Applied Ergonomics 48 (2015) 86e94

Contents lists available at ScienceDirect

Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo

Workefamily conflict and enrichment from the perspective of psychosocial resources: Comparing Finnish healthcare workers by working schedules Saija Mauno a, b, *, Mervi Ruokolainen b, Ulla Kinnunen b a b

€skyla €, Department of Psychology, Finland University of Jyva University of Tampere, School of Social Sciences and Humanities (Psychology), Finland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 June 2014 Accepted 17 November 2014 Available online

We examined workefamily conflict (WFC) and workefamily enrichment (WFE) by comparing Finnish nurses, working dayshifts (non-shiftworkers, n ¼ 874) and non-dayshifts. The non-dayshift employees worked either two different dayshifts (2-shiftworkers, n ¼ 490) or three different shifts including nightshifts (3-shiftworkers, n ¼ 270). Specifically, we investigated whether different resources, i.e. job control, managers' workefamily support, co-workers' workefamily support, control at home, personal coping strategies, and schedule satisfaction, predicted differently WFC and WFE in these three groups. Results showed that lower managers' workefamily support predicted higher WFC only among 3shiftworkers, whereas lower co-workers' support associated with increased WFC only in nonshiftworkers. In addition, shiftworkers reported higher WFC than non-shiftworkers. However, the level of WFE did not vary by schedule types. Moreover, the predictors of WFE varied only very little across schedule types. Shiftwork organizations should pay more attention to familyefriendly management in order to reduce WFC among shiftworkers. © 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.

Keywords: Workefamily conflict Workefamily enrichment Shiftwork

1. Introduction There is already evidence that shiftwork/nonstandard working hours, referring to working outside normal ‘office’ daytime hours 8.5, is associated with challenges in workefamily interface. One much used indicator of workefamily interface is workefamily conflict, which has so far received some attention in shiftwork research (Davis et al., 2008; Haines et al., 2008; Pisarski et al., 2008; Staines and Pleck, 1984). However, even less is known about what resources help shiftworkers in workefamily interface, for instance, to reduce their workefamily conflict. Our study focuses on this question by approaching workefamily interface via workefamily conflict and enrichment. Workefamily conflict (WFC) refers to a form of inter-role conflict in which role pressures from work and family are incompatible in some respect (Greenhaus and Beutell, 1985), whereas workefamily

€, Department of Psychology, Post * Corresponding author. University of Jyv€ askyla Box 35, 40014 Finland. Tel.: þ358 503186770. E-mail addresses: saija.mauno@jyu.fi (S. Mauno), mervi.ruokolainen@uta.fi (M. Ruokolainen), ulla.kinnunen@uta.fi (U. Kinnunen). http://dx.doi.org/10.1016/j.apergo.2014.11.009 0003-6870/© 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.

enrichment (WFE) describes positive workefamily interface by referring to the extent to which the quality of one's work (family) role enhances the quality of one's family (work) roles (Greenhaus and Powell, 2006). Both WFC and WFE are bidirectional experiences (from work to family and family to work) but here we concentrate on the work-to-family direction because schedule arrangements very likely affect more the experiences aroused at work (see Michel et al., 2011, for a meta-analysis). Earlier studies on the effects of shiftwork on workefamily interface have focused on negative phenomena, e.g. WFC, marital dissatisfaction, and positive experiences have been neglected. However, there is some evidence that the effects of shiftwork are not totally negative for employees (Merkus et al., 2012; Saksvik et al., 2011) or for their families (Mauno et al., 2014). Consequently, it is crucial also to investigate positive experiences as outcomes. WFE is one important indicator of life quality and role engagement (Kinnunen et al., 2014; McNall et al., 2010) and the phenomenon also deserves attention in shiftwork research. Here we examine WFC and WFE by comparing Finnish nurses (N ¼ 1634), working both dayshifts (n ¼ 874, non-shiftworkers) and non-dayshifts. The non-dayshift workers/shiftworkers worked either two different dayshifts (n ¼ 490, 2-shiftworkers) or three

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

different shifts including nightshifts (n ¼ 270, 3-shiftworkers). First, we compare whether WFC and WFE vary in these three groups. Second, we investigate whether different psychosocial (i.e. job control, managers' support, co-workers' support, control at home, personal coping strategies) resources are connected differently to WFC and WFE in these three groups. One additional resource factor we examine is schedule satisfaction. Specifically, we investigate whether these resources are equally beneficial in different schedules by comparing their relative contribution to WFC and WFE. We propose that employees with different work schedules may benefit differently from these resources in workefamily interface. If so, this needs consideration in work design and interventions, particularly in organizations with variable working hours. 1.1. Shiftwork as an antecedent of WFC and WFE The effects of shiftwork on workefamily interface began to attract the attention of workefamily interface scholars in the 1980s. A pioneer study by Staines and Pleck (1984) demonstrated that shiftwork (working weekends or variable days) was related to high WFC. Much later, Davis et al. (2008) showed that nightshift predicted higher levels of WFC. However, their study showed no association between working weekends, without nightshifts, and WFC or other negative family-related consequences. Barnett et al. (2008) published a study on working couples where the wife worked either dayshifts or evening shifts. They found that the wife's evening shift predicted her own elevated WFC, but not her husband's WFC. Haines et al. (2008) published a study on the effects of various shift types on WFC and depression, and found that shift schedules were related to high WFC, which, in turn, was linked to elevated depression. Pisarski et al. (2008) also indicated that WFC mediated the effects between poor organizational resources (support, control, team climate) and poor psychological well-being and physical symptoms in shiftworkers. Overall, these findings suggest that shiftwork predicts high WFC. Unfortunately, we found no studies on the effects of shiftwork on WFE or related constructs, e.g. workefamily facilitation. However, shiftwork presumably relates to low WFE because it is a proxy concept for WFC but describes positive workefamily interface instead of negative. Accordingly, our first hypothesis proposes: H1. Shiftworkers report more WFC and less WFE than do nonshiftworkers.

87

support or control at home, most likely affect conflict and enrichment in the family-to-work direction. However, occasionally crossdomain effects have also been observed, for example, spousal support has predicted low work-to-family conflict (Michel et al., 2011). Consequently, we also examine whether home-related resources, here control at home, are associated with WFC and WFE in the work-to-family direction (a cross-domain effect). Overall, numerous studies have confirmed that emotional support, especially that received from managers/supervisors (Lapierre and Allen, 2006; Mauno and Rantanen, 2013; Pisarski et al., 2008; Thomas and Ganster, 1995) but also from co-workers/teams (Mauno and Rantanen, 2013; Pisarski et al., 2008) and job control (Byron, 2005; Mauno and Rantanen, 2013; Mauno et al., 2006) are associated with low WFC but also with high WFE (Mauno and Rantanen, 2013; Voydanoff, 2004), particularly in the work-tofamily direction. Indeed, support and control are crucial resources in well-established job stress models (e.g. DemandeControleSupport model; Karasek and Theorell, 1990), and therefore also important in our study. On the other hand, control at home is also a valuable additional resource for workefamily interface (Mauno and Rantanen, 2013; Peeters et al., 2009), and is therefore included in our study. Even though these resources have gained some attention in shiftwork research (Pisarski et al., 2008; Pisarski and Barbour, 2014) earlier studies have not compared whether different forms of support and control are equally beneficial in different schedule arrangements in predicting WFC and WFE. We propose that support and control are even more important when schedule arrangements become more challenging, as in shiftwork. In other words shiftworkers may benefit more from these resources than do non-shiftworkers. Organizing one's family/private life can be more demanding in shiftwork arrangements, implying greater needs for coping resources (Davis et al., 2008; Mauno et al., 2014; Staines and Pleck, 1984; Winwood et al., 2006). Consequently, our second hypothesis proposes: H2a. The relationship between low levels of managers' and coworkers' support, job control and control at home and high WFC is stronger among shiftworkers than among non-shiftworkers, i.e. lacking support and control is more detrimental to shiftworkers' WFC. H2b. The relationship between high levels of managers' and coworkers' support, job control and control at home and high WFE is stronger among shiftworkers than among non-shiftworkers, i.e. support and control are more beneficial for shiftworkers' WFE.

1.2. Contextual resources at work and home as antecedents of WFC and WFE 1.3. Personal coping strategies as antecedents of WFC and WFE Research indicates that different contextual resources predict low WFC and high WFE (Byron, 2005; Kinnunen et al., 2014; Kossek et al., 2011; Michel et al., 2011). These contextual resources have most typically referred to various forms of support (emotional or instrumental) and control (sometimes called autonomy or flexibility) received either at work or home (Mauno and Rantanen, 2013; Peeters et al., 2009; Pisarski et al., 2008), which we also focused on. Support and control have generally been found to be crucial for successful workefamily interface (Byron, 2005; Kinnunen et al., 2014; Kossek et al., 2011; Michel et al., 2011) also among nurses (Mauno and Rantanen, 2013; Peeters et al., 2009; Pisarski et al., 2008), which is the target group in this study. Specifically, contextual resources can be expected to be most effective predictors for experiences arising in the same context. Thus, workrelated resources (e.g. managers' support) should show the most robust association with conflict and enrichment in the work-tofamily direction. Similarly, family-related resources, e.g. spousal

Studies on workefamily interface have recently paid attention to personal coping strategies as potential resources that might help an employee to reconcile work and family (Andreassi, 2011; Lapierre and Allen, 2006; Rantanen et al., 2011; Rotondo et al., 2003). A major distinction between personal coping strategies and contextual resources presented earlier (control, support) is that in the former case an employee him/herself is an active agent applying coping strategies, whereas contextual resources (support, control) are more inherently available at work or home. Research on coping has moved from exploring general coping strategies, e.g. problem-focused or avoidant coping (Andreassi, 2011; Lapierre and Allen, 2006; Rantanen et al., 2011; Rotondo et al., 2003), towards more detailed coping strategies, specifically encompassing an employee's cognitive and behavioral coping efforts applied when work and family demands need to be combined (Mauno et al., 2012; Neal and Hammer, 2007; Somech and Drach-Zahavy, 2007, 2012).

88

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

Sometimes these new coping strategies have been called ‘workefamily coping strategies’ (Mauno et al., 2012; Neal and Hammer, 2007; Somech and Drach-Zahavy, 2007, 2012), and we also use this term. Of various workefamily coping strategies, we focus on delegating at home and work (a person delegates some tasks to someone else at work/home) and prioritizing at work and home (a person prioritizes some tasks at work/home) (Mauno et al., 2012; Somech and Drach-Zahavy, 2007, 2012). These two coping strategies were selected because they describe concrete behavioral or cognitive efforts to cope with workefamily challenges. Earlier studies have also shown that they predict WFC (Mauno et al., 2012; Somech and Drach-Zahavy, 2007) and WFE (Mauno et al., 2012; Taskila-Rissanen et al., 2013), forming the present outcomes. However, we are not aware of studies comparing their relative contribution by schedule type. Nevertheless, it may well be, as with support and control discussed above, that these coping strategies are more crucial in more difficult schedule arrangements. Thus, our third hypothesis proposes: H3a. The relationship between high levels of delegating and prioritizing coping and low WFC is stronger among shiftworkers than among non-shiftworkers, i.e. less use of these coping strategies is more detrimental to shiftworkers' WFC. H3b. The relationship between high levels of delegating and prioritizing coping and high WFE is stronger among shiftworkers than among non-shiftworkers, i.e. these coping strategies are more beneficial for shiftworkers' WFE.

1.4. Schedule satisfaction as an antecedent of WFC and WFE Transactional stress theory (Lazarus and Folkman, 1984) suggests that the main thing in the stress and coping process is an individual's perception or subjective evaluation of a particular situation. Applying this idea to our study suggests that it is important to consider how an employee experiences his/her schedule arrangements. Thus, it may not matter if an employee works shifts if (s)he is totally satisfied with this arrangement, feeling positive about it. Research shows that satisfaction with scheduling is crucial to various positive outcomes (McNamara et al., 2011; Nabe-Nielsen et al., 2011a,b; Pisarski and Barbour, 2014), that might include WFC and WFE. Nevertheless, schedule satisfaction may not be so crucial among non-shiftworkers because integrating work and family life is probably easier when working only in daytime. Therefore, our fourth hypothesis proposes: H4a. The relationship between low schedule satisfaction and high WFC is stronger among shiftworkers than among nonshiftworkers. H4b. The relationship between high schedule satisfaction and high WFE is stronger among shiftworkers than among nonshiftworkers.

2. Method 2.1. Procedure and sample Data were collected in 2009 through an online survey among Finnish healthcare professionals (mainly nurses) using trade union membership registers. The response rate was 48% and the respondents represented well the target population in terms of sex, age, and education (Mauno and Rantanen, 2013; Ruokolainen et al., 2014).

The present sample consists of 1634 Finnish nurses working in three different schedules: (1) daytime workers (standard day hours from Monday to Friday; called non-shiftworkers), n ¼ 874; (2) two rotating dayshift workers (including morning and afternoon/evening shift, seven days a week, called 2-shiftworkers), n ¼ 490; and (3) three rotating shiftworkers (like the second group but their work included nightshift, called 3-shiftworkers), n ¼ 270. The sample was female-dominated (93%) and the respondents' mean age was 43 (SD ¼ 10) years. Vocational school/college education (94%) and permanent employment status (83%) predominated among the respondents. Typically they lived with a partner (85%) and had at least one child living at home (77%). Those having children at home had on average two children (M ¼ 2.3, SD ¼ 1.0). It is noteworthy that our sample was collected at national level (i.e. not at organizational level) and for this reason we did not have detailed information on how shifts were organized. Basically, Finnish shiftwork organizations can decide shift scheduling so long as they comply with the national labor laws and the European Union directives on working hours and scheduling. Because working in three shifts can be assumed to be the most stressful type of the shiftwork arrangements we studied, we provide some general information on working in three shifts in Finnish nursing. Most typically nursing in three shifts includes morning, evening and nightshifts in a predefined six-week period. A rest period between the shifts is at minimum 7 h and one free day should be allowed each week. The recommended rotation is morningeeveningenightshift, but other rotations (e.g. eveningemorning, morningenight) may also occur. Concerning nightshift, it is recommended that an employee works two or three consecutive nights (including a rest period between the nights) and nightshift should not exceed 8 h. Working in two shifts in nursing follows basically the same guidelines with the exception that nightshift does not typically exist. Furthermore, when working in two or three shifts these shifts should not vary day by day; instead it is recommended to work 2e4 days (8 h per shift) on the same shift. 2.2. Measures 2.2.1. WFC and WFE WFC was measured with four items assessing both strain- (2 items; e.g. “After work, I come home too tired to do some of things I'd like to do”) and time-based conflict (2 items; e.g. “The time I must devote to my job keeps me from participating equally in household responsibilities and activities”). The scale was based on the measure developed by Carlson et al. (2000). WFE was measured with four items from the scale developed by Carlson et al. (2006). Two items evaluated affective (e.g. “My involvement in my work puts me in a good mood and this helps me be a better family member”) and two developmental enrichment (e.g. “My involvement in my work helps me to acquire skills and this helps me to be a better family member”). The items in both scales were evaluated on a response scale from 1 (totally disagree) to 7 (totally agree). The Cronbach's alpha for WFC was 0.78 and 0.83 for WFE. 2.2.2. Contextual resources Managers' support was measured with four items (e.g. “In general, managers in this organization are quite accommodating of family-related needs”) from the scale developed by Thompson et al. (1999). The items used specifically elicit workefamily support, not overall support. Co-workers' support was measured with four items (e.g. “My colleagues support employees who temporarily want to reduce their working hours for private reasons”), again specifically assessing workefamily support. The scale was developed by Dikkers et al. (2007). The items in both scales were assessed on a 7point rating scale, ranging from 1 (totally disagree) to 7 (totally

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

89

agree). The Cronbach's alpha for managers' support was 0.91.and for co-workers' support 0.78. Job control was assessed via a 4-item scale developed in Finland (QPS Nordic Questionnaire, Elo et al., 2006). The items (e.g. “Can you decide your pace at work?”) were evaluated on a 5-point rating scale, ranging from 1 (hardly ever) to 5 (always). The Cronbach's alpha for job control was 0.76. We also measured perceived control at home as a contextual home-related resource on a scale with three items. The scale was similar to the job control scale but the word ‘work/job’ was replaced by the word ‘home’ (e.g. “At home, I can decide my own schedules”). Items were evaluated on a 5-point Likert-scale from 1 (totally disagree) to 5 (totally agree). The Cronbach's alpha for control at home was 0.78.

studied, this comparison was conducted only for psychosocial resources (Steps IIeIV) and only if the b-coefficient was significant in one/two groups but non-significant in other groups. Overlapping CIs indicate non-significant differences between the regression coefficients across the groups. In reporting the results of the regression analyses, we interpret standardized b-coefficients derived from the final step (IV), that is, after all predictors had been entered into the model.

2.2.3. Personal resources Workefamily coping strategies were evaluated with two subscales; prioritizing at work and home (4 items; e.g. “I eliminate the family/work duties that are the least important”) and delegating at work and home (6 items, e.g. “I manage my family/work duties by delegating some to others”). The scales were originally developed by Somech and Drach-Zahavy (2007) and used in Finnish studies (Mauno et al., 2012; Taskila-Rissanen et al., 2013). The Cronbach's alpha for prioritizing was 0.58 and for delegating 0.73. Schedule satisfaction was measured using one item “How satisfied are you with your working schedules?” using a rating scale from 1 (very dissatisfied) to 7 (very satisfied). Background factors, i.e. gender, age, education, parental status and contract type, were used as control variables in the statistical analyses (for categorization of the variables, see footnotes in Tables 1e5) as they have been shown to be linked to the workefamily interface (Byron, 2005).

Correlations between the variables by schedule type are shown in Table 1 (non-shiftworkers) and Table 2 (2- and 3-shiftworkers). ANCOVA results (M, SD, F values, and paired differences) for the three schedule groups are presented in Table 3. The groups differed in four variables out of nine. First, the groups differed in WFC: Nonshiftworkers reported lower WFC than shiftworkers, and the highest WFC was reported by the 2-shiftworkers. Second, the highest managerial support was reported by non-shiftworkers, and lowest by the 2-shiftworkers, whereas the 3-shiftworkers were between these two groups, but the group did not differ statistically from the 2-shiftworkers in paired comparisons. Third, the nonshiftworkers also perceived the highest job control, whereas lowest control was reported by the 2-shiftworkers, and all groups differed significantly from each other. Finally, the groups also differed in schedule satisfaction: the non-shiftworkers reported the highest schedule satisfaction, whereas the lowest satisfaction was reported by the 2-shiftworkers. However, the 2- and 3-shiftworkers did not differ from each other in schedule satisfaction in paired comparisons.

2.3. Statistical analyses Differences in WFC and WFE by schedule type were examined using ANCOVA with gender, age, education, parental status, and contract type as covariates, and WFC and WFE as dependent variables. We likewise examined differences in resources and schedule satisfaction by schedule type to ascertain whether the groups differed in these factors (as descriptive results). Next we examined whether resources predicted WFC and WFE differently vs. similarly in the three schedule types. We performed linear regression analysis (method enter) separately for the three schedule groups as follows: In the first step we controlled for the background factors (same factors as in ANCOVA analyses). In the second step we entered personal (prioritizing and delegating coping) and home-related (control at home) resources into the model, followed by job resources in the third step (managers' and co-workers' support, job control). Finally, in the fourth step, schedule satisfaction was entered into the equations. First we entered those predictors which we assumed to be weaker antecedents of WFC and WFE (i.e. background factors and personal and home-related resources) as they can be considered more distal predictors for WFC and WFE in the work-to-family direction, which we focused on. Entering weaker antecedents first into the regression equations enabled greater explanatory power for them. Overall, job resources can be assumed to be the strongest antecedents for WFC and WFE in the work-to-family direction (Byron, 2005; Kinnunen et al., 2014; Greenhaus and Powell, 2006). Schedule satisfaction, in turn, has been a really strong predictor in shiftwork research (Pisarski et al., 2008) and was therefore entered into the regression equation in the last step. We also computed confidence intervals (CI) for standardized regression coefficients [b ± (b/t)] in order to statistically compare the coefficients (only the coefficients relating to hypotheses 1e4) across the groups, which varied greatly in size. Of the predictors

3. Results 3.1. Correlations and mean comparisons

3.2. Predicting WFC by schedule type group The results of the regression analyses computed separately for the three groups are presented in Table 4 (WFC) and Table 5 (WFE). First we report the findings which were similar across the groups and then the group differences. Prioritizing coping (Step II) was significantly related to high WFC and schedule satisfaction (Step IV) to low WFC equally across the groups (CIs were overlapping). Furthermore, high control at home (Step II) and job control (Step III) were related to low WFC across the groups (CIs were overlapping). To summarize, in each group, perceived control at work and home and schedule satisfaction were beneficial resources predicting low WFC, whereas prioritizing coping was harmful as it predicted high WFC in each group. However, there were also some significant differences between the groups in predicting WFC. First, high managers' support (Step III) predicted low WFC only in 3-shiftworkers, as indicated by a nonsignificant b-coefficient for other groups (CIs were not overlapping between the groups). Second, high co-workers' support (Step III) was associated with low WFC only among nonshiftworkers but not in the groups of shiftworkers (CIs were not overlapping between the groups). To summarize the main differences between the groups: managers' support was a beneficial resource only for 3-shiftworkers, whereas co-workers' support mattered only for non-shiftworkers in terms of low WFC. 3.3. Predicting WFE by schedule type group As for WFC, certain resources showed equally beneficial effects across the groups in predicting WFE. Namely, in each group, high job control (Step III) predicted high WFE. Even though the b-coefficients were not significant for the groups of shiftworkers

90

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

Table 1 Pearson correlations of the study variables in non-shiftworkers (n ¼ 783e874). Variables

1

2

3

4

5

6

7

8

9

10

1. Sexa 2. Ageb 3. Educationc 4. Contract typed 5. Childrene 6. Delegating coping 7. Prioritizing coping 8. Control at home 9. Managers' support 10. Co-workers' support 11. Job control 12. Schedule satisfaction 13. WFC 14. WFE

e 0.09** 0.01 0.01 0.03 0.05 0.06 0.08* 0.01 0.02

11

12

e 0.06 0.32*** 0.39*** 0.08* 0.12*** 0.25*** 0.02 0.07*

e 0.07* 0.03 0.14*** 0.02 0.05 0.07* 0.08*

e 0.23*** 0.05 0.07* 0.01 0.02 0.04

e 0.09** 0.12*** 0.11** 0.01 0.00

e 0.21*** 0.01 0.23*** 0.27***

e 0.03 0.01 0.01

e 0.13*** 0.08*

e 0.61***

e

0.09* 0.03

0.03 0.00

0.25*** 0.03

0.08* 0.06

0.00 0.02

0.18*** 0.11**

0.06 0.06

0.11** 0.14***

0.41*** 0.42***

0.32*** 0.39***

e 0.40***

e

0.01 0.06

0.04 0.04

0.03 0.07*

0.03 0.07

0.01 0.03

0.04 0.06

0.24*** 0.03

0.17*** 0.06

0.17*** 0.22***

0.22*** 0.23***

0.19*** 0.23***

0.32*** 0.28***

13

e 0.18***

Note. *p  0.05, **p  0.01, ***p  0.001. a Women/men. b Low/high; a continuous variable. c Low/high; no vocational education/vocational school/vocational college/university degree. d Permanent/temporary. e No children/children.

(b ¼ 0.09 in both groups), CIs indicate that the groups did not differ as CIs were overlapping. Thus job control predicted high WFE in all groups. Schedule satisfaction (Step IV) was also positively related to WFE in each group (CIs were overlapping between the groups). In conclusion, job control and schedule satisfaction were equally beneficial for all employees irrespective of schedule type in terms of high WFE. We found two significant differences between the groups in predicting WFE. First, delegating coping (Step II) was associated with high WFE only among workers working in two shifts (CI for this group was not overlapping with the CIs of two other groups). Second, co-worker support (Step III) predicted high WFE more strongly among workers working in two shifts than in the other two groups (CIs were not overlapping between non-shiftworkers and workers working in two shifts). Interestingly, co-worker support was a significant predictor also in non-shiftworkers (b-coefficient was significant) and also in 3-shiftworkers (CI for this group overlapped with CI of 2-shift rotating shiftworkers). To summarize, the main group differences in predicting WFE: even though co-

worker support was a beneficial resource for all employees, it was the most robust predictor of WFE among 2-shiftworkers (compared to non-shiftworkers) and delegating coping was beneficial for WFE only among 2-shiftworkers, and not for others. Total explanation rates (R2) for WFC varied between 18% and 37%, being highest for 3-shiftworkers due to the particularly strong explanatory power of job resources (DR2 18%, especially managers' support) in this group. Total explanation (R2) for WFE varied between 12% and 19%, being lowest for 3- and for 2-shiftworkers, and in this latter group the highest explanatory rate was found for job resources (DR2 12%, especially co-workers' support).

4. Discussion and conclusions This study examined psychosocial resources as antecedents of WFC and WFE in Finnish nurses working in three different schedule arrangements. Our study contributes to shiftwork research from the viewpoint of workefamily interface by comparing the relative

Table 2 Pearson correlations of the study variables in 2-shiftworkers (n ¼ 394e490) and 3-shiftworkers (n ¼ 232e270). Variables

1

2

3

4

5

6

7

8

9

10

11

12

13

14

1. Sexa 2. Ageb 3. Educationc 4. Contract typed 5. Childrene 6. Delegating coping 7. Prioritizing coping 8. Control at home 9. Managers' support 10. Co-workers' sup. 11. Job control 12. Schedule satisf. 13. WFC 14. WFE

e 0.05 0.08 0.07 0.11 0.03 0.01 0.02 0.03 0.06 0.02 0.05 0.02 0.08

0.01 e 0.05 0.22*** 0.50*** 0.18** 0.09 0.13* 0.10 0.07 0.01 0.02 0.02 0.07

0.04 0.17*** e 0.02 0.09 0.02 0.10 0.03 0.06 0.03 0.03 0.04 0.13* 0.03

0.02 0.32*** 0.02 e 0.16** 0.06 0.03 0.00 0.02 0.10 0.07 0.01 0.01 0.05

0.03 0.52*** 0.11* 0.20*** e 0.11 0.20*** 0.16* 0.10 0.06 0.01 0.07 0.12 0.01

0.03 0.01 0.04 0.04 0.15*** e 0.16*** 0.02 0.14* 0.20*** 0.29*** 0.14* 0.04 0.13*

0.03 0.00 0.06 0.02 0.09 0.29*** e 0.00 0.03 0.04 0.04 0.07 0.26*** 0.07

0.14** 0.10* 0.07 0.03 0.20*** 0.03 0.11* e 0.23*** 0.08 0.12 0.14* 0.24*** 0.08

0.02 0.05 0.02 0.01 0.06 0.07* 0.06 0.09 e 0.56*** 0.36*** 0.46*** 0.44*** 0.22***

0.06 0.00 0.04 0.01 0.08 0.13** 0.09 0.05 0.51*** e 0.26*** 0.28*** 0.28*** 0.21***

0.19*** 0.02 0.04 0.04 0.06 0.02 0.05 0.10* 0.33*** 23*** e 0.35*** 0.31*** 0.17**

0.04 0.08 0.05 0.03 0.06 0.02 0.02 0.11* 0.48*** 0.36*** 0.41*** e 0.44*** 0.22***

0.08 0.06 0.00 0.01 0.03 0.00 0.14** 0.14** 0.27*** 0.20*** 0.36*** 0.40*** e 0.19**

0.06 0.03 0.01 0.00 0.05 0.16*** 0.07 0.10* 0.30*** 0.33*** 0.22*** 0.28*** 0.22*** e

Note. Correlations above the diagonal are for 2-shiftworkers and those below the diagonal are for 3-shiftworkers. *p  0.05, **p  0.01, ***p  0.001. a Women/men. b Low/high; a continuous variable. c Low/high; no vocational education/vocational school/vocational college/university degree. d Permanent/temporary. e No children/children.

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

91

Table 3 Differences between schedule type groups in key variables (variables with significant effects are bolded). Group 1 ¼ non-shiftworkers, Group 2 ¼ 2-shiftworkers, Group 3 ¼ 3shiftworkers. Variables

Group 1 M, SD

Group 2 M, SD

Group 3 M, SD

Df, F-value, p-value

WFC WFE Delegating coping Prioritizing coping Control at home Managers' support Co-workers' support Job control Schedule satisfaction

3.71, 4.06, 4.18, 3.16, 3.78, 4.84, 5.20, 2.79, 5.50,

4.52, 4.16, 4.30, 3.18, 3.69, 4.46, 5.15, 2.11, 4.74,

4.20, 4.08, 4.18, 3.19, 3.79, 4.51, 5.14, 2.24, 5.00,

1412 1411 1537 1538 1352 1429 1429 1455 1420

1.35 1.26 0.85 0.87 0.82 1.27 1.05 0.90 1.31

1.22 1.26 0.75 0.89 0.79 1.32 1.03 0.67 1.60

1.37 1.25 0.87 0.82 0.74 1.37 1.13 0.73 1.50

(2), (2), (2), (2), (2), (2), (2), (2), (2),

51.08, p < 0.001 0.80, p < 0.05 (ns) 2.22, p > 0.05 (ns) 1.21, p > 0.05 (ns) 0.25, p > 0.05 (ns) 11.18, p < 0.001 0.66, p > 0.05 (ns) 108.17, p < 0.001 35.18, p < 0.001

h2

Paired comparisons (p < 0.05)

0.07 0.00 0.00 0.00 0.00 0.02 0.00 0.13 0.05

1< ns ns ns ns 1> ns 1> 1>

2, 3, 2 > 3

2, 3 2, 3, 3 > 2 2, 3

Note. ANCOVA with sex (women/men), age (low/high; a continuous variable), education (low/high; no vocational education/vocational school/vocational college/university degree), contract type (permanent/temporary), children (no children/children) as covariates. Paired comparisons performed with Bonferroni test.

However, the second hypothesis, i.e. shiftworkers benefit more from managers' and co-workers' support, job control, and control at home in terms of low WFC (H2a) and high WFE (H2b) than nonshiftworkers, was only partially supported. Some of these antecedents functioned as similar predictors across the groups and major differences were found for job-related resources. Unexpectedly, the predictors moreover differed not only by schedule type but also by outcome. Altogether, the resources studied were more powerful predictors for WFC (18e37%) than for WFE (12e19%) in all groups. Specifically, we found that high levels of job control and control at home predicted low WFC across the groups and, contrary to our second hypothesis they were not more beneficial for shiftworkers. Thus feelings of control at work and home may have an equally beneficial effect on all employees (Byron, 2005; Mauno et al., 2006), not only for shiftworkers. However, schedule type moderated two of the relationships in line with our hypothesis: low managers'

importance of psychosocial resources across schedule types in predicting WFC and WFE. Consistent with our first hypothesis (H1), i.e., shiftworkers report more WFC and less WFE than non-shiftworkers, and earlier studies (Barnett et al., 2008; Davis et al., 2008; Pisarski et al., 2008; Staines and Pleck, 1984), WFC was more prevalent among shiftworkers than non-shiftworkers. However, WFC was most common among 2- and not among 3-shiftworkers, who also worked nights and the reasons for this would need further clarification. We found no differences in WFE across schedule types and thus the first hypothesis was only partially supported. It may be that WFE depends more on other factors than schedule arrangements, in contrast to WFC where they explained a lot of the variance (13%). Thus WFC and WFE have different antecedents (Greenhaus and Powell, 2006; Kinnunen et al., 2014) and both experiences need to be studied to gain a comprehensive picture of workefamily interface and its antecedents in different schedule arrangements.

Table 4 Predicting WFC by schedule type groups. Group 1 ¼ non-shiftworkers, Group 2 ¼ 2-shiftworkers, Group 3 ¼ 3-shiftworkers. Predictors

Group 1

Standardized b and CI

b

Group 2

I. Backgrounds Sexa Ageb Educationc Contract typed Childrene DR2

0.04 0.07 0.06 0.03 0.06 0.00

II. Personal resources Delegating copings Prioritizing coping Control at home DR2

0.02 0.22*** 0.14** 0.10***

0.06 to 0.02 0.18 to 0.26 0.18 to 0.10

0.01 0.16** 0.07 0.10***

0.05 to 0.03 0.11 to 0.21 0.12 to 0.02

0.02 0.22*** 0.12 0.10***

0.05 to 0.09 0.16 to 0.28 0.18 to 0.06

III. Job resources Managers' support Co-workers' support Job control DR2

0.06 0.11* 0.10* 0.05***

0.02 to 0.10 0.15 to 0.07 0.14 to 0.06

0.09 0.04 0.17*** 0.05***

0.15 to 0.03 0.09 to 0.01 0.22 to 0.12

0.25*** 0.02 0.12 0.18***

0.33 to 0.17 0.09 to 0.05 0.18 to 0.06

IV. Schedule satisfaction DR2 R2

0.21*** 0.03*** 0.18***

0.25 to 0.17

0.30*** 0.03*** 0.18***

0.34 to 0.22

0.25*** 0.05*** 0.37***

0.32 to 0.18

95% CI

b

Group 3 95% CI

0.08 0.04 0.02 0.01 0.02 0.00

b

95% CI

0.00 0.03 0.11 0.03 0.08 0.04**

Note. b ¼ standardized beta-coefficients derived from the IV step after all predictors have been entered to the model. CI ¼ 95% confidence interval for standardized beta-coefficients computed for steps II, III, IV. DR2 change in explanation rate in each step. R2 explanation rate for the total model including four steps. *p < 0.05, **p < 0.01, ***p < 0.001. a Women/men. b Low/high; a continuous variable. c Low/high; no vocational education/vocational school/vocational college/university degree. d Permanent/temporary. e No children/children.

92

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

Table 5 Predicting WFE by the schedule type groups. Group 1 ¼ non-shiftworkers, Group 2 ¼ 2-shiftworkers, Group 3 ¼ 3-shiftworkers. Predictors

Group 1

Standardized b and CI

b

I. Backgrounds Sexa Ageb Educationc Contract typed Childrene DR2

0.08* 0.09* 0.03 0.03 0.08* 0.02*

II. Personal resources Delegating copings Prioritizing coping Control at home DR2

0.01 0.04 0.05 0.02*

Group 2 95% CI

b

Group 3 95% CI

0.03 0.07 0.02 0.04 0.04 0.01

b

95% CI

0.07 0.10 0.01 0.04 0.08 0.01

0.04 to 0.02 0.00 to 0.08 0.02 to 0.08

0.14** 0.02 0.07 0.05***

0.09 to 0.19 0.04 to 0.08 0.02 to 0.12

0.01 0.02 0.06 0.02

0.12 to 0.08 0.10 to 0.06 0.02 to 0.14

III. Job resources Managers' support Co-workers' support Job control DR2

0.03 0.09** 0.13*** 0.07***

0.02 to 0.08 0.04 to 0.14 0.09 to 0.17

0.08 0.21*** 0.09 0.12***

0.02 to 0.14 0.15 to 0.27 0.03 to 0.15

0.06 0.14 0.09 0.07**

0.04 to 0.16 0.05 to 0.23 0.01 to 0.17

IV. Schedule satisfaction DR2 R2

0.19*** 0.03*** 0.14***

0.15 to 0.23

0.10 0.01 0.19***

0.04 to 0.16

0.14 0.02 0.12***

0.07 to 0.23

Note. b ¼ standardized beta-coefficients derived from the IV step after all predictors have been entered to the model. CI ¼ 95% confidence interval for standardized beta-coefficients computed for steps II, III, IV. DR2 change in explanation rate in each step. R2 explanation rate for the total model including four steps. *p < 0.05, **p < 0.01, ***p < 0.001. a Women/men. b Low/high; a continuous variable. c Low/high; no vocational education/vocational school/vocational college/university degree. d Permanent/temporary. e No children/children.

support predicted high WFC only in 3-shiftworkers, that is, among those nurses working nightshifts. Interestingly, mean comparisons revealed that overall shiftworkers reported less managers' support. It is possible that managers' and supervisors' roles are more crucial when three shifts are worked, which can be seen as the most challenging type of shiftwork. This finding may also relate to the fact that managers plan schedules and one unwelcome consequence may be that schedules including nightshifts are not optimal for an employee's workefamily integration, thus resulting in WFC. Some earlier studies have shown that nightshift is particularly challenging for an employee's family life (for a review see, Mauno et al., 2014), and for this reason managers' role, their formal and informal support, might have been a more valuable resource for those in the present study working in three shifts. On the other hand, low co-workers' support predicted high WFC only among non-shiftworkers, but not among shiftworkers, thereby confounding our second hypothesis. Why co-workers' support appeared more important in non-shiftwork merits further research because in this study the groups did not even differ in the level of co-worker support (mean comparisons). One explanation could be that, at least in nursing, work tasks differ across schedules, possibly affecting employees' needs and expectations regarding coworkers' support. Because our study does not offer irrefutable explanations why managers' and co-workers' support benefited employees working in different shifts in different ways, future studies should perform in depth analyses of the role of co-workers' and managers' support for WFC and WFE in different schedule arrangements, for example, by gathering interview data from shiftworkers who work in different shifts. Concerning WFE, very little support was found for the second hypothesis (H2b; shiftworkers benefit more from contextual resources than do non-shiftworkers). In contrast to the second hypothesis, we found that high job control and co-workers' support predicted high WFE in each group. Hence with WFE as the outcome,

these resources were not more beneficial for shiftworkers. However, the positive effect of co-workers' support was strongest among the 2-shiftworkers, which is difficult to explain in light of our data. However, it needs to be remembered that in all groups coworkers' support predicted high WFE. Earlier research reports that different forms of support and control are crucial resources associated with high WFE (Greenhaus and Powell, 2006; Kinnunen et al., 2014), and viewed in this light, our result corroborates those of earlier studies. The third hypothesis, i.e. the shiftworkers benefit more from personal coping strategies, i.e. prioritizing and delegating coping, than non-shiftworkers, was not supported regarding WFC (H3a). We found that prioritizing coping was associated with high rather than low WFC, and its harmful effect was present in each group. It is possible that reverse causality explains this result: WFC may result in prioritizing coping at work and home and not vice versa (Mauno et al., 2012; Somech and Drach-Zahavy, 2007) but we could not confirm this because the study design was cross-sectional. Overall, it makes sense that when a person has more roles and tasks it is important to prioritize. However, as a consequence of prioritizing a person neglects some responsibilities, which, in turn, may lead to a bad conscience, masking the potentially beneficial effects of prioritizing. As our study is unable to reveal the processes or mechanisms between prioritizing coping and WFC, future studies should examine this relationship more closely by ascertaining under what conditions prioritizing coping could be useful for employees. Moreover, delegating coping did not predict WFC in any group, thereby confounding the third hypothesis. Nevertheless, concerning WFE, we found that high delegating coping predicted high WFE, but only among 2-shiftworkers. Thus schedule type moderated this association, as we had expected in H3b, but not in both shiftworker groups. We have no explanation why delegating coping mattered only in this particular group. Overall, in interpreting these results, it

S. Mauno et al. / Applied Ergonomics 48 (2015) 86e94

is noteworthy that this is the first study to examine WFE by schedule type, focusing on its predictors. Consequently, more research evidence is needed to permit firmer conclusions in this respect. The fourth hypothesis, i.e., shiftworkers benefit more from schedule satisfaction than do non-shiftworkers in relation to WHC (H4a) and WFE (H4b), was not supported as schedule satisfaction predicted low WFC and high WFE in each group. However, it should be noted that mean comparisons indicated that non-shiftworkers reported higher schedule satisfaction than shiftworkers. Despite this mean difference, our findings suggest that the level of schedule satisfaction may not affect the relationship between schedule types, WFC, and WFE. Our results have implications for organizations; some of them are more shiftwork-specific and some of them may benefit all workers equally irrespective of their schedule types. Concerning shiftwork-specific implications mean comparisons revealed that shiftworkers reported not only higher WFC but also lower job control, managers' support and schedule satisfaction than did nonshiftworkers. More importantly, moderator analyses further revealed that lacking managers' support was related to high WFC only among those working in three shifts. These findings propose that shiftwork organizations should pay more attention to familyefriendly management practices planned for and targeted specifically at those working in three shifts. Good initiatives in this respect would include, for instance, familyefriendly training for managers and supervisors. A related issue is that in the present study shiftwork explained much variance in WFC (13%), calling attention to interventions that help to decrease WFC among shiftworkers, for example, providing different forms of organizational support and job control (e.g., schedule control) for shiftworkers. It is also essential to note that we found that some resources were beneficial for all workers irrespective of their schedule type. Most consistently, the findings indicated that schedule satisfaction and job control were associated with low WFC and high WFE in all workers. Thus, job control, including its different facets, needs to be improved in all organizations, not only in shiftwork organizations. In the healthcare sector, our focus here, managers and supervisors often have great authority over how work is organized and scheduled and hence their role is also decisive in developing job control and creating familyefriendly practices in workplaces. In addition, high control at home related to low WFC and high coworker support to high WFE in all workers, suggesting that these are additional valuable resources for all nurses regardless of their schedule type. This study also has a few shortcomings. First, the design was cross-sectional and therefore no reliable interpretations on causal relationships can be made. As already noted, WFC may result in higher prioritizing coping in contrast what we had hypothesized. Similarly, the level of WFC and WFE may determine the level of job control, support, and schedule satisfaction and not vice versa. However, theoretical models of WFC (Frone et al., 1992; Michel et al., 2011; Voydanoff, 2004) and WFE (Greenhaus and Powell, 2006) suggest that different psychosocial resources are likely to predict WFC and WFE and not vice versa, and we formulated our hypotheses on this basis. Second, the sample was female-dominated and derived from one field only, implying that our results should be replicated in more heterogeneous shiftwork samples. Third, indirect paths were not tested, e.g. schedule satisfaction may also act as a mediator between resources and WFC/WFE. However, this question was beyond our scope here. Fourth, some of the scales, e.g. prioritizing coping, had low reliability (