Perfectionism has been associated with social phobia, general psychopathology, and social and trait anxiety (Juster et al., 1995). Individuals high in perfectionism, particulary Concern Over Mistakes (CM), approach evaluative performance situations with a negative mindset. Compared to nonperfectionists, these individuals express more concern with how others will perceive their mistakes, show less confidence in their performance, and experience more debilitating anxiety (Frost & Henderson, 1991; Frost & Marten, 1990).

There have been few controlled attempts to investigate how the detrimental influence of perfectionism on performance might be ameliorated. The current study addresses this issue by comparing two approaches to remedying the interference of anxiety on a public speaking task. We administered either a cognitive restructuring procedure or a distraction task to students either high or low in CM, prior to giving a speech.

Overall, we expected that high CM subjects would experience more speech related anxiety than low CM subjects. In addition, we hypothesized that all subjects who received cognitive restructuring would experience less speech-related anxiety than those who were assigned to the distraction condition. Finally, we speculated that the benefits of cognitive restructuring would be more evident for subjects high in CM.

Subjects
Sixty subjects, identified as high in concern over mistakes = 27) and low in concern over mistakes (LCM: CM < = 17) according to the CM subscale of the Multidimensional Perfectionism Scale (MPS; Frost, Marten, Larhart, & Rosenblate, 1990), were obtained from a small, private women's college.

Measures
SUDS ratings:
Throughout the assessment, participants were asked to rate their levels of anxiety by providing SUDS ratings using the scale of 0 (not at all anxious) to 100 (extremely anxious) (Wolpe, 1982).
Pre-task measures: The Pre-Task Global Rating Form consisted of items designed to determine the subject's expectations of performance for themselves and others, and to ascertain how important it is for the subject to do well on the speech task (on a scale of 0: not at all to 100: extremely). On the Speech Related Concerns Questionnaire, subjects were told to rate on a seven point Likert-type scale ranging from 1 (not at all bothered) to 7 (very bothered) how bothered they were by thirteen particular thoughts associated with giving a speech (e.g. being boring, making mistakes, running out of things to say).
Post-task measures: The Speech Related Concerns Questionnaire was again administered. Using a 0 (not at all) to 100 (extremely) scale, subjects responded to a Post-Task Global Rating Form which measured 5 items, one of which concerned their anxiety level during the speech.
Audience ratings: Audience members were asked to rate the subject on perceived anxiety and the quality of the speech on a scale from 0 (not at all) to 100 (extremely).

Design
The experiment employed a 2 x 2 design with the first factor being Group (HCM vs. LCM) and the second factor being Condition (cognitive restructuring vs. distraction). A total of 15 subjects were assigned to each cell.

Experimental Manipulation
Cognitive Restructuring Condition Subjects were guided through a series of questions designed to help evaluate their feared predictions more realistically. In this phase, experimenters addressed the subject's probability overestimations, decatastrophized a feared outcome for the speech, and directed subjects to construct a coping thought to be used during the speech task.
Distraction Condition The distraction task required subjects to cross out all the e's that they found in a passage. This was designed to help keep the subject's mind off the upcoming speech.

Procedure
General Instructions: Upon subjects' arrival to the lab, experimenters explained the purpose of the SUDS ratings and obtained a baseline rating. Subjects were told of the duration of the speech (10 minutes) and were informed that they would be able to choose 3 of 5 possible speech topics including abortion, the health care system, mandatory seat belt laws, corporal punishment in the schools, and nuclear power. Experimenters then described the purpose of the "stop" sign, which could be used to stop the speech if subjects' anxiety became unberable. The Pre-Task Global Ratings Questionnaire and pre-task Speech Related Concerns Questionnaire were administered.
Preparation Phase: Experimenters obtained a SUDS rating at the beginning and end of this phase. Subjects were given 3 minutes to prepare for the speech.
Experimental Phase: Subjects were randomly assigned to either the cognitive restructuring or distraction phase. SUDS ratings were obtained at the beginning and end of this phase.
Speech Task Subjects were asked to provide written SUDS ratings at one-mintue intervals throughout the 10-minute speech task. Two audience members observed the speech. Audience members were instructed to show no emotion and hold a pad and pen to create the appearance of taking notes.
Post-Task Assessment: The post-task Speech Related Concerns Questionnaire and Post-Task Global Ratings Form were administered. At one minute intervals, experimenters requested SUDS ratings for five ratings, or until the subject reached her baseline anxiety rating (or lower), whichever came first. When questionnaires were completed, subjects were debriefed.

Expectations for the Speech: We conducted a series of one-way ANOVAs to compare group differences in Pre-Task Global Rating items. On the self-other difference variable, computed by subtracting subjects' ratings of how well they expected others to perform from their ratings of how well they expected themselves to perform, we found a significant Group effect. We also found a Group effect for pre-task importance ratings (See Table 1).

Speech-Related Anxiety Ratings: A 2 (HCM vs. LCM) x 2 (baseline vs. preparation phase) repeated measures ANOVA was conducted for subjects' SUDS ratings prior to and after learning about the speech task. A main effect of phase was revealed, F(1, 58) = 70.25, p < .001). No other significant effects were found.

A 2 (HCM vs. LCM) x 2 (cognitive restructuring vs. distraction) x 2 (pre- vs. post-experimental phase) repeated measures ANOVA of SUDS ratings revealed a significant main effect of phase, F(1, 56) = 7.91, p < .01, as well as a significant interaction between condition and phase, F(1, 58) = 4.53; p < .05. No other significant effects were noted.

We also conducted a 2 (HCM vs LCM) x 2 (cognitive restructuring vs. distraction) x 3 (experimental vs. task vs. post-task phase) analysis of covariance (ANCOVA) of SUDS ratings, controlling for baseline SUDS ratings. We found significant main effects of condition, F(1, 55) = 6.99; p < .05, and phase, F(2,112) = 59.37; p < .001. We found a marginally significant Group effect, F(1, 55) = 4.01; p = .05. Overall, subjects high in CM gave significantly higher SUDS ratings 9M = 47.02, SD = 22.61) than subjects low in CM (M = 34.75, SD = 21.09). No other significant effects were revealed.

Finally, a 2 (HCM vs. LCM) x 2 (cognitive restructuring vs. distraction) ANOVA of subjects' Post-Task Global Ratings revealed a significant Condition effect, F(1,56) = 4.29; p < .05, for subjects' retrospective ratings of how anxious they had been while speaking. Subjects who received cognitive restructuring reported feeling significantly less anxious (M = 49.33; SD = 23.48) than subjects assigned the distraction task (M = 61.17, SD = 23.40). There was a significant Group effect, F(1,56) = 9.03, p < .01. Overall, subjects high in CM recalled being significantly more anxious during the speech (M = 63.83, SD = 20.75) than subjects low in CM (M = 46.67, SD = 24.26). No significant interaction was found.

Speech Related Concerns: We conducted a 2 (HCM vs. LCM) x 2 (cognitive restructuring vs. distraction) x 2 (pre- vs. post-task) repeated measures ANOVA of subects' scores on the Speech Related Concerns Questionnaire . We observed a significant Group effect, F(1,56) = 19.61, p < .001, as well as a significant main effect of phase, F(1,56) = 5.25, p < .05. (See Table 2.) No other significant effects were noted.

Audience Ratings:We ran a 2 (HCM vs. LCM) x 2 (cognitive restructuring vs. distraction) ANOVA of audience perceptions of speakers' anxiety and quality of speech. No significant effects were revealed.

These data provide evidence of the utility of cognitive restructuring in reducing anxiety related to an evaluative task within a laboratory setting. All subjects' initial anxiety ratings in response to the speech task were similar; however, once the experimental phase was introduced, we found a consistent and significant effect of condition for the duration of the session. Specifically, subjects who received cognitive restructuring reported lower anxiety throughout the session than did those who performed the distraction task.

Although results revealed no differences in response to cognitive restructuring according to Group, there was a significant Group effect concerning expectations for the speech. HCM individuals expected that others would perform better on the task and also placed higher value on the speech than did LCMs.

In addition, while there was no significant difference by condition, there was a Group difference in concerns about the speech, such that HCMs were more bothered by thoughts related to the speech both before and after the task than LCM individuals. Therefore, although individuals high in CM did place more importance on the speech, expect others to perform better, and were more concerned by thoughts about the speech, they proved equally as receptive to cognitive restructuring as LCMs.

Audience ratings of speech quality and anxiety showed no significant difference by condition or group. Thus, despite anxiety differences found by condition or group. Thus, despite anxiety differences found by condition and group differences revealed in expectations for the speech and speech related concers, these effects were not evident in the subjects' performance. It is imporatant to note, however, that no inter-rater reliability checks were conducted on audience ratings.

The present study has important implications. It appears that there are methods that can effectively tap into perfectionists' maladaptive thought processes which typically have a detrimental effect on their performance. As revealed by our results, HCM individuals who received cognitive restructuring were given an effective coping mechanism which may have helped reduce their speech related anxiety. Future research is necessary to clarify the mechanisms which produced the present data. These results provide promising support for existing, yet untested, beliefs regarding the benefits of cognitive restructuring for perfectionists (Burns, 1980).

	HCM	LCM	F
	M	M	(1,58)
	SD	SD
Self-Other Difference	-19.33	-2.00	12.90***
	20.16	17.10
Importance of Speech	42.67	32.00	4.41*
	19.99	19.37

*** p < .001 ** p < .01 * p < .05

Table 2. Mean Ratings of Speech Related Concerns.

Phase	HCM	LCM
	M	M
	SD	SD
Pre-task	3.99	2.63
	1.23	1.10
Post-task	3.72	2.42
	1.39	1.14