Retaining Social Workers: The Role of Quality of Work and Psychosocial Safety Climate for Work Engagement, Job Satisfaction, and Organizational Commitment

ABSTRACT

The present study investigated how psychosocial safety climate (PSC), job demands (role conflict and work-family conflict), job resources (social support from superiors and social community at work), and assessments for quality of work relate to social workers’ work engagement, job satisfaction, and organizational commitment. The results of the questionnaire study (N = 831) showed that quality of work was strongly related to all three outcomes, whereas PSC was found to be related to social workers’ job satisfaction. The contribution of the study is discussed in relation to understanding the retention of social workers.

KEYWORDS:

• job satisfaction
• organizational commitment
• psychosocial safety climate
• quality
• retention
• work engagement

Introduction

In the last few decades, workload and stress-related health issues have increased in human service organizations (HSOs) resulting in high turnover rates (Dollard, Dormann, Boyd, Winefield, & Winefield, 2003; Mor Barak, Nissly, & Levin, 2001). In many Western societies, this development has been especially evident among social workers (Astvik & Melin, 2012; Giffords, 2009; Griffiths, Royse, Culver, Piescher, & Zhang, 2017; Tham & Meagher, 2008). Consequently, public social services face great difficulties with regard to the retention and the recruitment of personnel (Coffey, Dugdill, & Tattersall, 2004; Kim & Kao, 2014; SKL, Swedish Association of Local Authorities and Regions, 2015; 2016; Tham, 2007). The work environment of social workers is an important societal question as it affects the quality of the services provided and thus the welfare of society (Smith & Shields, 2013; Travis, Lizano, & Mor Barak, 2016).

To date, quite a substantial amount of research has examined factors relating to turnover among social workers (e.g., Kim & Kao, 2014; Kim & Stoner, 2008; Tham, 2007). However, though the public social services face high turnover rates, many social workers decide to remain in the profession because they are satisfied with their jobs as well as engaged and committed to their work (Frost, Höjer, Campanini, Sicora, & Kullburg, 2017; Hakanen, Ropponen, Schaufeli, & De Witte, 2018). Enhancing the understanding of the factors that contribute to positive work attitudes among social workers can add valuable insight for efforts and interventions aiming to increase retention (Collins, 2008).

Based on earlier research, and the theoretical framework of the job demands-resources model (Bakker & Demerouti, 2014), the present study investigates how two novel work environmental factors, psychosocial safety climate (Dollard & Bakker, 2010) and perceived quality of work (Berthelsen, Conway, & Clausen, 2017; Schaufeli & Taris, 2014) could predict indicators of retention among social workers: work engagement, job satisfaction, and organizational commitment.

Job demands-resources model

The basic proposition of the job demands-resource (JD-R) model is that work-related outcomes and health issues can be explained by attending to the health impairment process and the motivational process. Each process is triggered by job characteristics, categorized as job demands or job resources (Bakker & Demerouti, 2017). Job demands are defined as those job characteristics that need effort and can drain energy (e.g., quantitative demands, role conflict and work–family conflict), thus triggering the health impairment process. In contrast, job resources are defined as job characteristics that provide possibilities for achievements and support psychological needs (e.g., social support, feedback, and variation), thereby triggering the motivational process (Bakker & Demerouti, 2014). In addition, the JD-R model proposes that job demands and job resources can have an interactive effect (e.g., resources can buffer against the negative effect of demands, and demands can boost the positive effect of resources) on motivational and health-related outcomes (Bakker, Demerouti, & Euwema, 2005; Schaufeli & Taris, 2014). The JD-R model is an individual-level approach, as applied in the present study, but can also be applied to higher-level aggregations (e.g., work units, teams, or organizations: Schaufeli & Taris, 2014).

A vast amount of research has convincingly demonstrated that high levels of job demands and low/insufficient levels of job resources have negative effects on employees’ health, such as stress and burnout. In contrast, low/reasonable levels of job demands and high/sufficient levels of job resources have been found to have positive effects on outcomes such as engagement, job satisfaction, and organizational commitment (for overviews of JD-R research, see Bakker & Demerouti, 2014, 2017, 2018). For instance, job demands such as role conflict and work–family conflict has been related to unfavorable job attitudes, lower well-being, and ill health (Hall, Dollard, Tuckey, Winefield, & Thompson, 2010; Kim & Stoner, 2008;). In contrast, research has shown that job resources like social support and social community at work relates to positive job attitudes, higher levels of well-being, and better health (e.g., Bakker, Demerouti, & Sanz-Vergel, 2014; Bliese, Edwards, & Sonnentag, 2017). The JD-R model is a heuristic framework, which makes it applicable to different occupations. Because of its flexibility, the JD-R model is of practical importance for assessments that seek to generate insights to inform and guide the design of interventions in occupational health management and human resource (HR) management (Schaufeli & Taris, 2014). The antecedents for retention and turnover can differ between professional groups (Carmeli & Weisberg, 2006). For these reasons, though similar factors are included in other theoretical models such as the organizational social context (Glisson, Landsverk, Schoenwald, et al., 2008), the JD-R model is considered as a relevant theoretical framework for the current study as it advocates that job demands and resources should be specified for the occupational group being studied. Furthermore, in line with the positive organizational psychology approach (e.g., Bakker & Schaufeli, 2008; Di Fabio, 2017; Gable & Haidt, 2005), the present study focuses on the motivational process in the JD-R model.

Psychosocial safety climate

Psychosocial safety climate (PSC) is a construct that refers to organizational climate in terms of the extent the employees perceive that organizational policies, as well as management practices and procedures, prioritize and ensure the protection of their psychological health and safety (Dollard & Bakker, 2010; Law, Dollard, Tuckey, & Dormann, 2011). PSC can be integrated in the JD-R model, as the basic proposition is that PSC underpins levels of job demands and job resources (Afsharian, Zadow, Dollard, Dormann, & Ziaian, 2017; Bakker & Demerouti, 2018). Furthermore, attending to PSC is highly relevant because it complies with contemporary public policies for occupational health and safety (Leka, Jain, & Lerouge, 2017; Leka, Jain, Zwetsloot, & Cox, 2010; Potter et al., 2017; The Swedish Work Environment Authority, 2015). In essence, measures of work climate in an organization captures the perceptions of individuals (i.e., employees). When assessed and analysed on the individual level, measures of work climate are referred to as “psychological climate”—that is, how employees’ perception of the organizational climate has a psychological effect on health and well-being. However, employees can share their perceptions of the work climate. When this is the case, measures of work climate may be aggregated and analyzed on work unit or organizational level—and be referred to as “organizational climate” (Kuenzi & Schminke, 2009). Previous research has mainly approached PSC as a measure of organizational climate, but also in terms of psychological climate (Garrick et al., 2014). As the focus of the present study was to investigate how individual perceptions of different work environmental factors relate to important job attitudes indicative of social workers’ retention, perceptions of PSC were analyzed at the individual level (i.e., in terms of psychological climate).

Quality of work

Quality of work is central for employees and organizations. In HSOs, quality of work is closely related to the direct contact with clients or patients, which is even referred to as “the core of work” (Hasenfeld, 2010). For example, in health care organizations, quality of care has shown to be an important indicator of organizational performance (McHugh & Stimpfel, 2012), whereas turnover rate has a negative effect on quality of care (Castle & Engberg, 2005). Although it has been reported that perceptions of organizational efficiency seem to be important for how social workers’ evaluate service quality and that this in turn has an effect on social workers’ health and well-being, the direct effect of quality of work for social workers’ health and well-being has been largely neglected in previous research (Lee et al., 2013).

Social workers are generally intrinsically motivated and guided by a sense of altruism and idealism in their work (Smith & Shields, 2013). Yet, when confronted with an immense workload, social workers can either work faster, work more hours, or reduce the level of quality of the work performed. However, lowering the standards for quality of work has been reported to relate to health impairment and turnover among social workers (Astvik & Melin, 2012). Importantly, though social workers face different conditions in various welfare systems, reports suggest that moral concerns related to the quality of services has an effect on social workers’ retention in different European countries and the United States (Frost et al., 2017; Giffords, 2009). To our best knowledge, researchers have not yet addressed the importance of quality of work for social workers’ work engagement, job satisfaction, and organizational commitment. The present study contributes by investigating how perceived PSC and quality of work relate to work engagement, job satisfaction, and organizational commitment among social workers.

Indicators for retention: work engagement, job satisfaction, and organizational commitment

Research has demonstrated that work engagement, job satisfaction, and organizational commitment are important antecedents for retention (e.g., Cohen, 1992; Halbesleben & Wheeler, 2008; Kim & Kao, 2014; Lambert, Hogan, & Barton, 2001). However, though each concept is distinct and unique, they are intricately related. For instance, work engagement and job satisfaction can have a gain-cycle relationship (i.e., a positive reciprocal relationship over time; Guglielmi, Avanci, Chiesa, Marinari, Bruni, & Depolo, 2016), whereas, in turn, work engagement (Hakanen, Schaufeli, & Ahola, 2008) and job satisfaction (Srivastiva, 2013; Top, Tarcan, Tekingündüc, & Hekmet, 2013) have been reported to predict organizational commitment. Hence, individuals’ reasons for retention likely depend on an overall assessment that includes all aspects covered by the three concepts. Based on this consideration, and to gain a more complete understanding of factors that contribute to the retention of social workers, we chose to include work engagement, job satisfaction, and organizational commitment as indicators for retention in the present research.

Work engagement refers to a state where individuals’ experience high levels of energy and a strong identification with work and is defined and measured in terms of vigor, dedication, and absorption (Schaufeli & Bakker, 2004). Thus, it is not surprising that work engagement relates to employee retention (e.g., Halbesleben & Wheeler, 2008; Shuck, Reio, & Rocco, 2011). Overall, job resources are generally more important than job demands for the explanation of work engagement (Bakker, Albrecht & Leiter, 2011). In particular, social support from superiors and colleagues appear to be two key factors for enhancing work engagement (Bakker et al., 2014).

Global job satisfaction refers to an individuals’ overall evaluative attitude to the job and is distinguished from evaluations of satisfaction with specific aspects of the job (e.g., work tasks, pay), personal and work value correspondence (i.e., organizational commitment), and motivation (i.e., work engagement: see Judge, Weiss, Kammeyer-Mueller, & Hulin, 2017). Job satisfaction has been demonstrated to be important for retention (see, e.g., Freund, 2005; Hellman, 1997; Lambert et al., 2001), and to hold predictive validity for retention beyond the explanation given by organizational commitment (van der Heijden, van Dam, & Hasselhorn, 2009).

Organizational commitment refers to the extent to which individuals affectively identify with and are involved in the organization. Thus, organizational commitment is an affective state based on the perception of concordance between personal values and values distinctive for the employing organization (Jaskyte & Lee, 2009; Judge et al., 2017). Organizational commitment is especially dependent on job demands (e.g., role conflict) and organizational support, and considered as a strong predictor of retention (Cohen, 1992; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002).

Retention of social workers—the role of job demands, job-resources, and organizational factors

Professional tenure and public authority are two factors that seem to have a general impact on retention and turnover among social workers. Specifically, higher turnover rates have been reported among newly qualified social workers (Astvik & Melin, 2012; Barbee et al., 2009), and social workers practicing public authority (e.g., within the child and youth services and/or financial aid services: Coffey et al., 2004; Collins, 2008; SKL, Swedish Association of Local Authorities and Regions, 2015, 2016). Furthermore, previous research has explored how various work environmental factors relate to social workers retention and turnover. Next, we give a description of factors highlighted in this research.

Social work is a demanding occupation. However, it has been reported that role conflict and work–family conflict are two types of job demands that seems to be particularly significant for social workers’ retention and turnover (Jaskyte & Lee, 2009; Johnco, Salloum, Olson, & Edwards, 2014; Kim & Stoner, 2008). For example, role conflict and work–family conflict has been reported to predict social workers emotional exhaustion and work withdrawal (Travis et al., 2016) as well as job satisfaction (Lizano & Mor Barak, 2015).

As social work inevitably places high demands on employees, the importance of adequate job resources is vital. With regard to the retention of social workers, previous research notes the importance of job resources in terms of a supportive environment in general (i.e., social community at work) and social support from superiors (Frost et al., 2017; Johnco et al., 2014; McFadden, Campbell, & Taylor, 2014).

Social service departments are subjected to high organizational constraints compared to other occupational sectors (Coffey et al., 2004). Indeed, how social service organizations are perceived by the employed social workers has been demonstrated to have an effect on retention (Frost et al., 2017; Jaskyte & Lee, 2009; Lee et al., 2013; McFadden et al., 2014). Yet ultimately, work engagement, job satisfaction, and organizational commitment—as well as the decision to remain or to leave the job—is an individual issue for each social worker.

The present study

The aim of the present study was to introduce and investigate how two novel work environmental factors: PSC (Dollard & Bakker, 2010) and quality of work (Berthelsen et al., 2017; Schaufeli & Taris, 2014) relate to important indicators of retention among social workers: work engagement, job satisfaction, and organizational commitment. Building on the insights of previous research, the present study included predictor variables (i.e., job demands and job resources) reported to be of specific importance for social workers’ retention—to investigate if, and to what extent, PSC and quality of work add unique variance explained for the three outcome variables. Based on this rationale, specific hypotheses were stated for all predictor variables, with respect to the expected relations when accounting for the contribution of all predictor variables (i.e., the full regression model).

The following hypotheses were stated for the present study. First, PSC is suggested to reflect specific aspects of the psychological climate that underpins levels of job demands and job resources. Thus, we expected (Hypothesis 1) that PSC would be positively related to the three outcome measures work engagement, job satisfaction, and organizational commitment.

Second, as the job demands, role conflict and work-family conflict have been reported to relate to turnover among social workers, we expected (Hypothesis 2) role conflict and work–family conflict to be negatively related to work engagement, job satisfaction, and organizational commitment.

Third, job resources have been reported to be important for positive outcomes, and a supportive environment and experiencing support from managers are factors of specific relevance for social workers’ turnover and retention. Hence, we expected (Hypothesis 3) two aspects of social relations at work, namely, social community at work and social support from superiors, to be positively related to social workers work engagement, job satisfaction and organizational commitment.

Finally, previous research indicates that conditions that enable high-quality work are important for social workers’ health and well-being. Thus, we expected (Hypothesis 4) perceived quality of work to be positively related to work engagement, job satisfaction, and organizational commitment.

Methods

Participants and procedure

The data for the present study were collected as part of a workplace survey, conducted within the municipal social services in a larger Swedish city. An active dialogue between the researchers, representatives of the municipality (i.e., managers and the HR department), and representatives of the employees (i.e., labor unions) preceded the data collection. The municipality was provided with a link to a web survey, and invitations to participate in study were sent out by emails to all 1,044 social workers, by the municipality. The web survey took approximately 20 minutes to complete. In total, 831 social workers participated in the survey (80% participation rate). As the focus of the present study was on social workers working directly with clients, participants reporting a management position (n = 96), or who had not answered this question (n = 10), were excluded. Thus, the final sample consisted of 725 social workers (women = 85%, professional tenure: ˂1 year = 20%; 1–2 years = 20%; 2–5 years = 23%; 5–7 years = 8%; ˃7 years = 29%, practicing public authority = 76%). The present study was approved by the Regional Ethical Review Board, Lund Secretariat (dnr: 2015–476), and informed consent was obtained from all participants.

Materials and measures

The present study collected data primarily by use of the validated Swedish medium-length version of the Copenhagen Psychosocial Questionnaire (COPSOQ II: Berthelsen, Hakanen, & Westerlund, 2018; Berthelsen, Westerlund, & Kristensen, 2014; Pejtersen, Kristensen, Borg, & Bjorner, 2010). COPSOQ II includes a number of scales pertaining to employees’ perceptions and experiences of their work conditions and health. COPSOQ II is a generic and comprehensive instrument, which has been translated into more than 20 languages. Research from numerous countries has contributed to the validation of its use in research as well as for workplace assessments (Kristensen, 2010; Nübling, Burr, Moncada, & Kristensen, 2014). In addition, the data collection included scales concerning work engagement (Schaufeli, Bakker, & Salanova, 2006), quality of work (Berthelsen et al., 2017; COPSOQ International Network, 2018), and PSC (Dollard & Bakker, 2010).

Predictors

Psychosocial safety climate

The present study used the short PSC version (PSC-4: Berthelsen & Muhonen, 2017; Dollard & Bakker, 2010), which includes four of the 12 items from the full-length version. Participants rated each item on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). An item example is, “Senior management show support for stress prevention through involvement and commitment.”

Role conflicts

Role conflict (RC) was measured by three items from the COPSOQ II (Berthelsen et al., 2014; Pejtersen et al., 2010). The items were rated on five-point scales from 1 (to a very low degree) to 5 (to a very high degree). The scores are then converted to the scale 0 to 100 (i.e., 0, 25, 50, 75, 100). An item example is, “Are contradictory demands placed upon you at work?”

Work–family conflict

Work–family conflict (WFC) was assessed by three items from the COPSOQ II. The items were rated on four-point scales from 0 (no, not at all) to 3 (yes, definitely). Scores are then converted to the scale 0 to 100 (i.e., 0, 33, 67, 100). An item example is, “Do you feel that your work drains so much of your energy that it has a negative effect on your private life?”

Social support from superiors

Three items of the COPSOQ II measured social support from superiors (SSS) The items were rated on 5-point scales, ranging from 0 (never) to 4 (always). Scores are then converted to the scale 0 to 100 (i.e., 0, 25, 50, 75, 100). An item example is, “How often do you get help and support from your immediate superior, if needed?”

Social community at work

Social community at work (SCW) was measured by three items from the COPSOQ II. The items were rated on 5-point scales from 0 (never) to 4 (always). The scores are then converted to the scale 0 to 100 (i.e., 0, 25, 50, 75, 100). An item example is, “Do you feel part of a community at your place of work?”

Quality of work

Quality of work (QW) was assessed by two items, one developed and tested as part of the Swedish validation study of COPSOQ II (Berthelsen et al., 2017), and one developed by the International COPSOQ Network for inclusion in COPSOQ III (COPSOQ International Network, 2018). The items were rated on 5-point scales from 1 (to a very low degree) to 5 (to a very high degree). The scores are then converted to the scale 0 to 100 (i.e., 0, 25, 50, 75, 100). The two items were, “Are you satisfied with the level of quality in the services that is conducted at your work-place?” and “To what extent do you find it possible to perform your work tasks at a satisfactory quality?”

Dependent variables

Utrecht Work Engagement Scale

The Utrecht Work Engagement Scale (UWES) measures three dimensions of work engagement: vigor, dedication, and absorption (Schaufeli & Bakker, 2004). The present study used the short version with nine items (Schaufeli et al., 2006), rated on 7-point scales ranging from 0 (never) to 6 (always). An item example is, “I am enthusiastic about my work.”

Job satisfaction

Job satisfaction (JS) was measured by four items from the COPSOQ II. The items were rated on 5-point scales scored from 1 (very unsatisfied) to 5 (very satisfied). Scores are then converted to 0 to 100 (i.e., 0, 25, 50, 75, 100). An item example is, “Regarding your work in general, how pleased are you with your job as a whole, everything taken into consideration?”

Organizational commitment

Organizational commitment (OC) was measured by three items from the COPSOQ II. The items were rated on 5-point scales from 1 (to a very low degree) to 5 (to a very high degree). Scores are then converted to 0 to 100 (i.e., 0, 25, 50, 75, 100). An item example is, “Do you feel that your place of work is of great personal importance for you?”

Results

Data analyses

Data were analyzed with SPSS (ver. 24). To test the hypotheses, descriptive statistical analyses, correlational analyses, and hierarchical multiple regression analyses were conducted.

Descriptive statistics and correlations

Means, Standard Deviations, reliability scores (Cronbach’s alpha), and distributions are reported in Table 1. Cronbach’s alpha scores were satisfactory for all scales (Nunnaly, 1978).

Table 1. Descriptive statistics

	Mean (SD)	Cronbach’s alpha	Observed range	Skewness	Kurtosis
Psychosocial safety climate	7.7 (3.3)	.88	0–16	−.138	.117
Psychosocial safety climate	71.3 (19.4)	.83	0–100	−.885	.701
Social community at work	80.8 (14.0)	.80	16.67–100	−.570	.860
Role conflict	45.6 (17.1)	.68	0–100	−.067	.195
Work-family conflict	44.2 (28.2)	.82	0–100	.243	−.922
Quality of work	62.0 (18.3)	.76	0–100	−.262	.364
Work engagement	4.0 (0.9)	.93	0–6	−.527	.683
Job satisfaction	64.2 (16.6)	.77	0–100	−.353	.371
Organizational commitment	65.7 (18.9)	.67	8.33–100	−.335	−.136

Relationships between all scales and measures were analyzed by Pearson’s correlation (Table 2). Significant correlations were observed between all scales. Job demands (i.e., role conflict and work–family conflict) were negatively related to work engagement, job satisfaction, and organizational commitment. In contrast, PSC, job resources (i.e., social support from superiors and social community at work), and quality of work were positively related to work engagement, job satisfaction, and organizational commitment. Finally, correlations for the three outcome measures showed strong relations between work engagement, job satisfaction, and organizational commitment.

Table 2. Correlations

	1.	2.	3.	4.	5.	6.	7.	8.	9.
1. Psychosocial safety climate	—
2. Social support from superiors	.420	—
3. Social community at work	.204	.299	—
4. Role conflict	−.354	−.246	−.147	—
5. Work-family conflict	−.296	−.250	−.137	.363	—
6. Quality of work	.388	.346	.267	−.486	−.365	—
7. Work engagement	.316	.330	.264	−.383	−.312	.511	—
8. Job satisfaction	.412	.428	.267	−.429	−.389	.578	.562	—
9. Organizational commitment	.398	.481	.396	−.401	−.385	.601	.605	.706	—

Note. All correlations are significant at p ˂ .001.

Hierarchical multiple regression analyses

The data were screened for violations of assumptions. Scatterplots and normal P-plots for standardized residuals confirmed the assumptions of linearity, normality, and homogeneity of variance. Durbin-Watson statistics were acceptable (1.91–1.97), and variance inflation factors (VIFs) were less than 10 (1.59–1.61), indicating that the assumptions for independence of errors was met and that multicollinearity was not an issue.

Hierarchical regression analyses were performed (Table 3) to test the hypotheses and investigate the relationships between work characteristics and work environmental factors to the respective outcomes measures. Separate regression analyses were performed for each of the three outcomes: work engagement, job satisfaction, and organizational commitment. The regression analyses were conducted by entering the independent variables in five distinct steps. Step 1 controlled for the potential confounding effect of the background factors professional tenure (i.e., time in the profession) and public authority (dummy-coded as 1/0: 1 = practicing public authority). In Step 2, PSC was added to the model. Step 3 added job demands in terms of role conflict and work–family conflict, whereas Step 4 added the job resources social community at work and social support from superiors. Finally, in Step 5, quality of work was added.

Table 3. Hierarchical Multiple Regression Analyses

	Work engagement				Job satisfaction				Organizational commitment
	β	∆R^2	F Change	Adj R^2	β	∆R^2	F Change	Adj R^2	β	∆R^2	F Change	Adj R^2
Step 1.		.05	19.09***			.02	7.96***			.04	13.51***
Professional tenure	.04				−.04				−.07*
Public authority	−.08*				.02				−.03
Step 2.		.09	65.91***			.16	134.86***			.15	123.40***
Psychosocial safety climate	.04				.11***				.06
Step 3.		.08	32.86***			.12	58.73***			.11	52.32***
Role conflict	−.10**				−.13***				−.07*
Work-family conflict	−.08*				−.13***				−.12***
Step 4.		.05	24.70***			.05	29.30***			.13	79.95***
Social support superiors	.14***				.17***				.22***
Social community at work	.10**				.05				.18***
Step 5.		.06	62.09***	.33		.08	101.69***	.43		.09	125.39***	.51
Quality of work	.32***				.36***				.38***

Note. Standardized beta weights are from the fifth step of the regression analyses.

*p ˂ .05, **p ˂ .01, ***p ˂ .001

Work engagement

For work engagement, the full model (see adjusted R²) provided a quite substantial amount of variance explained (33%). When added to the model, each step provided a unique and significant contribution of variance explained (see: ∆R², F change). In the full model, all predictor variables—except for PSC—provided a significant contribution (see β) for social workers’ work engagement in the expected directions. Thus, the results provided support for Hypotheses 2, 3, and 4, but no support for Hypothesis 1.

Job satisfaction

With regard to job satisfaction, the model provided a considerable amount of variance explained (43%). Again, when added to the model, all steps provided a significant contribution of variance explained. In the full model, all predictor variables, except for social community at work, provided a significant contribution to the explanation of social workers’ job satisfaction. Hence, the results supported Hypotheses 1, 2, and 4 but provided only partial support for Hypothesis 3.

Organizational commitment

The regression analyses for organizational commitment explained a substantial amount of the variance (51%). Each of the hierarchical regression steps provided a significant contribution of variance explained. The full model demonstrated that all predictor variables significantly contributed to the explanation of social workers organizational commitment, with the exception of PSC, which was not found to hold significance in the full model. Thus, the results of the regression analysis for organizational commitment provided support for Hypotheses 2, 3, and 4 but only partial support for Hypothesis 1.

General discussion

Considering the high turnover rates among social workers in the public social services, the aim of the present study was to improve the understanding of factors contributing to the retention of social workers, by investigating the relationship between specific work environmental factors and positive work attitudes. Specifically, the present study focused on factors contributing to positive work attitudes among social workers, because such attitudes relate to retention (Halbesleben & Wheeler, 2008; Kim & Kao, 2014; Lambert et al., 2001; Meyer et al., 2002). Furthermore, in this investigation, we explored the importance of two novel work environmental factors for social workers work engagement, job satisfaction, and organizational commitment: PSC and quality of work.

Overall, the results provided partial support for our Hypothesis 1, stating PSC to be positively related to the three dependent variables work engagement, job satisfaction, and organizational commitment. In the full regression models, PSC did provide a significant contribution to the explanation of social workers’ job satisfaction—but not to the explanation of work engagement or organizational commitment. This result is somewhat surprising as previous research reports a positive relation between PSC and work engagement (Afsharian et al., 2017; Dollard & Bakker, 2010). Yet previous research on the relation between PSC and work engagement has investigated the effect of PSC on an aggregated group level (Dollard & Bakker, 2010), whereas the present study considers effects of PSC at the individual level. However, it may be discussed whether the effect of PSC is best understood as an antecedent for job resources and job demands (Dollard & Bakker, 2010), or in terms of a moderator in relation to job demands–resources and outcomes (Garrick et al., 2014). Our results show that PSC is important for social workers’ job satisfaction, but the importance of PSC for organizational commitment did not hold significance in the full regression model. This difference in the result is noteworthy because job satisfaction and organizational commitment are highly related concepts (e.g., r = .71, in the present study). Yet the observed differences show that the two concepts are distinct and unique (e.g., van der Heijden et al., 2009). One interpretation of our results is that they demonstrate that PSC in terms of psychological climate is important for job satisfaction, whereas PSC in terms of organizational climate is important for engagement (Afsharian et al., 2017).

The results of the present study corroborate the importance of specific job demands (i.e., role conflict and work–family conflict) and job resources (i.e., social community at work and social support from superiors) for social workers’ retention, reported by previous research (e.g., Frost et al., 2017; Johnco et al., 2014; Lizano & Mor Barak, 2015; Travis et al., 2016). Overall, confirming Hypotheses 2, our results demonstrate that job demands in terms of role conflict and work–family conflict have a negative effect on social workers’ work attitudes. The two job demands were most strongly related to social workers’ job satisfaction but also related to work engagement and organizational commitment. Moreover, in support of Hypothesis 3, the results showed that job resources in terms of social support from superiors and social community at work were positively related to work attitudes among social workers. This is line with the reports by previous research and underlines the importance of supportive management and collegial relations for social workers’ retention (Frost et al., 2017). In fact, the results show that support from superiors and social community at work were most strongly related to social workers organizational commitment, compared to work engagement and job satisfaction. Overall, social support from superiors was the strongest job resource predictor.

Finally, the present study provides new insights to the retention of social workers, by demonstrating the importance of quality of work for positive work attitudes among social workers. Confirming Hypothesis 4, perceived quality of work was found to be a strong predictor for all three positive work attitudes. These results are intriguing and accentuated by the fact that quality of work was included in the final step of the regression models. Our results confirm the notion reported by previous research, suggesting that quality of work is a factor that seems to be important for the retention of social workers (Astvik & Melin, 2012; Lee et al., 2013). Specifically, our findings suggest that when conditions for conducting high-quality work exist, this has a strong positive effect on social workers’ work engagement, job satisfaction, and organizational commitment. One possible interpretation of the observed importance for quality of work is that it reflects the extent that social workers perceive that it is possible to provide services that are in line with their own personal motives and convictions for the job (e.g., Smith & Shields, 2013). More specifically, if social workers perceive that prerequisites for high-quality work exist, this might be a key precursor for engagement, satisfaction, and commitment. Conversely, when conditions for quality of work are not present, this can have a deteriorating effect on the motivational process (Bakker & Demerouti, 2017).

Conclusions

In conclusion, the main finding was a consistent positive relationship between the quality of the work performed and the three important predictors for retention of social workers: work engagement, job satisfaction, and organizational commitment. Surprisingly we found that, in our model, PSC was only a significant predictor for job satisfaction.

Implications

The present study has some implications for practice as well as for future research.

Most importantly, our findings indicate that being able to conduct work of good quality is essential for social workers work engagement, job satisfaction and organizational commitment, which, in turn, are important for retention of staff. Therefore, addressing quality of work at different organizational levels is of high relevance for HR practices within the public social services. At a strategic and tactical level, it is relevant to ensure working conditions that enable social workers to provide good quality services for their clients. To do that managers need to analyze which factors that promote or prevent social workers to conduct work of good quality. Examples of factors to consider could be adequate staffing, relevant labor division among occupational groups, and providing opportunities for development and learning. For work groups, discussions aiming at a mutual understanding of what level of quality is realistic and satisfactory can contribute to increased role clarity and priority setting for the individual social worker. Besides, initiating such discussions of the core of work may also contribute to improved social support from superiors and work-related community in the work group, which are important job resources for health and well-being at the individual level, as shown in the present study. Furthermore, as turnover has a negative effect on quality (Castle & Engberg, 2005), it is important that HR practices in HSOs consider the association between the organizational context (e.g., climate) of the human service system, the nature (e.g., quality) of the services provided, and the well-being of the workers (Hemmelgarn, Glisson, & James, 2010; Lee et al., 2013). In this regard, we note that, even though the essence of quality in human services can be expected to be quite specific, HR practices in HSOs may adhere to recommendations from research on service quality at large—because the basic precursors for service quality seem to be rather general (see e.g., Aryee, Walumbwa, Seidu, & Otaye, 2016). Moreover, attending to workers’ evaluations of quality of work can inform and guide HR practices in HSOs, as an indication of actual service quality that corresponds to evaluations of quality by service recipients as well as to objective measures of quality in terms of both processes and outcomes (McHugh & Stimpfel, 2012).

However, it will be an important task for future research to investigate the complex interplay between working conditions and quality of work in social services further, by applying longitudinal and multilevel designs as well as register-based outcomes on quality of work.

Our expectations regarding the importance of PSC were only partly corroborated and to some extent in contrast to previous research. This calls for further studies to clarify the role of PSC. However, the findings of a positive relationship between PSC and job satisfaction support the relevance of managerial attention to develop, maintain, and actively communicate policies and practices to ensure that employees are confident that their psychological health is a valued and prioritized concern.

Finally, further research needs to investigate the potential effect of PSC and quality of work for actual retention and staff turnover.

Limitations and strengths

The present study has a number of limitations. One is that the data are cross-sectional, which prohibits causal claims about the directions of relationships. Future research should investigate causality by the use of a longitudinal design. Another limitation is that the study relies on self-reports with the potential risk of common method bias. Still, the correlation matrix (Table 2) makes the potential for common method error transparent. In addition, the different results for the three outcomes indicate a limited risk of common method bias.

We consider the use of validated instruments (PSC and COPSOQ II) as an advantage, whereas the operationalization of quality of work by only two items might be a potential disadvantage. However, the two items demonstrated a satisfactory reliability (α = .76), and an advantage is that they are generic, making future comparisons with other occupational groups possible. Still, we acknowledge the difficulties of defining and measuring quality and encourage future research to address this issue. One possibility may be to use a multifaceted approach and include both subjective (i.e., self-reports) and objective (e.g., normative criteria or client-evaluations) indicators in order to get a more complete picture for the quality concept (see, e.g., Geisler & Allwood, 2015).

The close relations to clients is suggested to be a fundamental characteristic, and even a defining feature of “the core,” of human service work (Hasenfeld, 2010). Based on this, we believe that the findings regarding quality of work can be generalized to social work in general, whereas the findings regarding PSC are expected to be more contextually dependent.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by AFA Insurance [130301; 170262]; the Swedish Research Council for Health, Working Life and Welfare [2016-07220]; Malmö University.