social withdrawal, it is
clouded by additional anhedonic features. There are many reasons conceivable for social withdrawal behavior. For example, in the context of Freeman et al.’s (2002) reasoning, social withdrawal may more likely be accompanied by interpersonal mistrust rather than by a lack of social pleasure. The use of a less affect-laden measure in the current study allowed us to examine social withdrawal in a more objective way.
Based on the above reasoning and as shown in Fig. 1, we hypothesize that 1) All insecure attachment styles (preoccupied, fearful and dismissing) and social withdrawal will predict (sub-) clinical paranoia. 2) The relationship between the avoidant attachment styles (fearful and
dismissing), but not the anxious or secure attachment styles, and (sub-)clinical paranoia will be mediated by social withdrawal. In testing these hypotheses, we will control for any demographic (e.g. age and gender) and clinical variables (e.g. duration of psychotic illness) found to predict paranoia. Associations have been found between depression and paranoia, as reviewed by Freeman (2007). In the current study, depression will therefore be assessed and controlled for by including an instrument for the diagnosis of depression (Andreasen, Flaum & Arndt, 1992).
Figure 1. Expected relationships and mediational effects between attachment styles, social withdrawal and (sub-)clinical paranoia.
Method
Participants and procedure
The current study utilized data from an existing dataset: the Genetic Risk and Outcome of Psychosis study (GROUP; for details, see Korver et al., 2012). In this multicenter cohort study, patients with psychotic disorders were selected from caseloads of psychiatric nurses,
psychologists and psychiatrists working in clinical centers in Amsterdam, Utrecht, Groningen and Maastricht. The inclusion criteria for the patient group were as follows: 1) aged between 16 and 50 years old, 2) diagnosed with non-affective psychotic disorder according to the criteria listed in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), 3) Dutch language proficiency and 4) able and willing to give written informed consent. Selected patients were informed of the purposes of the study and were asked permission to contact their first-degree relatives. Inclusion criteria for their siblings were 1) aged between 16 and 50 years old, 2) no lifetime psychotic disorder at baseline, 3) Dutch language proficiency, and 4) able and willing to give written informed consent.
Three measurements were administered over the course of six years. For cross-sectional and longitudinal analyses, the current study assessed attachment styles using data from the second measurement (T2) and social withdrawal and paranoia using data from measurement two and three (T2 and T3). The total test battery was administered in academic centers and psychosis departments in the Amsterdam, Groningen, Maastricht and Utrecht areas, or at the homes of patients who could not visit these locations. In the patient group, the test battery took two weekly two-hour sessions to complete and one three-hour session in the sibling group (Korver et al.,
2012). A sample of N=74 patients and N=66 siblings in the Amsterdam area was selected for use in the current study. As an add-on to the test battery, these participants had completed the
Relationship Questionnaire (Bartholomew & Horowitz, 1991).
Instruments
Attachment
Attachment anxiety and avoidance were assessed using the Dutch version of the Relationship Questionnaire (RQ; Bartholomew & Horowitz, 1991). The RQ can be found in Attachment 1 in this document. This four-item self-report questionnaire contains statements reflecting secure, fearful, preoccupied and dismissing attachment styles. The RQ can be used to measure attachment styles categorically as well as continuously. Continuous scores are acquired by having participants rate the degree to which all four statements apply to them on a seven-point Likert scale. Categorical use of the RQ requires participants to select one statement describing the attachment category most applicable to them. However, if they are unable to, the highest-rated attachment category from the continuous RQ-report is selected as the dominant category. Secure, fearful, preoccupied and dismissing attachment are reflected by scores of 1, 2, 3 and 4, respectively. In non-clinical samples, the RQ has shown validity and moderate test-retest
reliability (Griffin & Bartholomew, 1994a). However, to our knowledge, the RQ’s psychometric properties have not been reported for clinical populations.
Social withdrawal
Social withdrawal was assessed using the Social Engagement/Withdrawal scale of the Dutch version of the Social Functioning Scale (SFS; Birchwood et al., 1990), a self-report questionnaire measuring social functioning over the previous three months in multiple areas. The Social Engagement/Withdrawal scale includes items such as “How many hours of the day do you spend alone?”. A minimum of 0 and a maximum of 15 points can be scored on the SFS Social Engagement/Withdrawal scale. Lower scores on signify lesser social functioning compared to higher scores. The SFS has been found to have good reliability and was validated in a group of patients with schizophrenia and their family members (Birchwood et al., 1990).
Paranoia
In the patient sample, paranoia was assessed using the Suspiciousness item (PA-6) of the Dutch version of the Positive and Negative Syndrome Scales (PANSS; Kay, Flszbein & Opfer, 1987), a widely used interview for the severity of psychotic symptoms over the previous week. The Suspiciousness item is scored on a Likert scale from 1 (Absent) to 7 (Extreme). Scores above 4 are regarded as indicating clinical severity (Humphreys & Barrowclough, 2006). The PANSS has been found to have good internal consistency and good reliability (Korver-Nieberg et al., 2015) and was validated in a clinical sample consisting of people with schizophrenia (Kay, Flszbein & Opfer, 1987). However, these psychometric properties may be compromised due to the use of a single item.
In the sibling group, sub-clinical paranoia was assessed using the Suspiciousness scale of the Dutch version of the Structured Interview for Schizotypy-Revised (SIS-R; Vollema and
Ormel, 2000). This is a 31-item semi-structured interview and observational instrument for sub-clinical psychotic (= schizotypical) phenomena over the past three years. The Suspiciousness scale includes items such as “All in all, it is probably safer never to trust anyone” (Kendler, Lieberman & Walsh, 1989). Subjects are instructed to answer what applies to them ‘in general’, within a given time frame of 3 years. The researcher scores the items on a four-point Likert scale, ranging from absent (0 points) to severe (3 points), (Velthorst et al., 2012). The SIS-R has been found to be reliable and valid (Vollema & Ormel, 2000; Kendler, Lieberman & Walsh, 1989).
Depression
Depression was assessed using the Comprehensive Assessment of Symptoms and History (CASH; Andreasen, Flaum & Arndt, 1992), a structured interview for a variety of psychoses and affective disorders and is scored positive when individuals meet DSM-IV-TR criteria
(Andreasen, 1987; APA, 2013) for a current depressive episode. As mentioned by Andreasen (1987), the CASH has been psychometrically assessed and has been found to be reliable. However, to our knowledge, the results of these assessments have not been published.
Data analysis
All analyses were conducted using IBM SPSS 23. The RQ was used categorically to avoid participants falling into more than one attachment category and to thereby enhance the
interpretability of the current study’s results. As suggested by Hayes & Preacher (2014), we coded the four attachment categories using 3 dummy variables. The cross-sectional and prospective mediation models were tested using Hayes’ PROCESS for multicategorical predictors (version 2.15, Hayes, 2012) for IBM SPSS. First, we tested the mediational model cross-sectionally, separately for the group of patients with psychotic disorders and the group of their non-affected siblings. A cross-sectional mediation effect would be confirmed if statistically significant relationships between the RQ fearful and dismissive categories and the score on the PANSS Suspiciousness item and the SIS-R Suspiciousness scale at T2 weakened or disappeared when the score on the SFS Social Engagement/Withdrawal scale at T2 was entered as a
mediator.
Second, using linear regressions, we examined whether the scores on the RQ fearful and dismissive categories and on the SFS Social Engagement/Withdrawal scale (both at T2) predicted patients’ scores on the PANSS Suspiciousness item at three-year follow-up . If so, we would then test a prospective mediation model in the patient group, in which the score on the SFS Social Engagement/Withdrawal scale mediated this relationship. PROCESS’ dialog box (version 2.15, Hayes, 2012) allowed us to control for acute psychosis at baseline and other possible
confounders (e.g. gender). This mediation effect would be confirmed if the statistically significant relationships between the scores on the RQ fearful and dismissive categories at T2 and the score on the PANSS Suspiciousness item at T3 weakened or disappeared when the score on the SFS Social Engagement/Withdrawal at T2 was added as a mediator. This would assess the association between social withdrawal, reported three years earlier, and (sub-)clinical paranoia over the past week. However, in order to more clearly assess causality, we also prospectively examined this relationship in a shorter time frame (three months), using the T3 scores on both
the SFS Social Engagement/Withdrawal scale and the PANSS Suspiciousness item.
As an additional aim, we examined whether insecure attachment styles found to predict paranoia could distinguish between clinical and sub-clinical paranoia. This was done by coding a dummy variable for scores > 4 versus < 4 (Humphreys & Barrowclough, 2006) on the PANSS Suspiciousness item (at T2 or T3, depending on associations found).
Results Sample characteristics
Participants’ data was selected for use if 1) there was no more than 30% missing data 2) they endorsed one specific attachment category. The final sample consisted of N= 61 patients and N= 59 siblings. The mean duration of psychotic disorder in patients was 7.86 years (sd= 6.40). Patients were aged between 20 and 50 years (mean= 31.39 years, sd= 8.19) and siblings were aged between 19 and 56 years (mean= 32.08, sd= 9.02). Data on the PANSS
Suspiciousness item at T3 was available for N= 32 patients. Descriptives for demographic variables can be found in Table 1. Inspection of histograms and Kolmogorov-Smirnov tests showed that only the scores on the SFS Social Engagement/Withdrawal scale in siblings followed a normal distribution. Therefore, we applied bootstrapping in our analyses when possible.
Table 1. Demographic and clinical variables for patients and siblings Patients Siblings N (%) N (%) Gender Male 50 (82) 16 (27.1) Female 11 (18) 43 (72.9) Depression diagnosis Yes 3 (4.9) 1 (1.7) No 57 (93.4) 58 (98.3)
Three dummy variables for the four RQ categories were automatically coded by PROCESS. Endorsed most frequently, secure attachment was coded as a reference category (Field, 2013). Score descriptives for patients and siblings can be found in Table 2, scores arranged by attachment category are displayed in Table 3. As there was only one participant in the sibling group with a dominant preoccupied attachment style, this category was not analyzed further. There was a significant association between subject status (patient vs. sibling) and attachment style χ2 (1) = 13.32, p< .001, with patients endorsing secure attachment less
frequently and insecure attachment styles more frequently than siblings. Gender was found to predict the PANSS Suspiciousness item at T2, with women scoring lower than men. More
demographic variable categories were found to predict scores on the PANSS Suspiciousness item and the SIS-R Withdrawal scale. However, only one or two people belonged to the predictive categories. Therefore, these variables were not controlled for.
Table 2. Descriptives for questionnaire scores
Patients Siblings
Scale Min-max Mean SD Median Range Mean SD Median Range
SFS 0-15 10.05 2.70 10 4-15 12.84 2.36 13 4-15
PANSS T2 1-7 1.67 1.27 1 1-6
PANSS T3 1-7 1.94 1.34 1 1-6
Table 3. Descriptives for questionnaires by attachment category for patients and siblings Patients Siblings Scale Secure (N= 19, 32.8%) Fearful (N=12, 19.7% ) Preoccupied (N=10, 16.4%) Dismissing N= 19, 31.1%) Secure (N=39, 66.1%) Fearful (N=9, 15.3%) Preoccupied (N=1, 1.7%) Dismissing (N=10, 16.9%) SFS Social Engagement/ Withdrawal Mean 11.11 9.33 10.40 9.26 Mean 13.58 11.44 12 11.40 Min-max 0-15 SD 2.08 2.35 3.06 3.05 SD 1.87 2.56 2.91 Range 7-14 5-13 6-15 4-14 Range 5-15 6-14 4-14 PANSS Suspiciousness T2 Mean 1.35 1.92 1.50 1.95 Min-max 1-7 SD .93 1.08 1.58 1.51 Range 1-4 1-4 1-6 1-5 PANSS Suspiciousness T3 Mean 1.30 2.71 1.33 2.44 Min-max 1-7 SD .68 1.25 .52 1.88 Range 1-3 1-4 1-2 1-6 SIS-R Withdrawal Min-max 0-3 Mean .49 1.11 1 .44 SD .60 .93 .53 Range 0-2 0-2 0-1
Cross-sectional mediation analyses
First, we tested the cross-sectional mediation effect in patients while controlling for gender. Neither RQ category nor the score on the SFS Social Engagement/Withdrawal scale predicted the score on the PANSS Suspiciousness item at T2. Only dismissing attachment was found to predict the score on the SFS Social Engagement/Withdrawal scale, b= -1.89, se= .86, t= -2.20, 95% BCI [-3.61; -.17], p= .03. Based on these results, a cross-sectional mediation in patients was not confirmed.
In siblings, fearful attachment significantly predicted the score on the SIS-R
Suspiciousness scale b=.56, se=.26, t=2.18, 95% BCI [.04; 1.08], p=.03. However, the score on the SFS Social Engagement/Withdrawal scale did not predict the score on the SIS-R
Suspiciousness scale. Only dismissing attachment predicted SFS Social Engagement/Withdrawal score (b= -2.25, se= .82, t= -2.75, 95% BCI [-3.88; -.61], p= .01. A cross-sectional mediation effect in siblings was not confirmed. Fig. 2 shows the results of these analyses.
Figure 2. Results of cross-sectional associations in siblings (dotted arrow) and in both patients and siblings (solid arrow). *P<.05
Prospective mediation analyses
Next, linear regressions were conducted to analyze a longitudinal association between RQ category at T2 and the score on the PANSS Suspiciousness item at T3 in N= 32 patients. Fearful attachment at T2 significantly predicted the score on the PANSS Suspiciousness item at T3, b= 1.41, se= .52, 95% BCI [.33-2.34], p= .02 Because the score on the SFS Social
Engagement/Withdrawal scale at T2 did not significantly predict the score on the PANSS Suspiciousness item at T3, a further mediation analysis was unwarranted. We then assessed whether the score on the SFS Social Engagement/Withdrawal scale at T3 was associated with the PANSS Suspiciousness item at T3. A regression analysis using data available for N=29 patients was performed. No significant effects were found. Fig. 3 shows the results of these analyses.
As an additional analysis, we assessed whether fearful attachment could distinguish between clinical and non-clinical paranoia in patients, coded as a dummy variable for PANSS Suspiciousness scores < 4 and > 4. This was not confirmed based on the results of a Chi-square test.
Discussion
The current study aimed to assess relationships among insecure attachment styles, social withdrawal and paranoia in patients and siblings. The hypothesis that all three insecure
attachment styles predicted paranoia, was not confirmed. The only insecure attachment style to predict paranoia in siblings and patients (but only at T3), was fearful attachment, a style which consists of both anxious and avoidant strategies. Fearfully attached individuals reported more severe paranoia than securely attached individuals. The hypothesized mediational effect of social withdrawal on the relation between avoidant attachment styles and paranoia, was also not found. However, dismissing attachment style was cross-sectionally associated with social withdrawal in patients as well as siblings.
The male-to-female ratios in the patient and sibling sample were almost each other’s reverse (see Table 1). As Freeman et al. (2011) noted, the evidence on the relationship between gender and paranoia is equivocal. If it is the case that either males or females are more likely to report (sub-) clinical paranoia, then the power to detect paranoia in either our sibling or patient sample would have been compromised. Additionally, the most prominent caveat in the current study, the small sample size, decreased statistical power and generalizability. Therefore, the results should be interpreted with caution.
With regard to avoidant styles, the finding that only fearful attachment was associated with paranoia may be explained by its distinction from dismissing attachment. Inherent to both styles is the presence of negative views of others, but only fearfully attached individuals hold negative views of self. It has been argued that these negative views of self are activated by self-discrepancies, in which the coherence between actual self and self-ideals is threatened. In an
effort to protect against this confrontation, the blame for this discrepancy is placed on others. This external attribution then triggers thoughts and beliefs about others as dangerous and ill-intentioned (Bentall & Kaney, 1996). High self-esteem in dismissing individuals may make them resistant to such threats. Their negative views of others may simply follow from their lack of need for sociability. Paranoia would then more likely develop in a person who responds to personal threat with blame, fear and mistrust of others, as is the case in fearful attachment. Another explanation for the association between fearful attachment and (sub-) clinical paranoia is the following. As Green et al. (2008) theorized, paranoia exists in two forms: ideas of social reference (e.g. believing to be the subject of gossip) and ideas of persecution (e.g. believing that others intend harm). Korver-Nieberg et al. (2013) examined these types of paranoia separately and found that anxious attachment was associated with social reference paranoia, while avoidant attachment was associated with ideas of persecution. Fearful attachment, involving anxious as well as avoidant styles, should then be related to both types of paranoia. This was corroborated by the current study’s findings: fearful attachment was associated with the PANSS
Suspiciousness item, which covers ideas of persecution, as well as with the SIS-R Withdrawal
scale, reflecting both forms of paranoia.
The theoretical distinction in views of self and others may also explain the finding that dismissing, and not fearful attachment, predicted social withdrawal. Theoretically, this
withdrawal behavior should follow the dismissing person’s lack of interest and need for social relationships. Fearfully attached individuals, although wary of others and their intentions, still seek and crave social approval (Bartholomew & Shaver, 1998). Therefore, fearfully attached individuals may be apprehensive toward social withdrawal (Berry et al., 2006), as they balance their dependence on others with their fear of them. Dismissing individuals’ self-sufficiency may
allow them to withdraw from others without conflict.
Based on previous research, we expected preoccupied attachment, an anxious attachment style, to predict paranoia. We were unable to assess this in siblings, and it was not confirmed in patients. Preoccupied individuals, like fearful individuals, typically exhibit negative views of self (Bartholomew & Horowitz, 1991). These can be ameliorated by obtaining others’ approval, which the preoccupied individual (hyperactively) seeks. The question arises why fearful
attachment, also characterized by a search for social approval (Bartholomew & Shaver, 1998), does predict paranoia. Preoccupied and fearfully attached individuals may differ in their base rate of social interactions and size of social network. Compared to fearfully attachment, preoccupied attachment may be associated with larger social networks and more frequent social encounters. To our knowledge, social network size and exposure have not been compared between different attachment categories. Of course, factors much more elaborate than ‘friend count’ contribute to subjective social support. However, it remains plausible that positive views of others make the preoccupied individual more inclined to build a support system than the fearfully attached individual. Frequent exposure to social situations may provide evidence that other people are, at least sometimes, reliable and trustworthy (Moutoussis et al., 2007). A small social circle limits the occurrence of these opportunities and decreases the chances of paranoid thoughts being subjected to social correction, making paranoia more likely in fearful attachment (Freeman et al., 2002).
Our last hypothesis held that social withdrawal would mediate the relation between avoidant attachment and paranoia. Driven by attachment-related views of self and others, withdrawal would then precede paranoia, and not vice versa. This association could not be confirmed based on the current study’s results. As has been suggested in previous research (e.g.
Moutoussis et al., 2007), social withdrawal may function as a (maladaptive) coping mechanism. Withdrawal behavior relieves feelings of interpersonal threat aimed at the individual. If there is indeed a causal relationship in which paranoia follows social withdrawal, the current study’s findings would be unsurprising. However, the content of the SFS Social
Engagement/Withdrawal scale might also be a point of discussion. The factual nature of the items on this scale was thought to be an advantage, allowing us to assess social withdrawal as objectively as possible. However, these items reflected behavioral state over the past three months, while our conceptualization of social withdrawal was one that described a change in social behavior. Experience Sampling Methods may have provided a clearer chronology of social withdrawal behavior. A less intensive method might have been the use of a scale assessing change in social behavior over the past months. An example of an item tailored to Freeman et al.’s (2002) theory may have been “Compared to three months ago, how would you describe the amount of time you spent with others?”, with response options such as “It has decreased”, “It has remained stable” and “It has increased”. This type of measurement would more clearly indicate that the person had at some point initiated withdrawal behavior. Items probing the motivation behind this behavior change might then suggest whether social withdrawal functions as a mechanism for coping with ideas of interpersonal mistrust.
The current study had some limitations worth noting. First, although categorical use of the RQ was deemed most suitable for the interpretability of the current study’s results,
continuous measures are more commonly used and recommended. Attachment has been shown to better fit a dimensional distribution than a taxonomy (Fraley & Waller, 1998). Replication of the current study using continuous measures, such as those assessing anxious and avoidant attachment, may provide a more realistic representation of attachment and its relation to social
withdrawal and paranoia.
A second point of discussion is the possibility of overlap between insecure attachment and (sub-clinical) paranoia. Both constructs involve anxiety, avoidance, negative views of self and others, and, most notably, mistrust (Martin & Penn, 2001; Freeman, Garety & Kuipers, 2001; APA, 2013; Bartholomew & Horowitz, 1991). Although the relationship between attachment and psychosis has gained attention in the literature, the specific associations involving paranoia have been less researched. Therefore, it is uncertain if attachment and
paranoia can be seen as separate entities or as part of a single construct or spectrum. In the latter case, any associations found between attachment categories and paranoia would simply arise due to their commonalities, instead of an underlying theoretical link.
Lastly, the data on attachment was obtained using a self-report questionnaire, allowing for the possibility of biases such as socially desirable reporting. Moreover, there is evidence that attachment styles manifest themselves in ways that go beyond what participants are able to verbally report (Crowell, Fraley & Shaver, 1999). As has been noted by Berry, Barrowclough & Wearden, 1997, a richer operationalization of attachment might be attained by the use of multiple (informant) measures. Possible additional issues resulting from poor insight in patients with psychotic disorders (Bell et al., 2007) may hereby also be circumvented. A related suggestion for future research is to take into account differences in primary attachment relationships (Fraley & Davis, 1997). For example, parental relationships and relationships with clinicians may be more relevant than romantic relationships for patients.
An advantage of the current study was the availability of a sibling group with no history of psychotic disorder. This provided us with a reference group that was unaffected by illness-related symptoms, such as use of antipsychotic medication. The associations found between
fearful attachment and (sub-clinical) paranoia are in line with the continuum model (Freeman et al., 2005). This may imply that attachment is a relevant etiological factor in paranoia.
Conversely, secure attachment has been suggested to protect against psychosis (Mikulincer & Shaver, 2012). Studies, although low in psychometrical strength, have found secure attachment to be associated with later onset of psychosis, shorter hospitalization time and less severe symptomatology (Ponizovsky, Nechamkin & Rosca, 2007; Wilson & Costanzo, 1996). In the current study, patients with psychotic disorders were more likely to endorse insecure categories than their siblings. These findings seem to be in accordance with previous research.
The current study highlighted specific associations between insecure attachment styles on the one hand and social withdrawal and (sub-)clinical paranoia on the other. Future studies might examine whether fearful attachment is associated with both persecutory and social reference paranoia in patients with psychotic disorders. Mechanisms underlying the links between insecure attachment on the one hand and social withdrawal and paranoia on the other, may be explored by considering factors such as self-esteem, fear of rejection and need for approval.
References
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.
Andreasen, N. C. (1987). Comprehensive Assessment of Symptoms and History (CASH). Department of Psychiatry, University of Iowa College of Medicine.
Andreasen, N. C., Flaum, M., & Arndt, S. (1992). The Comprehensive Assessment of Symptoms and History (CASH): an instrument for assessing diagnosis and psychopathology. Archives of General Psychiatry, 49(8), 615-623
Bartholomew, K., & Horowitz, L.M. (1991). Attachment styles among young-adults: a test of a 4 category model. Journal of Personality and Social Psychology, 61, 226–244.
Bartholomew, K., & Shaver, P. R. (1998). Methods of assessing adult attachment. Attachment theory and close relationships, 25-45.
Bell, M., Fiszdon, J., Richardson, R., Lysaker, P., & Bryson, G. (2007). Are self-reports valid for schizophrenia patients with poor insight? Relationship of unawareness of illness to psychological self-report instruments. Psychiatry research, 151(1), 37-46.
Bentall, R. P., & Kaney, S. (1996). Abnormalities of self-representation and persecutory delusions: A test of a cognitive model of paranoia. Psychological Medicine, 26(06), 1231- 1237.
Berry, K., Barrowclough, C., & Wearden, A. (2007). A review of the role of adult attachment style in psychosis: unexplored issues and questions for further research. Clinical psychology review, 27(4), 458-475.
Berry, K., Barrowclough, C., & Wearden, A. (2008). Attachment theory: a framework for understanding symptoms and interpersonal relationships in psychosis. Behaviour Research and Therapy, 46(12), 1275-1282.
Berry, K., Wearden, A., Barrowclough, C., & Liversidge, T. (2006). Attachment styles, interpersonal relationships and psychotic phenomena in a non-clinical student sample. Personality and Individual Differences, 41(4), 707-718.
Birchwood, M., Smith, J. O., Cochrane, R., Wetton, S., & Copestake, S. O. N. J. A. (1990). The Social Functioning Scale. The development and validation of a new scale of social adjustment for use in family intervention programmes with schizophrenic patients. The British Journal of Psychiatry, 157(6), 853-859.
Bowlby, J. (1973). Attachment and loss: Separation: Anxiety and anger (Vol. 2).
Bowlby, J., 1988. A Secure Base: Clinical Applications of Attachment Theory. Basic Books, New York.
Brown, L. H., Silvia, P. J., Myin-Germeys, I., & Kwapil, T. R. (2007). When the Need to Belong Goes Wrong The Expression of Social Anhedonia and Social Anxiety in Daily
Life. Psychological Science, 18(9), 778-782.
Crowell, J. A., Fraley, R. C., & Shaver, P. R. (1999). Measurement of individual differences in adolescent and adult attachment.
Field, A. (2013). Discovering statistics using IBM SPSS statistics. Sage.
Fraley, R. C., & Davis, K. E. (1997). Attachment formation and transfer in young adults’ close friendships and romantic relationships. Personal relationships, 4(2), 131-144.
Fraley, R. C., & Waller, N. G. (1998). Adult attachment patterns: A test of the typological model.
Freeman, D. (2006). Delusions in the nonclinical population. Current Psychiatry Reports, 8(3), 191-204.
Freeman, D. (2007). Suspicious minds: the psychology of persecutory delusions. Clinical Psychology Review, 27(4), 425-457.
Freeman, D., Garety, P. A., & Kuipers, E. (2001). Persecutory delusions: developing the understanding of belief maintenance and emotional distress. Psychological medicine, 31(07), 1293-1306.
Freeman, D., Garety, P. A., Bebbington, P. E., Smith, B., Rollinson, R., Fowler, D., & Dunn, G. (2005). Psychological investigation of the structure of paranoia in a non-clinical
population. The British Journal of Psychiatry,186(5), 427-435.
Freeman, D., Garety, P.A., Kuipers, E., Fowler, D., & Bebbington, P.E. (2002). A cognitive model of persecutory delusions. British Journal of Clinical Psychology, 41, 331–347.
Freeman, D., McManus, S., Brugha, T., Meltzer, H., Jenkins, R., & Bebbington, P. (2011). Concomitants of paranoia in the general population.Psychological medicine, 41(05), 923 936.
Garety, P. A., Kuipers, E., Fowler, D., Freeman, D., & Bebbington, P. E. (2001). A cognitive model of the positive symptoms of psychosis. Psychological Medicine, 31(02), 189-195.
Green, C. E. L., Freeman, D., Kuipers, E., Bebbington, P., Fowler, D., Dunn, G., & Garety, P. A. (2008). Measuring ideas of persecution and social reference: the Green et al. Paranoid Thought Scales (GPTS). Psychological Medicine, 38(01), 101-111.
Griffin, D., & Bartholomew, K. (1994a). Metaphysics of measurement: The case of adult attachment. In K. Bartholomew, & D. Perlman (Eds.), Advances in personal
relationships: Vol 5. Attachment processes in adulthood (pp. 17−52). London: Jessica
Kingsley.
Hamilton, .E.(2000). Continuity and discontinuity of attachment from infancy through adolescence. Child Development, 71, 690−694
Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling. Retrieved from http://www.afhayes.com/ public/process2012.pdf
Hayes, A. F., & Preacher, K. J. (2014). Statistical mediation analysis with a multicategorical independent variable. British Journal of Mathematical and Statistical Psychology, 67(3), 451-470.
Humphreys, L., & Barrowclough, C. (2006). Attributional style, defensive functioning and persecutory delusions: Symptom‐specific or general coping strategy? British Journal of Clinical Psychology, 45(2), 231-246.
Kay, S. R., Flszbein, A., & Opfer, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophrenia Bulletin, 13(2), 261.
Kendler, K. S., Lieberman, J. A., & Walsh, D. (1989). The Structured Interview for Schizotypy (SIS): a preliminary report. Schizophrenia Bulletin, 15(4), 559
Kendler, K. S., McGuire, M., Gruenberg, A. M., & Walsh, D. (1995). Schizotypal symptoms and signs in the Roscommon Family Study: their factor structure and familial relationship with psychotic and affective disorders. Archives of General Psychiatry, 52(4), 296-303.
Korver, N., Quee, P. J., Boos, H., Simons, C. J., & Haan, L. (2012). Genetic Risk and Outcome of Psychosis (GROUP), a multi site longitudinal cohort study focused on gene
environment interaction: objectives, sample characteristics, recruitment and assessment methods. International Journal of Methods in Psychiatric Research, 21(3), 205-221.
Korver-Nieberg, N., Berry, K., Meijer, C., de Haan, L., & Ponizovsky, A. M. (2015).
Associations between attachment and psychopathology dimensions in a large sample of patients with psychosis. Psychiatry Research, 228(1), 83-88.
Korver-Nieberg, N., Fett, A. K. J., Meijer, C. J., Koeter, M. W., Shergill, S. S., de Haan, L., & Krabbendam, L. (2013). Theory of mind, insecure attachment and paranoia in
adolescents with early psychosis and healthy controls. Australian and New Zealand Journal of Psychiatry, 47(8), 737-745.
MacBeth, A., Gumley, A., Schwannauer, M., & Fisher, R. (2011). Attachment states of mind, mentalization, and their correlates in a first-episode psychosis sample. Psychology and Psychotherapy: Theory, Research and Practice, 84, 42–57.
Martin, J. A., & Penn, D. L. (2001). Social cognition and subclinical paranoid ideation. British Journal of Clinical Psychology, 40(3), 261-265.
Mikulincer, M., & Shaver, P. R. (2012). An attachment perspective on psychopathology. World Psychiatry, 11(1), 11-15.
Moutoussis, M., Williams, J., Dayan, P., & Bentall, R. P. (2007). Persecutory delusions and the conditioned avoidance paradigm: towards an integration of the psychology and biology of paranoia. Cognitive Neuropsychiatry, 12(6), 495-510.
Persons, J. B. (1986). The advantages of studying psychological phenomena rather than psychiatric diagnoses. American Psychologist, 41(11), 1252
Ponizovsky, A. M., Nechamkin, Y., & Rosca, P. (2007). Attachment patterns are associated with symptomatology and course of schizophrenia in male inpatients. American Journal of Orthopsychiatry, 77(2), 324.
Pos, K., Bartels-Velthuis, A. A., Simons, C. J., Korver-Nieberg, N., Meijer, C. J., de Haan, L., & van Os, J. (2015). Theory of Mind and attachment styles in people with psychotic
disorders, their siblings, and controls. Australian and New Zealand Journal of Psychiatry, 49(2), 171-180.
Read, J., & Gumley, A. (2010). Can attachment theory help explain the relationship between childhood adversity and psychosis. Telling stories, 51-94.
The Relationships Questionnaire. Retrieved January 3, 2016, from www.fetzer.org
Velthorst, E., Meijer, C., & GROUP Investigators. (2012). The association between social anhedonia, withdrawal and psychotic experiences in general and high risk
Velthorst, E., Nieman, D. H., Becker, H. E., van de Fliert, R., Dingemans, P. M., Klaassen, R., & Linszen, D. H. (2009). Baseline differences in clinical symptomatology between ultra high risk subjects with and without a transition to psychosis. Schizophrenia
Research, 109(1), 60-65.
Vollema, M. G., & Ormel, J. (2000). The reliability of the Structured Interview for Schizotypy Revised. Schizophrenia Bulletin, 26(3), 619.
Wilson, J. S., & Costanzo, P. R. (1996). A preliminary study of attachment, attention, and schizotypy in early adulthood. Journal of Social and Clinical Psychology, 5, 231–260.
Attachment 1. Relationship Questionnaire (Fetzer)
Scale:
Following are four general relationship styles that people often report. Place a checkmark next to the letter corresponding to the style that best describes you or is closest to the way you are
.
____ A. It is easy for me to become emotionally close to others. I am comfortable depending on them and having them depend on me. I don’t worry about being alone or having others not accept me.
____ B. I am uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I worry that I will be hurt if I allow myself to become too close to others.
____ C. I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don’t value me as much as I value them.
____ D. I am comfortable without close emotional relationships. It is very important to me to feel independent and self-sufficient, and I prefer not to depend on others or have others depend on me.
