Web-Based CBT for Perfectionism PDF

WEB-BASED CBT FOR PERFECTIONISM 1 Web-Based Cognitive-Behavioural Therapy for Perfectionism: A Randomized Controlled Trial Chantal A. Arpin-Cribbie, PhD 1; Jane Irvine, D.Phil 2; Paul Ritvo, PhD 2, 3 1 Department of Psychology, Laurentian University at Georgian College 2 3 Department of Psychology,...

WEB-BASED CBT FOR PERFECTIONISM 1 Web-Based Cognitive-Behavioural Therapy for Perfectionism: A Randomized Controlled Trial Chantal A. Arpin-Cribbie, PhD 1; Jane Irvine, D.Phil 2; Paul Ritvo, PhD 2, 3 1 Department of Psychology, Laurentian University at Georgian College 2 3 Department of Psychology, York University School of Kinesiology and Health Science, York University Published in Psychotherapy Research, 2012 Mar; 22(2):194-207 WEB-BASED CBT FOR PERFECTIONISM 2 Abstract This study assessed the effectiveness of a web-based cognitive behavioural intervention (CBT) in reducing perfectionism and psychological distress in post-secondary students. Participants assessed as high in perfectionism (N = 77) were randomized to 1 of 3 ten-week, web-based, intervention conditions (no treatment [NT], general stress management [GSM], or [CBT]). Results indicated the CBT condition was effective in reducing perfectionism, and supported a pattern of significantly greater improvement than observed in participants in the GSM or NT conditions. While both CBT and GSM demonstrated capacities to significantly reduce distress, for CBT participants, changes in perfectionism were significantly correlated with changes in depression and anxiety. Results offer support for the effectiveness of web-based CBT in positively affecting perfectionist-related problems. Given the considerable proportion of individuals who suffers from perfectionism related distress, the intervention’s apparent effectiveness, cost-effectiveness and ease of dissemination warrant future replication studies. Keywords: perfectionism, treatment, self-directed intervention, web-based, distress WEB-BASED CBT FOR PERFECTIONISM 3 Web-Based Cognitive-Behavioural Therapy for Perfectionism: A Randomized Controlled Trial Research suggests that perfectionism-related distress is a prevalent problem for students at post-secondary institutions (Grzegorek, Slaney, Franze, & Rice, 2004). While assessment methods and definitions may vary, many investigators and clinicians use cut-off scores on established scales, like the Perfectionism Cognitions Inventory (PCI, [Flett, Hewitt, Blankstein, & Gray, 1998]) to gauge the degree to which perfectionism is ‘problematic’. Perfectionism, in this regard, is indexed by the frequency of ‘automatic perfectionistic thoughts’ reflecting comparative evaluations between the currently perceived self and an ideal or ‘perfect’ self (Flett et al., 1998; Flett, Hewitt, Whelan, & Martin, 2007). These cognitive comparisons are, in turn, theoretically linked to attempts to avoid or reduce anxiety through the achievement of perfection which, when experienced as impossible, results in greater distress (Hewitt & Flett, 2002). Perfectionism indexed in this way is associated with lowered self-esteem (Hewitt, Flett, & Ediger, 1995), chronic headache (Bottos & Dewey, 2004), eating disorder (e.g., Cassin & Von Ranson, 2005), and both anxiety (e.g., Flett, Madorsky, Hewitt, & Heisel, 2002), and depressive disorders (e.g., Flett, Hewitt, Blankstein, & Mosher, 1995). Interventions to reduce perfectionism-related distress may thus be valuable in the near term, in directly reducing negative effects, and in the longer term, as perfectionism addressed effectively and in early developmental stages may spare individuals later, more severe problems. Post-secondary (university) students seem particularly vulnerable to perfectionismrelated distress. There currently exists a greater number of students in Canada opting to pursue some form of post-secondary education, and a significant lag in the capacity of Canadian universities to adequately support this trend (Frenette, 2006; Statistics Canada, 2006). As a WEB-BASED CBT FOR PERFECTIONISM 4 result, a large proportion of undergraduate students, particularly in their first years, are attending very large classes where conditions are more impersonal (Educational Advisory Committee, 2006; Gilbert, 1995). They continue, however, to perceive their environment and the conditions required for their success as more stringent and competitive than ever (Van Bavel, 2002). It is within the context of these environments and their perceived demands that post-secondary students can come to assume cognitive patterns that are likely to pose additional challenges with respect to their overall psychological well-being. Particularly during the initial university years, when unfamiliarity leads to increased stress and academic anxiety, the pursuit of perfection may be a tempting refuge while, simultaneously, pursuits result in reduced academic performance (Arthur & Hayward, 1997; Halgin & Leahy, 1989). Various conceptualizations explain how perfectionism can lead to poorer or better adaptation. The most dominate of these are the multidimensional theories of perfectionism outlined by Hewitt and Flett (1991), and by Frost, Marten, Lahart, and Rosenblate (1990). These theories account for both intrapersonal and interpersonal aspects in examining the costs and benefits of perfectionism in relation to individual adaptation. While perfectionism has been linked to adjustment and achievement, it has been more often associated with negative outcomes such as failure, guilt, and shame, and to serious pathologies like depression, alcoholism, anxiety, and personality disorders (Flett, Hewitt, Garshowitz, & Martin, 1997; Hayward & Arthur, 1998; Hewitt & Flett, 1991; Hewitt, Mittelstaedt, & Flett, 1990). In addition, perfectionism has recently been linked with anxiety sensitivity by Flett, Greene and Hewitt (2004) whose data suggest linkage with components of perfectionism that are more interpersonal in nature, such as perfectionistic self-presentation. In some individuals, perfectionism may thus act as a protective mechanism, guarding against others (or themselves) uncovering the validity of core negative WEB-BASED CBT FOR PERFECTIONISM 5 beliefs (e.g., I am defective). The cognitive salience of perfectionism, manifested as automatic thoughts revolving around needs to be perfect, is consistently predictive of psychological distress and thus deserving of intervention targeting. Given the proposition that anxiety may underlie problematic perfectionism, and evidence suggesting cognitive-behavioural treatments effectively reduce anxiety sensitivity (Otto & Reilly-Harrington, 1999), Flett et al. (2004) have suggested that individuals high in perfectionism may benefit from CBT interventions focusing on the fear of fear and tendencies towards frequent, uncontrollable thoughts about the need to be perfect. In addition to the inherent value of attempting to decrease overall perfectionism levels, it is important for clinical interventions to focus on modifying cognitive patterns, such as excessive rumination, frequently evidenced in those high in perfectionism (Flett, Hewitt, Blankstein, & Gray, 1998). Moreover, the rumination patterns that require targeting often focus on a need to constantly attain perfection. Though many studies have examined the negative correlates and associated adjustment difficulties experienced by those high in perfectionism, few have empirically tested interventions aimed at modifying the psychological distress related to perfectionism. Of the studies completed, results support CBT effectiveness in reducing perfectionism (Kutlesa, 2002; Pleva & Wade, 2006; Riley, Lee, Cooper, Fairburn, & Shafran, 2007). While these results are encouraging, other researchers soberly suggest perfectionistic beliefs are difficult to modify and are associated with poorer psychotherapy outcomes overall (Blatt & Zuroff, 2002). Their data, for example, suggest that when depression and eating-disordered behaviours are primary treatment targets, perfectionism may require separate targeting to achieve primary target improvements (Frost & DiBartolo, 2002). Furthermore, few of the promising intervention studies have been designed as WEB-BASED CBT FOR PERFECTIONISM 6 randomized trials, with most following limited case study or quasi-experimental designs. It is important to carefully ascertain whether promising findings hold up when tested rigorously. Thus, an important study goal was to replicate previous intervention studies demonstrating CBT effectiveness in reducing perfectionism and related distress, employing a more rigorous randomized clinical trial design. Use of the internet has become almost universal, especially among young adults (e.g., Skinner, Biscope, & Blake, 2003). There has also been an increasing number of studies developing and evaluating web-based counseling interventions (Ritterband et al., 2003). Furthermore, previous findings have suggested that individuals high in perfectionism can negatively react to treatment termination, due to a sense of personal failure, discontentment, and disappointment with themselves and the therapy (Blatt & Zuroff, 2002). Thus, the inherent need of these individuals to retain control and independence may result in unfavorable responses to externally imposed treatment terminations. As such, in the current study, a self-directed intervention was developed to modify problematic perfectionistic beliefs. Though minimal study has been devoted to such interventions in specifically influencing perfectionism, there is evidence of effectiveness for other behavioural health problems (e.g., Ritterband et al., 2003). Given that web-based interventions may gain momentum and possibly significance in the future treatment of psychobehavioural problems, study results may contribute to a better understanding of the effectiveness of self-directed therapeutic interventions. Thus, this study of the effectiveness of a web-based psychoeducational CBT intervention in decreasing levels of perfectionism and distress in a student population was conducted as a randomized controlled trial where students identified with high scores on a measure of perfectionism (Perfectionism Cognitions Inventory [PCI]) were randomly assigned to 1 of 3 WEB-BASED CBT FOR PERFECTIONISM 7 conditions: (a) a web-based CBT intervention aimed at modifying perfectionistic beliefs; (b) a web-based general stress management (GSM) condition that served as the alternative treatment condition; or (c) a no treatment (NT) wait-list control group that received no intervention between pre and post measures in order to control for natural history effects, regression to the mean, and the effects of repeated testing. It was hypothesized that perfectionism, as well as depression, anxiety, and cognitive vulnerability to negative affect, would decrease more from baseline to posttest for the CBT group than for either the GSM or NT groups. Changes in perfectionism were also hypothesized to be significantly related to changes in psychological distress, from baseline to posttest, for the CBT group. Method Participants Randomization. Of the 141 eligible participants (i.e. with PCI scores greater than one standard deviation above the mean), 83 initially consented to participate. Of the 83 participants who completed the baseline measures, six did not complete the posttest measures and were excluded from the analyses. The final participants (N = 77) were distributed across all three treatment groups with 29 participants in the CBT group, 26 in the GSM group, and 22 in the NT group. One-way ANOVAs were used to evaluate whether the three experimental groups (CBT, GSM, NT) differed on any of the perfectionism or distress measures at pretest. There were no significant differences between any of the groups at pretest (i.e., all ps >.05). Sample characteristics. No significant differences were noted across groups on the demographic data collected, and as such the demographic data will be reported for the entire sample. Participants ranged between 18 to 48 years of age, with a mean age of 20.14 (sd = 4.14), WEB-BASED CBT FOR PERFECTIONISM 8 and 70% of the participants were female. The sample consisted primarily of individuals who identified themselves as being single (97.4%), and who were in either the first (70.1%) or second (20.8%) year of their university program. Interventions Cognitive behavioural intervention. While many definitions of perfectionism have emphasized an individual’s establishing extremely high standards against which they gauge their performance, the findings of a series of studies by Frost et al. indicated that ‘Concern over Mistakes’ (CM) is a key component of this concept (e.g. Frost & DiBartolo, 2002; Frost et al., 1990). These studies have also noted that CM was the most closely associated dimension to psychopathological symptomatology in individuals. The salient feature of CM is the degree to which an individual is capable of accepting minor shortcomings in their performance while still recognizing the more general quality of overall success in their performance (Frost et al., 1990). Increasing concern over mistakes may be characterized by a dichotomist thinking style whereby one’s performance is viewed either as perfect or as being completely insignificant, with little or no middle ground. This cognitive dimension was also addressed by Ellis (2002), who suggested that the problem with dysfunctional perfectionism is the irrational importance attached to perfectionism. Essentially, Ellis suggests that it is not one’s desire to succeed or excel, or even to perform perfectly that poses a difficulty but refers to the problematic translation of this desire into a required non-negotiable need. Consequently, as evidenced in Frost et al.’s CM factor and Ellis’ formulation of the difficulties with dysfunctional perfectionism, what is clearly evidenced is the need to address this cognitive dimension as part of any cognitive intervention for the treatment of WEB-BASED CBT FOR PERFECTIONISM 9 perfectionism. More specifically, a cognitively-based intervention should try to reduce the distorted importance ascribed to perfectionism. Consequently, the CBT intervention included all materials found in the GSM intervention, in addition to materials aimed specifically at modifying perfectionistic beliefs and their related effects. The CBT component addressed the following: (a) Living in the real world (checking out your interpretations ); (b) Living in the world of ‘shoulds’ (examining and reevaluating expectations and the importance of personal choice ); (c) Working out your mind (recognizing how certain ways of thinking cause distress); (d) Dealing with negative moods (three skills for dealing with negative moods); (e) When a ‘want’ becomes ‘necessity’ (keeping perspective on desires); and (f) Dealing with academic and performance anxiety (helping you do and feel your best). General stress management. The GSM intervention utilized materials that covered the following: (a) Recognizing and dealing with stress (recognizing how stress uniquely ‘gets to you’ and learning what helps you to reduce stress); (b) Dealing with distractions and distractibility (seeing how stress gets you distracted and discovering what you can do to maintain focus); (c) Changing your stressors (learning to relax, including progressive relaxation and breath-focused relaxation); (d) Exercise (getting started and monitoring your progress); (e) Sleep (healthier sleep makes your brain work better); and (f) Meditation (maintaining awareness and balance). Outcome Measures Perfectionism was assessed using more than one measure. Multiple measures were employed in order to allow for the multifaceted representation of a construct clearly identified as multidimensional in nature. Thus, this would allow the researcher to identify how effectively the WEB-BASED CBT FOR PERFECTIONISM 10 intervention in question acted on particular components of perfectionism. It is felt that this was warranted, given previous research that linked certain facets of perfectionism differentially to various measures of psychological distress. Perfectionism. The Multidimensional Perfectionism Scale (MPS-F, [Frost et al., 1990]) is a 35-item questionnaire used to assess six dimensions of the nature of perfectionism. The subscale of interest in this study was concern over mistakes (CM). The CM subscale demonstrated high internal consistency (α =.87). Participants are asked to indicate to what extent they agree or disagree with statements on a 5-point Likert scale. The Multidimensional Perfectionism Scale (MPS-HF, [Hewitt & Flett, 1991]) is a 45item questionnaire used to assess three dimensions of perfectionist behaviour: (a) self-oriented perfectionism (SOP), (b) other-oriented perfectionism (OOP), and (c) socially prescribed perfectionism (SPP). Hewitt and Flett (1987) found alpha coefficients for the MPS-HF subscales of.86 for SOP,.82 for OOP, and.87 for SPP. Participants are asked to indicate to what extent they agree or disagree with a number of statements on a 7-point Likert scale. The Perfectionism Cognitions Inventory (PCI, [Flett, Hewitt, Blankstein, & Gray, 1998]) is a 25-item questionnaire used to assess the frequency of ‘automatic perfectionistic thoughts’. This measure indirectly gauges the extent to which an individual engages in cognitive evaluations between the ideal perfectionistic self and the current self. Participants are asked to indicate how frequently a given thought has occurred in the past week on a 5-point scale. Extensive evidence attests to the psychometric properties of the PCI in both student and clinical samples (Flett et al., 1998; Flett, Hewitt, Whelan, & Martin, 2007). The Almost Perfect Scale - Revised (APSR, [Slaney, Rice, Mobley, Trippi, & Ashby, 2001]) is a 23 item questionnaire used to assess three dimensions of perfectionism. The three WEB-BASED CBT FOR PERFECTIONISM 11 subscales assessed by the APS-R are: (a) discrepancy; (b) high standards; and (c) order. Of particular interest for the present study was the discrepancy subscale (DISCR) that examined the gap between perfectionistic standards and perceived level of attainment. The DISCR was of central interest in evaluating the effectiveness of a CBT intervention as one might expect to note a reduction over time on the discrepancy subscale as people high in perfectionism either lower their standards or become more accepting of themselves. This scale demonstrated high internal consistency with alpha coefficients ranging from.82 to.92. Participants were asked to describe their degree of agreement with a series of statements representing attitudes people may have toward themselves on a 7-point Likert scale. Depressive mood. The Center for Epidemiologic Studies-Depressed Mood Scale (CESD, [Radloff, 1987]) is a 20-item questionnaire used to assess depressive symptomatology. Alpha values of 0.85 for the general population, and 0.90 for psychiatric populations are indicative of the good internal consistency of the scale. Participants are asked to indicate the number that best describes the way they have felt on the given items, in the past week, on a 4-point scale. Anxiety. The Beck Anxiety Inventory (BAI, [Beck, Epstein, Brown, & Steer, 1988]) is a 21-item questionnaire used to assess clinical anxiety. This scale demonstrated high internal consistency with an alpha coefficient of.92. It has also demonstrated good concurrent validity as it has been found to correlate well with other measures of anxiety (Beck et al., 1988). Participants are asked to rate the degree to which they have been bothered by given symptoms during the previous week on a 4-point Likert scale. Cognitive vulnerability to negative affect. The Automatic Thoughts Questionnaire (ATQ, [Hollon & Kendall, 1987]) is a 30-item questionnaire used to assess automatic negative thoughts about the self by measuring the cognitive self-statements of an individual. The ATQ WEB-BASED CBT FOR PERFECTIONISM 12 examines four aspects of automatic thoughts: (a) personal maladjustment and desire for change; (b) negative self-concepts and negative expectations; (c) low self-esteem; and (d) helplessness. This scale demonstrated high internal consistency (α =.96). Participants are asked to rate the frequency of a given thought during the previous week on a 5-point Likert scale. Another measure of cognitive vulnerability used in the study was the Anxiety Sensitivity Index (ASI, [Reiss, Peterson, Gursky, & McNally, 1986]). The ASI is a 16-item questionnaire used to assess the fear of anxiety. Psychometric properties of the ASI have been found to be satisfactory, with test-retest reliability ranging from.75 -.85 and a Cronbach’s alpha of.88 (Reiss et al., 1986). Participants are asked to indicate, on a 5-point scale, the extent to which they believe that anxiety sensations cause embarrassment, illness, additional anxiety, and loss of control. Validity Checks Therapy credibility and expectancy validity check. The Credibility Expectancy Questionnaire (CEQ, [Devilly & Borkovec, 2000]) is a 6-item questionnaire used to assess both the cognitively based credibility of an intervention, and its more affectively based treatment expectancy. The CEQ showed high internal consistency for each of the two factors, with standardized alpha coefficients ranging from.79 -.90 and.81 -.86 for the expectancy and credibility factors respectively. Participants are asked to indicate, on either a 9 point scale or a percentage scale (in 10% increments), what they thought or felt about the intervention they received. As the general stress management intervention served as an alternative treatment condition it was important to compare the two interventions on measures of treatment credibility and expectancy. These analyses were necessary in order to be able to rule out differential levels WEB-BASED CBT FOR PERFECTIONISM 13 of credibility and expectancy (non-specific features of therapy) as alternative explanations for any differences in observed outcomes that might occur between the groups. Treatment adherence. A knowledge information questionnaire was developed to help the researchers evaluate adherence to the intervention protocols as a possible threat to the internal validity of the study. This measure assessed retention for specific details of both the CBT and general stress management intervention materials. The results of the knowledge questionnaire were not evaluated however as the psychometric properties of the scale indicated questionable validity. Procedure Students were asked to volunteer to participate in the study if they felt their academic or personal lives were negatively affected by perfectionism. They were advised that the study fully complied with the required research ethics protocols and were asked to provide written informed consent prior to participation. Potential participants were screened using the PCI and had to obtain a score greater than or equal to one standard deviation above the mean in order to be eligible for participation in the study. The PCI was selected as a screening tool because it correlates strongly with trait measures of perfectionism, and because of its emphasis on the cognitive aspects of perfectionism. Participants who were not comfortable with the English language or the use of web-based programs were excluded from the study. Participants (N = 297) were screened from introductory psychology classes at a large Canadian university. Those who met the requirements for participation in the study were contacted via email and offered the opportunity to take part in the study. Interested individuals came to campus to complete the baseline set of questionnaires and, upon completion, were randomly assigned to 1 of the 3 experimental conditions. Participants were provided with a WEB-BASED CBT FOR PERFECTIONISM 14 specific web address, based on group assignment. Both the CBT and GSM groups were instructed via the web on how to proceed through their respective web-based intervention protocols. The NT group was informed via their website that they would be asked to return in 11 weeks time (posttest) to complete a second set of measures. The duration of the intervention phase was 10 weeks. Approximately one week following the intervention period all study participants, regardless of group, came in to complete the post intervention questionnaire package. Results Sample Characteristics Pretest means relative to established norms. As expected, pretest means (see Table 1) indicated that relative to existing norms, the participants in this sample were significantly elevated on measures of perfectionism and distress. More specifically, mean scores on a number of these measures far exceeded the means for community, and in some cases psychiatric, samples (Flett et al., 2007; Gillis, Haaga, & Ford, 1995; Hewitt & Flett, 2004; MacDonald, Baker, Stewart, & Skinner, 2000; Reiss et al., 1986). Therapy credibility and expectancy. The CBT and GSM groups were evaluated in order to determine whether significant differences existed between them on measures of therapy credibility and expectancy. Two independent samples t-tests were conducted, one for credibility and one for expectancy. Results indicated there were no significant differences between the two intervention groups on measures of both credibility, t(53) =.818, p =.417, and expectancy t(53) = 1.136, p =.261. Data Screening WEB-BASED CBT FOR PERFECTIONISM 15 The scores on each item of each scale were initially screened for the presence of missing data. It was found that only 0.25% of the total data points were missing and stochastic regression imputation was used to replace missing values on individual items within each time period by regressing the item with missing data on the remaining items in the scale and then adding a random residual error. The total scale scores were then screened for the presence of nonnormality (separately by group given the nature of the analyses). No evidence of extreme nonnormality was found on any the measures. Pretest to Posttest Differences on Measures Paired samples t-tests were used to compare pretest and posttest scores on all scales for each of the three intervention groups. The results are presented in Table 1. For CBT participants, there were significant changes on all of the scales except the BAI. For GSM participants, significant changes were noted on only four of the scale scores (i.e., the ASI, the CM, the PCI, and the SOP). No significant changes were noted within the NT group. Covariation: Changes in Perfectionism and Changes in Distress within the CBT Group Correlations between changes in scale scores were computed using Pearson correlations on change scores (pre - post). Correlations for the CBT group are presented in Table 2. Only changes in OOP scores were not significantly correlated with any depression or anxiety measures. Tests of the Treatment Effect on Perfectionism Measures Means and standard deviations for each treatment group at pretest and posttest, on each of the perfectionism scales, are presented in Table 1. Analyses of covariance (ANCOVA), with the posttest scores as the dependent variable, the baseline scores as the covariate, and the treatment group as the independent variable, were used to determine if scores on the WEB-BASED CBT FOR PERFECTIONISM 16 perfectionism measures changed differentially from baseline to posttest across the treatment groups. Pairwise multiple comparisons with familywise error control were used to determine exactly where differences among the treatment groups existed. A summary of the ANCOVA results are presented in Table 3. For each of the SOP, SPP, PCI, CM, and DISCR, scores in the CBT condition decreased significantly more than NT scores. Further, scores in the CBT condition decreased significantly more than those in the GSM condition for the SOP, SPP, PCI, and CM. In addition, scores in the GSM condition decreased significantly more than those in the NT condition for the SOP, PCI, CM, and DISCR. Test of the Treatment Effect on Measures of Depression, Anxiety, and Cognitive Vulnerability to Negative Affect Means and standard deviations for each treatment group at pretest and posttest on each of the depression and anxiety scales are presented in Table 1. One-way ANCOVAs (pretest as covariate) were used to determine if changes in the depression, anxiety or cognitive vulnerability scores of participants from pretest to posttest were related to the treatment group allocation. A summary of the ANCOVA results are presented in Table 3. For both the BAI and the ATQ, CBT scores decreased significantly more than NT scores. Further, scores in the GSM condition decreased significantly more than those in the NT condition for the BAI. Clinical Significance of Pre-Post Perfectionism Results The clinical significance of an intervention relates to the capacity of the treatment to meet the standards for efficacy identified by clinicians. It differs from statistical significance by its reliance on external standards, related to the attainment of particular goals, which help determine the clinical benefits of a particular intervention. Currently, there does not exist any real agreement within the literature with regard to the identification of these key external standards. WEB-BASED CBT FOR PERFECTIONISM 17 One method that has been proposed by Jacobson and Truax (1991) suggests that clinically significant change could be assumed if a significant amount of improvement is observed, and a post-treatment score on a given measure is closer, after the treatment intervention, to the mean of the functional population than to that of the dysfunctional population. Using this conceptualization of clinically significant change, it was found that although the majority of the participants remained clinically unchanged, some notable clinically significant gains were observed, and these were primarily for a more modest proportion of the CBT group (see Table 4). Discussion Perfectionism related distress is an important consideration at post-secondary institutions given indications of the prevalence of this type of distress. Evidence suggests high levels of perfectionism are associated with elevated depression and anxiety levels, eating-disordered behaviors, reduced academic performance, and other problems. The study goal was thus to investigate whether a short term, web-based, CBT intervention could reduce levels of perfectionism and related distress in students with defined elevations. As expected, CBT participants scored significantly lower at posttest, relative to pretest, on nine of the measures of perfectionism, depressive symptomatology, and cognitive vulnerability to negative affect. Comparatively, GSM participants showed significant improvement on only four of the measures, whereas NT participants did not significantly improve on any of these same measures. Findings for the CBT group are consistent with previous studies that support the effectiveness of CBT in reducing perfectionism and distress in high perfectionism individuals (e.g. Kutlesa, 2002). Moreover, it is understandable that the GSM group experienced intervention benefits, WEB-BASED CBT FOR PERFECTIONISM 18 although to a more limited degree, as stress reduction interventions aim to reduce stress-reactive symptomatology and increase psychological well-being. The positive correlations established between Self-Oriented Perfectionism measures and symptoms of anxiety (e.g. Hayward & Arthur, 1998) and depression (e.g. Hayward & Arthur, 1998; Hewitt et al., 1990) suggest stress reduction interventions may offer such benefits. Individuals high in perfectionism have been described as highly stress responsive people who, in attempting to avoid or reduce anxiety and stress, pursue impossibly perfectionistic standards that result in greater stress (Hewitt & Flett, 2002). Thus, the ability to manage stress, without resorting to the pursuit of perfection, should directly counteract these stress exacerbating problems. Furthermore, given suggestions that anxiety underlies problematic perfectionism, and stress management interventions help reduce anxious mood symptoms , the findings regarding GSM benefits are consistent with previous studies where benefits were found related to overall stress reduction. The question is whether the beneficial effects of GSM, without the addition of specific, strategic cognitive changes will have effects as enduring as those associated with the CBT condition. The significant relationship between changes in perfectionism and changes in psychological distress for participants receiving the CBT intervention highlight that the improvements are not independent of one another. The fact that perfectionism was correlated with other forms of psychological distress at baseline, and that changes in perfectionism were related to changes in distress provides further evidence for the inter-relatedness of these states. This introduces the possibility that by effectively treating perfectionism, we may also be reducing the likelihood of experiencing other forms of distress. WEB-BASED CBT FOR PERFECTIONISM 19 As anticipated, participants receiving the CBT intervention demonstrated significantly more improvement on the perfectionism measures than participants in the GSM or NT groups, indicating the CBT intervention more effectively reduced perfectionism levels, with the pattern of improvements (excepting the DISCR scale), being significantly greater for the CBT condition than for the GSM condition. Nonetheless, while CBT participants improved significantly more than participants in the NT condition in anxiety and cognitive vulnerability to negative affect, they did not improve more than those in the GSM condition on these measures. Both conditions appear to have the capacity to effectively reduce psychological distress. It will be important to determine in subsequent research if CBT and GSM interventions both result in enduring benefits regarding reductions in perfectionism. The study findings are consistent with those of previous studies demonstrating decreased levels of perfectionism, depression, and anxiety following a brief CBT intervention (e.g., Kutlesa, 2002; Pleva & Wade, 2006; Riley et al., 2007). The present study corroborates earlier studies, adding weight to the cumulative evidence from controlled and uncontrolled intervention studies, that perfectionistic beliefs and attitudes can be modified. Given other evidence indicating problematic perfectionism is difficult to treat and associated with poorer psychotherapeutic outcomes, these promising findings require further replication. While the results suggest clear evidence of the utility of a web-based intervention based on CBT and GSM features, it is important to consider how the conclusions are qualified by observations that may be inferred from the posttest results. Although levels of depressive mood and negative automatic thoughts were reduced, levels of anxiety were still elevated at posttest and well above normative values. Still, it must be noted that there were significant reductions in anxiety sensitivity, and that anxiety sensitivity is a variable that has been related to various forms WEB-BASED CBT FOR PERFECTIONISM 20 of anxiety as well as to other indices of distress. This might suggest the need to include additional treatment components that focus more directly on reducing symptoms of anxiety and anxiety-related cognitions. In addition, although perfectionism was substantially reduced for CBT participants, overall levels of perfectionism cognitions and SOP were still relatively high in comparison to established norms and some participants still had clear elevations on perfectionism. Additional intervention would be beneficial given that the pattern of correlations found at posttest indicated that perfectionism was still associated with residual symptoms of depression and anxiety, despite the overall reductions in perfectionism. In evaluating the present intervention, it should be noted that it was delivered in such a way that participants could vary considerably in how much they made use of the intervention materials. While we attempted to assess adherence using a questionnaire developed specifically for this study, the psychometric properties of this scale were found to be questionable. It is nonetheless possible that both how often, and the duration of time participants focused on the material may have affected their responses to posttest measures. Given that the time spent on the intervention was not directly measured, and that the CBT protocol was lengthier than the GSM protocol, it is not possible to determine whether it was the content of the material or the amount of time spent on it that contributed most to the findings. It will be important for future studies to be able to assess and understand the way the materials are being used to better identify those factors most closely related to overall improvement. Furthermore, it is acknowledged that assessing treatment effects one-week post-treatment is a limitation. It remains to be determined if the beneficial effects would endure similarly in both conditions since one is targeted more at treating symptoms (GSM) while the other (CBT) is WEB-BASED CBT FOR PERFECTIONISM 21 geared to modify maintenance factors such as perfectionism. Future studies would need to examine the relative effectiveness of both conditions when assessed at different follow-up intervals. In conclusion, this study has made a significant contribution to the treatment of perfectionism using short-term CBT strategies. To the findings that elevated levels of perfectionism can result in more unfavourable responses to treatment interventions (Blatt & Zuroff, 2002) must be added evidence from the current study where highly beneficial perfectionism-specific outcomes suggest that the self directed nature of the intervention may reduce the negative reactivity related to control issues, which seem to be a factor in treating individuals high in perfectionism. The fact that this intervention is computer-based, part of a wave of computerized advances permitting new disseminations of psychological treatments, may be an advantage with this population. While the results do not allow us to explicitly identify the exact mechanism by which the cognitive restructuring strategies found in the CBT intervention had an effect, they do offer promising support for the effectiveness of web-based CBT interventions in dealing with perfectionistic concerns. Given the large proportion of individuals with perfectionism related concerns, and the cost-effectiveness and ease of dissemination of the current intervention, the researchers believe that self-directed interventions for perfectionism can play a significant role in the treatment of perfectionism in the near future. WEB-BASED CBT FOR PERFECTIONISM 22 References Arthur, N., & Hayward, L. (1997). The relationships between perfectionism, standards for academic achievement, and emotional distress in postsecondary students. Journal of College Student Development, 38, 622-632. Beck, A. T., Epstein, N., Brown, G. & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893-897. Blatt, S. J., & Zuroff, D. C. (2002). Perfectionism in the therapeutic process. In G. L. Flett, & P. L. Hewitt (Eds.), Perfectionism: Theory, Research, and Treatment (pp. 393 406). Washington, DC: American Psychological Association. Bottos, S., & Dewey, D. (2004). Perfectionists appraisal of daily hassles and chronic headaches. Headache: The Journal of Head and Face Pain, 44, 772-779. Cassin, S. E., & von Ranson, K. M. (2005). Personality and eating disorders: A decade in review. Clinical Psychology Review, 25, 895-916. Devilly, G. J., & Borkovec, T. D. (2000). Psychometric properties of the credibility expectancy questionnaire. Journal of Behavior Therapy and Experimental Psychiatry, 31, 73-86. Educational Advisory Committee. (2003). Meeting the challenges of larger classes. Retrieved October 27, 2006, from http://www.utoronto.ca/ota/LargerClasses.pdf Ellis, A. (2002). The role of irrational beliefs in perfectionism. In G. L. Flett, & P. L. Hewitt (Eds.), Perfectionism: Theory, Research, and Treatment (pp. 217 - 229). Washington, DC: American Psychological Association. Flett, G. L., Greene, A., & Hewitt, P. L. (2004). Dimensions of perfectionism and WEB-BASED CBT FOR PERFECTIONISM 23 anxiety sensitivity. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 22, 37-55. Flett, G. L., Hewitt, P. L., Blankstein, K., & Gray, L. (1998). Psychological distress and the frequency of perfectionistic thinking. Journal of Personality and Social Psychology, 75, 1363-1381. Flett, G. L., Hewitt, P. L., Blankstein, K., & Mosher, S. W. (1995). Perfectionism, life events, and depressive symptoms: A test of a diathesis-stress model. Current Psychology: Developmental, Learning, Personality, Social, 14, 112-137. Flett, G. L., Hewitt, P. L., Garshowitz, M., & Martin, T. R. (1997). Personality, negative social interactions, and depressive symptoms. Canadian Journal of Behavioural Science, 29, 28-37. Flett, G. L., Hewitt, P. L., Whelan, T., & Martin, T. R. (2007). The Perfectionism Cognitions Inventory: Psychometric properties and associations with distress and deficits in cognitive self-management. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 25, 255-277. Flett, G. L., Madorsky, D., Hewitt, P. L., & Heisel, M. J. (2002). Perfectionism cognitions, rumination, and psychological distress. Journal of Rational-Emotive & Cognitive-Behavior Therapy, 20, 33-47. Frenette, M. If you build it, will they come? University and college participation among local youth following the creation of a new university. Paper presented at: CEA 40th Annual Meetings; May, 2006; Concordia University, Montreal. Frost, R. O., & DiBartolo, P. M. (2002). Perfectionism, anxiety, and obsessive- WEB-BASED CBT FOR PERFECTIONISM 24 compulsive disorder. In G. L. Flett, & P. L. Hewitt (Eds.), Perfectionism: Theory, Research, and Treatment (pp. 341 - 371). Washington, DC: American Psychological Association. Frost, R. O., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449-468. Gilbert, S. (1995). Quality education: Does class size matter? Retrieved October 27, 2006, from http://www.aucc.ca/_pdf/english/publications/researchfile/1995-96/vol1n1_e.pdf Gillis, M. M., Haaga, D. A. F., & Ford, G. T. (1995). Normative values for the Beck Anxiety Inventory, Fear Questionnaire, Penn State Worry Questionnaire, and Social Phobia and Anxiety Inventory. Psychological Assessment, 7, 450-455. Grzegorek, J. L., Slaney, R. B., Franze, S., & Rice, K. G. (2004). Self-criticism, dependency, self-esteem, and grade point average satisfaction among clusters of perfectionists and nonperfectionists. Journal of Counseling Psychology, 51, 192-200. Halgin, R. P., & Leahy, P. M. (1989). Understanding and treating perfectionistic college students. Journal of Counselling and Development, 68, 222-225. Hayward, L., & Arthur, N. (1998). Perfectionism and post-secondary students. Canadian Journal of Counselling, 323, 187-199. Hewitt, P. L., & Flett, G. L. (2004). The Multidimensional Perfectionism Scale: Technical manual. Toronto: Multi-Health Systems. Hewitt, P. L., & Flett, G. L. (1991). Perfectionism in the self and social contexts: Conceptualization, assessment, and association with psychopathology. Journal of Personality and Social Psychology, 60, 456-470. Hewitt, P. L., Flett, G. L., & Ediger, E. (1995). Perfectionism traits and perfectionistic WEB-BASED CBT FOR PERFECTIONISM 25 self-presentation in eating disorder attitudes, characteristics, and symptoms. International Journal of Eating Disorders, 18, 317-326. Hewitt, P. L., Mittelstaedt, W. M., & Flett, G. L. (1990). Self-oriented perfectionism and generalized performance importance in depression. Individual Psychology, 46, 67-73. Hollon, S. D., & Kendall, P. C. (1987). Automatic thoughts questionnaire. In K. Corcoran, & J. Fischer (Eds.), Measures for clinical practice: A sourcebook (pp. 103104). New York: Collier Macmillan. Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12-19. Kutlesa, N. (2002). A group intervention with university students who experience difficulties with perfectionism. (Doctoral dissertation, University of Calgary, 2002). Dissertation Abstracts International, 64, (2-A), 398. MacDonald, A. B., Baker, J. M., Stewart, S. H., & Skinner, M. (2000). Effects of alcohol on the response to hyperventilation of participants high and low in anxiety sensitivity. Alcoholism: Clinical and Experimental Research, 24, 1656-1665. Otto, M. W., & Reilly-Harrington, N. A. (1999). The impact of treatment on anxiety sensitivity. In S Taylor (Ed.), Anxiety Sensitivity: Theory, Research and Treatment of the Fear of Anxiety (pp. 321 – 336). Mahwah, NJ: Erlbaum. Pleva, J., & Wade, T. D. (2006). Guided self-help versus pure self-help for perfectionism: A randomised control trial. Behaviour Research and Therapy, 45, 849-861. Radloff, L. S. (1987). Center for epidemiologic studies: Depressed mood scale. In K. WEB-BASED CBT FOR PERFECTIONISM 26 Corcoran, & J. Fischer (Eds.), Measures for clinical practice: A sourcebook (pp. 118119). New York: Collier Macmillan Inc. Reiss, S., Peterson, R. A., Gursky, D. M., & McNally, R. J. (1986). Anxiety sensitivity, anxiety frequency , and the prediction of fearfulness. Behaviour Research and Therapy, 24, 1-8. Riley, C., Lee, M., Cooper, Z., Fairburn, C. G., & Shafran, R. (2007). A randomized controlled trial of cognitive-behaviour therapy for clinical perfectionism: A preliminary study. Behaviour Research and Therapy, 45, 2221-2231. Ritterband, L. M., Gonder-Frederick, L. A., Cox, D. J., Clifton, A. D., West, R. W., & Borowitz, S. M. (2003). Internet interventions: In review, in use, and into the future. Professional Psychology: Research and Practice, 34, 527-534. Skinner, H., Biscope, S., & Poland, B. (2003). Quality of internet access: Barrier behind internet use statistics. Social Science & Medicine, 57, 875-880. Slaney, R. B., Rice, K. G., Mobley, M., Trippi, J., & Ashby, J. S. (2001). The revised Almost Perfect Scale. Measurement and Evaluation in Counseling and Development, 34, 130-145. Statistics Canada. Census of population: Earnings, levels of schooling, field of study and school attendance. Available at: http://www.statcan.ca/Daily/English/030311/d030311a.htm. Accessed July 15, 2006. Van Bavel J. (2002) Applying for graduate studies in psychology: A 12-step program. Psynopsis: Canada’s Psychology Newspaper, 24, 12. WEB-BASED CBT FOR PERFECTIONISM 27 Table 1 Comparison of Means (Standard Deviations) on Each of the Scales at Pretest and Posttest for Each of the Treatment Groups Pretest Posttest Scale CBT Stress No Treatm. Self-Oriented Perfectionism 85.49 (9.62) 86.18 (10.97) 84.37 (12.15) 73.20** 80.31** (10.98) (14.08) 85.17 (14.53) Other-Oriented Perfectionism 65.98 (13.68) 64.39 (9.80) 64.83 (13.52) 60.45* (11.18) 64.63 (9.11) 62.50 (11.28) Socially Prescribed Perfectionism 64.83 (13.87) 67.97 (13.66) 65.92 (14.55) 55.52** 68.03 (10.84) (14.42) 67.76 (13.25) Concern over Mistakes 29.43 (6.94) 30.41 (7.00) 30.21 (7.87) 23.34** (5.02) 30.23 (8.59) Discrepancy 61.95 (12.76) 58.83 (13.73) 61.04 (14.11) 51.75** 55.92 (11.39) (15.97) 64.95 (12.79) Perfectionism Cognitions 66.14 (15.55) 68.83 (10.53) 69.75 (12.50) 50.24** 60.15** (15.72) (17.10) 70.36 (12.35) Depressed Mood 23.93 (9.53) 25.28 (11.37) 27.67 (11.62) 19.53* (8.08) 24.65 (12.70) 27.00 (9.62) Beck Anxiety Inventory 16.23 (10.86) 18.14 (10.76) 16.00 (8.98) 14.14 (8.53) 16.38 (10.87) 19.73 (12.36) Anxiety Sensitivity 44.33 (12.49) 42.84 (11.30) 43.54 (11.39) 39.85* (11.80) 38.31* (13.51) 41.18 (11.29) Automatic Thoughts 72.03 (22.76) 77.59 (28.37) 83.31 (27.22) 57.76** 71.04 (20.22) (32.53) 83.00 (26.55) CBT Stress No Treatm 27.73** (6.40) Note. *, ** indicate significant change from pretest to posttest. * = p <.05, ** = p <.01. WEB-BASED CBT FOR PERFECTIONISM 28 Table 2 Correlations Between Changes in Perfectionism Scores and Changes in Depression, Anxiety, and Vulnerability Scores in the CBT Group Beck Anxiety Inventory Anxiety Sensitivity Depressive Mood Automatic Thoughts Perfectionism Scale Self-Oriented Perfectionism.44*.41*.31.42* Other-Oriented Perfectionism.22.34.09.15 Socially-Prescribed Perfectionism.44*.45*.47**.60** Concern over Mistakes.36.44*.37.37* Discrepancy.29.45*.25.51** Perfectionism Cognitions.64**.49**.51**.67** Note. * p <.05, ** p <.01. WEB-BASED CBT FOR PERFECTIONISM 29 Table 3 Test of the Treatment Effect on Outcome Measures ________________________________________________________________________ CBT-GSM F p η 2 p CBT-NT p GSM-NT p Perfectionism Self-Oriented Perfectionism 9.05