Reaching below the surface: Automatic perfectionism associations in students and persons suffering from an eating disorder

Reaching below the surface: Automatic perfectionism associations in students and persons suffering from an eating disorderKathleen De CuyperProf. Dr. Dirk Hermans (supervisor), Prof. Dr. Guido Pieters, Prof. Dr. Laurence Claes (cosupervisors) This doctoral project aims at investigating the psychometric properties of newly developed measures of perfectionism. In the literature, the construct of perfectionism has been mainly studied at the self-reported level, differentiating between an achievement striving and evaluative concerns dimension, which are inter-correlated. However, the self-knowledge of people shows considerable blind spots. Therefore, questionnaires assessing aspects of perfectionism have to be complemented with indirect measurement tasks. Typically for these tasks is that they assess the research object without relying on introspection and self-report with respect to the to-be-measured construct.This doctoral project includes a theoretical and an empirical part. First, in a systematic and narrative review, the studies using indirect measurement tasks to assess the implicit self-concept of personality are discussed. This review informs implicit social cognition researchers, personality researchers and clinicians about the advances that have been made in the construction of valid implicit personality measures.The empirical part of this dissertation consists of the assessment of aspects of perfectionism in students and in persons suffering from an eating disorder (ED) at three levels: at the level of self-report, of behaviour, and of automatic associations. First, at the level of self-report, the factor structure and construct validity of the Dutch Multidimensional Perfectionism Scale of Hewitt and Flett (1991) were investigated in a large student sample. Our study confirmed the original three-factor structure, at least when a method factor that refers to the negatively keyed items was included in the three-factor model. Second, at the level of overt behaviour, the construct validity of the extended letter-detection task (Tallis, Eysenck & Mathews, 1991) and the drawing task were examined in students and in persons suffering from an ED. Both tasks discriminated between the two samples from which is known that they differ with respect to the two self-reported dimensions of perfectionism. However, in contrast to the hypothesis, the tasks did not directly correlate with the self-reported dimensions of perfectionism. Third, at the level of automatic associations, the psychometric qualities of the Aiming at Perfection Single Category Implicit Association Test (Karpinsky & Steinman, 2006) and the Failing/Suffering Implicit Association Test (F/S IAT; Greenwald, McGhee, & Schwartz, 1998) were investigated in students and in ED patients. (1) The Aiming at Perfection SC-IAT measured the extent to which people - at an automatic level - evaluate others asmore achievement striving compared to themselves. Evidence for the construct validity is mostly lacking at the moment. However, the more others were evaluated as more achievement striving compared to the selves, the worse was the exam performance of the students at the end of the academic year, and the slower the anorexia nervosa patients gained weight during their inpatient treatment. (2) The Failing/Suffering IAT assessed the extent to which people evaluate failing as more negative than suffering, at an automatic level. The positive correlations between this IAT and the accuracy and time spent on the extended letter-detection task, in the students and in the ED patients, showed evidence for the construct validity of the Failing/Suffering IAT. Moreover, the Failing/Suffering IAT correlated positively with perfectionistic results on the drawing task in the ‘automatic processing’ condition of an experimental study, as was hypothesized. Regarding the predictive validity, the more failing was evaluated as more negative than suffering, the more ED symptomatology was reported after two months in the patients with short illness duration, and the slower the anorexia nervosa patients gained weight. Finally, the self-reported dimensions of perfectionism predicted self-reported outcome after a follow-up period of seven months. After controlling for the overlap between the two dimensions, the achievement striving dimension showed a positive effect on ED symptomatology. The evaluative concerns dimension had a negative effect on ED and other psychopathological symptoms.The construct and the predictive validity results of the perfectionism measures studied at the three levels, showed that perfectionism assessed as a feature and assessed as behaviour has adaptive and maladaptive consequences. Examining intervening variables between the perfectionism measures and criterion variables in future studies might give further insight in the valence of this concept.status: publishe