Age-related associations between work over-commitment and zest for work among Swedish employees from a cross-sectional and longitudinal perspective

Abstract

BACKGROUND: In aging societies, zest for work may be pivotal when deciding to stay occupationally active longer. Psychosocial work stress is a prevalent public health problem and may have an impact on zest for work. Work over-commitment (WOC) is a personal coping strategy for work stress with excessive striving and a health risk. However, the long-term effect of WOC on zest for work is poorly understood.

OBJECTIVE: To investigate the age-related associations of work over-commitment with zest for work.

METHODS: During 1996–1998 and 2000–2003, predominantly industrial workers (n = 2940) participated in the WOLF-Norrland study and responded to a questionnaire referring to socio-demographics, WOC, zest for work, effort-reward imbalance proxies, and mental health. Age-adjusted multiple logistic regressions were performed with original and imputed datasets.

RESULTS: Cross-sectionally, work overcommitted middle-aged employees had an increased prevalence of poor zest for work compared to their contemporaries without WOC (OR: 3.74 [95%-CI 2.19; 6.40]). However, in a longitudinal analysis associations between onset of ‘poor zest for work’ and the WOC subscales ‘need for approval’ (OR: 3.29 [95%-CI 1.04; 10.37]) and ‘inability to withdraw from work’ (OR: 5.14 [95%-CI 1.32; 20.03]) were observed.

CONCLUSION: The longitudinal findings among older employees could be relevant regarding the expected need to remain occupationally active longer.

Keywords

Work stress WOC ERI coping

1 Introduction

In globally aging societies with shrinking work forces zest for work may become an important motivator to staying occupationally active longer. The basic idea behind ‘zest for work’ goes back to Super 1953 [1]: work has an influence on both work and life satisfaction that depends on the extent to which work satisfies the person’s expectations through values, outlet for abilities, and personality traits. The research on zest for work is sparse but the concept comes close to the global concept of job satisfaction, which is defined as the extent to which persons like or dislike their jobs [2]. Zest, in general, has been shown to be correlated to work satisfaction [3]. Work-related factors might have an effect on zest for work. Psychosocial stress such as high job demands are predictive of negative emotions towards work, deteriorated health, poor work functioning, increased risk of accidents/injury at work, increased spells of sick absence, and reduced zest for work [4]. If the work environment is rewarding a person should, theoretically speaking, attribute a positive value to work and have zestful feelings towards work on the way to work [5]. Reduced job satisfaction could be caused by high work stress; role stressors, such as role conflict and role ambiguity, were negatively related to affective commitment mediated through job satisfaction [6].

Work-stress creates negative emotions at work, with an emphasis on stressful events and on reward/punishment aspects [7], and creates small margins for maneuver for reciprocal and supportive relationships between employers and employees. The well-tested stress model, Effort-Reward-Imbalance (ERI), introduced by Siegrist [8] emphasizes two sources of stress: the extrinsic and intrinsic components of the model. While the ‘extrinsic’ component describes an imbalance between rewards gained and efforts spent at work, which can introduce chronic work-related stress, the ‘intrinsic’ component work over-commitment (WOC) describes an aberrant coping style. Overcommitted employees are assumed to misjudge external demands and their internal coping abilities lead to exhaustion and detrimental effects on health in the long term [8]. Another characteristic of employees prone to WOC is a tendency to make large efforts at work that they perceive as not being appropriately rewarded [8, 9]. An imbalance between large efforts and low rewards at work was associated with a risk of job dissatisfaction [10, 11]. This higher risk for job dissatisfaction may be particularly true for employees prone to WOC since it has been shown that WOC and ERI are highly associated [12].

There may be age-related differences in the associations between WOC and zest for work. While this aspect has not yet been tested for the association between WOC and zest for work, Siegrist [8] assumed an age-dependent change of WOC in the sense of a ‘coping career’. This concept of coping career traces back to the original assumptions of Matschinger and coworkers [12] where they suggested that younger employees prone to WOC under estimate demands and overestimate coping potential while older employees do the opposite. While the coping style of younger employees can be successful for them, the one chosen by older employees can be detrimental to their health. However, this assumption has never been tested empirically. We therefore studied age-related associations between WOC traits and low zest for work. The employee may have a misperception about the relationship between his investment and the expected rewards, which may lead on the one hand to low zest temporarily, or on the other hand to a chronic state of low zest for work.

WOC could influence zest for work temporally but also over several years. Even reversed causation may apply. Both aspects could be relevant regarding time-dependent consequences of low zest for work, such as a lower productivity and sickness absence or the choice of sustainable preventive strategies. We therefore investigated the associations between WOC and zest for work, cross-sectionally as well as longitudinally. Little is known about the relationship between personality traits of overcommitted persons and zest for work and the commonly used short version of WOC contains little information [9]. We used the comprehensive long version of WOC and investigated total WOC scores as well WOC subscales regarding the following hypotheses:

WOC is positively associated with low zest for work in the cross-sectional multiple regression analyses.

WOC is positively associated with low zest for work in the longitudinal multiple regression analyses.

The predictive value of WOC on low zest for work differs depending on age.

2 Material and methods

2.1 Design and participants

Data from the WOLF-Norrland study (WOLF-N) was analyzed. The WOLF-N study is a prospective cohort study aimed to investigate the effects of psychosocial work factors and life style on health. All Occupational Health Service units (OHS) in Västernorrland and Jämtland, two provinces in Northern Sweden, were invited to participate between 1996 and 1998 – period 1 (T1). The OHS offered a health examination to all of their employees who were provided with questionnaires. Fifteen OHS organizations participated and the personnel were trained in standardised measurement procedures. The participants were re-invited to the follow-up, between 2000 and 2003 – period 2 (T2). Up to three reminders were sent in the follow-up study, unless the participant actively declined to participate. The study population at T1 included 5092 employees in the study area. The response rate was 93% (4715 subjects). At T2, the response rate was 71% (3637 participants). The mean time elapsed between T1 and T2 was five years (±11 months). Analyses were based on 2940 subjects who were occupationally active at both times.

Ethical approval was obtained from the Ethical Committee at Karolinska Institutet, Stockholm, Sweden.

2.2 Measurements

2.2.1 Dependent variable – low zest for work

One overall question was used to assess negative emotions towards work while travelling to work, ‘low zest for work’: “How do you usually feel when you are on your way to work?” The response options were answering yes (1) or no (0) to ‘I feel happy and satisfied at the thought of the work that awaits me’, ‘I have a quite positive feeling for work’,’ Neither positive nor negative feelings about work’ (combined and coded as 0 for zest for work), ‘Feelings of some discomfort towards work’, and ‘Feelings of strong aversion towards work’ (combined and coded as 1 for low zest for work). This question has been used earlier by Rubenowitz [13].

2.2.2 Independent variable – work over-commitment

WOC is the intrinsic component of the ERI-model. We used the original scale [9] with 29 items to get a deeper insight into the different facets of work overcommitment in relation to zest for work. Respondents were asked to indicate their degree of agreement on a four-point Likert scale worded ‘strongly disagree’ to ‘strongly agree’. After recoding items with reversed response options, the indices were summed and divided into low and high WOC. In the total WOC index, scores ranged between 29 and 116, with high scores indicating a critical style of coping with work demands. The 29 items were grouped into the four WOC indices: ‘need for approval’, ‘competitiveness’, ‘disproportional irritability’, and ‘inability to withdraw from work’. The indices were added and divided into low and high, with high values representing a negative coping style.

‘Need for approval’ comprised the questions, “I usually take criticism very seriously”, ”I get angry with myself when I cannot completely resolve a problem at work”, “I get especially frustrated when my work is not properly appreciated”, “I only feel successful when I perform better than I expected”, “Other people have confidence in my ability to handle difficult tasks”, and “The slightest compliment really boosts my confidence”. The internal consistency with a Cronbach’s alpha of 0.62 at T1 was reasonable.

‘Competitiveness’ constituted the questions “I am fueled by ambition”, “I enjoy proving certain people wrong”, “Always being a little better or faster than others is a sort of game to me”, “I do not let others do my work”, and “I get furious when anybody questions my competence”. The Cronbach’s alpha of 0.48 at T1 was low.

‘Disproportionate irritability’ consisted of the questions “Even the slightest interruption bothers me”, “I get upset with others more often than I should”, “I get easily overwhelmed by time pressures at work”, “I can get furious if someone does not understand me the first time”, “I do everything possible to be in control”, “I do not usually get annoyed when my work routine is interrupted”, “I always want more than I can get”, and “I am always mentally prepared to do what needs to be done next”. Cronbach’s alpha of 0.56 at T1 was low.

‘Inability to withdraw from work’ contained the questions “If something needs to be done right I would rather do it myself”, “I can get very upset when someone keeps me from what I am supposed to be doing”, “I start thinking about work as soon as I get up in the morning”, “When I get home, I can easily relax and forget all about work”, “People close to me say I sacrifice too much for my job”, “My private life come first, and then work”, “Work is usually on my mind when I go to bed”, “Every once in a while, I like when others keep me from working”, and “If I put off doing something that needs to get done today, I will have trouble sleeping at night”. Cronbach’s alpha of 0.68 at T1 was reasonable.

2.3 Confounding variables

The analyses were adjusted for personal factors (sex and education), effort-reward imbalance proxy [ERI-proxy), and mental ill health (GHQ-12) [14]. Investigating the predictive value of WOC for zest for work: zest for work was adjusted at baseline (T1) to control for reversed causation. Education was divided into three categories (low, primary and secondary school; middle, advanced secondary school (Swedish ‘gymnasium’); and high, university education).

2.4 Effort–reward imbalance proxy (ERI-proxy)

A detailed description of the effort-reward imbalance model and its validation has previously been published [8, 9]. Individuals scoring in the upper tertile of the ERI ratio have an imbalance between efforts spent (high costs at work) and rewards gained (monetary and non-monetary) at work, which, according to the model, may result in perceived stress. The ERI instrument was represented in the questionnaire provided at T2. However, the internal response rate was low. To compensate for that, proxy questions were used. Proxies underestimate the associations being studied, as highlighted in a study on the same population by Fahlén and coworkers [15]. The proxies were therefore weighted to approximate the original ERI measure. The weighting factor was calculated by the differences in point estimates between the original and proxy measures of ERI available in a subsample of the data set in the logistic regression analysis used to predict zest for work.

2.5 Mental ill-health

The General Health Questionnaire [5] investigates psychiatric morbidity, primarily depression and anxiety. The 12-item General Health Questionnaire (GHQ12), which measures depressive symptoms in particular, was used at T2 in the WOLF-F study. This information was not available at T1. Symptoms of ‘mental ill-health’ were classified as more than five symptoms.

2.6 Statistical analyses

Internal consistency of the four WOC indices was measured using Cronbach’s alpha. Multiple logistic regression analyses were used to calculate odds ratios (OR) and 95% confidence intervals (95% -CI) for the associations between zest for work and WOC index, or its subscales: need for approval, competitiveness, disproportional irritability, and inability to withdraw from work. All analyses were controlled for sex, education, ERI-proxy, and severe mental ill-health. All analyses were stratified based on age. SPSS 21 was used for all statistical analyses [16].

2.7 Handling the effects of missing data

Complete case analysis (CC) can be a source of bias if data are partially missing or missing in a non-random way, leading to a loss of power [17]. Single imputation (SI) methods tend to underestimate variability. As a state of the art technique, multiple imputations (MI) can compensate for this disadvantage of SI by generating, analyzing, and finally pooling several plausible imputed data sets. Therefore, multiple imputations with the fully conditional specification (FCS) algorithm in SPSS were used to replace missing data. FCS is a specific MI procedure that enables the imputation of parametric and non-parametric continuous and categorical items [18]. The results for CC and MI are presented in Tables 2 and 3.

3 Results

The majority of participants in all age groups were male (83.9% : Table 1). At T1, the mean age was 42 years with SD 8.9. The age range of participants was19–62 years, with one outlier under 19 years. In the whole sample, 15.2% were highly educated and 45.1% had a low education. The younger age groups had a larger proportion of highly educated employees and those with university degrees than the older age groups. Middle-aged employees had the highest percentage of employees with high scores on total WOC. For the WOC subscales, ‘disproportional irritability’ and ‘inability to withdraw from work’ was highest in the older age groups; this declined from the middle age group to the youngest age group. We tested interaction with age groups and found a tendency for significance, p = 0.06, for age for the ‘the inability to withdraw from work’. All other p-values for interaction were >0.1.

Table 1
Characteristics of respondents (n = 2940)

Characteristics Age group [years]

19–34 35–49 50–69

(n = 580) (n = 1495) (n = 864) p-values^*

Zest at T1 (n) (570) (1495) (862)

Low [%] 4.7 5.4 3.8 0.23

Zest at T2 (n) (570) (1479) (844)

Low [%] 4.3 5.5 3.4 0.08

WOC at T1 (n) (508) (1296) (742)

High [%] 28.3 43.6 40.4 <0.0005

Need for approval at T1 (n) (569) (1447) (832)

High [%] 37.3 35.0 37.1 0.49

Competiveness at T1 (n) (567) (1463) (840)

High [%] 46.0 45.5 44.0 0.24

Disproportional irritability at T1 (n) (569) (1476) (841)

High [%] 40.9 47.7 55.3 <0.0005

Inability to withdraw from work at T1 (n) (523) (1356) (784)

High [%] 39.0 45.3 55.5 <0.0005

Weighted ERI Proxy ratio at T1 (n) (563) (1417) (808) <0.0005

Low [%] 45.5 36.3 36.6

Middle [%] 31.4 29.3 28.7

High [%] 23.1 34.4 34.7

Ill mental health at T2 (n) 441 840 402

[%] 0.7 1.4 0.7 <0.0005

Gender at T1 (n) (579) (1493) (864) 0.12

Male [%] 86.4 84.1 82.3

Education at T1 (n) (579) (1488) (862) <0.0005

Low (n) (61) (676) (585)

[%] 10.5 45.4 67.9

Middle (n) (398) (591) (174)

[%] 68.7 39.7 20.2

High (n) (120) (221) (103)

[%] 20.7 14.9 11.9

Characteristics	Age group [years]
Zest at T1 (n)	(570)	(1495)	(862)
Low [%]	4.7	5.4	3.8	0.23
Zest at T2 (n)	(570)	(1479)	(844)
Low [%]	4.3	5.5	3.4	0.08
WOC at T1 (n)	(508)	(1296)	(742)
High [%]	28.3	43.6	40.4	<0.0005
Need for approval at T1 (n)	(569)	(1447)	(832)
High [%]	37.3	35.0	37.1	0.49
Competiveness at T1 (n)	(567)	(1463)	(840)
High [%]	46.0	45.5	44.0	0.24
Disproportional irritability at T1 (n)	(569)	(1476)	(841)
High [%]	40.9	47.7	55.3	<0.0005
Inability to withdraw from work at T1 (n)	(523)	(1356)	(784)
High [%]	39.0	45.3	55.5	<0.0005
Weighted ERI Proxy ratio at T1 (n)	(563)	(1417)	(808)	<0.0005
Low [%]	45.5	36.3	36.6
Middle [%]	31.4	29.3	28.7
High [%]	23.1	34.4	34.7
Ill mental health at T2 (n)	441	840	402
[%]	0.7	1.4	0.7	<0.0005
Gender at T1 (n)	(579)	(1493)	(864)	0.12
Male [%]	86.4	84.1	82.3
Education at T1 (n)	(579)	(1488)	(862)	<0.0005
Low (n)	(61)	(676)	(585)
[%]	10.5	45.4	67.9
Middle (n)	(398)	(591)	(174)
[%]	68.7	39.7	20.2
High (n)	(120)	(221)	(103)
[%]	20.7	14.9	11.9

^*Differences among the age groups, calculated by χ² test.

The percentages of employees with high scores on ERI-proxy were similar in the middle- and older-aged employees, but were much lower in the younger employees. The percentage of mental ill-health was lower in older employees. When calculating the rank correlation Kendall’s tau Beta between low zest at T2 and mental ill-health (GHQ12) at T2, significant relationships were observed for age groups 19–34 (t = 0.20; p < 0.005) and 35–49 (t = 0.15; p < 0005), and a tendency at age 50–69 (t = 0.09; p = 0.085), (data not shown).

H1 was partly supported (see Table 2) since associations were found between high total WOC and low zest for work in the cross-sectional analyses for the middle (OR_MI = 3.22, 95% CI 2.09–5.49) and older (OR_MI = 3.05, 95% CI 1.36–6.87) age groups but not the younger age groups, even after adjustment for sex, education, and ERI- proxy. Among the older and younger age groups, the differences in WOC subscales indicated a tendency for a high need for approval with low zest for work in comparison to the middle age group. While disproportional irritability was associated with low zest for work in the middle aged (OR_MI = 2.28, 95% CI 1.27–4.09), an opposite tendency was found for competitiveness. Inability to withdraw from work was associated with low zest for work among middle (OR_MI = 1.93, 95% CI 1.11–3.36) and older age groups (OR_MI = 3.32, 95% CI 1.15–9.59).

Table 2

Multiple logistic regression cross-sectional analyses of the association between high WOC and low zest for work, stratified by age group (n = 2940)

	Low zest for work
Item	Age group [years]
	19–34		35–49		50–67
	(young)		(middle)		(older)
	OR (95% -CI)		OR (95% -CI)		OR (95% -CI)
	CC	MI	CC	MI	CC	MI
	n = 481	n = 580	n = 1221	N = 1495	n = 687	n = 864
WOC^‡	0.98	1.19	3.75^***	3.22^***	2.46^*	3.05^**
	(0.36;2.67)	(0.49;2.90)	(2.19;6.40)	(2.09;5.49)	(1.02;5.97)	(1.36;6.87)
Need for approval^‡	2.07	2.14^T	1.37	1.32	2.26^T	2.28^T
	(0.81;5.33)	(0.90;5.06)	(0.80;2.33)	(0.80;2.17)	(0.90;5.66)	(0.99;5.24)
Competitiveness^‡	1.11	1.19	0.58^T	0.63^T	0.75	0.68
	(0.45;2.74)	(0.51;2.77)	(0.33;1.02)	(0.37;1.05)	(0.31;1.86)	(0.31;1.51)
Disproportional	0.49	0.71	2.43^**	2.28^**	0.54	0.64
Irritability^‡	(0.18;1.36)	(0.29;1.78)	(1.27;4.63)	(1.27;4.09)	(0.21;1.40)	(0.27;1.50)
Inability to withdraw from work^‡	0.88	0.73	2.70^**	1.93^*	2.45	3.32^*
	(0.33;2.35)	(0.30;1.79)	(1.44;5.06)	(1.11;3.36)	(0.82;7.31)	(1.15;9.59)

OR = Odds ratio. CI = 95% confidence interval. CC = Complete Case analysis. MI = Multiple Imputation. Multiple logistic regression analyses adjusted for sex, education, and ERI proxy tertiles weighted. ^‡Low values are the reference category. ^TTendency for significance (p < 0.1); ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

H2 was not supported (see Table 3): high total WOC was not associated with low zest for work in longitudinal analyses, after adjustment for sex, education, effort reward imbalance, and mental ill health.

H3 was, however, partly supported (see Table 3) in multivariable adjusted analyses, since age-related associations were found: ‘inability to withdraw from work’ was associated with onset of low zest for work 5 years later among older employees (OR = 3.78, 95% CI 1.20–11.93). This association between ‘inability to withdraw from work’ and low zest for work 5 years later among older employees remained when excluding participants with low zest for work at baseline (OR: 5.14 [95%-CI 1.32; 20.03]). Furthermore, there was an association between ‘need for approval’ and onset of low zest for work among older subjects (OR: 3.29 [95%-CI 1.04; 10.37]) (Table 4).

Table 3

WOC and its subscales at baseline (T1) as predictors for low zest for work at follow-up (T2), stratified by age group

	Low zest for work at T2
	Age group [years]
Predictor	19–34		35–49		50–67
	(young)		(middle)		(older)
	OR (95% -CI)
WOC at T1	CC^a	MI^b	CC^a	MI^b	CC^a	MI^b
	(n = 481)	(n = 580)	(n = 1221)	(n = 1495)	(n = 687)	(n = 864)
High WOC index	0.58	0.61	1.24	1.37	0.86	0.87
	(0.7;1.99)	(0.18;2.05)	(0.70;2.19)	(0.78;2.38)	(0.34;2.28)	(0.35;2.25)
Need for approval	0.85	0.89	1.68^T	1.55	2.42^T	2.01
	(0.29;2.44)	(0.31;2.55)	(0.93;3.04)	(0.87;2.76)	(0.89;6.61)	(0.76;5.32)
Competitiveness	1.03	1.03	1.43	1.41	0.57	0.56
	(0.40;2.68)	(0.40;2.65)	(0.79;2.59)	(0.79;2.53)	(0.22;1.50)	(0.23;1.54)
Disproportional irritability	1.22	1.25	1.02	1.09	0.40	0.54
	(0.44;3.41)	(0.45;3.46)	(0.53;1.96)	(0.57;2.08)	(0.14;1.16)	(0.20;1.46)
Inability to withdraw from work	0.93	0.90	0.69	0.76	3.78^*	3.19^*
	(0.34;2.57)	(0.33;2.48)	(0.37;1.29)	(0.41;1.41)	(1.20;11.93)	(1.03;9.83)

OR = Odds ratio. CI = 95% confidence interval. Logistic regression (OR, 95% -CI) adjusted for sex, education, zest for work, and weighted ERI-Ratio proxy at T1, mental ill health by GHQ 12 at T2. CC = complete case analysis. MI = multiple imputation. WOC = work overcommitment. ^aComplete case analysis. ^bMultiple imputation. ^TTendency for significance (p < 0.1); ^*p < 0.05.

Table 4

Final model for WOC and its subscales at baseline (T1) as predictors for onset of low zest for work at follow-up (T2), stratified by age group and excluded low zest for work at baseline

	Low zest for work at T2
	Age group [years]
Predictor	19–34	35–49	50–67
	(young)	(middle)	(older)
	OR (95% -CI)
WOC at T1	CC	CC	CC
	(n = 481)	(n = 1221)	(n = 687)
High WOC index	1.58 (0.42;5.92)	0.91 (0.47;1.73)	0.84 (0.31;2.28)
Need for approval	0.80 (0.25;2.62)	1.63 (0.84;3.19)	3.29^* (1.04;10.37)
Competitiveness	1.07 (0.38;2.99)	0.98 (0.49;1.94)	0.57 (0.19;1.76)
Disproportional irritability	1.05 (0.35;3.19)	1.09 (0.53;2.23)	0.40 (0.14;1.16)
Inability to withdraw from work	1.15 (0.39;3.41)	0.71 (0.35;1.44)	5.14^* (1.32;20.03)

OR = Odds ratio. CI = 95% confidence interval. Logistic regression (OR, 95% -CI) adjusted for sex, education and weighted ERI-Ratio proxy at T1, mental ill health by GHQ 12 at T2. CC = complete case analysis. WOC = work overcommitment. Significance ^*p < 0.05.

4 Discussion

4.1 Cross-sectional findings

The purpose of this study was to investigate the association between work over-commitment and zest for work. In a cross-sectional sample of occupationally active employees (mainly industrial workers) in Sweden, high work over-commitment was associated with low zest for work in middle and older aged employees. This finding was in agreement with the findings of Rauschenbach and coworkers [19], we observed more significant association of WOC subscales with zest in cross-sectional analyses, thus providing more evidence for temporary effects.In our study, this cross-sectional association was mostly pronounced in the middle-aged group and in the WOC-subscale ‘disproportional irritability’ and in the middle- and older-aged groups ‘inability to withdraw from work’. Conceptually, work over-commitment is similar to type-A behavior [20], which includes aggressive and angry traits. This relationship can therefore be expected. In an Australian community sample, work over-commitment strongly moderated the influence of ERI on expressions of anger in road traffic [21] and driving anger increased among those with high work over-commitment [22].

Inability to withdraw from work was associated with low zest for work in the middle and older age groups in the present study. This finding reflects the unsuccessful ability of the older employees prone to work over-commitment to cope, potentially reflecting misjudgment of their ability to manage demands combined with an underestimation of their ability to cope. This may result from lack of acceptance of the situation or not being given the opportunity by the employer to scale down work. In one study, it was found that older age is associated with acceptance the circumstances and that there is a link between age acceptance and lower negative effect [23]. This age-related association also went in the opposite direction: employees in the middle-aged group with high scores on the WOC scale ‘competitiveness’ had a tendency to lower the risk for low zest for work, which indicated being competitive could have a protective effect in this age group. The competitiveness subscale in WOC was inspired by the theory of type-A personality, and some type-A behavior components are considered to be protective or represent healthy competitive style. This is often seen in leadership which is associated with willingness to compete and can predict low job strain and high job control [24]. However, according to Matschinger and coworkers [12], WOC-coping career might be more successful early in the occupational career among middle-aged employees.

The overall prevalence of high work over-commitment among young employees in Sweden was lower than in young Japanese students of a similar age. Japanese students have an equally high prevalence of work over-commitment, about 40%, as older Swedish employees. However, the higher prevalence of WOC among students in Japan than in young Swedish employees may be due to the difference in the character of work involved, and these groups are therefore not comparable.

None of the WOC subscales were associated with low zest for work in the younger age group. High need for approval and low zest for work were not related among the youngest group. This finding is in agreement with the theory of Matschinger and coworkers [12] since the coping style of young employees prone to WOC was found to be successful and not associated with low zest for work. On the contrary, older participants (>50 years), with high values ‘for need for approval’ suffered from low zest for work when they did not receive supervisor attention in terms of appreciation of their work and limited setting, meaning they had a greater need to be recognized by their superiors despite their older age. To the best of our knowledge there are no studies on age-related associations of need for approval and zest for work. There are concerns about the health of older employees. In Japan, job dissatisfaction is strongly associated with mental health problems among employees older than 55 years [25]. The strong association among psychosocial work conditions, especially high work over-commitment, among middle and older age groups, rather than a younger age group, deserves further attention [26].

4.2 Longitudinal findings

In the longitudinal analysis, only the WOC subscale ‘inability to withdraw from work’ was predictive of low zest for work five years later among workers older than 50 years. This association persisted after adjusting for mental ill health. This observation was in agreement with the coping career theory by Matschinger and coworkers [12], where younger employees develop a coping behavior with a strong sense of control and underestimation of demands and whose good health, professional training, and achievement motivation lead to increased responsibility and job involvement. This coping behavior does not appear to cause a distressing experience in this age group, and may explain why the younger group did not report having low zest for work. Later in the coping career when the distressing experiences increase, signs of fatigue and feelings of failure become prominent, as does awareness of coping failure, which is possibly expressed through loss of zest for work. Overcommitted older workers had poorer well-being than their non-overcommitted colleagues. A prospective Belgian study showed negative effects of work over-commitment on psychological health, and a predictive value of high WOC for depression one year later [27]. Stress and negative emotions towards work due to dissatisfaction have negative effects on well-being [28]. Improvement in older employees’ well-being at work is a preventive measure that could help them from leaving the job early [29, 30]. Job satisfaction is found to be negatively associated with intention to leave the work place [6]. Therefore, the personality and the needs of older employees need to be understood in order to have a sustainable healthy older workforce.

4.3 Strengths and limitations

One strength of the study is the use of the original comprehensive version with 29 items describing four indexes of personality-based work over-commitment. In a review of the ERI model [20], the original version of WOC predicts adverse health outcomes, even when controlled for effort-reward imbalance; however, only one study by Bakker and coworkers [30] used the original WOC questionnaire. The reliability of the long version of the WOC questionnaire has been tested with reasonable Cronbach’s alpha values [8], which are found in this study for the two indices, ‘need for approval’ and ‘inability to withdraw from work’. The strengths of this study were a large sample size and a high response rate in a sample intended to include all employees in the area under study. This study contributes to the research on work over-commitment and zest for work. This is an area where only few studies exist but which seems to be very relevant as the labor force grows older and extending working life is a frequent subject of public debate. With the long version of WOC, it was possible to investigate the role of personality traits in work overcommitted persons in relation to zest for work. The longitudinal study design enabled investigation of the predictive value of WOC dimensions on zest for work, an area where there is a lack of information.

Beside these merits the study has some limitations: One drawback of the study was that the data are several years old and the working conditions may have changed since 1996–1998 and 2000–2003, respectively. Since not much has been published on zest for work in the last decades, yet we think that the study is a valuable extension of existing knowledge. Another drawback is relatively small number of subjects with low zest for work, possibly due to a selection of healthy workers. This could have introduced selection bias. In certain sub-analyses for the WOC scales, there might have been insufficient power to recognize relevant effects on zest for work; there were only tendencies towards significance for several associations. The overall prevalence of low zest for work was about 5%, with the highest prevalence in the middle-aged and the lowest prevalence among the older employees. One possible explanation of these differences might be that younger and middle-aged workers must share their time between family life and work, whereas, older workers have more time for work. These findings could also be due to the effect of studying healthy workers.

Another limitation was the low internal reliability, especially in the competitiveness index. Since a low Cronbach’s alpha indicates that the questions included in the index measure different concepts, the results involving the competitive index in particular should be interpreted with caution. Cronbach’s alpha was also low for the disproportionate irritability index. It should be remembered that this study took place in Northern Sweden where the culture makes disproportional emotional expressions uncommon. Information can be lost through missing data, which can be a considerable source of bias when ignored [32]. List-wise deletion or complete case analysis is the default setting for treating missing data in many software packages and means that a subject is excluded from data analysis when a value is missing for any variable of this observation unit. This may be justified only if values are missed completely at random. Otherwise, subjects with missing values are different in at least one variable from complete cases and are not representative of the whole sample. Potential bias can be introduced in complete case analysis (CC; see e.g. Sterne et al. [33]). Multiple imputations are a state-of-the-art technique for treating missing data and procedures that use multiple imputations produce estimates of missing values with higher accuracy than single imputation procedures, because of good estimation of standard errors [34, 35]. The limitations presented through missing data and the potential bias this introduced were highlighted in the comparison of the CC and MI (FCS = fully conditional specification method) analyses. Furthermore, the overall generalization of the results was limited by the lack of female employees in the data set, a common sample composition in mainly industrial workers, and the use of proxies for ERI instead of the original questionnaire.

As the WOC partly derives from the Type-A behavior style, ‘high need for control’ indicates that people with high scores tend to overestimate their capabilities of achieving the goals they set for themselves, which leads to work dissatisfaction [36]. Therefore, coping through work over-commitment could be a strategy used by certain personality types. Personality traits with a biological basis were assumed to be related to vulnerability to illness. It was found by Wirtz and coworkers [37] that depressive symptoms related to over-commitment at work are being moderated through serotonin 5-HTT gene. They concluded that this lends credence to the role played by genetic factors in behavioral outcomes such as the personality trait over-commitment at work. Such personality types are often regarded as relatively stable characteristics as they are seen to be significantly determined by genetic factors. We have also found some evidence that WOC seems to be stable under changes of working conditions [38]. Under the assumption of over-commitment being a personality trait ongoing investigations of genetic and environmental factors influencing WOC are needed including coping with it.

Regardless of the contributing factors to work over-commitment, the outcome is the same: employees may have a negative feeling towards work on the way to work. In our study one overall question was used to assess negative emotions towards work while travelling to work, ‘low zest for work’: “How do you usually feel when you are on your way to work?”. A similar question was used in the Copenhagen Burnout Inventory (CBI) [39] “Are you exhausted in the morning at the thought of another day at work? Odagiri et al. [40] found associations between burnout measured by CBI and high work over-commitment.

The challenge lies in understanding the psychosocial imbalance within the organization that contributes to work over-commitment, and to intervene in time. For example it is likely that components of WOC like ‘inability to withdraw from work’ are affected by the content as well as the conditions of the work, which has to be investigated in future studies. One recent study found that personality traits are predictors of enterprising and social vocational interests [5]. Job satisfaction suggests a positive evaluation of work and it is an important predictor of turnover intention and absenteeism [41]. People who are not satisfied with their work have a higher intention to leave work. The recognition of early signs of over-commitment in employees, such as need for approval and an inability to withdraw from work, is important. Behavior-oriented and condition-oriented prevention in the work place can therefore have a positive effect, but this requires further investigation. The results from our study could be helpful in the prevention of injury among the aging workforce. Low zest for work may signal increased risk of ill health and absence due to sickness. Positive attitude and rewards for older workers could increase their enthusiasm for work and stop unhealthy work over-commitment. Employers can also implement programs to assist and encourage older workers to pull back from work at a certain time, for example to introduce 8-4 work schedule or flexible working hours, as a standard procedure to minimize the inability to withdraw from work.

In conclusion, there is an association between high work over-commitment and low zest for work in middle- and older-aged employees. However, only perceived inability to withdraw from work predicted low zest for work five years later in older employees. These are important findings in times of ageing societies. Low zest for work can be affected by poor psychosocial working conditions, which can result in work-related stress. This is problematic in employees who are prone to work over-commitment, especially among older workers. Awareness about societal, managemental, and individual levels in the meaning of psychosocial aspects of work for health is important, as low zest for work may signal increased risk of ill health and absence due to sickness. Future research focusing on rewards and limited setting for older workers should determine whether organizational management could be used to reduce work over-commitment as a coping style and its determinants as risks factors for low zest for work. The future research could include understanding the needs of the millennial generation in the workplace to help understand how to make work places sustainable for them as they continue to work into older years.

Conflict of interest

The authors declare that they have no conflict of interest.

Footnotes

Acknowledgments

The study received a small grant from Forskiningsrådet för hälsa, arbetsliv och välfärd (FORTE), Sweden, to initiate this project.

References

Super

. A theory of vocational development. Am Psychol 1953;8:186–95.

Spector

. Job satisfaction, Application, Assessment, Causes, and Consequences. Thousand Oaks, California: SAGE publications. 1997.

Peterson

, Park

, Hall

, Seligman

MEP

. Zest and work. J Organiz Behav 2009;30:161–72.

Josephson

, Vingård

. Zest for work? Assessment of enthusiasm and satisfaction with the present work situation and health – a 1.5-year follow-up study. Work 2007;29:225–31.

Berings

, De Fruyt

, Bouwen

. Work values and personality traits as predictor of enterprising and social vocational interests. Pers Indivi Differ 2004;6:349–64.

Antón

. The impact of role stress on workers’ behaviour through job satisfaction and organizational commitment. Int J Psychol 2009;44:187–94.

Weiss

. Deconstructing job satisfaction. Separating evaluations, beliefs and affective experiences. Hum Resour Manage R 2002;12:173–94.

Siegrist

. Adverse health effects of high-effort/low-reward conditions. J Occup Health Psych 1996;1:27–41.

Siegrist

, Starke

, Chandola

, Godin

, Marmot

, Niedhammer

, et al. The measurement of effort-reward imbalance at work: European comparisons. Soc Sci Med 2004;58:1483–99.

10.

De Jonge

, Bosma

, Peter

, Siegrist

. Job strain, effort-reward imbalance and employee well-being: A large-scale cross-sectional study. Soc Sci Med 2000;50:1317–27.

11.

, Yang

, Cheng

, Siegrist

, Cho

S-Il

. Effort–reward imbalance at work and job dissatisfaction in Chinesehealthcare workers: A validation study. Int Arch Occ Env Health 2005;78:198–204.

12.

Matschinger

, Siegrist

, Dittmann

. Type A as a coping career—toward a conceptual and methodological redefinition. In Schmidt

T.H.

, Dembroski

T.M.

, & Blumchen

(Eds.). Biological and psychological factors in cardiovascular disease. Berlin: Springer; 1986:104–26.

13.

Rubenowitz

. Företagshälsovården måste samverka aktivt med produktionen. Arbete, Människa och Miljö 1989;4:237–3.

14.

Goldberg

, Williams

. A user’s quide to the GHQ. Windsor: NFER-Nelson; 1988.

15.

Fahlén

, Knutsson

, Peter

, Åkerstedt

, Nordin

, Alfredsson

, et al. Effort-reward imbalance, sleep disturbance and fatigue. Int Arch Occ Env Health 2006;79:371–8.

16.

SPSS (Statistical Package of the Social Sciences), Chicago, Il; 60606; 2003.

17.

Enders

. Applied Missing Data Analysis (Methodology in the Social Sciences). 1st ed. New York, London: Guilford Press; 2010.

18.

Raghunathan

, Lepkowski

, Van Hoewyk

, Solenberger

. A multivariate technique for multiply imputing missing values using a sequence of regression models. Surv Methodol 2001;27:85–95.

19.

Rauschenbach

, Hertel

. Age differences in strain and emotional reactivity to stressors in professional careers. Stress Health 2011;27:48–60.

20.

Van Vegchel

, de Jonge

, Bosma

, Schaufeli

. Reviewing the effort-reward imbalance model: Drawing up the balance of 45 empirical studies. Soc Sci Med 2005;60:1117–31.

21.

Hoggan

, Dollard

. Effort–reward imbalance at work and driving anger in an Australian community sample: Is there a link between work stress and road rage? Accident Anal Prev 2007;39:1286–95.

22.

Lau

. Effort-reward imbalance and overcommitment in employees in a Norwegian municipality: A cross sectional study. J Occup Med Toxicol 2008;3:1–11.

23.

Siegrist

, Wege

, Puhlhofer

, Wahrendorf

. A short generic measure of work stress in the era of globalization: Effort reward imbalance. Int Arch Occ Env Health 2009;82:1005–13.

24.

Hintsa

, Jokela

, Pulkki-Råback

, Keltikangas-Järvinen

. Age- and cohort-related variance of type-A behavior over 24 years: The young finns study. Int J Behav Med 2014;21:927–35.

25.

Tatsuse

, Sekine

. Job dissatisfaction as a contributor to stress-related mental health problems among Japanese civil servants. Ind Health 2013;51:307–18.

26.

Kuoppala

, Lamminpää

, Husman

. Work health promotion, job well-being, and sickness absences—A systematic review and meta-analysis. J Occup Environ Med 2008;50:1216–27.

27.

Godin

, Kittel

, Coppieters

, Siegrist

. A prospective study of cumulative job stress in relation to mental health. BMC Public Health 2005;5:67.

28.

Shimazu

, Schaufeli

. Is workaholism good or bad for employee well-being? The distinctiveness of workaholism and work engagement among Japanese employees. Ind Health 2009;47:495–502.

29.

Ilmarinen

, Rantanen

. Promotion of work ability during ageing. Am J Ind Med Supplement 1999;36:21–3.

30.

Sinisammal

, Belt

, Härkönen

, Möttönen

, Väyrynen

. Managing well-being at work during s—expert viewpoints. Open J Saf Sci Tech 2012;2:25–31.

31.

Bakker

, Killmer

, Siegrist

, Schaufeli

. Effort-reward imbalance and burnout among nurses. J Adv Nurs 2000;31:884–91.

32.

Schafer

, Graham

. Missing data: Our view of the state of the art. Psychol Methods 2002;7:147–77.

33.

Sterne

, White

, Carlin

, Spratt

, Royston

, Kenward

, et al. Multiple imputation for missing data in epidemiological and clinical research: Potential and pitfalls. BMJ 2009;338:157–60.

34.

Tanner

, Wong

. The calculation of posterior distributions by data augmentation. J Am Stat Assoc 1987;82:528–40.

35.

. Missing data analysis using multiple imputation: Getting to the heart of the matter. Circ Cardiovasc Qual Outcomes 2010;3:98–105.

36.

Ward

, Eisler

. Type A behavior, achievement strivings, and a dysfunctional self-evaluation system. J Pers Soc Psychol 1987;53:318–26.

37.

Wirtz

, Siegrist

, Schuhmacher

, Hoefels

, Maier

, Domes

, Schwab

, Zobel

. The association between overcommitment at work and depressive symptoms is moderated by the polymorphic region of the 5-HTT gene. Psychiatry Research 2013;2:199–200.

38.

du Prel

J-B

, Runeson-Broberg

, Westerholm

, Nordin

, Knutsson

, Alfredsson

, Fahlen

, Peter

. Work overcommitment – Is it a trait or a state? Submitted 01/2016.

39.

Kristensen

, Borritz

, Villadesn

, Christensen

. The Copenhagen Burnout Inventory: A new tool for the assessment of burnout. Work & Stress 2005;19(3):192–207.

40.

Odagiri

, Shimomitsu

, Ohya

, Kristensen

. Overcommitment and high effort are strongly associated with burnout among Japanese nurses. International Journal of Behavioural Medicine 2004;11(Suppl):2014.

41.

Dobreva-Martinova

, Villeneuve

, Strickland

, Matheson

. Occupational role stress in the Canadian forces: Its association with individual and organizational well-being. Can J Behav Sci 2002;34:111–21.