Abstract
The pursuit of excellence is encouraged at all levels of our society. Such ambitions are particularly validated in the academic community, and may promote individual achievement and societal progress. Taken to an extreme, however, striving towards perfectionism may become an obsession. The question may well be asked, what is the cost? While the answer is necessarily complex, and needs to consider the biopsychosocial system as a whole, it is possible to narrow this question and examine the cost in terms of human distress. One index of such distress is suicide ideation and behaviour.
Suicide ideation is recognised as a major public health issue with a particular impact upon young people in Australia. An Australian-based study reported that 11.7% of young men and 9.7% of young women reported some degree of suicide ideation in the few weeks prior to testing [1]. A large study of completed suicides among university students reported that the largest number of suicides occurred in the 20–24-year-old age group, and among graduate students [2]. Although the overall suicide rate for younger students in the USA was reported as being half that of a matched sample in the general population, students aged 25 and over had a significantly higher risk than younger students. These studies have resulted in an increased interest in the identification of modifiable risk factors. Based upon the current research on suicide ideation and suicidal behaviour [3], it may be hypothesised that one such variable is perfectionism. This study investigates whether particular dimensions of perfectionism are related to suicide ideation among a sample of Australian university students.
Recent research has focused upon perfectionism as a trait with potentially negative consequences. Evidence suggests that perfectionistic individuals experience increased negative affect before, during and after evaluative tasks, judge their work as lower in quality than non-perfectionists, and report the quality of their work should have been better [4]. Furthermore, people high in perfectionism reportedly experience an increased frequency and greater range of symptoms of psychopathology than those with low levels of perfectionism, and are more seriously troubled by severity of procrastination [5].
Individuals with high perfectionism and high levels of self-criticism are apparently also vulnerable to experiences of failure, reacting with increased levels of depression [3,4,6]. Specifically, perfectionistic attitudes have been associated with current depression, and predictive of depressive severity beyond prior depression [7].
According to both Burns [8] and Hollender [9], suicide is one potential outcome for the perfectionistic individual. Burns [8] suggested that perfectionistic thinking patterns may partly explain elevated suicide rates among some professional groups, while Hollender [9] pointed to the perfectionist's failure to reach unattainable goals as a factor in depression, which if persistent, could lead to suicide.
Support for an association between perfectionism and suicide threat was recently demonstrated in a series of studies of both psychiatric patients and college students [10]. While it was initially reported that only socially prescribed perfectionism (expectation that others hold these standards for the individual) was significantly correlated with suicide ideation and intent in psychiatric patients, later studies suggested suicide ideation was correlated with both self-oriented (high personal standards) and socially prescribed perfectionism [11].
There is evidence to support a differential relationship between perfectionism and suicidal preoccupation from the results of canonical correlational analysis reported by Adkins and Parker [12]. In this study, college students were assessed on both covert and overt measures of suicide preoccupation together with the Multidimensional Perfectionism Scale (MPS) [4]. Significant associations were demonstrated for the overt measures of suicidal preoccupation, but not the covert measures of suicidal themes. As the Concern about Mistakes (CM) and Doubts about Actions (D) subscales were related to suicidal preoccupation, these dimensions were defined as elements of ‘passive perfectionism’, whereas the Parental Criticism (PC), Personal Standards (PS), and Parental Expectations (PE) were not related to suicidal preoccupation and thought to reflect ‘active perfectionism’.
According to this framework, passive perfectionism indicates an inordinate fear of mistakes, a tendency to second-guess decisions and procrastination. On the other hand, active perfectionism does not indicate a risk of suicidal preoccupation and appears to act as a motivating, rather than inhibiting, factor in achievement strivings [12].
In this context, multidimensional conceptualisations of perfectionism as a construct comprised of two distinct global factors, maladaptive evaluative concerns and positive achievement strivings. Based upon this rationale, the purpose of this study is to investigate the relationship between the dimensions of perfectionism and suicide ideation in the context of an Australian university student population. Students were assessed on total perfectionism, which consists of five perfectionism dimensions, psychological distress and a measure of suicide ideation. It was expected that high levels of perfectionism would be associated with suicide ideation, with this relationship influenced by one or more negative perfectionism dimensions which can be characterised as ‘passive perfectionism’.
In the first instance, a significant and positive relationship is predicted between the presence of suicide ideation and perfectionism. It is also expected that one or more perfectionism dimensions will be differentially related to increased psychological distress and suicide ideation. Second, it is predicted that a significant and positive relationship between increased levels of perfectionism and suicide ideation will also be evident.
Method
Participants
The sample consisted of 405 university students (101 males; 301 females; 3 of unspecified gender) who volunteered to participate in the study. The age of participants ranged from 17 to 54 (mean = 22.7, SD = 6.4), with 95% enrolled as full-time students and 5% part-time. The students were predominantly Caucasian.
Instruments
The multidimensional perfectionism scale
The MPS [4] is a 35-item scale comprised of six subscales designed to tap the dimensions and nature of perfectionism. The MPS produces an overall perfectionism score in addition to the six subscale scores: Personal Standards (PS), Concern over Mistakes (CM), Doubts about Actions (D), Parental Criticism (PC), Parental Expectations (PE) and Organisation (O). It consists of statements which are rated on a five-point scale, from 1 (strongly disagree) to 5 (strongly agree).
Convergent validity of the MPS has been demonstrated [4,8,13,14]. There is also consistent evidence of concurrent validity in the three studies of the MPS and the other multidimensional measure of perfectionism developed by Hewitt and Flett [5,15–17]. An internal reliability for the MPS has been reported as 0.90 (Cronbachs α) with scores for the subscales ranging from 0.77 to 0.93 [4].
In the present study, a 29-item, five-subscale version of the MPS in which the Organisation sub-scale is omitted was used, as this subscale score is not required for the calculation of the total perfectionism score. All subscales were scored so that higher scores reflected increased levels of perfectionism. In this sample, Cronbach's α for the MPS was 0.94 while the subscales ranged in value from 0.79 to 0.92.
The General Health Questionnaire-28 (GHQ-28) [18] is a self-administered screening instrument designed to detect psychological distress in both clinical and non-clinical settings. Split-half reliability for the total scale has been reported as 0.95 [18]. A version of this instrument has been validated in Australian populations [19]. Suicide ideation was assessed on the basis of responses to four items which were extracted from the GHQ-28 in order to derive a measure of suicide ideation [1]. These were: ‘Have you recently felt that life isn't worth living?’, ‘Have you recently thought of the possibility that you might do away with yourself?’, ‘Have you recently found yourself wishing you were dead and away from it all?’ and ‘Have you recently found that the idea of taking your own life kept coming into your mind?’
As demonstrated by Goldney et al. [1], scoring the GHQ-28 questions in a Likert manner (0, 1, 2, 3) represents increasing intensity of suicide ideation for each question. Summing the scores for each question produces a score on a scale of suicide ideation, which can then be used for comparison of participants' scores on other instruments. In addition to a score for suicide ideation, a total score on the GHQ-28 was calculated for each respondent, with higher values indicating a higher level of psychological distress.
Although there is an absence of psychometric properties for this measure in the literature, significant associations between the suicide ideation questions of the GHQ-28 with measures of depression, hopelessness, anomie and self-esteem have been previously reported in an Australian university student population [1]. The internal reliability of the GHQ-28 in this sample was 0.90 (Cronbach's α), while the subscale for suicide ideation was 0.89.
Procedure
Students were surveyed during university lecture periods. The purpose of the study was explained and students were informed that participation was voluntary. Participants were asked to read the information form preceding the questionnaire before completing the survey. The total response rate was 80.5%, with 10 questionnaires invalidated due to incomplete responses.
Results
Data analysis
The data were grouped according to absence or presence of suicide ideation for analysis of variance on all dependent measures. Follow-up analysis of the suicide ideation scale involved comparisons between perfectionist groups, which were calculated on per-centile values of the upper and lower thirds of MPS total scores.
Prevalence of suicide ideation
Frequencies of responses to the individual suicide questions were assessed and tabulated by gender (Table 1) for comparative purposes with the previously published data [1]. With the Likert scoring method, responses to the four questions were consistent, with 81.7% to 83% of all subjects definitely denying thoughts of suicide.
Responses to questions on suicide ideation by gender and total sample
Slightly fewer women than men reported suicidal thoughts, with a maximum of 16% of women and 25.7% of men indicating some degree of suicide ideation. Scores for all four questions were then summed to derive a score for each case on a scale of suicide ideation, which was used to create a dichotomous variable for the presence of suicide ideation (0 = no, 1 > = yes).
Both suicide ideation groups were cross-tabulated by gender for proportional differences. Chi-squared analysis demonstrated statistical significance by gender (χ2 = 4.78, df = 1, 398, p = < 0.05, with results indicating that males (35%) were more likely than females (23.8%) to report suicide ideation. Of the total sample, 26.6% gave at least one positive response to one or more questions, indicating the presence of suicide ideation.
Suicide ideation and perfectionism
Mean scores for all measures for both suicide ideation groups are presented in Table 2. As analysis for age and gender effects on GHQ-28 total scores and MPS total scores were non-significant, subsequent analyses were collapsed to include the total sample.
Means, standard deviations, univariate F-ratios and probability values for Multidimensional Perfectionism Sub-Scale scores and differences between groups
Separate one-way, between-groups ANOVAS were conducted to determine if the suicide ideation groups differed on total GHQ-28 and total MPS scores. As expected, there were significant between group differences in both GHQ total scores (F = 132.57, df = 1, 373, p < 0.001) and MPS total scores (F = 4.86, df = 1, 387, p < 0.05). The suicide ideation group had higher levels of both psychological distress and perfectionism than those reporting an absence of suicide ideation. Means and standard deviations are shown in Table 2.
To determine if the suicide ideation groups differed on perfectionism dimensions, MANOVA was conducted with the five-subscale scores of the MPS as the dependent variables. With the use of Wilks' criterion, the main effect of suicide ideation was significant (F = 3.55, df = 5, 383, p < 0.01). To control for family-wise error, follow-up univariate tests were set at an adjusted alpha level of 0.01. These tests revealed significant group differences on two perfectionism dimensions, CM and D. Between-group differences were non-significant for PC, PE and PS. Means, standard deviations and results of univariate F-tests are shown in Table 2. These results show higher mean scores for the suicide ideation group on all subscale scores except PS.
The final analysis involved comparisons between low, moderate and high perfectionism groups on scores for the suicide ideation scale. All ANOVA results indicated suicide ideation scores differed significantly between perfectionist groups (Kruskal-Wallis χ2 = 9.59, df = 2, n = 389, p < 0.01).
There were no significant differences on suicide ideation scores between either the low and moderate perfectionist groups (Kruskal-Wallis χ2 = 2.82, df = 1, n = 258, p > 0.05) or the moderate and high perfectionist groups (Kruskal-Wallis χ2 = 2.57, df = 1, n = 261, p > 0.05).
Discussion
Prevalence of suicide ideation
The prevalence of suicide ideation over the few weeks prior to testing was similar to the figures reported by Goldney et al. [1], as assessed both for individual questions and for the results overall (17.5% of women and 20.2% of men in Goldney's sample vs 16.3% of women and 25.7% of men in the current study reported some degree of suicide ideation). It should be noted that it is feasible to score the data using a binary system, which results in stricter response criteria and hence lower prevalence scores. The finding that males were more likely than females to report the presence of suicide ideation is consistent with a previous study of Australian university students [20], which reported a higher proportion of males in category 2 (high level of suicide ideation). The results of the present study support the utility of the GHQ-28 for the assessment of suicide ideation and provide a basis for further comparisons in future Australian studies.
Perfectionism and suicide ideation
The presence of suicide ideation was associated with significantly increased scores on a measure of psychological distress (GHQ-28 total scores), a global measure of perfectionism (MPS total scores) and two perfectionism dimensions, providing support for the hypothesis that the suicide ideation group would report higher levels of perfectionism and increased psychological distress. Furthermore, significant differences between low, moderate and high perfectionist groups were found on the scale of suicide ideation, indicating support for the second hypothesis that there is a significant and positive relationship between increased levels of perfectionism and suicide ideation. This result is consistent with previous studies undertaken in the USA [10–12,21] and supportive of earlier theorising [8,22].
Dimensions of perfectionism
Concern over Mistakes and Doubts about Actions are two dimensions of perfectionism and of key interest. These dimensions reflect negative reactions to mistakes and a tendency to doubt the quality of one's performance [4,23]. As previously noted, in their post-hoc formulation of the relationship between perfectionism dimensions and suicidal preoccupation, Adkins and Parker have suggested that perfectionism creates ‘impediments to action’ in those with high levels of passive perfectionism. In contrast, those with high levels of active perfectionism are motivated to achieve, rather than paralysed or inhibited by the perfectionism [12].
According to Frost et al. [4], Doubts about Actions reflects a tendency to feel that projects are not completed to satisfaction. This dimension has been described as a vague sense of doubt about the quality of one's performance [8,22], and is assessed in the MPS by items from the Maudsley Obsessive-Compulsive scale of Rachman and Hodgson, 1980 [24]. Although not an area of inquiry in the present study, it should be noted that obsessive-compulsive features have been previously emphasised as core characteristics of the perfectionistic personality, and perfectionism has been seen as a core component included in the diagnosis of obsessive-compulsive personality disorder [25].
To return to the original question, ‘the cost of being perfect’, we acknowledge that striving toward high standards is admirable, but there appears to be a point, beyond which, the person's striving takes on the characteristic of a trait with an obsessional dimension, termed perfectionism. We agree with Adkins and Parker [12], who have suggested that perfectionism, as a trait, predisposes individuals to suicide ideation and preoccupation. The current findings confirm that while perfectionism in the generic sense is valued in tertiary students, passive perfectionists (defined as perfectionists who are extremely afraid of mistakes, tend to procrastinate and frequently second-guess their decisions) are at increased risk for suicide preoccupation. In contrast, those defined as active perfectionists are not at increased risk of suicide ideation. Given the significant associations between suicide ideation and the Concern over Mistakes and Doubts about Actions dimensions, and that students scoring high in overall perfectionism differed to both moderate and low perfectionist groups on suicide ideation, the results of the present study are supportive of the distinction.
Conclusion
These initial findings relating a multidimensional model of perfectionism as a meaningful variable in suicide ideation have been undertaken in a non-clinical Australian sample. The findings reveal a significant relationship between the presence of suicide ideation and particular dimensions of perfectionism. Future research is needed to gain a better understanding of the complex interrelationship between personality and temperament, environmental factors and self-destructive behaviour.