Fear of Relationship Loss: Attachment Style as a Vulnerability Factor in Job Burnout.

OBJECTIVE: To investigate the correlation between attachment styles and various burnout risk groups ("relaxed", "wornout", "challenged" and "burnout") and whether attachment styles suitably discriminate between individual burnout risk groups.
METHOD: The study involved 2,320 participants (1,668 women and 652 men), who completed an adrenal burnout syndrome questionnaire, a performance-based self esteem scale, a work addiction risk test and a relationship questionnaire.
RESULTS: A one-way analysis of variance confirmed attachment style differences between burnout risk groups. The challenged and burnout groups differed from the relaxed and wornout groups by having a significantly lower secure attachment style score and a higher insecure (avoidant and preoccupied) attachment style score. The canonical discriminant analysis showed that the predictors (secure, preoccupied and avoidant attachment styles) can be used to appropriately classify 85.4% of respondents in the predicted burnout risk groups.
CONCLUSION: The study confirmed the hypothesis that two insecure attachment styles (i.e. avoidant and preoccupied) predominate in the challenged and burnout groups, and that a secure attachment style predominates in the relaxed and wornout groups. Burnout syndrome can thus be conceived as the result of excessive and compulsive efforts to retain a relationship that is perceived as insecure or to reduce (excessive) fear of losing this relationship.

Introduction

Burnout and Wornout

Burnout can be conceived of as a process resulting from unsuccessful (workaholic) efforts to maintain a stable self-image through achievements, which are a component part of one’s self-esteem. Burnout is a decompensation, which is the final result of the process of self-exhaustion through overcompensated activity (1).

“Wornout” is an expression denoting a feeling of cognitive and physical exhaustion, and emotional reactions to this condition without the associated signs of anxiety and fear (2), or exhaustion that does not result from attempts to create or maintain self-esteem (3).

In terms of burnout risk, a distinction can be made between the following four groups: “relaxed”, “challenged”, “wornout” and “burnout”. Hallsten et al. (3) distinguished between these groups based on the following two criteria: a high or low level of burnout and high or low performance-based self-esteem. Pšeničny and Perat (4) in turn use three criteria: a high or low level of burnout, high or low performance-based self-esteem, and workaholism.

Compulsive motivation that ultimately leads to exhaustion and burnout originates from a deficit in self-esteem that is labile and dependent on achievements and external validation (performance-based self-esteem). The fear of losing self-esteem can be so strong that it leads to excessive working (and emotional) over-investment. This over-investment is also an indicator of poorer self-regulation when it exceeds all limits and hard work is replaced by its compulsive form (i.e. workaholism), which ultimately poses a threat to physical and mental health. These two traits, which may also indicate a personality disorder in most individuals at risk of burnout, can therefore be considered a risk factor for the development of burnout syndrome. These risk-posing traits are only weakly expressed in relaxed and wornout individuals (4).

Attachment Styles and Fear of Relationship Loss

In his attachment theory, Bowlby (5, 6) explains that feelings of (in)security are the consequence of (un) successful interactions between a child and his/her primary caregiver (most often the mother). These repeating patterns of behaviour in interaction with the caregiver transform into mental representations. These specific mental representations are referred to as attachment styles, which constitute one’s basic mental model for understanding the social world. Moreover, an attachment style is an implicit and automated method of organising stimuli. However, it is subject to perceptual and cognitive errors.

Bartholomew and Horowitz (7) empirically validated four attachment styles and categorised them into a secure attachment style (representations of both the self and the other are predominantly positive) and three insecure attachment styles: a preoccupied attachment style (a negative representation of self and a positive representation of the other), a fearful-avoidant attachment style (the representations of self and of the other are relatively negative) and a fearful-dismissing attachment style (the representation of self is positive and the representation of the other is negative).

Adult attachment styles and mental representations both influence the formation of the models of self, of other and of relationships with others - that is, they also influence the (subconscious) expectations about relationships with others and our understanding of the social environment. Individuals with an implicit insecure attachment style in general experience relationships as less secure and stable (8, 9).

According to the sociometer theory (10), threats to one’s relational value, such as social rejection, exclusion and criticism, are the most acute modern-day stressors because they undermine the feeling of social value, esteem and status (11). Research confirms the correlation between social stressors and job burnout (12), between burnout and hypersensitivity to rejection and criticism (4), and between an insecure attachment style and hypersensitivity to rejection and abandonment as a predictor of burnout (13).

Attachment Style, Regulation of Stress Response and Work Environment

Early experiences of secure or insecure emotional attachment play a crucial role in the development of an individual’s personality. Among other things, they are key to an individual’s ability to regulate emotions and cope with stressful situations, and have a significant effect on how an individual responds to the threat of losing an important attachment object (14).

The experience of insecure attachments results in a higher level of neuroticism or a generally weaker ability to regulate stress responses (15), and in various forms of personality traits that increase personal vulnerability, such as perfectionism, emotional lability, etc. (16) and workaholism (1, 17).

Compared to those with an insecure attachment style, employees with a secure attachment style express higher general satisfaction, enjoy their work more, and are less worried about relationships at work. They experience less stress and cope with it more successfully, create a better work-life balance, and know how to seek help. Insecurely attached individuals are more afraid of rejection and poor performance, tend to be more over-involved, and feel less important and valued at work. Some avoid social contact (18, 19, 20, 21).

Several studies have confirmed a correlation between attachment styles and stress and burnout at work. Their authors report a negative correlation between a secure attachment style and burnout symptoms, and positive correlations between burnout and insecure attachment styles (i.e. preoccupied and avoidant) (15, 22, 23, 24). These correlations are understandable because people become attached to their work and the workplace, as well as to the people they encounter in this important living environment. The same notions and emotional responses, including expectations and fears, as exist in any other attachment relationship therefore come to the fore within this social context. Hence, for most people work and the work environment are important sources of security or insecurity. People who grew up experiencing insecure attachment, however, have at-risk personality traits that, in certain conditions, drive them into a process of exhaustion and may increase their work-related or other engagements endlessly (1, 4).

Research Problem

We explored whether the relaxed and wornout groups differ from the challenged and burnout groups in terms of attachment style.

Hypotheses

This study’s basic premise was that insecure attachment affects the development of traits that pose a risk of burnout (i.e. performance-based self-esteem and workaholism). Individuals with a higher level of at-risk traits were therefore also expected to have more pronounced insecure attachment styles. The following working hypotheses can be specified based on these premises:

H The average score on various attachment styles (secure, preoccupied, avoidant and dismissing) differs significantly by burnout risk group (independent variable: relaxed, challenged, wornout and burnout).

H A significant correlation is expected between various attachment styles (secure, preoccupied, avoidant and dismissing) and burnout risk groups (relaxed, challenged, wornout and burnout).

H The set of attachment styles (secure, preoccupied, avoidant and dismissing) produces a suitable distinction between the burnout risk groups.

Method

Participants and Instruments

Participants

The convenience sample included participants that completed the questionnaires posted on the website of the Institute for Human Resource Development between January 2018 to April 2019. A total of 2,320 individuals (1,668 women and 652 men) completed the questionnaires. Fourteen per cent of them were unemployed (including students and seniors), 69% were in employment, 7% were self-employed and 10% held managerial positions. The youngest respondent was 18 years old and the oldest was 69 (M=38.50, SD=11.33).

Instruments

The Adrenal Burnout Syndrome Questionnaire or ABSQ (25) comprises four scales: body symptoms (45 items), emotional symptoms (94 items), behavioural symptoms (61 items) and cognitive symptoms (46 items). Cronbach’s α=0.962. The outcomes are the variable “average adrenal burnout syndrome rate” or ABSRa (with scores from 0 to 3), hereinafter referred to as “burnout,” and the categorical variable “burnout classes” or ABSCl (0=no symptoms; 1=mild; 2=medium; 3=strong burnout). The latter was further converted into the dichotomous variable “burnout categories” or ABSCat (low level=no symptoms; high level=1, 2, 3).

The participants were then further divided into four burnout risk groups, whereby the variable “risk” (ABSRi) was obtained. The classification procedure is described below under “H1 testing.”

The Performance-Based Self-Esteem Scale or PBSE Scale (3) comprises four statements and measures performance-based self-esteem. The responses were rated on a five-point Likert scale (1=Fully disagree, 2=Somewhat disagree, 3=Neither agree nor disagree, 4=Somewhat agree, 5=Fully agree). The questionnaire’s reliability measured using Cronbach’s alpha was 0.872. The outcome is the variable “average test score” or, as we called it, “performance-based self-esteem” or PBS (with scores from 1 to 5). This variable was further converted into the dichotomous variable “performance-based self-esteem categories” or PBSCat (low level=1–2.45; high level =2.46–5). The criterion used was the average score reported by the authors of this scale.

The Work Addiction Risk Test or WART (26) comprises 25 items and measures the tendency for compulsive hard work or workaholism. Cronbach’s α=0.966. The responses were rated on a four-point Likert scale (1=Never true, 2=Sometimes true, 3=Often true, 4=Always true). The outcome of the test is the variable “workaholism” or WORKHL (with scores ranging from 25 to 100). The average score reported by Robinson is 47, which was used as the criterion for converting this variable into the dichotomous variable “workaholism categories” or WORKCat (low level: 25–47; high level=48–100).

Based on the results of these three questionnaires, the respondents were divided into four burnout risk groups: relaxed, challenged, wornout and burnout. Three classification criteria were used: a low or high level of burnout (ABSCat), low or high performance-based self-esteem (PBS Cat), and a low or high level of workaholism (WORKCat), as the last two variables are the main predictors of burnout (1).

The variable “risk” (ABSRi) was thus obtained, with four groups of respondents at various risk of burnout: the relaxed group (a low level of performance-based self-esteem, workaholism and burnout), challenged (a high level of performance-based self-esteem and/or workaholism and/or a low level of burnout), the wornout group (a low level of performance-based self-esteem and workaholism, and a high level of burnout), and the burnout group (a high level of performance-based self-esteem and/or workaholism and burnout).

The Relationship Questionnaire or RQ (7) comprises four short paragraphs describing a prototypical attachment pattern as it applies in close adult peer relationships (Bartholomew & Horowitz, 1991). Participants first select (forced choice) one of the four attachment styles (secure, preoccupied, avoidant and dismissing) and then rate each one on a seven-point Likert scale. These scores form the profile of an individual’s attachment style. The outcome is the categorical variable “attachment style” (STYLE) and four continuous variables: “secure attachment style” (SECAS), “preoccupied attachment style” (PRE-AS), “avoidant attachment style” (AVO-AS), and “dismissing attachment style” (DIS-AS; scores for all these ranging from 1 to 7). The last four variables were combined into an aggregate variable “attachment styles” (ASTYLES).

Procedure

The PBSE scale and WART test were translated into Slovenian by Andreja Pšeničny and Mitja Perat with the author’s permission, after which they were back-translated by an English specialist.

Participants completed the questionnaires posted on the website of the Institute for Human Resource Development between January 2018 and April 2019. Their personal data was protected in accordance with the Slovenian Personal Data Protection Act.

Results

Descriptive Sample Statistics

Attachment styles by category (forced choice): the secure attachment style was selected by 31.1% of participants, the preoccupied style by 19.3%, the avoidant by 27.5% and the dismissing by 22%. The differences were statistically significant (χ2 (3)=79.02, p<0.01)

The differences in the test scores between the low- and high-level categories (see the Methods section for the variables used) are shown in Table 3.

Table 3

Frequencies, descriptive statistics and differences in test scores by category “low level” and “high level” for burnout, performance-based self-esteem and workaholism.

		N	M	SD	t	t-test df	p
ABSCat	Low	1,130	0.06	0.11	−66.19	1778.04	0.00**
	High	1,190	1.22	0.58
PBSCat	Low	431	1.69	0.41	−60.15	120.50	0.00**
	High	2,694	3.74	0.66
WORKCat	Low	335	37.84	7.28	−41.21	133.521	0.00**
	High	1,225	67.85	12.77

Statistically significant difference p<0.01

Notes: ABSCat=adrenal burnout syndrome categories; PBSCat=performance-based self-esteem categories; WORKCat=workaholism categories

Based on the results shown in Table 3, participants were divided into four burnout risk groups. Table 4 shows the descriptive statistics for the selected parameters by individual ABSRi category.

Table 4

Descriptive statistics (mean scores and standard deviations) for performance-based self-esteem, workaholism and burnout by burnout risk group (ABSRi), and the results of the one-way analysis of variance.

		M	M	F df(3.3389)	p
Performance-based self-esteem	Relaxed	19.16	5.60	1,489.78	0.00**
	Challenged	41.55	10.50
	Wornout	20.49	4.47
	Burnout	43.82	9.11
Workaholism	Relaxed	39.24	8.91	712.45	0.00**
	Challenged	63.37	17.46
	Wornout	44.50	7.22
	Burnout	67.47	13.75
Burnout	Relaxed	.03	.08	1,686.04	0.00**
	Challenged	.03	.02
	Wornout	.81	.29
	Burnout	1.29	.60

Statistically significant difference p<0.01

Table 5 shows the final classification of participants into ABSRi categories.

Table 5

Participant classification into burnout risk groups (ABSRi variable).

	Low ABSCat	High ABSCat
Low PBSCat and WORKCa	Relaxed	Wornout
	832 (35.9%)	328 (14.1%)
High PBSCat and WORKCat	Challenged	Burnout
	258 (11.1%)	902 (38.9%)

Notes: ABSCat=adrenal burnout syndrome categories; PBSCat=performance-based self-esteem categories; WORKCat=workaholism categories

H1 Testing

Proceeding from the assumption that an insecure attachment style is associated with burnout risk, the challenged and burnout group ought to have a higher score for insecure attachment styles (preoccupied, avoidant and dismissing) and a lower secure attachment style score than the relaxed and wornout groups.

To check this, we carried out a repeated measures analysis of variance. Because Mauchly’s test showed that sphericity (χ2(2)=189.60, p<0.01) was violated, we corrected the degrees of freedom using Greenhouse–Geisser sphericity estimates.

The results confirmed our hypothesis for two insecure attachment styles (i.e. preoccupied and avoidant) and the secure attachment style.

The challenged and burnout group showed a significantly lower secure attachment style score and a significantly higher score for two insecure attachment styles (i.e. preoccupied and avoidant) than the relaxed and wornout (Figure 1, Table 6). A post-hoc comparison of difference also shows no statistically significant differences between the relaxed and wornout groups in any attachment style. The dismissing attachment style is significantly lower among the relaxed group than among the burnout group (p<0.05), whereas no significant differences in this style can be observed between the other groups. The first part of the hypothesis that burnout risk groups also differ by the degree of expression of secure and insecure attachment styles can be confirmed for the secure and two insecure attachment styles (i.e. preoccupied and avoidant), whereas it can only partly be confirmed for the dismissing style.

Figure 1

Comparison of attachment style mean scores (secure, preoccupied, avoidant and dismissing) between burnout risk groups (relaxed, challenged, wornout and burnout).

Table 6

Summary of one-way repeated measures analysis of variance: attachment style mean score (ASTYLES variable) by group (relaxed, challenged, wornout and burnout; (ABSRi variable).

Variability source	SS	df	MS	F	p
Between individuals
Cross section	464.632	3	154.877	56.215	0.00**
Error	6,380.744	2,316	2.755
Within individuals
ASTYLES	1,482.684	2.627	564.354	108.028	0.00**
ASTYLES*ABSRi	3321.341	7.882	421.400	80.664	0.00**
Error	31,787.110	6,084.654	5.224

Statistically significant difference p<0.01

If it is primarily insecurely attached individuals who burn out, one would expect a considerably higher percentage of those that chose an insecure attachment style among the challenged and burnout group than among the relaxed and wornout groups.

As anticipated, more than a half of the relaxed and wornout individuals reported a secure attachment style, whereas only one-fifth reported the same among the challenged, and only every thirtieth among the burnout group. The avoidant attachment style predominated in these two groups (Table 7). The differences were statistically significant (χ2(9)=1,208.619, p<0.01). Based on this, the second part of the hypothesis can also be confirmed.

Table 7

Comparison of the percentage of participants with secure and insecure attachment styles (ASTYLES variable) between burnout risk groups (ABSRi variable).

			Attachment styles (ASTYLES)
			Secure	Preoccupied	Avoidant	Dismissing
Burnout risk groups	Relaxed	N	446	56	34	296
		%	53.6%	6.7%	4.1%	35.6%
	Challenged	N	53	49	102	54
		%	20.5%	19.0%	39.5%	20.9%
	Wornout	N	193	48	15	72
		%	58.8%	14.6%	4.6%	22.0%
	Burnout	N	30	295	488	89
		%	3.3%	32.7%	54.1%	9.9%

Note: No cells (0.0%) have an expected number below 5. The minimum expected number is 49.82

H2 Testing

In the final stage, we also explored whether the set of attachment styles studied also enabled appropriate classification of participants into the burnout risk groups defined based on three criteria (i.e. performance-based self-esteem, workaholism and burnout). A canonical discriminant analysis with a gradual inclusion of variables was used. The variable “dismissing attachment style” (DIS-AS) was excluded from the analysis.

All three functions are statistically significant (Table 8). The first function is moderately positively correlated with the preoccupied and avoidant attachment styles. The second function is very strongly positively correlated with secure attachment, and the third function is very strongly positively correlated with the remaining predictors, with the preoccupied style being inversely proportionate to this function (Table 9). Based on these three predictors, originally 85.4% of participants were classified into appropriate burnout risk groups (Table 10). The third hypothesis can thus be confirmed.

Table 8

Testing the equality of centroids by burnout class (ABSCl).

Functions tested	Wilks’s lambda	χ2	df	p
1 / 3	0.12	4,921.07	12	0.00**
2 / 3	0.80	507.64	6	0.00**
3	0.95	11.53	2	0.00**

Statistically significant difference p<0.01

Table 9

Relative influence of predictors on the functions’ values (standardised coefficients) and the correlation between the “ASTYLES” variables and functions (structural matrix).

	Standardised coefficientsFunctions			Structural matrixFunctions
	1	2	3	1	2	3
Secure style	−0.42	−0.92	0.15	−0.38	0.92*	−0.05
Preoccupied style	0.88	0.35	−0.61	0.34	0.31	−0.85*
Avoidant style	0.94	0.40	0.47	0.57	0.12	0.78*
Dismissing style	−0.03	0.06	0.27	−0.00	−0.01	0.41*

Maximum absolute correlations between each variable and individual discriminant function

Table 10

Percentages of appropriately classified members of individual burnout risk groups.

		Predicted group members
		Relaxed	Challenged	Wornout	Burnout
Original group members	Relaxed	86.3%	4.1%	9.6%	0.0
	Challenged	0.0	91.1%	8.9%	0.0
	Wornout	0.0	10.7%	89.3%	0.0
	Burnout	0.0	15.9%	2.7%	81.5%

Discussion

Like many previous studies, this study also confirmed that individuals who developed insecure attachment styles due to specific circumstances during their childhood burn out more often and more strongly than those that were securely attached to their parents or caregivers. The new finding presented in this study is that a distinction can also be made between the wornout and burnout individuals based on their attachment styles. What the burnout and wornout groups have in common is a general feeling of overtiredness and exhaustion, but unlike the burnout group, the wornout do not show any significant psychopathological characteristics and symptoms (3, 4) argue that performance-based self-esteem is the psychopathological characteristic that distinguishes the wornout from the burnout. Pšeničny and Perat (4) also added workaholism as the third criterion. The same criteria are used to distinguish between the relaxed and the challenged. Due to these risky personality traits, we consider the “challenged” to be potential candidates for burnout.

The study confirmed the hypothesis that a secure attachment style predominated among wornout and relaxed individuals and that insecure attachment styles were more common among the burnout and challenged. In addition, attachment styles proved to be an important predictor for classifying participants into these four categories (over 85% of participants were appropriately classified).

Even in more strenuous and stressful circumstances, more securely attached individuals will experience only (normal) signs of exhaustion (wornout). Unlike burnout, this does not lead to decompensation because these individuals will not neglect their needs due to excessive fear of losing an important relationship. In our opinion, securely attached individuals will respond reciprocally: they will take criticism, rejections and threats into consideration, and increase their investment only to a reasonable extent because they are able to emotionally cope with potential loss.

In individuals with preoccupied and avoidant attachment styles, the fear of losing an important relationship, including interpersonal relationships at work and employment itself, may be such a strong stressor that it can function as compulsive motivation and trigger over-investment (workaholism) or a search for constant validation through performance (performance-based self-esteem) with a single goal in mind: to retain this relationship. Adverse psychological as well as objective work circumstances and inappropriate interpersonal relationships in the workplace can be a real trigger for this fear for all employees, except that it is disproportionately strong in those who carry an insecure attachment style and triggers defensive over-engagement. Burning out and burnout can thus be conceived as the result of excessive and compulsive efforts to retain a relationship that is perceived as insecure, or to reduce (excessive) fear of losing this relationship.

Because the burnout group shows characteristics indicating the presence of psychopathological factors, and hence differs from the wornout group in this regard, this finding should be taken into account in selecting measures for preventing and treating these conditions. In addition to measures aimed at improving work conditions and acquiring more suitable skills and knowledge (which can successfully reduce wornout), individuals who are burning out also require professional help.

Long-term (development-oriented) psychotherapy has a beneficial effect on changing a person’s attachment style by simultaneously and inversely proportionately increasing the secure attachment style and reducing insecure attachment (27). Various forms of training, coaching and counselling, as well as shorter forms of therapy (e.g. cognitive behavioural therapy) may also be helpful for managing the symptoms and constructing short-term coping strategies for better functioning in difficult work situations.

This information is also important for employers in managing burnout. Specifically, burnout risk can be reduced in individuals identified as challenged (burnout candidates) by taking into account their specific vulnerabilities and taking appropriate action, increase their feelings of secure employment and decrease burnout triggers.

Limitations

This study’s limitations lie in the unrepresentative nature of the sample and the instruments used. The convenience sample included in the study does not necessarily reflect the percentage of burnout, wornout and challenged individuals within a population. The Relationship Questionnaire used has its own limitations, particularly in terms of the validity of its content. Therefore, it cannot be used reliably in clinical practice.

Conclusion

The study confirmed the hypothesis that two insecure attachment styles (i.e. avoidant and preoccupied) predominate in the challenged and burnout groups, and that a secure attachment style predominates in the relaxed and wornout groups. Burnout syndrome can thus be conceived as the result of excessive and compulsive efforts to retain a relationship that is perceived as insecure or to reduce (excessive) fear of losing this relationship.

Table 1

Burnout risk groups.

Burnout risk	Performance-based self-esteem and/or workaholism
	Low	High
Low	Relaxed	Challenged
High	Wornout	Burnout

Table 2

Descriptive statistics for the testing results (mean scores and standard deviations).

Variable	M	SD
Burnout (ABSRa)	0.65	0.72
Performance-based self-esteem (PBS)	34.21	14.09
Workaholism (WORKHL)	56.59	17.98
Secure style (SEC-AS)	4.12	2.19
Preoccupied style (PRE-AS)	3.29	2.05
Avoidant style (AVO-AS)	3.63	2.24
Dismissing style (DIS-AS)	3.59	2.01