پرسشنامه کمال گرایی بالینی: شواهد بیشتر برای گرفتن دو عامل عزم و نگرانی کمال گرایی

The construct of clinical perfectionism, conceptualized as a one-dimensional construct (Shafran, Cooper, & Fairburn, 2003), has drawn considerable debate because of the associated critique of multidimensional perfectionism’s relevance for clinical theory and research. Hence Dickie, Surgenor, Wilson, and McDowall’s (2012) finding that the Clinical Perfectionism Questionnaire (CPQ), designed to measure clinical perfectionism, was two-factorial and thus multidimensional makes an important contribution to the debate. The present study aimed to replicate Dickie et al.’s finding examining the CPQ’s factorial structure in 316 university students. In addition, the study examined the CPQ’s convergent correlations with dispositional perfectionism, perfectionism cognitions, and perfectionistic self-presentation. CPQ total scores showed large-sized positive correlations with dispositional perfectionism and perfectionism cognitions demonstrating convergent validity. However, the study confirmed that the CPQ was two-factorial with Factor 1 mainly capturing perfectionistic strivings and Factor 2 mainly capturing perfectionistic concerns, thus questioning the CPQ’s construct validity. Together with Dickie et al.’s findings, the present study’s findings suggest that—if the CPQ is a valid measure of clinical perfectionism as conceptualized by Shafran et al. (2003)—the construct of clinical perfectionism and its dimensionality need to be reconsidered.

1.1. Multidimensional perfectionism Perfectionism is a personality disposition characterized by striving for flawlessness and setting exceedingly high standards of performance accompanied by overly critical evaluations of one’s behavior (Frost et al., 1990 and Hewitt and Flett, 1991). Over the past 20 years, research has produced converging evidence that perfectionism has many facets and is best conceptualized as a multidimensional characteristic (see Enns & Cox, 2002, for a review). In particular, two main dimensions have been differentiated: perfectionistic strivings (also termed personal standards perfectionism) comprising perfectionistic personal standards and a self-oriented striving for perfection and perfectionistic concerns (also termed evaluative concerns perfectionism) comprising concern about making mistakes, feelings of discrepancy between one’s standards and performance, and fears of negative evaluation and rejection by others if one fails to be perfect (see Stoeber & Otto, 2006, for a review). 1.2. Clinical perfectionism and the CPQ Questioning the relevance of multidimensional perfectionism for clinical theory and practice, Shafran, Cooper, and Fairburn (2002) put forward the alternative concept of clinical perfectionism defined as “the overdependence of self-evaluation on the determined pursuit of personally demanding, self-imposed, standards in at least one highly salient domain, despite adverse consequences” (p. 778). Moreover, Shafran and colleagues stressed that clinical perfectionism was not multidimensional perfectionism ( Shafran et al., 2003) in opposition to the widely accepted view that perfectionism is best conceptualized as multidimensional and that multidimensional conceptions of perfectionism also capture clinically relevant aspects of perfectionism (e.g., Hewitt, Flett, Besser, Sherry, & McGee, 2003). To measure clinical perfectionism, Fairburn, Cooper, and Shafran (2003) developed the 12-item Clinical Perfectionism Questionnaire (CPQ; see Table 1). Yet, even though the CPQ has been employed in clinical research for 10 years (e.g., Shafran, Lee, & Fairburn, 2004), little is known about the CPQ’s reliability and validity. Only recently studies have been published examining the CPQ’s reliability and validity. The findings are mixed. On the one hand, Steele, O’Shea, Murdock, and Wade (2011) and Chang and Sanna (2012) found the CPQ total scores to show good reliability (Cronbach’s αs = .83) and convergent validity displaying large-sized positive correlations 1 with personal standards perfectionism, evaluative concerns perfectionism, and the subscales of the Hewitt–Flett Multidimensional Perfectionism Scale (HF-MPS). Moreover, the CPQ showed incremental validity predicting variance in depressive symptoms above the variance explained by personal standards, evaluative concerns perfectionism, and the HF-MPS subscales. On the other hand, Steele et al. (2013) found the CPQ total scores to show only satisfactory reliability (α = .70). Furthermore, Dickie, Surgenor, Wilson, and McDowall (2012) found Item 8 to show positive correlations with half of the CPQ items and negative correlations with the other half. They also found an overall low item-total correlation for Item 8 and thus suggested excluding the item from the CPQ. More importantly, when employing exploratory factor analysis (EFA) on the CPQ items, Dickie and colleagues found the CPQ to show a two-factorial structure which—if clinical perfectionism is conceptualized as one-dimensional, not multidimensional—raised questions about the CPQ’s construct validity. Table 1. Clinical Perfectionism Questionnaire (CPQ) items: Item analysis and exploratory factor analyses. EFA 1 EFA 2 Items: Over the past month,... CITC F1 F2 F1 F2 1. Have you pushed yourself really hard to meet your goals? .42 .73 –.19 .73 –.27 2. Have you tended to focus on what you have achieved, rather than on what you have not achieved? (R) .02 .21 –.47 .22 –.36 3. Have you been told that your standards are too high? .46 .64 .08 .55 .01 4. Have you felt a failure as a person because you have not succeeded in meeting your goals? .50 .27 .78 .09 .79 5. Have you been afraid that you might not reach your standards? .50 .33 .71 .17 .65 6. Have you raised your standards because you thought they were too easy? .28 .42 .13 .32 .07 7. Have you judged yourself on the basis of your ability to achieve high standards? .55 .47 .56 .34 .46 8. Have you done just enough to get by? (R) –.06 –.32 .59 –.35 .50 9. Have you repeatedly checked how well you are doing at meeting your standards (for example, by comparing your performance with that of others)? .53 .62 .32 .52 .21 10. Do you think that other people would have thought of you as a “perfectionist”? .49 .71 .02 .66 –.07 11. Have you kept trying to meet your standards, even if this has meant that you have missed out on things? .53 .74 .05 .71 –.05 12. Have you avoided any tests of your performance (at meeting your goals) in case you failed? .24 .13 .50 .05 .37 Note. N = 316. Underlined words as in the original ( Fairburn et al., 2003). CITC = corrected item-total correlation; EFA 1 = exploratory factor analysis (EFA) using principal components analysis and varimax rotation (following Dickie et al., 2012); EFA 2 = EFA using principal axis factoring and promax rotation (following Russell, 2002); F1 = Factor 1, F2 = Factor 2. EFA 2 loadings are from the pattern matrix, and |loadings| > .30 are boldface. (R) = reverse-scored item. Answers to Items 2 and 8 were reversed when computing CITCs, but were not reversed in the EFAs to aid the interpretability of positive vs. negative loadings (see 3.2). Table options 1.3. Limitations of Dickie et al.’s (2012) study Dickie et al.’s (2012) study is the first to explore the factorial structure of the CPQ and thus makes an important contribution to research on clinical perfectionism and the question of whether clinical perfectionism is one- or multidimensional. However, the study had a number of limitations. First, the study excluded Item 8 from the initial EFA (because of the low item-total correlation) and Item 7 from the final EFA (because it displayed substantial loadings on both factors). Consequently, the factorial structure of the CPQ including all 12 items still remained to be explored. Second, the study employed principal components analysis (PCA) as factor extraction method and only reported factor loadings for the two-factorial solution after orthogonal varimax rotation. However, PCA is not regarded a proper factor analysis, and orthogonal rotation—assuming uncorrelated factors—is not recommended if factors are expected to be correlated (e.g., Fabrigar et al., 1999 and Russell, 2002). Dickie et al. followed up their orthogonal rotation with an oblique rotation and found a small, but significant correlation (r = .24). However, the two subscales they derived from their factor solution showed a larger correlation (r = .39), and their EFA did not include Item 7 (which showed substantial loadings on both factors). Hence an EFA including all 12 CPQ items with oblique rotation could be expected to show higher factor correlations. Finally, Dickie et al. following the results of their final EFA computed two CPQ subscales (Subscale 1 comprising Items 1, 3, 6, 9, 10, and 11; Subscale 2 comprising Items 2, 4, 5, and 12) and investigated the subscales’ convergent validity regarding perfectionistic strivings (personal standards) and perfectionistic concerns (concern over mistakes and doubts about actions). Subscale 1 showed a larger correlation with perfectionistic strivings than with perfectionistic concerns whereas Subscale 2 showed a larger correlation with perfectionistic concerns than with perfectionistic strivings. However, they did not investigate the convergent validity of the CPQ total scores. Hence the convergent validity of the CPQ remained to be further explored. 1.4. The present study Replicability is an essential criterion for psychological research to ascertain that empirical findings are reliable and valid. Because the CPQ is the only questionnaire currently available to measure clinical perfectionism, it was important to replicate Dickie et al.’s (2012) finding of a two-factorial structure and further investigate the nature of the two factors. Furthermore, it was important to examine the factorial structure of the CPQ and the convergent validity of the CPQ total scores including all 12 items because—except for Dickie et al.’s study—all studies using the CPQ to measure clinical perfectionism computed CPQ total scores based on responses to all 12 items. Consequently, the present study examined the factorial structure of the CPQ including all 12 items and investigated the convergent validity of the CPQ total scores in relation to measures capturing different aspects of dispositional perfectionism and multidimensional perfectionism cognitions. Because in the limitation section of their article Dickie et al. discussed the possibility of self-presentation influencing the CPQ scores, the present study also included measures of perfectionistic self-presentation differentiating promotion-focused (perfectionistic self-promotion) and prevention-focused self-presentation (nondisplay and nondisclosure of imperfection) to examine the CPQ scores’ relationships with perfectionistic self-presentation.