Perfectionism and Personality Disorders as Predictors of Symptoms and Interpersonal Problems

Introduction

Perfectionism reflects an unflinching pursuit of unusually high levels of achievement, for example, in the areas of financial, athletics and/or academia. In general, perfectionism has been considered, a multidimensional, rather than singular, personality trait. Frost and colleagues (1990) suggested perfectionism is composed of six semi-independent elements including 1) setting high standards for performance, 2) having negative reactions to projected punishments for errors 3) having feelings of inferiority when anticipating an error, 4) negatively perceiving parents critiques of any flaws, 5) doubting one’s performance, and 6) being overly concerned with organization and order. Hewitt and Flett (1991) suggested perfectionism is made up of three elements: self-oriented perfectionism, which involves both setting high standards for achievement and self-criticizing when there is a perception these standards are unmet; other-oriented perfectionism, which involves setting unrealistically high standards for other people, and socially prescribed perfectionism, which involves a perception that others hold unrealistically high expectations about the self.

Perfectionism may play a causal role in the development of different forms of symptom disorders (Egan, Wade & Shafran, 2011; Harvey et al., 2004; Sassaroli et al., 2008). For instance, persons with eating disorders exhibit heightened levels of perfectionism (Bardone-Cone, Wonderlich, Frost et al., 2007) that may predate the onset of illness (Lilenfield et al., 2006). Perfectionism is correlated with levels of depression and suicidal ideation (Hewitt, Flett, Sherry & Caelian, 2006) and mood instability in bipolar disorders (Alloy, Abramson, Walshaw et al., 2009). Heightened perfectionism has also been found in individuals experiencing general anxiety (Frost & DiBartolo, 2002), social anxiety (Juster et al., 1996) and obsessive compulsive disorder (Obsessive Compulsive Cognitions Working Group, 1997).

Perfectionism may also contribute to social dysfunction among persons with a psychiatric condition. The link between perfectionism and symptoms of depression has been suggested to be mediated by negative social interactions, avoidant coping and the perception of a lack of social support (Dunkey, Zuroff & Blankstein, 2003). A relationship between concern about mistakes and doubts concerning actions with general symptomatology was found to be mediated by maladaptive coping (Park et al., 2010).

Perfectionism may also play a role in many personality disorders ([PDs¹] Ayearst et al., 2012; Benjamin, 1996). Obsessive compulsive PD [OCPD] often involves setting high standards and battling procrastination, which is a result of doubt and an association between self-oriented perfectionism and compulsivity. This has been found in non-clinical (Hewitt & Flett, 1991; Sherry et al., 2007) and clinical (Grilo, 2004) samples. Perfectionism also appears in narcissism in the form of high standards for both oneself and others and harsh criticism of others (McCown & Carlson, 2004). Persons with depressive PD excessively blame themselves for any setbacks and are judgmental of others (Huprich, Porcerelli, Keaschuk, Binienda & Engle, 2008). Heightened socially prescribed perfectionism has also been found in borderline, avoidant, and dependent PD (Hewitt & Flett, 1991; Hewitt, Flett & Turnbull, 1992). Persons with avoidant PD are self-critical and expect to reach unreasonably high standards of performance in social situations, and they fear criticism, which may correspond to internalized criticism by their parents. Individuals with passive-aggressive PD tend to be overly critical of others and themselves. Their concern about making mistakes often hampers the ability to make decisions.

Personality disorders and perfectionism may interact and intensify one another (Sherry et al., 2007). The presence of both perfectionism and OCPD is a risk factor for the development of eating disorders (Goodwin, Haycraft, Willis & Meyer, 2011), and the combination of perfectionism and narcissism has been linked to unneeded cosmetic surgery (Fitzpatrick et al., 2011).

Perfectionism has also been linked to social dysfunction, even in those without diagnosed psychopathologies. In non-clinical samples maladaptive perfectionism (the triad consisting of excessive concern about making mistakes, doubts about which action to take and expected harsh criticism from others) has been linked to greater levels of hostile-dominant and friendly submissive forms of interpersonal problems (Slaney et al., 2006), poorer quality of marital relationships (Haring, Hewitt & Flett, 2003), interpersonal sensitivity, hostility, paranoia (Hill et al., 2004), and interpersonal aversive behavior (Habke & Flynn, 2002).

Though evidence suggests that perfectionism is related to symptoms and social dysfunction many questions about it remain unanswered: For example, is maladaptive perfectionism linked to overall suffering beyond what is accounted for by the association with specific symptoms? In the area of PD correlations of perfectionism with personality pathology have been explored mostly in non-clinical samples (e.g. Sherry et al., 2007); possible links with PD, and the correlations of both with symptoms and relational dysfunctions, need to be explored in a treatment-seeking population.

Understanding the role of perfectionism in PD in clinical populations may capture aspects of the PD that are not fully accounted for in DSM (American Psychiatric Association, 2013) and yield more targeted therapies.

We assessed perfectionism, symptoms and PD, and interpersonal functioning in both treatment populations and community samples. First, we predicted that higher levels of maladaptive perfectionism (concern over mistakes, doubts about action, and [anticipated?] parental criticism) would appear in a clinical sample compared to the community sample. Second, we expected that perfectionism would be linked with global psychopathology in both groups. Third, we estimated levels of perfectionism would be correlated to the number of PD traits in both groups. Fourth, we predicted that perfectionism would be correlated with heightened levels of interpersonal problems in both groups. Fifth, we anticipated that concern over mistakes, doubts about actions and parental criticism would make a unique contribution in predicting general psychopathology and interpersonal problems, beyond the one explained by the frequency of PD traits.

Methods

Participants

The clinical sample (n=93) was composed of outpatients from a variety of private treatment centres, and they were receiving either pharmacotherapy or cognitive behavioural therapy. Exclusion criteria were diagnoses of psychotic disorder, bipolar I disorder, or alcohol or drug dependence. The community sample (n=100) was recruited through advertisements in schools and other institutions or by referral of other participants. The mean ages and male to female ratios for the clinical and nonclinical participants respectively were 35.57 (sd = 10.28); 41 male/52 female and 32.95 (sd = 11.39); 31 male/69 female.

Measures

The Assessment of DSM-IV Personality Disorders Italian Version. (ADP-IV; Pedone et al., 2005; Schotte, De Doncker, Vankerckhoven, Vertommen, & Cosyns, 1998) is a 94-item questionnaire, which assesses PD according to the DSM-IV criteria. For each criterion, the patient self-rates its typicality. The ADP-IV allows categorical and dimensional diagnostic formats. In this study, we only used the dimensional trait scores. Psychometric properties have been reported elsewhere (Schotte et al., 1998; Pedone et al., 2005).

The Frost Multidimensional Perfectionism Scale. ([FMPS] Frost et al., 1990) is a 35-item self-report questionnaire which includes six subscales: concern over mistakes, doubts about actions, personal standards, parental expectations, parental criticism, and organization. The concern-over-mistakes subscale refers to negative reaction to mistakes and perceptions of even minor errors as failure. Doubts about actions refers to an over repeated doubting about the quality of one’s performance. Personal standards describe the tendency to set excessively high standards. Parental expectations and parental criticism refer to perceiving one’s parents as having high expectations or being excessively critical (Frost et al., 1990). The FMPS has been reported to have acceptable reliablility and validity (Enns & Cox, 2002).

The Symptom Checklist-90-R. ([SCL-90-R] Derogatis, 1977) is a 90-item self-report inventory which assesses nine primary symptom dimensions. For the purposes of this study we were interested in the estimate of global psychopathology.

The Inventory of Interpersonal Problems (IIP-47) Pilkonis, Kim, Proietti & Barkham, 1996; Italian version Ubbiali, Donati & Chiorri, 2011) is a 47 item self-report scale that assesses interpersonal problems, and consists of five subscales: interpersonal sensitivity, interpersonal ambivalence, aggression, need for social approval, and lack of sociability.

Analyses

Analyses were planned in four phases. First, we planned to compare the perfectionism scores of the clinical and community samples. Second, we planned to examine whether perfectionism was linked to demographic variables. Third, we planned to correlate perfectionism scores with PD traits at the cluster and disorder level, global symptoms, and social function levels, controlling for demographics if necessary. Given the large number of correlations we adjusted the alpha level to p > .01. Finally, to determine whether perfectionism was linked to symptoms and social function independent of PD traits, we planned stepwise multiple regressions, two each for the clinical and for the nonclinical groups. One regression predicting the general symptoms severity index from the total of cluster A, B and C traits in a first step. This was followed by the perfectionism subscale scores in the second regression, predicting the summary score for social function from the total of cluster A, B and C traits in a first step followed by the perfectionism subscale scores.

Results

Mean and standard deviations of perfectionism for the clinical and community groups are presented in Table 1. The clinical sample had higher perfectionism scores on the scales for personal standards, excessive concerns over mistakes, parental criticism, parental expectations, and doubts about action. There were no differences in the organization scale and so this scale was not further considered. Correlations and t-tests indicated that, in the clinical group, age was not significantly related to perfectionism scores; however, women had higher perfectionism scores (p > .05) than men on the five subscales under investigation. For the community group, men and women did not differ on perfectionism scores, and age was not related to perfectionism. Partial correlations were then performed exploring the relationship of perfectionism with number of PD traits by cluster and diagnosis, psychopathology and interpersonal problems for the clinical group with gender entered as a covariate. As revealed in Table 2, a broad pattern of significant relationship appeared, with higher levels of multiple aspects of perfectionism predicting heightened levels of PD traits, poorer social function, and higher psychopathology.

The correlations between the same variables for the community sample were next conducted and are presented in Table 3. Since demographics were not related to perfectionism in the community sample, no covariates were included in these analyses. Here again, a broad pattern of significant relationships revealed that among the nonclinical group multiple aspects of perfectionism predicted height ened level of PD traits, poorer social function, and great levels of psychopathology.

Finally, we examined whether the association of perfectionism in the clinical and nonclinical groups were related to social function and symptoms independent of PD traits. Four stepwise multiple regressions were performed. In the first regression IIP total scores of the clinical group were predicted in two steps: PD traits from clusters B and C were entered in the first step, followed by the perfectionism subscales. In the second regression the SCL90 GSI of the clinical group was predicted by the same two steps as in the first regression. Greater levels of cluster B traits and greater scores on the doubts-about-actions subscale were both uniquely related to poor levels of social function total for the clinical group (Table 4). Greater levels of cluster C traits and greater scores on the excessive-concern-overmistakes subscale were related to heightened symptoms for the clinical group.

The final two regressions repeated the procedures of the first two only for the community sample. In these analyses social functioning was predicted by greater levels of both cluster B and C traits and higher concern over mistakes for the community sample. Symptoms for the community sample were predicted by greater levels of cluster A personality disorder traits followed by higher concern over mistake (Table 5).

Discussion

In this study we examined whether participants in a clinical sample demonstrated heightened perfectionism relative to participants in a community sample and whether perfectionism broadly correlated with levels of psychopathology, the number of PD traits, and report of interpersonal problems in both samples. Analyses revealed that the clinical sample had higher levels of concern over mistakes, doubts about actions, parental criticism, personal standards and parental expectations. Perfectionism was also found to correlate with the number of Axis II PD traits in both groups. Levels of perfectionism were also linked with overall psychopathology. Individuals who reported worrying about criticism and having trouble making decisions appear to be prone to higher levels of emotional discomfort.

Within the clinical group concern over mistakes was associated with eight out of 10 DSM IV PD traits, doubts about actions were correlated with five out 10 types of PD traits, mostly from cluster C, parental criticism was correlated with six PD traits. Regarding the relationships between perfectionism and interpersonal, we found fear of making mistakes was related to many kinds of interpersonal dysfunctions, ranging from the opposite tendencies to avoidance of social contacts or attacks, and inability to cooperate and being easily hurt by negative comments. Over concern about making errors was linked with a longing for approval. Doubts about actions were almost as highly and generally correlated with interpersonal issues, with the exception of aggressive tendencies. Parental criticism was only associated with lack of sociability, suggesting an avoidance mechanism of social contacts aimed at protecting the self against the impact of negative comments. All these correlations persisted even after being controlled for the possible effects of gender differences.

In the community sample results were even more striking: Concern over mistakes, doubts about actions, and parental criticism were linked to symptom severity, almost any PDs, and interpersonal problems. Even in persons not seeking for treatment perfectionism this seems to be associated with distress and social dysfunction, which is consistent with the findings of Park and colleagues (2010).

It’s important to note that our results only partially support our hypothesis that perfectionism would be linked to symptoms and interpersonal problems beyond correlation with PD traits. In the clinical group, the level of social dysfunction was related to the number of cluster B traits and with over concern about mistakes. By contrast, general symptom severity was correlated with the number of cluster C traits and doubts about action. In the non-clinical, community group worrying about making errors was significantly linked with the overall level of psychiatric symptoms and interpersonal problems, even when the number of PD traits were controlled for statistically. In the clinical group concern over mistakes was a predictor of problems with cluster B traits only. One interpretation is that having a negative image of oneself may lead to a fear of making mistakes, which in turn leads to insecurity in relationships and social problems. Doubts about actions may also contribute to the symptom aspects in individuals who have prominent cluster-C features. It may be that those persons are prone to symptoms, such as obsessions or depression, because they are inclined to ruminate about which actions to take. However, these findings are weaker than we expected. We would have predicted that both concern over mistakes and doubts about actions were predictors of symptoms and interpersonal problems in the different clusters. Future studies should investigate such a relation. Overall, our findings open the idea that PD and perfectionism form a toxic interaction leading to more severe symptoms and poorer interpersonal functioning.

One interpretation of our findings is that high levels of perfectionism present a risk of developing heightened levels of psychopathology and to interpersonal problems. Maladaptive perfectionism appears as a problem for many, with heightened levels of PD traits. Striving for perfection, fearing criticism when making mistakes, and doubting one’s actions may be elements of many PDs that are not yet acknowledged by current classifications (Ayearst et al., 2012). For example, narcissism was associated not only with both high standards and parental expectations (as we expected), but also with concern over mistakes, which paints a portrait of a person concealing feelings of failure (McHoskey et al., 2001; Sherry et al., 2006). The associations of perfectionism with specific PDs need to be confirmed in larger samples, where many people meet full a PD-specific diagnosis (our analysis has been conducted at the level of PD traits only). The prototype of many personality disorders, such as avoidant and borderline, may end up including perfectionism.

Avenues are open for future research regarding the interactions between perfectionism and maladaptive functioning. Persons with personality dysfunctions, and perhaps tendencies to social withdrawal or antagonism, may, as a consequence, become perfectionistic as a strategy to cope with social challenges. On the other hand, learnt perfectionistic schemas may pave the way for personality dysfunctions as the person’s sole focus on meeting unrealistically standards makes impossible to adjust to the complex need of social relating. The question remains open about the possible path leading to symptoms and poor social functioning. Perfectionism may reflect a tendency to maladaptive behaviors and regulatory strategies that underlie personality pathology and eventually lead to symptoms and social problems. It is possible that some cognitive and affective processes at the foundations of many PD may lead to perfectionistic coping strategies, which, in turn, lead to distress and maladaptation. Future studies are needed to explore such links.

There were unexpected findings. One we have already noted is that perfectionism in predicting symptoms beyond the PD traits was less than expected. We found that in the clinical sample doubts about action predicted symptoms only in persons with cluster-C PD traits. The parental-criticism aspect (which was related to symptoms and social problems) did not explain a significant portion of the variance of symptoms and interpersonal problems in persons with heightened levels of PDs. Moreover, and largely against expectations, the associations of perfectionism and OCPD were marginal and limited to concern over mistakes, though doubts about actions as considered a prominent feature of this disorder. Future research should investigate correlations of OCPD and multidimensional perfectionism in clinical populations meeting full criteria for the disorder. Also, associations of parental criticism and concern over mistakes with paranoid features was unexpected, though some studies found such a relationship (Hill et al., 2004).

There are limitations to this study. One, was that the correlational nature of the study precluded any implications about causality. Our findings may be considered preliminary at best in this domain due to lack of statistical power. Our study relied on self-reports, which may have led to inflated correlations due to shared-method variance. Also some PDs may have been under-or over-diagnosed because of the self-report assessment measure. Results need to be replicated with studies using diagnostic assessment interviews.

Moreover, patients were assessed at various times—both at intake and during treatment; this may have biased the results in any direction depending on time spent in therapy and effectiveness of treatment. Confounding factors, such as time in treatment, type of treatment administered, or type and dosage and medication for the clinical group, were not collected. The therapists were from different orientations of cognitive therapies and there were differences in years of expertise and adherence to treatment. Finally, we analysed patients only at the level of personality traits; therefore, to better explore possible correlations of perfectionism with specific disorders, future work will need to focus on individuals who meet full diagnostic criteria for psychopathologies. There is the need to assess individuals before they enter therapy to avoid any therapy effects on the target variables.

Overall our findings are consistent with previous studies in community samples, and they expand knowledge in the treatment-seeking population, with the unexpected exception of obsessive-The correlations of perfectionism and PD we found suggest a careful treatment option includes analyzing processes of perfectionism and PD together. Great care needs to be paid to the therapeutic relationship. It is essential to minimize the risk of having persons with PD and perfectionism view the therapist a harsh critic and therapy assignments as performances to be judged. Based on our finding, it appears that perfectionism is one prominent target for refining PD treatments.