Changes in Leadership Behaviors Predict Changes in Job Satisfaction and Well-Being in Low-Skilled Workers

Abstract

The purpose of this longitudinal study of 255 members of the low-skilled workforce was to enhance insight into the nature of the relations between specific supervisor behavior (social support, positive feedback, task-related communication) and employee well-being. Data were analyzed using latent change models focusing on interindividual change and change–change associations over time. Our results indicated that interindividual differences in the intraindividual change in perceived supervisor behavior were related to changes in indicators of well-being across a 6-month period. These results provide further evidence for longitudinal associations between leader behavior and employee outcomes as well as the necessity of designing specific interventions for low-level managers.

Keywords

health-promoting leadership low-skilled workforce well-being job satisfaction latent change

About 16% of the German workforce is low skilled (Lyly-Yrjänäinen, 2008), including formally qualified employees who work in positions that do not require formal training. These educated employees in the low-skilled workforce are often immigrants with meager German language skills or workers who lack an occupational qualification that is accepted in Germany. Therefore, low-skilled workforces are characterized by a high proportion of immigrant workers (Hoppe, 2011; Seebaß & Siegert, 2011). On the other hand, workers who have completed their vocational educations are also often employed in low-skilled jobs due to poor employment opportunities in their regular professions. Low-skilled jobs have comparatively unfavorable general working conditions such as high physical demands, job insecurity, rotating shift work, and time pressure (e.g., Borrell, Muntaner, Benach, & Artazcoz, 2004; Lipscomb, Loomis, McDonald, Argue, & Wing, 2006; Niedhammer, Sultan-Taïeb, Chastang, Vermeylen, & Parent-Thirion, 2012; Rydstedt, Devereux, & Sverke, 2007; Schreuder, Roelen, Koopmans, & Groothoff, 2008). These potentially well-being–impairing working conditions of nonprofessional occupations appear in combination with a lack of complexity and skill variety and a comparative lack of resources such as autonomy or social support (Borrell et al., 2004; Eurofond, 2007; Morgeson & Humphrey, 2006; Rydstedt et al., 2007). Furthermore, immigrant workers, who are overrepresented in low-skilled jobs relative to professional occupations, tend to receive little social support and appreciation from supervisors and colleagues in the workplace (Dalgard, Thapa, Hauff, McCubbin, & Syed, 2006; Hoppe, 2011; Wadsworth et al., 2007). This working environment contributes to general inequalities in health outcomes between occupational groups (Christensen, Labriola, Lund, & Kivimäki, 2008; Schrijvers, van de Mheen, Stronks, & Mackenbach, 1998), with the result that low-skilled workers are a high-risk group not only for cardiovascular diseases (Steptoe & Marmot, 2002) but also for impaired general well-being (Batinic, Selenko, Stiglbauer, & Paul, 2010). Despite their comparatively more adverse working conditions and their vulnerability to impaired health, the low-skilled workforce has been given little scientific and organizational attention (Busch, Staar, Aborg, Roscher, & Ducki, 2010).

One of the few resources that low-skilled workers may encounter is positive supervisor behavior consisting of support and appreciation (Bakker & Demerouti, 2007). Until now, health-promoting leadership research has not focused on definite workgroups and their specific requirements, and this research has particularly neglected unskilled or semiskilled workers. In the following, we first describe concrete leadership behavior, which may work as a resource for this target group. We collected longitudinal data on these perceived leader behaviors by first-level supervisors and workers’ job satisfaction and job-related well-being to test our hypotheses that change in leadership behavior comes along with change in employees’ well-being. This longitudinal relationship is a strong inference of causality.

Health-Promoting Leadership Behavior

There is an increasing interest in determining the association between leadership behavior and employee well-being (for reviews, see Gregersen, Kuhnert, Zimber, & Nienhaus, 2011; Kuoppala, Lamminpää, Liira, & Vainio, 2008; Skakon, Nielsen, Borg, & Guzman, 2010), whereas many health-promoting leadership researchers are focused on confirming the health-related effects of exiting leadership theories such as the leader–member exchange approach (e.g., Gerstner & Day, 1997; Volmer, Niessen, Spurk, Linz, & Abele, 2011) and transformational leadership (e.g., Arnold, Turner, Barling, Kelloway, & McKee, 2007; Kelloway, Turner, & Barling, 2012; Tafvelin, Armelius, & Westerberg, 2011). Nevertheless, research has not yet determined which concrete supervisor behavior is recommendable and should be included in leadership intervention programs with the purpose of increasing employee well-being. Shamir, House, and Arthur (1993) point out in their theoretical article that there is a need for research on the explicit effects of the processes and preconditions of particularly charismatic or transformational leadership on employees. As a result, the understanding of leadership has been expanded in recent empirical investigations, and researchers have identified work and worker characteristics, such as meaningfulness, opportunities for development, employee trust, and self-efficacy, as mediators of the association between transformational leadership and employee well-being (Arnold et al., 2007; Kelloway, Turner, et al., 2012; Nielsen & Munir, 2009; Nielsen, Randall, Yarker, & Brenner, 2008). Unfortunately, workers in low-skilled jobs hardly ever experience these work characteristics: They tend to experience low task demands, and therefore, they are not given many chances to achieve self-efficacy at work (Felfe & Schyns, 2002; Schyns & von Collani, 2010). They are also given few opportunities for development at work (Lyly-Yrjänäinen, 2008) and few reasons to perceive their work as meaningful (Hackman & Oldham, 1976). These mediator effects as well as less emergence and effectiveness of transformational leadership in the lower echelons of management in general (Bruch & Walter, 2007; Shamir & Howell, 1999) indicate that we have to go beyond transformational leadership and mediator research when it comes to low-skilled workers. In a qualitative study with low-skilled workers and first-level supervisors, Winkler, Busch, and Duresso (2013) found that social support, positive feedback, and task-related communication by supervisors were concrete behaviors that were able to improve the job satisfaction and well-being of employees in low-skilled jobs. This finding was subsequently confirmed in a cross-sectional study on low-skilled workers and their supervisors (Winkler, Busch, Clasen, & Vowinkel, 2014).

Social support provided by the supervisor is the most studied psychosocial resource in the workplace (Cohen & Wills, 1985; de Lange, Taris, Kompier, Houtman, & Bongers, 2003; Gilbreath & Benson, 2004; van Dierendonck, Haynes, Borrill, & Stride, 2004). Social support provided by the supervisor is associated with job satisfaction and well-being in low-skilled employees and immigrants (Hoppe, 2011; Pelfrene et al., 2010). As stated by the substitutes for leadership theory (Kerr & Jermier, 1978), supportive leadership matters, especially when specific individual, job, and organizational characteristics (e.g., professional orientation of the subordinates, intrinsically satisfying tasks) are absent. Consequently, supportive leadership seems to be an adequate behavior for promoting well-being in low-skilled workforces.

Task-related communication from supervisor has been confirmed to relate to job-related well-being of low-skilled workers (Winkler et al., 2013; Winkler et al., 2014). Members of first-level management differ from middle and upper management members in their role, responsibilities, span of control, work tasks, education, and training experience (Kerr, Hill, & Broedling, 1986; Kraut, Pedigo, Mckenna, Dunnette, & Kraut, 1989; Priestland & Hanig, 2005). Accordingly, first-level managers show different behavior and have other ways of influencing their subordinates. Hales (2005) explored the role of frontline managers in the United Kingdom and found that the core tasks or responsibilities of these low-level managers consisted of holding brief meetings for staff (especially in semiskilled and unskilled workforces), acting as a communication channel up and down, and giving praise for good work. Thus, communication and information are central aspects of leadership behavior, particularly in multicultural work environments (Amason, Allen, & Holmes, 1999; Chang & Tharenou, 2004; Sadri & Tran, 2002). Task-related communication from the supervisor is positively related to subordinates’ job satisfaction and negatively related to burnout (Kim & Lee, 2009; Miles, Patrick, & King, 1996).

The importance of task-related communication as a resource for subordinates has also been confirmed in populations other than low-skilled workers: After controlling for communication, Fix and Sias (2006) found no additional incremental prediction of job satisfaction by leader–member exchange. Various dimensions of communication—in particular, job-relevant communication from a superior—are positively associated with subordinate job satisfaction and negatively related to subordinate burnout (Kim & Lee, 2009; Miles et al., 1996).

Finally, positive feedback by supervisor has been found to relate positively to low-skilled worker’s well-being (Winkler et al., 2013; Winkler et al., 2014). A supervisor who communicates positive feedback to the individual employee enhances self-esteem as important aspect of well-being of the employee (Ashford & Cummings, 1985; Kelloway, Weigand, McKee, & Das, 2012). Low-skilled workers rarely receive signs of appreciation and esteem at work (Rydstedt et al., 2007; Winkler et al., 2013) although it is considered an important social resource (van Vegchel, de Jonge, Bakker, & Schaufeli, 2002). Positive feedback enhances subordinates’ individual job satisfaction irrespective of nationality (Lam, Yik, & Schaubroeck, 2002; Van De Vliert, Shi, Sanders, Wang, & Huang, 2004). Positive feedback is further part of the positive leadership construct, which is a predictor of job satisfaction and well-being at work after controlling for transformational leadership and is especially effective when transformational leadership is perceived to be low (Kelloway, Weigand, et al., 2012).

Longitudinal Leadership Research

Shamir (2011) discusses the role of time in leadership theory and research, arguing that “leadership is not an a-temporal phenomenon” (p. 308). Most studies about the effects of leadership behavior have neglected the temporal characteristic by employing cross-sectional designs while simultaneously recommending longitudinal designs (Hunter, Bedell-Avers, & Mumford, 2007; Yammarino, 2013). Existing results of longitudinal studies suggest a reciprocal and dynamic relation between leadership and subordinates’ well-being instead of a one-way process of influence (Nielsen et al., 2008; van Dierendonck et al., 2004). The main point of conducting a longitudinal study is to analyze change. Therefore, most longitudinal studies investigate the association between values of two variables measured at different points in time or examine lagged relations controlled for previous values of the respective constructs (e.g., van Dierendonck et al., 2004; Volmer et al., 2011). With these autoregressive methods, change is not assessed directly but rather is modeled as unexplained variance in the prediction of the same variable measured in the past. To examine whether change in supervisor behavior triggers change in subordinates’ well-being, the first step that needs to be taken is to use a model that is able to address these changes explicitly.

We applied a repeated-measures design in an attempt to confirm that change in supervisor behavior predicts change in employee’s well-being over time. If individual variation in the perception of a supervisor exists over time, this behavior should be able to influence the well-being of a subordinate. To analyze intraindividual changes, individual differences in stability and change need to be included in order to gain a more complete and person-oriented understanding of the dynamics of perceived leadership behavior and employee outcomes. The aim of this study was to shine new light on the changing nature of supervisor behavior and well-being at work. This procedure allowed us to take a more dynamic perspective and to take a step closer to being able to draw a conclusion about a causal direction.

Additionally, we wanted to determine what interval of time would be adequate for studying changes in perceptions of leadership behavior as well as changes in employee well-being and job satisfaction over time. An examination of existing longitudinal studies in occupational health and leadership research illustrates a lack of theory to support choices of time periods between measurement points as well as mostly reasons of convenience for choosing the length of the time period (DeLange, Taris, Kompier, Houtman, & Bongers, 2004; Shamir, 2011; Zapf, Dormann, & Frese, 1996).

The Present Study

In the present study, we consider both, the specific demands regarding supervisor behavior of low-skilled workforces and the causal relationships between changes in supervisor behavior and changes in the well-being of the employees. The first objective of this study was to investigate the patterns of changes in perceived supervisor behavior and employee well-being during 3-month and 6-month intervals. The study design should be able to identify a meaningful time interval for perceiving significant changes in perceived supervisor behavior and employee well-being.

Research Question: Are there meaningful changes in, and relationships between perceived supervisor behavior and employees’ work-related well-being across 3-month and 6-month time intervals?

The main purpose of this study was then to examine intraindividual changes in concrete daily supervisor behaviors observable by employees and realizable by supervisors of low-skilled workforces in order to predict intraindividual changes in the job satisfaction and well-being of the workers. By examining existing research on health-promoting leadership, we identified social support, positive feedback, as well as task-related communication as relevant concrete leadership behaviors that could potentially affect the low-skilled workforce.

Hypothesis 1a: Change over time in employees’ perceptions of social support given by the supervisor will be positively associated with change in employees’ reported well-being (job satisfaction, job-related affective well-being).

Hypothesis 1b: Change over time in employees’ perceptions of positive feedback given by the supervisor will be positively associated with change in employees’ reported well-being (job satisfaction, job-related affective well-being).

Hypothesis 1c: Change over time in employees’ perceptions of supervisor’s task-related communication will be positively associated with change in employees’ reported well-being (job satisfaction, job-related affective well-being).

Hypothesis 2: Change over time in employees’ perceptions of social support, positive feedback, and task-related communication by the supervisor will predict change in employees’ reported well-being (job satisfaction, job-related affective well-being).

Method

Procedure and Participants

The data were assessed as part of an extensive research project on unskilled and semiskilled workers¹ in three German companies (A: catering, B: service production, C: manufacturing). The criterion for inclusion in the study was employment in a job that had no formal education requirements, regardless of the actual qualifications of the individual employees. All employees who worked in the participating sections of the companies were invited to participate in the study. Therefore, employees participated voluntarily in our longitudinal field study comprising three waves of data collection that occurred at 3-month time intervals. Participants filled out the questionnaire during regular working hours in break rooms and received financial incentives for their repeated participation (50€). Paper-and-pencil questionnaires were matched across time by personal codes. The current analyses were based on N = 255 participants (A: n = 89, B: n = 32, C: n = 134) who did not participate in any intervention program; 44% were female; the mean age was 40.1 years (SD = 11.8, range = 16-62 years). About 35% of the respondents reported having no formal qualification, 61% had completed a vocational education, and 2% even held a university degree. The sample comprised 47.6% immigrants—that is, workers who were born abroad (34.1% first generation) or who had at least one parent born abroad (13.5% second generation) originating mainly from Turkey (13.1%), Poland (5.6%), Kazakhstan (4.0%), and Russia (3.2%). The respondents had worked for the company for an average of 8.8 years (SD = 9.3); 88% of the respondents worked full-time, and 64% had a contract of employment of indefinite duration. Of all participating employees, 64% worked in shifts, and 66% also worked at night.

Measures

At each measurement occasion, the participants were administered the same questionnaire. We carefully selected existing scales and items and linguistically simplified them to ensure the comprehensibility for all participants as well as to target the specific needs of low-skilled workers. To consider the large number of immigrants in our sample, we used a bilingual questionnaire with linguistically simplified items in German and in one additional language (Polish, Russian, or Turkish). Moreover, we offered the participants the opportunity to obtain personal assistance. The response format for all scales was a 4-point Likert-type scale.

Supervisor behavior was assessed by three latent behavior scales as used in Winkler et al. (2014): Social Support, Positive Feedback, and Task-Related Communication.

Social support was measured with a linguistically simplified version of the German Social Support scale by Frese (1989) also used by Hoppe (2011) for low-skilled workers consisting of three items (e.g., “My boss asks me whether I have problems or trouble at work”; “My boss makes it easier for me to deal with my work”; 1 = totally disagree, 4 = totally agree). Cronbach’s alphas were .82, .78, and .83 for T1, T2, and T3, respectively.

Beside social support we assessed positive feedback and task-related communication. The items for positive feedback and task-related communication were based on the results of a qualitative study with low-skilled workers and their supervisors (Winkler et al., 2013) and the dimension job-relevant communication by Miles et al. (1996). Positive Feedback is a three-item scale concerning how often the immediate supervisor provides positive feedback about the employee’s work (e.g., “My boss tells me if I have done something well”; “My boss tells me how I can improve my work”; 1 = totally disagree, 4 = totally agree). Cronbach’s alphas were .78, .79, and .80 for T1, T2, and T3, respectively. The scale Task-Related Communication was composed of three items about how often the supervisor is approachable at the worksite and provides work-relevant information (e.g., “How often does your boss inform you about relevant things?”; “How often do you feel your boss is approachable?”; 1 = very rarely/never, 4 = very often/always). Cronbach’s alphas were .73, .70, and .73 for T1, T2, and T3, respectively.

Well-being was represented using two latent factors: job satisfaction and affective job-related well-being. To measure job satisfaction we linguistically simplified the job satisfaction scale of the Copenhagen Psychosocial Questionnaire (Nübling, Stößel, Hasselhorn, Michaelis, & Hofmann, 2005). We included five items about the satisfaction with the work in general, colleagues, physical demands, supervisor, and work tasks (e.g., “How satisfied are you with your work in general?”; “How satisfied are you with your colleagues?”; 1 = very unsatisfied, 4 = very satisfied). The first item (job satisfaction in general) was supported with scale anchors represented by varying degrees of smiling or frowning faces (Kunin, 1955). For theoretical and empirical reasons (Williams & O’Boyle, 2008), we combined the five items into two parcels: overall job satisfaction and a composite of satisfaction with various job facets. Both indicators have advantages as well as shortcomings. The first (reflective) indicator was the item asking how satisfying the work is in general, complemented with smiley faces. Such a global index of overall job satisfaction has been reported to be a reliable and valid measure of job satisfaction (Dolbier, Webster, Mccalister, Mallon, & Steinhardt, 2005; Wanous, Reichers, & Hudy, 1997). The second indicator was conceptualized as a composite of satisfaction with specific facets of one’s work—a multidimensional construct usually requiring a formative indicator model (Law & Wong, 1999; MacKenzie, Podsakoff, & Jarvis, 2005). Even if distinguishing between these measurement models is highly recommended, they have several shortcomings (i.e., identification, interpretation), and there are contradictory recommendations about how to use them (Bollen & Bauldry, 2011; Edwards, 2011; Howell, Breivik, & Wilcox, 2007). Hence, we decided to avoid formative measurement in favor of using a parsimonious model that would be suitable for the less differentiated conceptualization of the job satisfaction of blue-collar workers (Hu, Kaplan, & Dalal, 2010). Cronbach’s alphas for the second indicator were .61, .60, and .72 for T1, T2, and T3, respectively. The mean value of these items formed the second indicator of job satisfaction.

The Affective Job-Related Well-Being scale was a modified version based on the WHO-Five Well-Being Scale (Brähler, Mühlan, Albani, & Schmidt, 2007) including two items about energetic behavior and two items about positive feelings at work (e.g., “Do you feel vital and active at work?”; “Are you in a good mood at work?”; 1 = not at all, 4 = very). Cronbach’s alphas were .71, .79, and .83 for T1, T2, and T3, respectively.

Statistical Analysis

Structural Equation Modeling

All analyses were implemented using Mplus Version 6.12 and maximum likelihood parameter estimation (Muthén & Muthén, 2010). To assess the overall fit of the models, we examined the following fit indices: the comparative fit index (CFI), the root mean square error of approximation (RMSEA), the standardized root mean square residual (SRMS), and the chi-square test statistic as well as the chi-square to degrees of freedom ratio. We interpreted the indices according to the recommendations by Hu and Bentler (1999).

Missing Data

In field studies, it is quite common to find that the response rate varies between measurement occasions. There were 143 (56%) participants who provided complete data across all three measurement points, whereas 112 (44%) participants did not participate at all three occasions. More specifically, 42 participants were absent at the first measurement point, 51 at the second, and 65 at the third. The demographic characteristics of these two groups were compared. We conducted chi-square tests to test for significant differences between the complete and the partial complete cases based on gender, χ²(1) = 0.00, p = .98; immigration background, χ²(2) = 0.66, p = .72; and education, χ²(3) = 5.53, p = .14; and t tests based on age (t = 0.69, p = .49) and tenure (t = 3.77, p = .00). We found no significant differences (p > .05) between continuous participants and those who completed the questionnaire only one or two times on any variables except for tenure. Employees who participated across all three measurement points had worked for the company for significantly longer (M = 11 years) than participants who did not complete all three measurement points (M = 6 years). To handle the existing missing data, we estimated all models with the full information maximum likelihood approach (Arbuckle, 1996; Little & Rubin, 2003), which is generally the recommended approach for handling missing data (Ployhart & Ward, 2011; Schafer & Graham, 2002) and is “likely to produce similar analysis results” (Enders, 2010, p. 336) as multiple imputation. When applying full information maximum likelihood, data should be either missing completely at random (MCAR) or missing at random. Data are MCAR when the missing data are completely unrelated to any variables or covariates in the study. This assumption was evaluated with Little’s (1988) chi-square MCAR test on the study variables and categorical covariates in SPSS/PASW 18 statistical software. The nonsignificant result, χ²(1,751) = 1800.67, p = .20, indicated that there were no differences between respondents with partial and complete data on the variables included; therefore, we used Graham’s (2003) saturated correlates approach to incorporate tenure as an auxiliary variable.

True Change Analyses

To examine interindividual differences in true intraindividual change in supervisor behavior and employee well-being, we specified a latent true change model (McArdle, 2009; Steyer, Eid, & Schwenkmezger, 1997; Steyer, Partchev, & Shanahan, 2000). These models are the result of the reparameterization of the usual longitudinal model in such a way that the true change scores between two measurement occasions are specified as latent variables as illustrated for one construct (job satisfaction) in Figure 1 (Δjob satisfaction_{T2 − T1}, Δjob satisfaction_{T3 − T1}). This procedure of modeling change on the latent level has the chief advantage of being able to estimate change uncontaminated by measurement error. Additionally, these latent variables may be discussed as other exogenous or endogenous latent variables in structural equation modeling. For example, one can test whether the variance in latent change is significantly larger than zero, indicating reliable interindividual differences in intraindividual change. Given reliable variance in change, one can interpret correlations and regressions of more latent change factors (reciprocal change over time). We additionally modeled indicator-specific method factors to separate indicator-specific variance from error variance as recommended by Eid, Schneider, and Schwenkmezger (1999). The first indicator of each construct acts as a reference indicator without an indicator-specific factor loading. As displayed for job satisfaction in Figure 1, we fixed the loadings to 1 at each time point and did not allow correlations between the indicator-specific factors and the latent state factors of the same construct.

Figure 1.

Measurement model specifying latent change with indicator-specific factors.

There are several theoretical and methodological reasons for analyzing longitudinal data with a true intraindividual change analysis instead of the traditional statistical procedures for longitudinal data that are used in leadership research (i.e., autoregressive cross-lagged models). The overall aim of longitudinal studies is to analyze change over time, which is directly estimated and examined in true intraindividual change models. Autoregressive cross-lagged models are basically fixed-effect models that estimate the coefficients as the same for all studied participants; therefore, these procedures are not applicable for assessing assumed differences in intraindividual changes (Curran & Hussong, 2002; Hertzog & Nesselroade, 2003; Rogosa, 1980). It is not possible to directly specify change or to regress change on change in these popular autoregressive cross-lagged models. Using latent change variables, we have the opportunity to examine individual differences in stability and change over time. Correlations of ratings assessed at different time points conceal these important intraindividual processes.

Results

Measurement Invariance: Longitudinal Validity

To model latent mean differences, it is essential that the measures that are used represent the same underlying construct equally well across measurements. Therefore, longitudinal measurement invariance has to be established to ensure that the latent factors have equivalent measurement properties across time (Meredith & Horn, 2001). Thus, we needed to determine whether strong measurement invariance would hold across time for the latent factors in our sample. This type of invariance requires that factor loadings and intercepts of the manifest indicators are invariant over time. Additionally, we ran a series of longitudinal confirmatory analyses to ensure the construct validity of the five studied variables.

As illustrated in Table 1, we compared one model that included all indicators and their latent variables without restrictions against one model with equal factor loadings of the indicators across time and another model with equal intercepts of the indicators across time. The results of the comparative fit analyses indicated strong longitudinal invariance of our parameters; consequently, all factor loadings and intercepts were constrained to equality over time in the baseline model. Furthermore, the five-factor solution fit the data best, including strong (>.50) and significant (p < .001) factor loadings and thus supporting the discriminant validity of the construct.

Table 1.

Goodness-of-Fit Measures of Longitudinal Measurement Model: First-Order Longitudinal Factor Models.

Model	AIC	BIC	χ² (df)	SRMR	CFI	RMSEA [CI]	Δχ² (df)
Factor loadings free	18142.69	19151.95	1109.10 (795)	.063	.935	.039 [.034, .045]
Equal factor loadings	18131.07	19069.50	1137.48 (815)	.065	.933	.039 [.034, .045]	28.38 (20) ns
Equal intercepts (five-factor model)	18121.11	18988.72	1167.52 (835)	.065	.931	.040 [.034, .045]	30.04 (20) ns
Four-factor model (a)	18139.52	18854.86	1271.93 (878)	.070	.919	.042 [.037, .047]	104.41 (43)***
Four-factor model (b)	18211.37	18926.71	1343.78 (878)	.067	.904	.046 [.041, .050]	176.26 (43)***
Three-factor model	18257.86	18852.80	1458.27 (912)	.072	.887	.048 [.044, .053]	290.75 (77)***
Two-factor model	18263.24	18769.64	1513.65 (937)	.075	.881	.049 [.045, .054]	346.13 (102)***
One-factor model	19617.61	20046.10	2912.02 (959)	.334	.597	.089 [.086, .093]	1744.50 (124)***

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; SRMR = standardized root mean square residual; CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval; df = degrees of freedom; ns = not significant. n = 255. Four-factor model (a): social support and task-oriented communication combined, four-factor model (b): social support and positive feedback combined, three-factor model: supervisor behavior combined, two-factor model: supervisor behavior combined and outcomes combined.

Model comparison *p < .05. **p < .01. ***p < .001.

The baseline model contained therefore the indicators of social support, positive feedback, task-related communication, job satisfaction, and affective job-related well-being, and therefore, it contained the latent variables: initial levels, change factors, and indicator-specific factors. All error terms of latent variables were allowed to covary with each other.

Baseline Model

The fit of this model was determined to be good, χ²(705) = 951.17, RMSEA = .037 [.031, .043], CFI = .949, SRMR = .052, especially when considering the relatively small sample size used to test our complex and quite restrictive model (Hu & Bentler, 1999). The first objective (research question) of this study was to investigate patterns of change of supervisor behavior as well as employee well-being across time intervals of 3 and 6 months. The intercorrelations between the latent variables as well as the latent means and variances are displayed in Table 2. The mean change was not significantly different from zero for Δtask-related communication_{T3 − T1}, indicating no general change in this variable across the mentioned time frame, whereas the significant negative means of the true difference scores indicated a mean-level decrease in perceived leadership behavior, job satisfaction, and well-being over time in general. In confirming the significance of the variances of the true difference scores, we concluded that the variability in the change between participants for all latent variables except for job satisfaction in the first 3 months was reliable (Δjob satisfaction _{T2 − T1}).

Table 2.

Estimated Correlation Matrix Between Latent Variables and Latent Change Variables.

	Mean (SE)	Variance (SE)	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
Latent variable
1. Social support_T1	0.00 (—)	0.64 (0.09)***
2. ΔSocial support_{T2 − T1}	−0.27 (0.06)***	0.50 (0.09)***	−.57***
3. ΔSocial support_{T3 − T1}	−0.17 (0.07)*	0.62 (0.10)***	−.55***	.66***
4. Positive feedback_T1	0.00 (—)	1.06 (0.12)***	.73***	−.36***	−.27**
5. ΔPositive feedback_{T2 − T1}	−0.26 (.07)***	0.63 (.10)***	−.40***	.72***	.37***	−.44***
6. ΔPositive feedback_{T3 − T1}	−0.24 (0.08)**	0.92 (0.14)***	−.45***	.43***	.70***	−.53***	.43***
7. Task-related communication_T1	0.00 (—)	0.51 (0.07)***	.72***	−.46***	−.46***	.57***	−.25**	−.32***
8. ΔTask-related communication_{T2 − T1}	−0.13(0.05)**	0.20 (0.05)***	−.37**	.77***	.37**	−.35**	.43**	.33*	−.56***
9. ΔTask-related communication_{T3 − T1}	−0.10 (0.06)	0.42 (0.08)***	−.50***	.44***	.75***	−.38***	.22	.55***	−.58***	.41**
10. Job satisfaction_T1	0.00 (—)	0.17 (0.04)***	.31**	−.08	−.22	.40***	−.18	−.27*	.42***	−.08	−.24
11. ΔJob satisfaction_{T2 − T1}	−0.10 (0.04)*	0.06 (0.04)	.01	.20	.71**	.22	.08	.22	−.17	−.02	.45	.08
12. ΔJob satisfaction_{T3 − T1}	−0.13 (0.04)**	0.09(0.04)*	−.20	.51**	.82***	−.18	.56**	.61***	−.17	.08	.56**	−.14	.46
13. Affective job-related well-being_T1	0.00 (—)	0.30 (0.04)***	.29***	−.13	−.11	.33***	−.17	−.24**	.27**	−.10	−.13	.72***	.12	.03
14. ΔAffective job-related well-being_{T2 − T1}	−0.08 (.04)*	0.13 (.03)***	.01	.12	.13	.07	.30*	.24	.13	−.07	−.04	−.03	.62*	.23	−.22
15. ΔAffective job-related well-being _{T3 − T1}	−0.09 (0.04)*	0.16 (0.03)***	−.06	.22	.33**	−.04	.25*	.35**	.15	−.09	.16	−.03	.40	.42*	−.17	.62***
Control variable
Age			−.19*	.27**	.19	−.09	.11	.07	−.15	.11	.09	−.14	.32	.26	.17	.04	.20
Gender (0 = female, 1 = male)			−.16*	.20*	.09	−.14	.05	.12	−.10	.24*	.14	−.22	.03	.13	−.16	.02	−.05
Migration background			.01	.01	−.06	.01	.12	−.11	.00	.05	.06	−.19*	.02	.11	.13	−.09	−.02
Tenure			−.02	−.06	.01	−.02	.11	−.07	−.02	.15	.00	.12	−.34	.06	−.07	.08	−.00
Employment (0 = permanent, 1 = temporary)			−.03	.04	−.12	−.11	.19	.03	.00	−.13	−.03	−.08	−.11	−.15	−.12	−.10	−.01

Note. n = 255; Δ indicates the difference between the subscripted measurement points of the true variable.

p < .05. **p < .01. ***p < .001.

The latent variables at T1 were significantly positive correlated; coefficients ranged between r = .27 and r = .73. The initial level of supervisor behavior (T1) was significantly correlated with change (Δ_{T2 − T1} and Δ_{T3 − T1}) in the single constructs. These correlations indicate that participants with high scores at the first measurement occasion tended to show greater decline over time. The well-being variables did not show these correlations.

To specify our final model, which included directed paths from supervisor behavior to well-being, we inspected the correlations between change in supervisor behavior and change in employee well-being (Hypotheses 1a, 1b, 1c). There was only one significant correlation between changes in supervisor behavior and changes in employee job satisfaction or well-being in the first 3 months (Δpositive feedback_{T2 − T1} with Δaffective job-related well-being_{T2 − T1}). Inspection of the change scores for the 6-month interval (Δ_{T3 − T1}) showed that all supervisor behavior change variables were correlated with change in employee job satisfaction (r = .56-.82). Also, social support and positive feedback were correlated with change in affective job-related well-being (r = .33 and .35, respectively). The correlation coefficient between intraindividual change in perceived leadership and employee well-being is significant for the 6 months period. It is not significant for all relations for the 3-month period. Changes in supervisor behavior at 3 months were also correlated with changes in well-being indicators at 6 months: The changes in social support and positive feedback across the first time period were associated with job satisfaction after the second time period (r = .51 and .56), and change in positive feedback was associated with affective job-related well-being after the second time period (r = .25).

Latent Regression Analysis

In the three models that we used to test Hypothesis 2, the dependent latent change variables (job-related affective well-being, job satisfaction) were regressed on the independent latent change variables (supervisor behavior). Each supervisor behavior was modeled separately, and each model included both change variables (ΔT2 − T1, ΔT3 − T1; Table 3). The results of these regressions indicated significant coefficients for predicting job satisfaction from all supervisor behaviors for the 6-month period. For job satisfaction, social support had the highest coefficient of determination. The regressions predicting job-related affective well-being from positive feedback were significant for both periods, whereas the coefficients for social support as well as task-related communication were not significant in any model.

Table 3.

Single Standardized Regression Coefficients for Perceived Leadership Resources on Subordinates’ Job-Related Affective Well-Being and Job Satisfaction.

	Job-related affective well-being						Job satisfaction
	ΔT2 − T1			ΔT3 − T1			ΔT2 − T1			ΔT3 − T1
	Coefficient (SE)	t	R²	Coefficient (SE)	t	R²	Coefficient (SE)	t	R²	Coefficient (SE)	t	R²
Social support
ΔT2 − T1	.22 (.13)	1.67	.05	.06 (.18)	0.30	.12	.31 (.24)	1.30	.10	−.11 (.21)	−0.54	.74
ΔT3 − T1	—		.05	.31 (.17)	1.85	.12	—		.10	.34 (.21)	4.49***	.74
Positive feedback
ΔT2 − T1	.31 (.13)	2.51*	.10	.15 (.14)	1.12	.15	.14 (.24)	0.63	.02	.29 (.17)	1.73	.38
ΔT3 − T1	—		.10	.29 (.12)	2.45*	.15	—		.02	.43 (.16)	2.80**	.38
Task-related communication
ΔT2 − T1	.05 (.17)	0.28	.00	−.18 (.18)	−0.97	.05	.23 (.30)	0.76	.05	−.30 (.24)	−1.21	.30
ΔT3 − T1	—			.22 (.13)	1.66		—			.61 (.20)	3.15**

Note. n = 255; Δ indicates the difference between the subscripted measurement points for the true variable.

p < .05. **p < .01. ***p < .001.

In an additional model, we combined the three supervisor behaviors into one composite measure (from the scale means for each measurement occasion). We chose the composite measure to display a parsimonious model and to avoid the difficulty of isolating the single impacts of each behavior in one regression model. The high intercorrelations of the variables describing the supervisor behavior would otherwise lead to unstable estimations of single regression coefficients for the correlated variables (Farrar & Glauber, 1967). Based on preliminary regression analyses with the demographic and focal variables, we included control variables by regressing the initial level of the respective variables on the control variables in the latent regression model. As a result of including only significant paths, supervisor behavior was regressed on age, whereas the well-being indicators were regressed on immigrant background (dummy coded: 0 = German, 1 = immigrant) and gender (dummy coded: 0 = female, 1 = male). The standardized regression coefficients of the composite measure model, χ²(351) = 481.07, RMSEA = .039 [.030, .047], CFI = .954, SRMR = .066, are displayed in Figure 2. The overall model fit of this model was very good, especially when excluding the controls, and did not result in changes in the relations among the latent variables, χ²(263) = 314.30, RMSEA = .028 [.013, .039], CFI = .982, SRMR = .049. The common contribution for Δjob satisfaction_{T3 − T1} was R² = .61 and for Δaffective job-related well-being_{T3 − T1} was R² = .11.

Figure 2.

Latent regression model with significant standardized path coefficients.

Discussion

The current study addressed several understudied areas in organizational psychology. First, this study addressed the work situation of workers in semiskilled and unskilled jobs, in particular their main psychosocial resources at work, namely, health-promoting leadership behavior of their direct supervisors. Second, it focused on leadership behavior of the lower echelon of management, which has also been rather neglected in leadership research. Third, in our study we investigate interindividual differences in intraindividual change in perceived supervisor behavior and intraindividual change in well-being indicators across time to test causal relationships. The present investigation demonstrates that latent change models can be applied to analyze longitudinal data of supervisor behavior and employee well-being proceeded on the assumption that only perceived change of supervisor behavior can lead to change in subordinates’ well-being. We conceptualized small examples of daily leadership behavior (social support, positive feedback, and task-related communication behavior), which could be assumed to be easily operationalized by first-line managers in order to support their subordinates. Findings verified intraindividual changes in the perception of supervisor behavior and well-being as well as the positive association between changes in these perceived supervisor behaviors and changes in well-being over time. These findings allow us a new perspective on the effect of supervisor behavior for the target group of low-skilled workers.

Theoretical Implications

Including two time intervals allowed us to analyze interindividual differences in intraindividual changes 3 months as well as 6 months later. We found significant mean changes and reliable variability in change in most latent variables. Exceptions were the nonsignificant mean change in task-related communication after 6 months and the nonsignificant variance in job satisfaction change in the first 3 months. These results indicate that information-providing behavior by supervisors is relatively stable at the group level, and there are no significant differences between the individual workers in their change in job satisfaction across a 3-month time interval. The low variance in change in job satisfaction in general can be traced back to the measurement that included only two indicators. For further investigations, we recommend that job satisfaction be measure with at least three indicators when calculating intraindividual change variables.

Since longitudinal investigations were conducted, there exists the debate about the appropriate time lag between measurement occasions (Shamir, 2011). A good choice of time lag is essential especially in field studies without planned interventions. The explicit formulation of change allowed a more detailed view to change of leader behavior as well as job satisfaction and well-being over different time spans. Analyzing change–change associations indicated that a 3-month time interval was too short for finding relations between latent change in supervisor behavior and employee well-being, whereas we found associations between supervisor behavior change in the first 3 months and across the whole time interval with job satisfaction change after 6 months. In summary, it can be stated that the relationship between changes is more stable when analyzing a 6-month time period. This finding is very useful for implications to design longitudinal field studies regarding associations between leadership and employee outcomes. With regard to determining a time interval that is appropriate for analyzing change, we recommend at least a 6-month interval to also eliminate the possibility of memory effects. However, if the time interval between the measurement occasions is larger, one might expect a greater number of dropouts as well as confounding organizational changes, especially in this target group of low-skilled workers.

To examine Hypotheses 1a, 1b, and 1c, we differentiated between job satisfaction and job-related affective well-being as indicators of well-being. We confirmed our hypotheses except for the association between change in task-related communication and change in affective job-related well-being. We found that change in task-related communication was associated with change in employee job satisfaction, but it was not associated with change in affective job-related well-being. Comparing the separate regression analyses of the three supervisor behaviors, change in social support seemed to be the best predictor of change in job satisfaction, whereas change in positive feedback best predicted change in job-related affective well-being. According to these results, we assume that the supervisor’s communication and social support do not affect the affective component of well-being, whereas they are related to both cognitive and affective components of attitudinal job-satisfaction–related outcomes (Kaplan, Warren, Adam, & Thoresen, 2009). Task-related communication could be classified as instrumental support (Cohen & Wills, 1985; Semmer et al., 2008). The effects of instrumental support depend on the subjective relevance that the supervisor’s behavior has for the recipient.

Including immigration background as a control variable in the latent regression model revealed significant negative effects on job satisfaction. Immigrants in our sample were less satisfied with their work than native workers. According to these differences, in further studies concerning low-skilled workers, it seems necessary to provide detailed comparisons between immigrants and native workers from the same workgroups (Jørgensen et al., 2011; Olesen et al., 2012; Ortega, Carneiro, & Flyvholm, 2010).

Limitations

The current findings must be interpreted with the due consideration of several limitations. The high proportion of missing participants at each measurement occasion is the main limitation of longitudinal investigations such as ours (Ployhart & Ward, 2011). We assume that the main reasons for these incomplete cases are that employees were on vacation or taking sick days during one or more of the measurement occasions. Additionally, our sample comprised a large number of temporary employees, a finding that is characteristic of the analyzed target group. Therefore, some employees began the study at the second measurement occasion or dropped out after the first. To deal with these missing data, we used advanced estimation techniques. The second weakness of this study was that we relied solely on self-report data. Both supervisor behavior and employee well-being were reported through the perceptions of the employees, although all items were simple and specific (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). However, it bears mentioning that the mean changes in all the integrated variables were negative across time. Consequently, it could be assumed that the repeated measurements per se could have had an effect on the responses due to motivational biases. The repeated invitation to fill out the same questionnaire could influence the rating in a negative direction. It is recommended to examine the motivation with supplementary questions. Another possible explanation for decreasing mean ratings in general could be that workers in low hierarchical levels, in particular, are not used to filling out questionnaires and could fear negative consequences of a negative rating. Experiencing no individual consequences from taking the survey and building trust in the research team could result in more honest answers and subsequently worse ratings of their work situation and their immediate supervisor. Further studies could include items to control for the fear of negative consequences. Beside these explanations, there could be other changes (e.g., economic situation) at work influencing both the leader behavior and well-being in general. For further studies explicit focusing change we suggest to add an open question like “Did anything in the past 3 months change at your work influencing your well-being?”

Because we relied exclusively on self-reported measures, we cannot rule out the possibility of shared method variance (Spector, 2006). Nevertheless, to examine intraindividual change in subjective perceptions, we had to rely on self-report measures (Chan, 2009).

The measures that we used were chosen carefully and modified to suit the perceptions of work characteristics as well as the language abilities of immigrant respondents and other respondents who lacked much education in low-skilled jobs. Even if these modifications threaten validity, we found them to be unavoidable in order to assess data from our target group. The scales we used lack explicit validation especially tests of convergent validity of the used simplified items with existing measures. Simultaneously, these tailor-made questionnaires are responsible for new insights into an understudied and vulnerable workforce. We do not recommend using complicated wordy questionnaires for low-qualified and immigrant workers. Nevertheless, a larger sample of low-skilled employees would have strengthened our conclusions in general.

Another limitation to be mentioned is the parceling strategy we used for job satisfaction. We chose an indicator structure to describe job satisfaction with two indicators measuring overall job satisfaction. This procedure has several advantages as well as shortcomings (Williams & O’Boyle, 2008). Using parcels instead of items as indicators results in a smaller covariance matrix and therefore less parameter estimates and a better model fit. The focus of our study was the relation among the latent leader behaviors and the studied outcomes rather than relations between items of job satisfaction. In this case, parceling is recommended.

The concept of intraindividual change provides a new approach to leadership research. Although we could confirm the hypothesized association between change in perceived supervisor behavior and employee well-being, we recommend that future research examine potential moderating and mediating variables such as work climate and that future research include neutral control variables to investigate the effects of repeated measurements.

Our results suggest a causal relationship between leadership behavior and employee well-being although studies have shown an association between supervisor behavior and employee well-being in the form of a dynamic reciprocal process or a feedback loop (van Dierendonck et al., 2004). It is easily conceivable that employees with high well-being at work will rate their supervisors more positively and that changes in well-being affect changes in supervisor perception. Nevertheless, longitudinal change associations provide stronger inferences of causality than cross-sectional associations or longitudinal simple rating associations. Latent change models are an adequate possibility to analyze longitudinal associations as precondition for causality in field studies. Intervention studies with experimental designs are for sure even stronger. Despite these limitations, the major strengths of this study include the use of longitudinal three-wave data, a focus on the specific target group, as well as the use of sophisticated statistical methods.

Practical Implications

The present study has particular implications for health-promoting interventions for low-skilled workers. These workers have been found to show low motivation for participating in training and health-promoting activities. The reasons are multifaceted. Low self-efficacy, negative attitudes toward healthy behavior, perceptions of low behavioral control, and limited reading and writing skills are a few examples (Busch et al., 2010). Immigrants have even more problems participating in trainings because of language and cultural barriers. The participation rates of immigrants are 50% lower than those of natives (Ambos, 2005). Successfully reaching this target group is difficult. Based on the results of our study, offering training interventions to the direct supervisors of low-skilled workers may provide an effective way to promote the health of these workers. Until now, managers of the lowest hierarchical level are only seldom the ones who participate in leadership interventions. In addition to the finding that levels of supportive and appreciative supervisor behavior are associated with the outcomes of their subordinates, the results of this study showed that changes in these behaviors lead to changes in subordinates’ well-being. They should be taught to be aware that their behavior influence the well-being of their subordinates and that their behavior in general can serve as a protective factor for workers in low-skilled jobs. We suggest that supervisors be taught the value of positive interactions with their subordinates at work in general and, more specifically, how to behave in a health-promoting manner. This study confirmed the effects of simple behaviors that can easily be applied at work on a daily basis.

Footnotes

Acknowledgements

The authors wish to express their gratitude to Dr. Tobias Koch (research associate, Methods and Evaluation, Freie Universität Berlin) for his helpful suggestions with regard to data analysis using structural equation modeling.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by the German Federal Ministry of Education and Research, Project Registration No. 01EL0803, duration 2009-2013, project leader C. Busch.

Notes

Author Biographies

Eva Winkler is a PhD Candidate in Work and Organizational Psychology at Hamburg University. She was research assistant in the ReSuDi research project 2011-2013. She currently works for a service company in Hamburg in the department for occupational health and safety. Her main areas of research areoccupational health psychology, health-promoting leadership, first-level supervisors and low-skilled multicultural workforces.

Christine Busch was 2009-2013 the project leader of the research project ReSuDi as well as 2006-2009 project leader of the predecessor project ReSum [Stress- and Resource Management for Low-Qualified Workers: The Development of a Multiplier Concept]. She works since 1998 as research associate in the department of Work- and Organizational Psychology at Hamburg University.Her research interests include especially occupational health psychology, stress management intervention programs and their evaluation.

Julia Clasen is Professor of Work- and Organisational Psychology at Euro-FH - University of Applied Sciences, Hamburg. She is interested in occupational health psychology, self-employment and workplace flexibility.

Julia Vowinkel graduated from Eberhard Karl’s Universtiy of Tübingen with a major in “Work and Organizational Psychology” and “Clinical Psychology”. She enrolled as a PhD/doctoral student at the University of Hamburg afterwards and started to work as a research assisstant in the “ReSuDi” project. Her current research interests are well-being at work, occupational stress and health, cross-cultural research and methodology as well as immigrant employees.

References

Amason

Allen

M. W.

Holmes

S. A.

(1999). Social support and acculturative stress in the multicultural workplace. Journal of Applied Communication Research, 27, 310-334. doi:10.1080/00909889909365543

Ambos

(2005). Nationaler Report: Geringqualifizierte und berufliche Weiterbildung—empirische Befunde zur Weiterbildungssituation in Deutschland. Retrieved from http://www.die-bonn.de/weiterbildung/Literaturrecherche/details.aspx?ID=3210

Arbuckle

J. L.

(1996). Full information estimation in the presence of incomplete data. In Marcoulides

G. A.

Schumacker

R. E.

(Eds.), Advanced structural equation modeling: Issues and techniques (pp. 243-277). Mahwah, NJ: Lawrence Erlbaum.

Arnold

K. A.

Turner

Barling

Kelloway

E. K.

McKee

M. C.

(2007). Transformational leadership and psychological well-being: The mediating role of meaningful work. Journal of Occupational Health Psychology, 12, 193-203. doi:10.1037/1076-8998.12.3.193

Ashford

S. J.

Cummings

L. L.

(1985). Proactive feedback seeking: The instrumental use of the information environment. Journal of Occupational Psychology, 58, 67-79. doi:10.1111/j.2044-8325.1985.tb00181.x

Bakker

A. B.

Demerouti

(2007). The job demands-resources model: State of the art. Journal of Managerial Psychology, 22, 309-328. doi:10.1108/02683940710733115

Batinic

Selenko

Stiglbauer

Paul

K. I.

(2010). Are workers in high-status jobs healthier than others? Assessing Jahoda’s latent benefits of employment in two working populations. Work & Stress, 24, 73-87. doi:10.1080/02678371003703859

Bollen

K. A.

Bauldry

(2011). Three Cs in measurement models: Causal indicators, composite indicators, and covariates. Psychological Methods, 16, 265-284. doi:10.1037/a0024448

Borrell

Muntaner

Benach

Artazcoz

(2004). Social class and self-reported health status among men and women: What is the role of work organisation, household material standards and household labour? Social Science & Medicine, 58, 1869-87. doi:10.1016/S0277-9536(03)00408-8

10.

Brähler

Mühlan

Albani

Schmidt

(2007). Teststatistische Prüfung und Normierung der deutschen Versionen des EUROHIS-QOL Lebensqualität-Index und des WHO-5 Wohlbefindens-Index [Testing and standardization of the German version of the EUROHIS-QOL and WHO-5 quality-of life-indices]. Diagnostica, 53, 83-96. doi:10.1026/0012-1924.53.2.83

11.

Bruch

Walter

(2007). Leadership in context: Investigating hierarchical impacts on transformational leadership. Leadership & Organization Development Journal, 28, 710-726. doi:10.1108/01437730710835452

12.

Busch

Staar

Aborg

Roscher

Ducki

(2010). The neglected employees: Work-life balance and stress management intervention program for low-qualified workers. In Houdmont

Leka

(Ed.), Contemporary occupational health psychology: Global perspectives on research and practice (pp. 98-123). New York, NY: John Wiley.

13.

Chan

(2009). So why ask me? Are self-report data really that bad? In Lance

C. E.

Vandenberg

R. J.

(Eds.), Statistical and methodological myths and urban legends: Doctrine, verity and fable in the organizational and social sciences (pp. 311-338). New York, NY: Routledge.

14.

Chang

Tharenou

(2004). Competencies needed for managing a multicultural workgroup. Asia Pacific Journal of Human Resources, 42, 57-74. doi:10.1177/1038411104041534

15.

Christensen

K. B.

Labriola

Lund

Kivimäki

(2008). Explaining the social gradient in long-term sickness absence: A prospective study of Danish employees. Journal of Epidemiology & Community Health, 62, 181-183. doi:10.1136/jech.2006.056135

16.

Cohen

Wills

T. A.

(1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310-357.

17.

Curran

P. J.

Hussong

A. M.

(2002). Structural equation modeling of repeated measures data: Latent curve analysis. In Moskowitz

D. S.

Hershberger

S. L.

(Eds.), Modeling intraindividual variability with repeated measures data (pp. 59-85). Mahwah, NJ: Erlbaum.

18.

Dalgard

O. S.

Thapa

S. B.

Hauff

McCubbin

Syed

H. R.

(2006). Immigration, lack of control and psychological distress: Findings from the Oslo Health Study. Scandinavian Journal of Psychology, 47, 551-558. doi:10.1111/j.1467-9450.2006.00546.x

19.

de Lange

A. H.

Taris

T. W.

Kompier

M. A.

Houtman

I. L.

Bongers

P. M.

(2003). “The very best of the millennium”: Longitudinal research and the demand-control-(support) model. Journal of Occupational Health Psychology, 8, 282-305. doi:10.1037/1076-8998.8.4.282

20.

de Lange

A. H.

Taris

T. W.

Kompier

M. A.

Houtman

I. L.

Bongers

P. M.

(2004). The relationships between work characteristics and mental health: Examining normal, reversed and reciprocal relationships in a 4-wave study. Work & Stress, 18, 149-166. doi:10.1080/02678370412331270860

21.

Dolbier

C. E.

Webster

J. A.

Mccalister

K. T.

Mallon

M. W.

Steinhardt

M. A.

(2005). Reliability and validity of a single-item measure of job satisfaction. American Journal of Health Promotion, 19, 194-198.

22.

Edwards

J. R.

(2011). The fallacy of formative measurement. Organizational Research Methods, 14, 370-388. doi:10.1177/1094428110378369

23.

Eid

Schneider

Schwenkmezger

(1999). Do you feel better or worse? The validity of perceived deviations of mood states from mood traits. European Journal of Personality, 13, 283-306.

24.

Enders

C. K.

(2010). Applied missing data analysis. New York, NY: Guilford Press.

25.

Eurofond. (2007). Fifth European Working Conditions Survey (European Foundation for the Improvement of Living and Working Conditions, Ed.). Luxembourg: Author. Retrieved from http://www.eurofound.europa.eu/surveys/ewcs/index.htm

26.

Farrar

D. R.

Glauber

R. R.

(1967). Multicollinearity in regression analysis: The problem revisited. Review of Economics and Statistics, 49, 92-107.

27.

Felfe

Schyns

(2002). The relationship between employees’ occupational self-efficacy and perceived transformational leadership-replication and extension of recent results. Current Research in Social Psychology, 7(9), 137-162.

28.

Fix

Sias

P. M.

(2006). Person-centered communication, leader-member exchange, and employee job satisfaction. Communication Research Reports, 23, 35-44. doi:10.1080/17464090500535855

29.

Frese

(1989). Gütekriterien der Operationalisierung von sozialer Unterstutzung am Arbeitsplatz [Psychometric criteria of the operationalization of social support at work]. Zeitschrift für Arbeitswissenschaft, 43, 112-121.

30.

Gerstner

C. R.

Day

D. V.

(1997). Meta-analytic review of leader-member exchange theory: Correlates and construct issues. Journal of Applied Psychology, 82, 827-844. doi:10.1037//0021-9010.82.6.827

31.

Gilbreath

Benson

P. G.

(2004). The contribution of supervisor behaviour to employee psychological well-being. Work & Stress, 18, 255-266. doi:10.1080/02678370412331317499

32.

Graham

J. W.

(2003). Adding missing-data-relevant variables to FIML-based structural equation models. Structural Equation Modeling, 10, 80-100.

33.

Gregersen

Kuhnert

Zimber

Nienhaus

(2011). Führungsverhalten und Gesundheit—Zum Stand der Forschung [Leadership behavior and health—The current state of research]. Gesundheitswesen, 73, 3-12.

34.

Hackman

J. R.

Oldham

G. R.

(1976). Motivation through the Design of Work: Test of a Theory. Organizational Behavior and Human Performance, 16, 250–279.

35.

Hales

(2005). Rooted in supervision, branching into management: Continuity and change in the role of first-line manager. Journal of Management Studies, 42, 471-506. doi:10.1111/j.1467-6486.2005.00506.x

36.

Hertzog

Nesselroade

J. R.

(2003). Assessing psychological change in adulthood: An overview of methodological issues. Psychology and Aging, 18, 639-657. doi:10.1037/0882-7974.18.4.639

37.

Hoppe

(2011). Psychosocial working conditions and well-being among immigrant and German low-wage workers. Journal of Occupational Health Psychology, 16, 187-201. doi:10.1037/a0021728

38.

Howell

R. D.

Breivik

Wilcox

J. B.

(2007). Reconsidering formative measurement. Psychological Methods, 12, 205-218. doi:10.1037/1082-989X.12.2.205

39.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

40.

Kaplan

Dalal

R. S.

(2010). An examination of blue- versus white-collar workers’ conceptualizations of job satisfaction facets. Journal of Vocational Behavior, 76, 317-325. doi:10.1016/j.jvb.2009.10.014

41.

Hunter

S. T.

Bedell-Avers

K. E.

Mumford

M. D.

(2007). The typical leadership study: Assumptions, implications, and potential remedies. Leadership Quarterly, 18, 435-446. doi:10.1016/j.leaqua.2007.07.001

42.

Jørgensen

M. B.

Rasmussen

C. D. N.

Carneiro

I. G.

Flyvholm

M.-A.

Olesen

Ekner

. . . Holtermann

(2011). Health disparities between immigrant and Danish cleaners. International Archives of Occupational and Environmental Health, 84, 665-674. doi:10.1007/s00420-010-0607-2

43.

Kaplan

S. A.

Warren

C. R.

Adam

Thoresen

C. J.

(2009). A note on the relationship between affect(ivity) and differing conceptualizations of job satisfaction: Some unexpected meta-analytic findings. European Journal of Work and Organizational Psychology, 18(1), 29-54.

44.

Kelloway

E. K.

Turner

Barling

(2012). Transformational leadership and employee psychological well-being: The mediating role of employee trust in leadership. Work & Stress, 26, 39-55.

45.

Kelloway

E. K.

Weigand

McKee

M. C.

Das

(2012). Positive leadership and employee well-being. Journal of Leadership & Organizational Studies, 20, 107-117. doi:10.1177/1548051812465892

46.

Kerr

Hill

K. D.

Broedling

(1986). The first-line supervisor: Phasing out or here to stay? Academy of Management Review, 11, 103-117.

47.

Kerr

Jermier

J. M.

(1978). Substitutes for leadership: Their meaning and measurement. Organizational Behavior and Human Performance, 22, 375-403.

48.

Kim

Lee

S. Y.

(2009). Supervisory communication, burnout, and turnover intention among social workers in health care settings. Social Work in Health Care, 48, 364-385. doi:10.1080/00981380802598499

49.

Kraut

A. I.

Pedigo

P. R.

Mckenna

D. D.

Dunnette

M. D.

Kraut

(1989). The role of the manager: What’s really important in different management jobs. Academy of Management Executive, 3, 286-293.

50.

Kunin

(1955). The construction of a new type of attitude measure. Personnel Psychology, 8, 65-78.

51.

Kuoppala

Lamminpää

Liira

Vainio

(2008). Leadership, job well-being, and health effects: A systematic review and a meta-analysis. Journal of Occupational and Environmental Medicine, 50, 904-915. doi:10.1097/JOM.0b013e31817e918d

52.

Lam

S. S. K.

Yik

M. S. M.

Schaubroeck

(2002). Responses to formal performance appraisal feedback: The role of negative affectivity. Journal of Applied Psychology, 87, 192-201. doi:10.1037//0021-9010.87.1.192

53.

Law

K. S.

Wong

(1999). Multidimensional constructs in structural equation analysis: An illustration using the job perception and job satisfaction constructs. Journal of Management, 25, 143-160. doi:10.1177/014920639902500202

54.

Lipscomb

H. J.

Loomis

McDonald

M. A.

Argue

R. A.

Wing

(2006). A conceptual model of work and health disparities in the United States. International Journal of Health Services, 36, 25-50. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16524164

55.

Little

R. J. A.

(1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83, 1198-1202.

56.

Little

R. J. A.

Rubin

D. B.

(2003). Statistical analysis with missing data. Sociological methods research (Vol. 18). New York, NY: Wiley.

57.

Lyly-Yrjänäinen

(2008). Who needs up-skilling? Low-skilled and low-qualified workers in the European Union. Retrieved from http://www.eurofound.europa.eu/pubdocs/2008/97/en/1/EF0897EN.pdf

58.

MacKenzie

S. B.

Podsakoff

P. M.

Jarvis

C. B.

(2005). The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions. Journal of Applied Psychology, 90, 710-730. doi:10.1037/0021-9010.90.4.710

59.

McArdle

J. J.

(2009). Latent variable modeling of differences and changes with longitudinal data. Annual Review of Psychology, 60, 577-605. doi:10.1146/annurev.psych.60.110707.163612

60.

Meredith

Horn

J. L.

(2001). The role of factorial invariance in modeling growth and change. In Collins

L. M.

Sayer

A. G.

(Eds.), New methods for the analysis of change (pp. 203-240). Washington, DC: American Psychological Association. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2001-01077-007&lang=es&site=ehost-live

61.

Miles

E. W.

Patrick

S. L.

King

W. C.

(1996). Job level as a systemic variable in predicting the relationship between supervisory communication and job satisfaction. Journal of Occupational and Organizational Psychology, 69, 277-293.

62.

Morgeson

F. P.

Humphrey

S. E.

(2006). The Work Design Questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91, 1321-1339. doi:10.1037/0021-9010.91.6.1321

63.

Muthén

L. K.

Muthén

B. O.

(2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Muthén & Muthén.

64.

Niedhammer

Sultan-Taïeb

Chastang

J.-F.

Vermeylen

Parent-Thirion

(2012). Exposure to psychosocial work factors in 31 European countries. Occupational Medicine, 62, 196-202. doi:10.1093/occmed/kqs020

65.

Nielsen

Munir

(2009). How do transformational leaders influence followers’ affective well-being? Exploring the mediating role of self-efficacy. Work & Stress, 23, 313-329. doi:10.1080/02678370903385106

66.

Nielsen

Randall

Yarker

Brenner

S.-O.

(2008). The effects of transformational leadership on followers’ perceived work characteristics and psychological well-being: A longitudinal study. Work & Stress, 22, 16-32. doi:10.1080/02678370801979430

67.

Nübling

Stößel

Hasselhorn

H. M.

Michaelis

Hofmann

(2005). Erfassung psychischer Belastungen—Erprobung eines Messinstrumentes (COPSOQ) [Measuring psychological stress and strain at work: Evaluation of the COPSOQ Questionnaire in Germany]. Schriftenreihe der Bundesanstalt für Arbeitsschutz und Arbeitsmedizin. Bremerhaven, Germany: Wirtschaftsverlag NW.

68.

Olesen

Carneiro

I. G.

Jørgensen

M. B.

Flyvholm

M.-A.

Rugulies

Rasmussen

C. D. N.

. . . Holtermann

(2012). Psychosocial work environment among immigrant and Danish cleaners. International Archives of Occupational and Environmental Health, 85, 89-95. doi:10.1007/s00420-011-0642-7

69.

Ortega

Carneiro

I. G.

Flyvholm

M.-A.

(2010). A descriptive study on immigrant workers in the elderly care sector. Journal of Immigrant and Minority Health, 12, 699-706. Retrieved from http://cat.inist.fr/?aModele=afficheN&cpsidt=23105225

70.

Pelfrene

Vlerick

Mak

R. P.

Smet

Kornitzer

Backer

G. D.

(2010). Scale reliability and validity of the Karasek “Job Demand-Control-Support” model in the Belstress study. Work & Stress, 15, 297-313.

71.

Ployhart

R. E.

Ward

A.-K.

(2011). The “quick start guide” for conducting and publishing longitudinal research. Journal of Business and Psychology, 26, 413-422. doi:10.1007/s10869-011-9209-6

72.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J.-Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903. doi:10.1037/0021-9010.88.5.879

73.

Priestland

Hanig

(2005). Developing first-level leaders. Harvard Business Review, 83, 112-120.

74.

Rogosa

(1980). A critique of cross-lagged correlation. Psychological Bulletin, 88, 245-258.

75.

Rydstedt

L. W.

Devereux

Sverke

(2007). Comparing and combining the demand-control-support model and the effort reward imbalance model to predict long-term mental strain. European Journal of Work and Organizational Psychology, 16, 261-278. doi:10.1080/13594320601182311

76.

Sadri

Tran

(2002). Managing your diverse workforce through improved communication. Journal of Management Development, 21, 227-237. doi:10.1108/02621710210420291

77.

Schafer

J. L.

Graham

J. W.

(2002). Missing data: Our view of the state of the art. Psychological Methods, 7, 147-177. doi:10.1037//1082-989X.7.2.147

78.

Schreuder

K. J.

Roelen

C. A. M.

Koopmans

P. C.

Groothoff

J. W.

(2008). Job demands and health complaints in white and blue collar workers. Work, 31, 425-432.

79.

Schrijvers

C. T.

van de Mheen

H. D.

Stronks

Mackenbach

J. P.

(1998). Socioeconomic inequalities in health in the working population: The contribution of working conditions. International Journal of Epidemiology, 27, 1011-1018. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10024196

80.

Schyns

von Collani

(2010). A new occupational self-efficacy scale and its relation to personality constructs and organizational variables. European Journal of Work and Organizational Psychology, 11, 219-241.

81.

Seebaß

Siegert

(2011). Migranten am Arbeitsmarkt in Deutschland [Migrants in the German Labor Market] (Working Paper No. 36). Nürnberg, Germany. Retrieved from http://www.bamf.de/SharedDocs/Anlagen/DE/Publikationen/WorkingPapers/wp36-migranten-am-arbeitsmarkt-in-deutschland.pdf?__blob=publicationFile

82.

Semmer

N. K.

Elfering

Jacobshagen

Perrot

Beehr

T. A.

Boos

(2008). The emotional meaning of instrumental social support. International Journal of Stress Management, 15, 235-251. doi:10.1037/1072-5245.15.3.235

83.

Shamir

(2011). Leadership takes time: Some implications of (not) taking time seriously in leadership research. Leadership Quarterly, 22, 307-315. doi:10.1016/j.leaqua.2011.02.006

84.

Shamir

House

R. J.

Arthur

M. B.

(1993). The motivational effects of charismatic leadership: A self-concept based theory. Organization Science, 4, 577-594. doi:10.1287/orsc.4.4.577

85.

Shamir

Howell

J. M.

(1999). Organizational and contextual influences on the emergence and effectiveness of charismatic leadership. Leadership Quarterly, 10, 257-283.

86.

Skakon

Nielsen

Borg

Guzman

(2010). Are leaders’ well-being, behaviours and style associated with the affective well-being of their employees? A systematic review of three decades of research. Work & Stress, 24, 107-139. doi:10.1080/02678373.2010.495262

87.

Spector

P. E.

(2006). Method variance in organizational research: Truth or urban legend? Organizational Research Methods, 9, 221-232. doi:10.1177/1094428105284955

88.

Steptoe

Marmot

(2002). The role of psychobiological pathways in socio-economic inequalities in cardiovascular disease risk. European Heart Journal, 23, 13-25.

89.

Steyer

Eid

Schwenkmezger

(1997). Modeling true intraindividual change: True change as a latent variable. Methods of Psychological Research Online, 2, 21-33.

90.

Steyer

Partchev

Shanahan

(2000). Modeling true intra-individual change in structural equation models: The case of poverty and children’s psychosocial adjustment. In Little

T. D.

Schnabel

K. U.

Baumert

(Eds.), Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples (pp. 109-126). Mahwah, NJ: Erlbaum.

91.

Tafvelin

Armelius

Westerberg

(2011). Toward understanding the direct and indirect effects of transformational leadership on well-being: A longitudinal study. Journal of Leadership & Organizational Studies, 18, 480-492. doi:10.1177/1548051811418342

92.

Van De Vliert

Shi

Sanders

Wang

Huang

(2004). Chinese and Dutch interpretations of supervisory feedback. Journal of Cross-Cultural Psychology, 35, 417-435. doi:10.1177/0022022104266107

93.

van Dierendonck

Haynes

Borrill

Stride

(2004). Leadership behavior and subordinate well-being. Journal of Occupational Health Psychology, 9, 165-175. doi:10.1037/1076-8998.9.2.165

94.

van Vegchel

de Jonge

Bakker

Schaufeli

(2002). Testing global and specific indicators of rewards in the effort-reward imbalance model: Does it make any difference? European Journal of Work and Organizational Psychology, 11, 403-421. doi:10.1080/13594320244000265

95.

Volmer

Niessen

Spurk

Linz

Abele

A. E.

(2011). Reciprocal relationships between leader-member exchange (LMX) and job satisfaction: A cross-lagged analysis. Applied Psychology, 60, 522-545. doi:10.1111/j.1464-0597.2011.00446.x

96.

Wadsworth

Dhillon

Shaw

Bhui

Stansfeld

Smith

(2007). Racial discrimination, ethnicity and work stress. Occupational Medicine, 57, 18-24. doi:10.1093/occmed/kql088

97.

Wanous

J. P.

Reichers

A. E.

Hudy

M. J.

(1997). Overall job satisfaction: How good are single-item measures? Journal of Applied Psychology, 82, 247-252. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9533198

98.

Williams

L. J.

O’Boyle

E. H.

(2008). Measurement models for linking latent variables and indicators: A review of human resource management research using parcels. Human Resource Management Review, 18, 233-242. doi:10.1016/j.hrmr.2008.07.002

99.

Winkler

Busch

Clasen

Vowinkel

(2014). Leadership behavior as a health-promoting resource for workers in low-skilled jobs and the moderating role of power distance orientation. Zeitschrift für Personalforschung, 28, 96-116.

100.

Winkler

Busch

Duresso

(2013). Gesundheits förderliches Führungsverhalten bei un-und angelernten, kulturell diversen Belegschaften [Health-promoting leadership behavior in unskilled and semiskilled, culturally diverse workforces]. Gruppendynamik und Organisationsberatung, 44, 153-169. doi:10.1007/s11612-013-0206-2

101.

Yammarino

(2013). Leadership: Past, present, and future. Journal of Leadership & Organizational Studies, 20, 149-155. doi:10.1177/1548051812471559

102.

Zapf

Dormann

Frese

(1996). Longitudinal studies in organizational stress research: A review of the literature with reference to methodological issues. Journal of Occupational Health Psychology, 1, 145-169. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9547043

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB