The Self-Generated Stress Scale: Development,Psychometric Features,and Associations With Perfectionism,Self-Criticism,and Distress

Abstract

In the current article, we describe the development and validation of a self-report measure of self-generated stress and its associations with measures of perfectionism, self-criticism, and distress. The Self-Generated Stress Scale is a seven-item inventory that taps the tendency to see oneself as someone who generates and adds to existing personal stress (i.e., making a challenging situation worse or turning a life problem into a bigger problem). Psychometric analyses with data from three samples of university students showed that the Self-Generated Stress Scale has one factor and acceptable internal consistency. Analyses established that self-generated stress is associated with trait perfectionism, self-criticism, dependency, and self-silencing, as well as indices tapping cognitive perfectionism and perfectionistic self-presentation. Self-generated stress was also associated with distress and psychosomatic health symptoms. Regression analyses established that scores on the Self-Generated Stress Scale predict unique variance in distress and negative affect beyond the variance attributable to self-criticism and other measures of stress. Overall, our findings attest to the further use of the Self-Generated Stress Scale in various contexts and highlight that certain perfectionists suffer from a perceived tendency to make their lives more stressful. The implications of these findings are discussed along with directions for future research.

Keywords

perfectionism self-criticism dependency stress generation depression

One prevailing view of psychopathology and distress is that people have a vulnerability or diathesis that makes them susceptible and this vulnerability is activated following the experience of significant life stress, especially if this stress reflects themes of personal importance that are central to one’s self-definition. The general view that existed for many years in the research literature is that stress is uncontrollable and it is something that just happens to an individual who is the victim of misfortune. Seminal work by Hammen (1991) changed this view. Hammen (1991) introduced the notion that certain people might actually be responsible for some or most of the stress in their lives because they have engaged in behavior and made decisions that made their lives more stressful. That is, they generated stress for themselves and it is this tendency that is believed to contribute to the onset or persistence of depression. This view sees stress as dependent on the actions and tendencies of the person who undergoes this stress. For instance, the person who creates conflicts with other people would be high in stress generation.

The concept of stress generation is highly relevant to perfectionism. Hewitt and Flett (1993) first suggested that the association among perfectionism, stress, and depression is due, in part, to a tendency for perfectionism to be associated with stress generation. Hewitt and Flett (2002) provided an expanded conceptual account that identified perfectionism and stress generation as one of four ways that stress contributes to the link between perfectionism and depression. Perfectionists consistently generate stress for themselves through the pursuit of unrealistic goals, but they can also generate stress through the conflict inherent in holding other people up to impossible standards or making choices while under pressure to live up to perfectionistic expectations imposed on the self. We observed further that “Perfectionistic behavior can generate stress that stems, in part, from perfectionists’ tendencies to stringently evaluate themselves and others, focus on negative aspects of performance, and experience little satisfaction” (Hewitt & Flett, 2002, p. 259).

There have been countless studies of stress generation over the past three decades (Hammen, 2006, 2018; Liu & Alloy, 2010). Typically, stress generation is assessed in one of two ways. Hammen (1991) has assessed stress generation with the semi-structured UCLA Stress Interview. Participants are queried about the stressful experiences in their lives. Raters then review their responses and determine the degree to which this person is someone who has created stress in his or her own life. This type of stress is known as “dependent stress” because it is seen as depending on the actions and decisions of the person who reported the stress (i.e., the stress generator). Another approach is to have a team of raters assess which stressors seemingly reflect dependent stress after the participant has completed a stressful events checklist.

A relatively small proportion of studies have examined perfectionism and stress generation with these types of methods, and the obtained results are consistent with the proposed link between perfectionism and stress generation. La Rocque, Lee, and Harkness (2016) established links between self-oriented perfectionism and achievement-related stress, as well as socially prescribed perfectionism and interpersonal stress in a sample of 503 university students. Stress generation was assessed by having a team of raters evaluate the stressful events reported on an events checklist. Mandel, Dunkley, and Starrs (2018) conducted longitudinal research on perfectionism and stress generation with a sample of 145 community adults. Stress generation was evaluated 4 years after the study began when participants took part in the interview developed by Hammen (1991). Mandel et al. (2018) found that a composite measure of self-critical perfectionism that included socially prescribed perfectionism and self-criticism was associated with the generation of interpersonal stressors. Another study of chronic fatigue patients focused on self-critical perfectionism represented by a measure of trait self-criticism (Luyten et al., 2011). Five raters evaluated responses to a 114-item checklist measure of daily hassles. It was confirmed that self-critical perfectionism was associated significantly with the generation of daily hassles. Finally, Hewitt and Flett (2002) described supplementary analyses of the life events evaluated by a team of raters in an earlier study by Hewitt, Flett, and Ediger (1996). It was reported that both self-oriented and socially prescribed perfectionism were associated with events rated as possibly dependent, but these dimensions were not associated with independent events that did not reflect stress generation.

Given the manner in which these stress generation assessments are conducted, it cannot be assumed that the person who was interviewed or who reports stressful events on a checklist is necessarily aware of her or his role in generating stress. Some people may be highly attuned to their role in stress generation, whereas others may be oblivious to the ways in which they have generated stress for themselves. The current research described in this article reflects our interest in focusing on the extent to which people are actually cognizant of their contributions to stress. Do people see themselves as “stress generators” by creating stress and adding to their existing stress? To our knowledge, a self-report measure of stress generation has not been developed, though there is some evidence indicating that people can identify stressors that are dependent on their own actions (Scallion & Cummings, 2018). There are several reasons why a self-report measure of self-generated stress would be a useful addition to the literature. First, and foremost, it would be meaningful in clinical and counseling settings to be able to assess whether the person who is a stress generator actually has a sense of contributing to life problems and other stressors. It is almost certain that people who are rated as generating stress vary considerably in the extent to which they are aware of this tendency and are cognitively focused on this tendency. It could also be conceivable that some people do not play a substantial role in stress generation but, nevertheless, perceive themselves as being stress generators.

Second, the ability to conduct a stress interview with raters requires resources that may not be available to certain investigators. The creation of a brief self-report measure alleviates this problem, albeit with the caveat that interviewer-based assessments are obviously more desirable in most instances.

Finally, the availability of such a measure would represent another means of testing hypothesized associations stemming from conceptual models, such as the proposed link between stress generation and perfectionism (Hewitt & Flett, 2002). This issue was a central theme addressed in the current research. However, in doing so, we took a broader approach to perfectionism and stress generation by examining perfectionism not only in terms of the trait dimensions, but also in terms of cognitive perfectionism (i.e., automatic thoughts about needing to be perfect) and perfectionistic self-presentation. We expected that the links with self-reported self-generated stress would apply to these various ways of conceptualizing and assessing perfectionism. The experience of frequent automatic thoughts may result in a cognitive preoccupation and degree of internal pressure that makes adaptive responding less likely, and a proposed link would be in keeping with the general association between rumination and stress generation (Stroud, Sosoo, & Wilson, 2018). Perfectionistic self-presentation may create a distance and a lack of authenticity that complicates matters for the already stressed perfectionist.

We describe the development and characteristics of the Self-Generated Stress Scale (SGSS) below. We included several other measures that should be part of the nomological network of the stress generation construct. For instance, self-criticism was a focus because people who acknowledge their role in stress generation may be harsh toward themselves. The construct tapped by the SGSS likely overlaps to some degree with a self-blaming attributional style, but our specific emphasis is on awareness of the personal tendency to make challenging circumstances more stressful.

Parenthetically, it should also be noted that we set out to develop a brief measure with item content that can be broadly applicable to people from various backgrounds, even though our work focuses on university students. The SGSS is designed to tap a perspective that we believe can be found in people in general and that may be implicated in the perceived experience of chronic stress and elevated levels of various forms of distress and dysfunction (e.g., personality disorder, addiction, nonsuicidal self-injury).

Method

Participants and Procedure

This study was based on three samples of university students. The first sample had 247 students (155 women, 91 men, and one person who did not report gender). Their mean age was 20.3 years (SD = 3.7). The majority of the participants (70%) were in their first year of university, with 17% in their second year.

The second sample consisted of 120 university students (86 women, 34 men). Their mean age was 20.4 years (SD = 3.9). Many of the participants (58.3%) were in their first year of study, with 23.3% in their second year.

The third sample consisted of 228 university students (157 women, 71 men) with a mean age of 18.9 years (SD = 2.3). First-year students took part in a larger project examining adjustment among individuals making the transition to university. Participants were approached during the first 6 weeks of the fall semester and again in February. The current data are from this second assessment. Participants were paid US$10 for their participation at Time 2. Overall, 75% of the participants had completed high school during the previous year.

Participants in the first sample were administered the following self-report questionnaires after providing their consent.

SGSS

The instructions for this scale are shown in the appendix along with the items in the final version. Initially, a pool of 15 items was created. The scale length was decreased to seven items following the results of our initial confirmatory factor analysis and closer inspection of item wording. Items were rejected if they seemed to have too much of an attributional focus. A representative item from the final version is “The ways I have dealt with stress in my life have often resulted in extra stress for me.” Two items intended originally to be reverse-scored items can be retained as filler items if desired: “My stress is mostly due to other people and is not because of me” and “The stress in my life is mostly due to random events and is not due to me.” Items are rated on a scale from 1 (strongly disagree) to 5 (strongly agree).

Multidimensional Perfectionism Scale (MPS)

This 45-item scale measures three trait perfectionism dimensions: self-oriented (e.g., “One of my goals is to be perfect in everything I do”), other-oriented (e.g., “If I ask someone to do something, I expect it to be done flawlessly”), and socially prescribed (e.g., “Success means that I must work even harder to please others”) (Hewitt & Flett, 1991). The psychometric properties of the MPS have been well established in student and clinical samples (Hewitt & Flett, 1991, 2004).

Perfectionistic Self-Presentation Scale (PSPS)

This 27-item scale assesses a style that reflects the need to seem perfect rather than be perfect (Hewitt et al., 2003). It assesses perfectionistic self-promotion (e.g., “I try always to present a picture of perfection”), nondisplay of imperfection (e.g., “I will do almost everything to cover up a mistake”), and nondisclosure of imperfection (e.g., “I should always keep my problems to myself”). This scale possesses good psychometric properties (Flett & Hewitt, 2015; Hewitt et al., 2003).

Perfectionism Cognitions Inventory (PCI)

This 25-item inventory assesses the frequency of perfectionistic automatic thoughts over the past week (Flett, Hewitt, Blankstein, & Gray, 1998). Sample items include “People expect me to be perfect” and “Why can’t things be perfect?” Extensive research has shown that the PCI possesses good reliability and validity (Flett & Hewitt, 2015; Flett et al., 1998).

Center for Epidemiological Studies-Depression Scale (CES-D)

This 20-item self-report measure assesses the frequency of depressive symptoms experienced over the past week (Radloff, 1977). Sample items include “I felt sad” and “I had crying spells.” The CES-D possesses good reliability and validity (Radloff, 1977).

UCLA Loneliness Scale

This 20-item measure assesses levels of perceived loneliness (Russell, Peplau, & Cutrona, 1980). A sample item is “How often do you feel left out?” This measure possesses sufficient reliability and validity (Russell et al., 1980).

The participants in our second sample completed the SGSS, the Hewitt and Flett (1991) MPS, the PSPS, and the CES-D. The following self-report scales were also administered over the Internet to participants:

Frost Multidimensional Perfectionism Scale (FMPS)

This 35-item scale assesses concern over mistakes (e.g., “People will probably think less of me if I make a mistake”), personal standards (e.g., “I set higher goals than most people”), parental expectations (e.g., “Only outstanding performance is good enough in my family”), parental criticism (e.g., “I never felt like I could meet my parents’ standards”), doubts about actions (e.g., “I usually have doubts about the simple everyday things I do”), and organization (Frost, Marten, Lahart, & Rosenblate, 1990). We dropped the organization subscale to be consistent with prior research. The FMPS subscales possess good psychometric properties (Flett & Hewitt, 2015).

Inventory of College Students’ Recent Life Experiences (ICSRLE)

We used an abbreviated 21-item version of the ICSRLE to assess daily hassles of university students (Kohn, Lafreniere, & Gurevich, 1990). Respondents rate the extent to which each hassle applies to their lives over the last month on a scale from 0 (not at all part of my life) to 3 (very much part of my life). Sample items include “Struggling to meet the academic standards of others,” “Social rejection,” and “Not enough time to meet your obligations.” The ICSRLE has good psychometric properties (Kohn et al., 1990).

The participants in our third sample were administered the SGSS, along with the following personality measures and indices of adjustment:

The McGill Revision of the Depressive Experiences Questionnaire (MDEQ)

This 48-item scale assesses self-criticism (e.g., “There is a considerable difference between how I am now and how I would like to be”) and dependency (e.g., “After an argument, I feel very lonely”) (Santor, Zuroff, & Fielding, 1997). The subscales have good reliability and validity (Santor et al., 1997; Santor, Zuroff, Mongrain, & Fielding, 1997).

Coping Inventory for Stressful Situations (CISS)

The CISS taps coping dispositions (Endler & Parker, 1994, 1999). It requires respondents to indicate how frequently they engage in various activities when they encounter difficult, stressful, or upsetting situations. It assesses three factors: task-oriented (16 items: e.g., “Schedule my time better”), emotion-oriented (16 items: e.g., “Blame myself for procrastinating”), and avoidance-oriented (16 items) coping, which measures distraction (eight items: e.g., “Go out for a snack or meal”) and social diversion (five items: e.g., “Visit a friend”). The psychometric properties and evidence of construct validity are summarized in Endler and Parker (1999).

The Positive Affect and Negative Affect Schedule (PANAS)

The PANAS has 10 adjectives assessing trait positive affect (e.g., “excited”) and trait negative affect (e.g., “upset”) (Watson, Clark, & Tellegen, 1988). Participants rated positive and negative emotions/feelings according to how much they generally experience them. Scores on the PANAS have demonstrated high test–retest reliability and acceptable convergent validity (Watson et al., 1988).

SUNYA Revision of the Psychosomatic Symptom Checklist (PSC)

The PSC measures 17 psychosomatic symptoms (e.g., “headaches,” “fatigue,” and “dizziness”; Attanasio, Andrasik, Blanchard, & Arena, 1984). Respondents rate the frequency and intensity of symptoms. However, we examined only the frequency scores. This instrument has demonstrated reliability and validity in university students (Attanasio et al., 1984).

The Hopkins Symptom Checklist–Abbreviated (HSCL-A)

This 21-item measure has subscales assessing negative affectivity (labeled general feelings of distress; e.g., “feeling blue”), somatic distress (e.g., “pains in the lower part of your back”), and performance difficulty (e.g., “difficulty in speaking when you are excited”) (Green, Walkey, McCormick, & Taylor, 1988). It also provides a total score. Green et al. (1988) reported that the subscales represent distinct factors with internal consistencies of .75 or greater among students.

The Silencing The Self Scale (STSS)

This 31-item inventory has four factors tapping silencing the self (e.g., “I don’t speak my feelings in an intimate relationship when I know they will cause disagreement”), externalized self-perception (e.g., “I tend to judge myself by how I think other people see me”), care as self-sacrifice (e.g., “Caring means putting the other person’s needs in front of my own”), and divided self (e.g., “Often I look happy enough on the outside, but inwardly I feel anxious and rebellious”) (Jack & Dill, 1992). The overall score was used in this study. Jack and Dill (1992) established that this scale has good psychometric properties.

Results

Descriptive Statistics

The SGSS had acceptable internal consistency, with alphas of .81 for Sample 1, .79 for Sample 2, and .85 for Sample 3. The means for self-generated stress were 22.11 (SD = 5.33) for Sample 1, 23.49 (SD = 5.14) for Sample 2, and 21.10 for Sample 3 (SD = 5.87). Tests of sex differences in mean scores found no significant differences, except in the second sample. Women reported significantly higher self-generated stress than men, t = −2.51, p = .014.

Confirmatory Factor Analysis

A confirmatory factor analysis was conducted for the first sample with AMOS Version 18 software using maximum likelihood estimation procedures to examine the model fit of the SGSS items. A one-factor model of self-generated stress was tested with the final seven items selected for inclusion. This model was an acceptable fit to the data, χ²(14) = 44.06, p = .00, comparative fit index (CFI) = .93, Tucker–Lewis index (TLI) = .90, standardized root mean square residual (SRMR) = .05, root mean square error of approximation (RMSEA) = .09, 90% confidence interval (CI) = .06, .13, p_close = .011. As can be seen in Table 1, all items had factor loadings of .45 or higher. The results suggest that our new scale is unidimensional, as expected.

Table 1.

Factor Loadings for the Items of the Self-Generated Stress Scale—Sample 1 (N = 247).

Items	Factor loadings
I have a habit of putting myself in situations that are more stressful than they need to be	.57
The ways I have dealt with stress in my life have often resulted in extra stress for me	.63
I seem to create a lot of problems for myself	.71
Much of the stress I experience is due to the choices I make	.60
At times, I seem to have a knack for making a stressful situation much worse	.74
Part of my stress reflects my decision to be associated with certain people	.45
My mistakes tend to make things a lot more stressful for me	.61

Correlational Analyses

Table 2 displays the correlations among self-generated stress, trait and self-presentational perfectionism dimensions, perfectionistic cognitions, depression, and loneliness in our first sample. The results indicated that self-generated stress was positively correlated with self-oriented and socially prescribed perfectionism. The correlation with socially prescribed perfectionism was significantly stronger than the correlation between self-generated stress and self-oriented perfectionism (z = −3.80, p = .001). In addition, self-generated stress was positively associated with all PSPS facets and the PCI, as well as with depression and loneliness. Further correlational analyses showed that socially prescribed perfectionism, all facets of perfectionistic self-presentation, and perfectionistic cognitions were positively correlated with both depression and loneliness.

Table 2.

Correlations Among Self-Generated Stress, Dimensions of Trait and Self-Presentational Perfectionism, Perfectionistic Cognitions, Depression, and Loneliness—Sample 1 (N = 247).

Measures	1	2	3	4	5	6	7	8	9	10
1. SGSS	—
2. Self	.16*	—
3. Other	.10	.54**	—
4. Social	.47**	.36**	.38**	—
5. Promote	.33**	.57**	.42**	.56**	—
6. Ndisplay	.43**	.36**	.30**	.48**	.79**	—
7. Ndisclose	.35**	.37**	.35**	.57**	.59**	.60**	—
8. PCI	.40**	.50**	.34**	.55**	.57**	.50**	.54**	—
9. CES-D	.53**	.11	.06	.45**	.33**	.38**	.38**	.45**	—
10. UCLA	.39**	.10	.11	.41**	.28**	.38**	.49**	.29**	.59**	—

Note. SGSS = Self-Generated Stress Scale; Self = self-oriented perfectionism; Other = other-oriented perfectionism; Social = socially prescribed perfectionism; Promote = perfectionistic self-promotion; Ndisplay = nondisplay of imperfection; Ndisclose = nondisclosure of imperfection; PCI = Perfectionism Cognitions Inventory; CES-D = Center for Epidemiological Studies-Depression Scale; UCLA = UCLA Loneliness Scale.

p < .05. **p < .01, two-tailed.

Table 3 presents the correlations for the second sample. Self-generated stress was positively correlated with self-oriented perfectionism, socially prescribed perfectionism, and all PSPS facets. Self-generated stress was also positively correlated with other-oriented perfectionism and all FMPS dimensions. The highest correlation involving self-generated stress was with the FMPS doubts about actions subscale. Finally, self-generated stress was positively associated with daily hassles and depression.

Table 3.

Correlations Between Self-Generated Stress and Dimensions of Trait Perfectionism, Perfectionistic Self-Presentation, Daily Hassles, and Depression—Sample 2 (N = 120).

	Self-generated stress
Self-oriented perfectionism	.39**
Other-oriented perfectionism	.25**
Socially prescribed perfectionism	.38**
Concern over mistakes	.40**
Personal standards	.39**
Parental expectations	.36**
Parental criticism	.21*
Doubts about actions	.51**
Perfectionistic self-promotion	.34**
Nondisplay of imperfection	.33**
Nondisclosure of imperfection	.23*
Daily hassles	.44**
Depression	.41**

p < .05. **p < .01, two-tailed.

All trait and self-presentational dimensions of perfectionism were correlated positively with depression in our second sample. Moreover, all of the perfectionism dimensions except for nondisclosure of imperfection were positively correlated with daily hassles. These results are available upon request. Finally, scores on the measure of daily hassles were positively correlated with depression (r = .41, p < .01).

Table 4 displays the correlations for our third sample. There were robust associations between SGSS scores and both self-criticism and dependency. In addition, there was a significant link between SGSS scores and overall levels of self-silencing. The pattern of results with the coping measure showed that the SGSS was associated significantly with emotion-oriented coping, but not associated with task-oriented or avoidance-oriented coping.

Table 4.

Correlations Between Self-Generated Stress and Dimensions of Personality, Coping, Adjustment, and Affect—Sample 3 (N = 228).

	Self-generated stress
Dependency	.54**
Self-criticism	.62**
Positive affect	−.20**
Negative affect	.49**
Task-oriented coping	−.01
Emotion-oriented coping	.32**
Avoidance-oriented diversion	.10
Avoidance-oriented social distraction	.04
Hopkins—general distress	.42**
Hopkins—performance difficulties	.40**
Hopkins—somatic	.30**
SUNYA psychosomatic symptoms	.32**
Self-silencing	.46**

p < .05. **p < .01, two-tailed.

Analyses with the affect and adjustment measures showed that there was a negative association between the SGSS and positive affect, but a much stronger positive association between the SGSS and negative affect. The SGSS was also associated with all three HSCL-A measures and the SUNYA measure of psychosomatic symptoms.

Regression Analyses

A hierarchical multiple regression analysis was performed with data from the second sample to evaluate whether self-generated stress would predict depression beyond the variance attributable to another stress measure (i.e., daily hassles). The measure of daily hassles was entered into the first predictor block, followed by self-generated stress (see Table 5). The results revealed that daily hassles significantly predicted 16.8% of the variance in depression scores, F(1, 118) = 23.91, p < .001. Self-generated stress predicted an additional 6.8% of the variance, F(2, 117) = 18.08, p < .001.

Table 5.

Summary of Hierarchical Multiple Regression for Variables Predicting Depression—Sample 2 (N = 120).

Variable	R ²	∆R²	B	SE B	95% CI
Step 1	.168***
Daily hassles			.26***	.07	[.15, .41]
Step 2		.068**
Self-generated stress			.38**	.13	[.11, .62]

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

Regression analyses were also conducted with the data from Sample 3 to establish whether the SGSS could account for unique variance in the affect and adjustment variables beyond the considerable variance attributable to self-criticism and dependency. This was a stringent test of our new measure given the extensive associations between SGSS scores and both self-criticism and dependency. The results are summarized briefly here due to space considerations. When positive affect was the outcome variable, an initial block composed of self-criticism and dependency accounted for a significant 4% of the variance, with self-criticism being a significant individual predictor. Subsequent entry of the SGSS accounted for an additional 2% of the variance, t = −2.41, p < .05. When negative affect was the outcome variable, self-criticism and dependency accounted for a significant 21.3% of the variance, with self-criticism once again being a significant individual predictor. The SGSS accounted for an additional 5.3% of the variance, t = 4.02, p < .001.

The same analysis conducted with the measure of psychosomatic symptoms found that the SGSS did not predict beyond the 11.6% of the variance predicted by the MDEQ measures. However, this was not the case with the HSCL-A measures. For instance, when total HSCL-A scores comprised the outcome variable, self-criticism and dependency accounted for a significant 23% of the variance, with self-criticism once again being a significant individual predictor. Subsequent entry of the SGSS accounted for an additional 2.3% of the variance, t = 2.81, p < .001.

Mediation Analysis

A structural equation model was tested for the first sample with AMOS Version 18 software using maximum likelihood estimation procedures to investigate whether self-generated stress would mediate the associations that socially prescribed perfectionism and perfectionistic cognitions have with distress (see Figure 1). This model is based generally on hypotheses advanced by Hewitt and Flett (2002). In the hypothesized model, the predictors were socially prescribed perfectionism and perfectionistic cognitions, the mediator was self-generated stress, and the outcome was a latent distress factor comprising both depression and loneliness. The correlation between socially prescribed perfectionism and perfectionistic cognitions was accounted for in this model. The model was a good fit to the data, χ²(2) = 5.23, p = .073, CFI = .99, TLI = .96, SRMR = .02, RMSEA = .08, 90% CI = [.00, .17], p_close = .198. Both socially prescribed perfectionism and perfectionistic cognitions were positively associated with self-generated stress, which was, in turn, linked positively with distress. Socially prescribed perfectionism and perfectionistic cognitions were both still positively associated with distress after taking the mediator into account.

Figure 1.

Mediated model of socially prescribed perfectionism, perfectionistic cognitions, self-generated stress, and distress—Sample 1.

We evaluated whether the indirect effects of socially prescribed perfectionism and perfectionistic cognitions on distress through self-generated stress were significant. To test the indirect effects, 2,000 bootstrap samples were generated to obtain 95% bias-corrected CIs. The 95% CI for the mediated effect of socially prescribed perfectionism on distress through self-generated stress was .05 to .17, and the 95% CI for the mediated effect of perfectionistic cognitions on distress through self-generated stress was .01 to .08. Because these CIs do not contain zero, the indirect effects were both significant. Therefore, self-generated stress mediated the links that socially prescribed perfectionism and perfectionistic cognitions have with distress.

Discussion

The current research was designed to create a psychometrically sound self-report measure of self-generated stress and then evaluate its associations with various measures representing the perfectionism construct and poor adjustment. Our results with data from three university student samples established that it is feasible and meaningful to assess individual differences in self-reported self-generated stress. A brief seven-item measure consisting of one factor with acceptable internal consistency was developed. It can be inferred from our results that there is a subset of university students who have a clear sense that they create stress in their lives and they have a history of exacerbating their stressors and life problems in ways that will prolong the stress experience and could make life stressors and problems seem overwhelming. It is likely that these students have low self-esteem given the established role of chronically low self-esteem in stress generation (Schiller, Hammen, & Shahar, 2016). It follows that students who perceive that they generate stress may not attempt to cope or problem-solve due to a sense that their efforts will prove to be self-defeating because they will only make things worse. This view ought to promote substantial avoidance coping and procrastination in dealing with life concerns.

The correlational results from our first two samples attested to the proposed association between trait perfectionism and self-generated stress first proposed by Hewitt and Flett (1993, 2002). It was found that both self-oriented and socially prescribed perfectionism were associated significantly with self-reported self-generated stress, though the association with self-oriented perfectionism in our first sample was quite modest. It was also found that this link with trait perfectionism extended to concern over mistakes and doubts about actions; these doubts and concerns become more understandable when they are found among someone who perceives that she or he tends to make stressors and problems worse rather than better.

Additional results indicated that this link between self-generated stress and perfectionism extended beyond trait perfectionism. A tendency to experience frequent automatic thoughts about needing to be perfect and to be high in various facets of perfectionistic self-presentation was also associated with reports of self-generated stress. It is likely that the people who are highly focused on projecting an image of perfection will also be people who are not willing or able to discuss with other people those times when they have made stressful circumstances even more stressful. It is also likely that to the extent that these people actually do make things more stressful in interpersonal circumstances, this could add to the objective social disconnection believed to plague certain perfectionists (for a discussion, see Hewitt, Flett, & Mikail, 2017).

Other results obtained in this research were consistent with the predictions outlined by Hewitt and Flett (2002). Mediational analyses yielded support for the contention that self-generated stress mediates the links that exist between distress and both socially prescribed perfectionism and cognitive perfectionism in the form of frequent automatic thoughts about needing to be perfect. This was not tested in our first sample for self-oriented perfectionism given the modest link found between self-oriented perfectionism and self-generated stress. Clearly, there are several possible reasons why self-reported self-generated stress should act as a mediator for socially prescribed perfectionism and perfectionistic automatic thoughts. Vulnerable individuals who report self-generated stress will be exceptionally hard on themselves and this may include an acute awareness of being discrepant from standards and becoming even more discrepant as a result of reacting in ways that add to their already high levels of stress. They will likely be somewhat disengaged from attempts to cope and problem-solve and they will likely also anticipate and expect that their tendency to generate stress will continue into the future.

There are clear implications for the perfectionist who reports a tendency to generate stress. Most notably, a perceived tendency to make stressful situations even more stressful could be regarded as a form of self-regulatory failure that is in opposition to the goals and ambitions of the perfectionist. In addition, self-awareness of self-generated stress likely adds to the pressure experienced by many perfectionists. Clearly, the person who sees his or her life as full of situations that have only been made worse is not someone who is having the perfect life. Similarly, the person who needs to seem perfect but who makes life more stressful is also someone who is failing to meeting the goal of projecting the desired image of being perfect.

The results from our third sample provided a revealing overview of the type of student who feels like he or she has a tendency to generate stress. As expected, the association with perfectionism extended to a self-critical form of perfectionism. Robust positive associations were found between scores on our new measure and indices of trait self-criticism and dependency. The student who reports elevated self-generated stress is also someone who tends to have an emotion-oriented coping style and who engages in self-silencing. This stress orientation is also linked characteristically with diminished positive affect and greater negative affect, along with other forms of distress. Other results point to the potential health implications of seeing oneself as the type of person who generates stress for themselves and perhaps for other people.

As expected, our correlational results established that students with higher SGSS scores also tended to have more daily hassles. Similarly, we would expect that students with high SGSS scores would also report elevated levels of major life events stress, life pressures, and negative social interactions if these types of stress were assessed. Importantly, the results of our regression analysis suggested that perceived self-generated stress is unique in that SGSS scores accounted for unique variance in depression scores beyond daily hassles. Future research examining self-generated stress with the SGSS needs to examine the daily lives of respondents to confirm that these reported associations are in keeping with the life experiences of these individuals. A vital issue for further research is to ascertain whether people with high scores on the SGSS persist in perceiving themselves as having played a role in stressors that most other people would deem to be beyond their personal control or sphere of influence. That is, self-perceived stress generators could conceivably have an overdeveloped sense of their responsibility when it comes to life stress.

Further analyses showed that elevated SGSS scores not only were associated with depression, but they were also associated with higher levels of loneliness. This finding points to perceived self-generated stress as a potential contributor to loneliness and experiences of isolation and social disconnection. However, given that these associations were found with cross-sectional data, it cannot be presumed that self-generated stress is a vulnerability factor for depression and loneliness. Longitudinal research is needed and it seems, at least on the surface, that self-reported self-generated stress is a factor that ought to be associated with various forms of distress in a cyclic, sequential manner.

Although there were many novel elements of the current research, it is also evident that several issues must be examined in future research. First, and foremost, more work is needed on the psychometric properties of the SGSS. This inventory needs to be evaluated to determine such things as its test–retest reliability and degree of overlap with response style measures. Here, we would expect that there is not much contamination due to social desirability and high scorers on the SGSS will have relatively little self-deception. Research is also needed to better understand the precipitants of self-generated stress. One key possibility that merits further consideration is being unable to control hostility and the expression of anger adds to stress generation (Sahl, Cohen, & Dasch, 2009). The generalizability of the current findings in community and clinical samples also needs to be evaluated. Parenthetically, we should reiterate that although we restricted our focus to university students, the SGSS items were worded so that they can be administered to people from various backgrounds and stages of life, so this proposed extended focus and analysis of the SGSS is clearly warranted. Indeed, a key issue for future research is to evaluate whether there is a discernible tendency to perceive self-generated stress that stretches across the lifespan. Overall, much can be learned through applying a developmental lens to self-reported self-generated stress.

In summary, we developed a brief self-report measure of self-generated stress and found that scores on this new measure were associated with trait perfectionism, cognitive perfectionism, and perfectionistic self-presentation. Other results showed that elevated SGSS scores are linked robustly with self-criticism and dependency and people who see themselves as stress generators also tend to engage in self-silencing. Not surprisingly, self-perceived stress generation was associated with distress, psychosomatic symptoms, and negative affect, and there were substantial indications of incremental validity in that self-reported stress generation can account for unique variance in these measures of distress and negative affect. People who acknowledge their role in stress generation are clearly in need of stress counseling and perhaps conflict management skills, but their difficulties likely go beyond this because they have incorporated into their identity the sense that there is something about themselves that tends to make a bad situation worse. This tendency would suggest the need for therapeutic interventions designed to provide a more positive identity and self-view built around such themes as unconditional self-acceptance and self-compassion. As we have indicated elsewhere (Hewitt et al., 2017), these interventions may need to be extensive if it turns out that seeing oneself as a stress generator is rooted in early life experiences and feedback from significant others that have instilled and reinforced this unsettling and unfavorable view of the self.

Footnotes

Appendix

Author Contributions

G.L.F. conceived and created the scale, facilitated data collection, designed and conducted data analyses, and participated in the writing of the manuscript; P.L.H. provided key conceptual insights; and T.N. conducted data collection, designed and conducted analyses, and participated in the writing of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The first author was supported by a Canada Research Chair in Personality and Health.

ORCID iD

Taryn Nepon

References

Attanasio

Andrasik

Blanchard

E. B.

Arena

J. G.

(1984). Psychometric properties of the SUNYA revision of the Psychosomatic Symptom Checklist. Journal of Behavioral Medicine, 7, 247-257.

Endler

N. S.

Parker

J. D.

(1994). Assessment of multidimensional coping: Task, emotion, and avoidance strategies. Psychological Assessment, 6, 50-60. doi:10.1037/1040-3590.6.1.50

Endler

N. S.

Parker

J. D.

(1999). Coping Inventory for Stressful Situations (CISS): Manual. Toronto, Ontario, Canada: Multi-Health Systems.

Flett

G. L.

Hewitt

P. L.

(2015). The assessment of perfectionism. In Boyle

G. J.

Saklofske

D. H.

Matthews

(Eds.), Measures of personality and social psychological constructs (Vol. 2, pp. 595-618). London, England: Elsevier.

Flett

G. L.

Hewitt

P. L.

Blankstein

K. R.

Gray

(1998). Psychological distress and the frequency of perfectionistic thinking. Journal of Personality and Social Psychology, 75, 1363-1381.

Frost

R. O.

Marten

Lahart

Rosenblate

(1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449-468.

Green

D. E.

Walkey

F. H.

McCormick

I. A.

Taylor

A. J.

(1988). Development and evaluation of a 21-item version of the Hopkins Symptom Checklist with New Zealand and United States respondents. Australian Journal of Psychology, 40, 61-70.

Hammen

(1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100, 551-561.

Hammen

(2006). Stress generation in depression: Reflections of origins, research, and future directions. Journal of Clinical Psychology, 62, 1065-1082.

10.

Hammen

(2018). Risk factors for depression: An autobiographical review. Annual Review of Clinical Psychology, 14, 1-28.

11.

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.

12.

Hewitt

P. L.

Flett

G. L.

(1993). Dimensions of perfectionism, daily stress, and depression: A test of the specific vulnerability hypothesis. Journal of Abnormal Psychology, 102, 58-65.

13.

Hewitt

P. L.

Flett

G. L.

(2002). Perfectionism and stress in psychopathology. In Flett

G. L.

Hewitt

P. L.

(Eds.), Perfectionism: Theory, research, and treatment (pp. 255-284). Washington, DC: APA Press.

14.

Hewitt

P. L.

Flett

G. L.

(2004). Multidimensional Perfectionism Scale: Technical manual. Toronto, Ontario, Canada: Multi-Health Systems.

15.

Hewitt

P. L.

Flett

G. L.

Ediger

(1996). Perfectionism and depression: Longitudinal assessment of a specific vulnerability hypothesis. Journal of Abnormal Psychology, 105, 276-281.

16.

Hewitt

P. L.

Flett

G. L.

Mikail

S. F.

(2017). Perfectionism: A relational approach to assessment, treatment, and conceptualization. New York, NY: Guilford.

17.

Hewitt

P. L.

Flett

G. L.

Sherry

S. B.

Habke

Parkin

Lam

R. W.

. . . Stein

M. B.

(2003). The interpersonal expression of perfection: Perfectionistic self-presentation and psychological distress. Journal of Personality and Social Psychology, 84, 1303-1325.

18.

Jack

D. C.

Dill

(1992). The Silencing the Self Scale: Schemas of intimacy associated with depression in women. Psychology of Women Quarterly, 16, 97-106.

19.

Kohn

P. M.

Lafreniere

Gurevich

(1990). The Inventory of College Students’ Recent Life Experiences: A decontaminated hassles scale for a special population. Journal of Behavioral Medicine, 13, 619-630.

20.

La Rocque

C. L.

Lee

Harkness

K. L.

(2016). The role of current depression symptoms in perfectionistic stress enhancement and stress generation. Journal of Social & Clinical Psychology, 35, 64-86. doi:10.1521/jscp.2016.35.1.64

21.

Liu

R. T.

Alloy

L. B.

(2010). Stress generation in depression: A systematic review of the empirical literature and recommendations for future study. Clinical Psychology Review, 30, 582-593.

22.

Luyten

Kempke

Van Wambake

P. V.

Claes

Blatt

S. J.

Van Houdenhove

(2011). Self-critical perfectionism, stress generation, and stress sensitivity in patients with chronic fatigue syndrome. Psychiatry, 74, 21-30.

23.

Mandel

Dunkley

D. M.

Starrs

C. J.

(2018). Self-critical perfectionism, daily interpersonal sensitivity, and stress generation: A four-year longitudinal study. Journal of Psychopathology and Behavioral Assessment, 40, 701-713. doi:10.1007/s10862-018-9673-7

24.

Radloff

L. S.

(1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.

25.

Russell

Peplau

L. A.

Cutrona

C. E.

(1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472-480.

26.

Sahl

J. C.

Cohen

L. H.

Dasch

K. B.

(2009). Hostility, interpersonal competence and daily dependent stress: A daily model of stress generation. Cognitive Therapy and Research, 33, 199-210.

27.

Santor

D. A.

Zuroff

D. C.

Fielding

(1997). Analysis and revision of the Depressive Experiences Questionnaire: Examining scale performance as a function of scale length. Journal of Personality Assessment, 69, 145-163.

28.

Santor

D. A.

Zuroff

D. C.

Mongrain

Fielding

(1997). Validating the McGill revision of the Depressive Experiences Questionnaire. Journal of Personality Assessment, 69, 164-182.

29.

Scallion

L. M.

Cummings

J. A.

(2018). Comparison of team and participant ratings of event dependence: Inferential style, cognitive style, and stress generation. Journal of Social & Clinical Psychology, 37, 697-724.

30.

Schiller

Hammen

C. C.

Shahar

(2016). Links among the self, stress, and psychological distress during emerging adult: Comparing three theoretical models. Self and Identity, 15, 302-326.

31.

Stroud

C. B.

Sosoo

E. E.

Wilson

(2018). Rumination, excessive reassurance seeking, and stress generation among early adolescent girls. Journal of Early Adolescence, 38, 139-162.

32.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070.