Introduction

In recent years, paid employment has become more important in organizing third-sector activities (e.g., Wijkström 2011). These activities have been assumed to attract especially ideologically oriented employees who share the employer organization’s values and goals (e.g., Kim and Lee 2007; Light 2002). Some have even suggested that nonprofit employees are willing to sacrifice their own well-being for the sake of others (Mann 2006; Perry and Wise 1990). Thus, this kind of motivation can be expected to produce energetic and enthusiastic employees who do not give up even in the face of challenges. They can be expected to experience high work engagement.

Employees with high work engagement are energetic and determined in their work. They are proud of and dedicated to work, which they see as rewarding, and are sometimes so immersed in it that they can lose track of time (Schaufeli et al. 2002). In empirical studies, this kind of attitude to work is usually explained by job demands and resources. Job demands such as time demands, the unpredictability of work, and conflicts between work and family have been shown to decrease work engagement. On the other hand, job resources such as control over work and social support from supervisors and colleagues have been found to increase work engagement (Bosman et al. 2005; Hakanen et al. 2008; Huynh et al. 2012; Mauno et al. 2007).

Previous studies, however, have not been fully able to capture third-sector employees’ work engagement. Instead, they have concentrated on volunteers (Huynh et al. 2012) rather than paid employees, and have disregarded public service motivation theory which suggests that third-sector employees have a special ideological orientation to work (see e.g., Perry and Hondeghem 2008). This orientation makes work more meaningful for employees (Mann 2006; Perry and Wise 1990), and therefore should result in high work engagement. This article bridges the research gap by analyzing third-sector paid employees and by extending the previous work engagement models by examining ideological orientation to work. The literature section provides a more detailed discussion of the work engagement in the third sector, conceptual model, and related hypothesis. This is followed by the empirical section which answers two research questions. First, it determines whether third-sector employees report higher work engagement than employees generally in the work engagement studies. Second, it analyzes antecedents of the work engagement. The results show that third-sector employees experience higher work engagement than employees generally in the work engagement studies. According to the analysis, this is associated job demands, and resources and their ideological orientation to work.

Work Engagement and Third Sector

The concept of work engagement represents a relatively new research tradition called positive psychology. The aim of positive psychology was to encourage focus on people’s strengths and find out how everyday life could be made more fulfilling instead of concentrating on mental illnesses and inabilities (Seligman and Csikszentmihalyi 2000). Work engagement was developed as a positive, yet distinct from for the concept of burnout (Demerouti et al. 2001). Contrary to burnout, which is characterized by exhaustion, cynicism, and inefficiency (Maslach et al. 2001), work engagement was seen as a positive factor related to a state of mind which consists of vigor, dedication, and absorption (Schaufeli et al. 2002, p. 74). Vigor is defined as an energetic and persistent attitude to work even in the face of difficulties. Dedication is characterized by strong involvement in work, a sense of significance, enthusiasm, inspiration, pride, and challenge. The last of the three dimensions, absorption, is characterized by total concentration in work, the employee perceiving time as passing quickly and finding it difficult to detach him/herself from work (Ibid.).

This kind of energetic and enthusiastic attitude to work is presumably prevalent in third-sector organizations. Third-sector organizations are mission-driven entities that do not aim to maximize profits. This approach permeates their organizational culture and identity and thus has been suggested to serve as a selector to attract employees who have an ideological orientation to work (Mirvis and Hackett 1983; Rose-Ackerman 1996). This includes, for example, the assumption that employees in the third sector share their employer organizations’ ideas and vision (Rose-Ackerman 1996). Since third-sector organizations enable employees to promote goals and values that they share with their employer organizations, employees are more likely to consider their work to be meaningful. In the empirical studies on the subject, third-sector employees have been shown to be more satisfied and committed to their work in comparison to employees in either private or the public sector (e.g., Borzaga and Tortia 2006; Mirvis and Hackett 1983).

However, not only do third-sector organizations select ideologically oriented employees, but they also use managerial tools to encourage employees’ involvement in the organization’s objectives (Brown and Yoshioka 2003; Mirvis and Hackett 1983). According to previous research, these tools include, for example, a democratic governance structure that gives employees more opportunities to express themselves freely on the job (e.g., Mirvis and Hackett 1983). In the relevant empirical studies, third-sector work has been shown to offer more variety, challenge, and job autonomy in comparison to work in either the public or the private sector. On the other hand, as a downside of challenging and autonomous work seems to be that employees do not receive as much direct feedback regarding their job performance compared to employees in the other sectors (Borzaga and Tortia 2006; McMullen and Schellenberg 2003; Mirvis and Hackett 1983; Ruuskanen et al. 2012.). Furthermore, owing to project funding, there is always an on-going battle of funding during the same time that the client’s needs have to be met. This is potential source of work insecurity and high time demands (Cunningham 2001; McMullen and Schellenberg 2003).

Thus, the third-sector operating environment includes two potential sources for high work engagement: employees’ ideological orientation to work, which makes work meaningful, and managerial tools that are designed to increase employees’ level of involvement in the work. On the other hand, this working environment has the potential to limit work engagement, if time demands and employment insecurity prevent employees from performing their work to their desired standard.

Previous studies, however, do not provide a clear answer as to whether the third-sector operating environment encourages high or low work engagement. Schaufeli et al. (2006) found that white-collar workers in the nonprofit sector report less work engagement than managers, police officers, or educators. From their comparison, however, it is not apparent whether they understand nonprofit as public or third-sector employees. Therefore, more reliable assessment of the work engagements level have to be looked for from other motivational constructs which are closely related to work engagement. These include “intrinsic motivation” which is closely related to vigor, “commitment” which has conceptual similarity with dedication, and “flow” which is closely related to absorption (Mauno et al. 2007). Of these intrinsic motivation and commitment are familiar from the previous third-sector research. It has been observed that third-sector employees show stronger intrinsic motivation and are more committed to their organization compared with public or private sector employees (e.g., Borzaga and Tortia 2006; De Cooman et al. 2011). Thus, based on the third-sector literature, it would seem that third-sector employees could report a higher work engagement on average than employees in the other sectors. The following hypothesis is formulated:

Hypothesis 1

Third-sector employees report higher work engagement than employees generally in the work engagement studies.

Toward Work Engagement’s Conceptual Model in the Third Sector

Regardless of whether third-sector employees report high or low work engagement, it is important to understand why. Therefore, the second aim of this article is to analyze work engagement’s antecedents. This will commence by presenting previous theoretical models of work engagement and showing how they disregard third-sector employees’ ideological orientation to work.

Most of the previous work engagement studies are based on psychological models that are designed to study employees’ well-being. Some of the best known models are job demand-control (JDC) Model (Karasek 1979), effort-reward imbalance (ERI) model (Siegrist 1996), and job demands-resources (JD-R) model (Demerouti et al. 2001; Bakker and Demerouti 2007). Karasek’s JD-R and Siegrist’s ERI models are so-called balance models. Karasek (1979) assumes that well-being is result of high job demands and high control over work, whereas Siegrist (1996) suggests that it is a result of a balance between (a) the efforts that employees put into their work and (b) potential rewards such as salary, esteem rewards or status control, which they receive as compensation. Critics against these models have focused particularly on their simplicity. The models concentrate on only a few job characteristics, thus disregarding many other aspects of working life that can contribute to employees’ well-being (e.g., Bakker and Demerouti 2007; De Jonge and Kompier 1997). Criticism is taken into account in the more recently developed JD-R model, which, for this reason, is more flexible in terms of job characteristics (Bakker and Demerouti 2007; Hakanen and Roodt 2010). In the JD-R model, job characteristics are divided into job demands and resources. Job demands refer to physical, psychological, social, and/or organizational aspects of the job that require physical and/or psychological effort from an employee (Demerouti et al. 2001). Examples are time and work pressure, role conflicts, and emotional demands. Job resources, in turn, refer to the physical, psychological, social, and/or organizational characteristics of the job which help the employee achieve goals, reduce job demands, or stimulate personal growth and development. Examples of resources are career opportunities, social support, role clarity, and job control. In the JD-R model, job demands are assumed to decrease work engagement, whereas job resources are assumed to increase it (Demerouti et al. 2001; Schaufeli and Bakker 2004).

The JD-R model can be used to analyze third-sector managerial practices such as job autonomy, but it is too limited to reveal the whole picture of third-sector work engagement. It does not take into account employees’ ideological orientation. Therefore, this article extends the JD-R model toward public service motivation theory. It suggests that third-sector employees identify with their job because they want to serve the public interest or some other philosophical goals and values which they share with their employer organization (Mann 2006; Perry and Vandenabeele 2008; Rainey and Steinbauer 1999). This makes work more meaningful and therefore it is assumed to lead to many positive outcomes (Perry and Wise 1990) such as work engagement.

Previously public service motivation theory has been used to explain, for example, high job performance, job satisfaction, and organizational commitment (e.g., Alonso and Lewis 2001; Brewer and Selden 2000; Crewson 1997; Naff and Crum 1999). Most of these studies, however, have concentrated on public servants with whom the concept was initially associated (see definitions in Perry and Wise 1990; Rainey 1982). Only recently has it been noted that third-sector employees also share similar concern with services that enrich communities and civil society (e.g., Mann 2006; Park and Word 2012). As a difference to the public sector, however, it seems to be that third-sector employees are not that interested in policy making (Word and Carpenter 2013) or servicing the public and public good, but instead their service motivation is more related to the local level (Lee and Wilkins 2011).

Despite these differences third-sector employees’ service motivation can be assumed to motivate employees to perform better and increase their involvement in organizations activities (Leiter 2008; Perry and Wise 1990). In this article, it is further assumed to increase work engagement. Following on from the discussion above, this article proposes an extended JD-R model to study third-sector employees’ work engagement (see Fig. 1). As in the JD-R model, job demands are assumed to decrease work engagement, whereas job resources are assumed to increase it (Halbesleben 2010). An entirely new element is employees’ ideological orientation to work, which includes public service motivation and value congruence. Ideological orientation is considered as distinct from job resources, since previously, Leiter (2008) argued that considering values as a job resource neglect its implications for involvement and efficacy. In the model, both dimensions of ideological orientation are presumed to increase work engagement. More detailed discussion of variables and hypotheses associated with the figures are presented in the following chapters.

Fig. 1
figure 1

Extended JD-R model with the ideological orientation component

Full size image

Job Demands and Resources in the Third Sector

The original part of the JD-R model covers job demands and resources (see Fig. 1). Of these job demands have been suggested to interfere with the employees’ perceived reciprocity and thus result in reduced work engagement (Bosman et al. 2005; Maslach et al. 2001). Specifically, in terms of job demands, this article concentrates on time demands, unpredictability of work, and employment insecurity. The first of these, time demands, resembles quantitative workload that is too much work in a limited amount of time (see, e.g., Karasek and Theorell 1990). Unpredictability of work, also called qualitative insecurity, resembles a threat of changes in important job features. Employment insecurity, also called quantitative insecurity, in turn, resembles threat of job loss. Previous studies from the third sector suggest that third-sector organizations may have difficulties in offering employees secure jobs with reasonable time demands, because of the discontinuity associated with their project funding (Cunningham 2001; McMullen and Schellenberg 2003). Higher time demands, unpredictability, and insecurity, compared to the public and private sectors, has been further observed among Finnish third-sector employees (see Ruuskanen et al. 2012). Since employees cannot predict upcoming work tasks or whether they have workplace at all, these will cause stress and decreased work engagement (see, e.g., Sverke et al. 2002). In the empirical studies high workload and employment insecurity have been observed to reduce work engagement (e.g., Bosman et al. 2005; Halbesleben 2010; Mauno et al. 2007). These studies, however, did not analyze the other dimension of insecurity. Since both insecurities can be considered as stressors (Sverke et al. 2002) and both are important threats in third-sector work (Ruuskanen et al. 2012), the unpredictability of work will also be assumed to reduce work engagement. Based on the discussion above the following hypothesis is formulated:

Hypothesis 2

Job demands decrease work engagement. Specifically, time demands, unpredictability of work, and employment insecurity are negatively associated with work engagement.

In terms of work engagement, the most important job characteristics, however, are related to available job resources (Halbesleben 2010). They help employees to achieve work goals and to stimulate personal growth and development. Furthermore, job resources may also reduce job demands (Schaufeli and Bakker 2004). Specifically, as regards job resources, this article focuses on job control and social support from supervisor and colleagues. First, job control refers to employees’ decision latitude and abilities to use their own skills at work (Karasek and Theorell 1990). Previous studies have observed that third-sector organizations offer their employees more autonomy and opportunities to express themselves freely on the job than are available in either public or private sector organizations (Borzaga and Tortia 2006; McMullen and Schellenberg 2003; Ruuskanen et al. 2012). Second, this article analyzes social support, which refers to various types of support such as assistance that employees receive from others within the organization. Previous studies suggest that third-sector employees do not receive as much social support from their supervisors, or direct feedback regarding their job performance, compared to employees in the public and private sectors. In turn, social support from colleagues does not differentiate the sectors (Mirvis and Hackett 1983; Ruuskanen et al. 2013). Both types of social support, however, are important for employees since encouragement and discussion of problems help employees to achieve their work goals and increase work engagement (e.g., Schaufeli and Bakker 2004). In the previous empirical studies job control and social support have been shown to increase paid (e.g., Mauno et al. 2007) and voluntary (Huynh et al. 2012) employees’ work engagement. Thus, based on the discussion above the following hypothesis is formulated.

Hypothesis 3

Job resource increases work engagement. Specifically, job control, social support from supervisor, and social support from colleagues increase work engagement.

Third-Sector Employees’ Ideological Orientation as an Extension of the JD-R Model

In addition to job demands and resources, this article includes third-sector employees’ ideological orientation as part of the JD-R model (see Fig. 1). Positive attitudes to work help employees to consider their work as more meaningful and thus motivate them to perform better and increase their involvement in organizational activities (Leiter 2008; Perry and Wise 1990). Specifically, ideological orientation includes public service motivation and value congruence.

Public service motivation refers to employee orientation to deliver services to people with the intention of doing good for others and society (Perry and Hondeghem 2008).This is reflected, for example, in Light’s (2002) observations that third-sector employees considered “accomplishing something worthwhile” as more important reason to take their current job than employees in the for-profit or public sector. This kind of attitude makes work more meaningful and can lead to many positive outcomes such as better job performance, higher job satisfaction, organizational commitment (Perry and Wise 1990), and work engagement.

Nevertheless, previous studies also suggest that implementation of public service motivation depends on employer organization (Moynihan and Pandey 2007; Taylor 2007). In order for public service motivation to increase positive outcomes, employees and an organization’s values have to fit together and employees must see that their values are being realized in everyday life (Bright 2008; Wright and Pandey 2008). Therefore, public service motivation has to be considered alongside value congruence. Value congruence refers to common values with the employer organization and, according to Leiter (2008), it motivates employees to pursue shared objectives and thus increase their work engagement. Based on the discussion above the following hypothesis is formulated:

Hypothesis 4

Employees’ ideological orientation to work increases work engagement. Specifically, public service motivation and value congruence increase work engagement.

Data and Methods

Data for this article were collected from Finland, which is one of the Nordic countries. Traditionally, in the Nordic countries, the share of paid labor has been relatively small because of the third-sector organizations’ small share in the welfare service production (Salamon et al. 2003). In recent years, however, the distinct roles of a strong welfare state and civil society have started to fragment. In particular, municipalities, which are the basic providers of welfare services in Finland, have increasingly turned welfare service production over to third-sector organizations (e.g., Julkunen 2000; Möttönen and Niemelä 2005). Increasing role in the welfare services has further increased the importance of paid employees in the Nordic countries (e.g., Wijkström 2011). For example, in Finland, full-time equivalent (FTE) employees have increased by 62 % in the third sector since 1995. In 2011, third-sector organizations employed approximately 77,000 FTE employees, which represented 5 % of total employment in the Finnish economy (Statistics of Finland 2012).

Previously, however, Finnish third-sector employees did not receive much attention in academic studies. One reason could be the difficulty of collecting a sample of employees. In Finland, third-sector paid work is divided between a number of organizations. According to the National board of Patents and Registration of Finland, there are approximately 133,000 associations (Register of Associations 2011) and 2,800 foundations (Register of Foundations 2011). Most of third-sector organizations are small or they do not have paid employees. Collecting a representative sample of third-sector employees is therefore difficult.

In the “Paid work in the third sector” project, employees were approached through their trade unions (Ruuskanen et al. 2012). Union density among employees is high in Finland (70 %) compared with other OECD countries (OECD 2012), which means that trade organizations’ membership registers are good resources for reaching third-sector employees. Four main trade unions were identified in expert interviews and selected for the study. These were: Akava Special Branches (Akavan erityisalat), the Trade Union for the Public and Welfare Sectors (JHL), the Federation of Special Service and Clerical Employees (ERTO), and the Union of Professional Social Workers (Talentia).

The survey was conducted in 2011 using a combination of Internet and postal surveys. The Internet survey was carried out among all trade unions. The trade unions sent a link to a web-based survey to all members who had registered their e-mail address and were employed by an association or foundation. In the registers of ERTO and JHL, there were more employees whose e-mail addresses were not available. These employees were considered to differ from the majority (e.g., lower level of education, older). Therefore, the survey was supplemented with a mail survey in these two unions. Respondents to the mail survey were chosen by systematic random sample among members whose e-mail addresses were not available.

The response rate for the whole dataset was 22 % (e-mail 21 %, mail survey 41 %) and the net sample was 1,412 (N). The response rate was quite low but not unusual in the last few decades (Anseel et al. 2010). In particular, response rates in Internet surveys have generally been lower than in mail surveys (Shih and Fan 2008). Possible reasons include the increased volume of unsolicited “junk” mail and obsolete e-mail addresses. Also, in this study, the majority of surveyed members did not respond to the e-mail message: only 30 % opened the link. Nevertheless, despite the low response rate and the decision to target the survey at the four trade organizations, the sample appeared to be fairly representative. In a comparison between Statistic Finland’s (2012) information on FTE employees and the numbers of employees in this dataset, the percentages of different fields were mostly consistent (social and health care: StatFi 44 % vs. this 47 %; activities of membership organizationFootnote 1: StatFi 15 % vs. this 16 %; culture, sport, and recreation: StatFi 11 % vs. this 14 %; business and professional associations: StatFi 8 % vs. this 7 %). Education and research was the only field under-represented in the dataset (StatFi 16 % vs. this 6 %).

Measures

Dependent Variable

To measure work engagement, the validated Finnish version of the shorter Utrecht Work Engagement Scale was used (Hakanen 2002; Schaufeli et al. 2002). The UWES instrument consists of nine items, which form three underlying dimensions. Vigor is described by three statements: “at my work, I feel bursting with energy”, “at my job, I feel strong and vigorous”, and “when I get up in the morning, I feel like going to work”. Dedication is described by the following statements: “I am enthusiastic about my job”, “my job inspires me”, and “I am proud of the work that I do”. The last of the three dimensions, absorption, is described by these statements: “I feel happy when I am working intensely”, “I am immersed in my work”, and “I get carried away when I’m working”. The response scale is rated on a 7-point frequency-based scale (1 = never to 7 = every day). The Cronbach alpha for the scale is 0.94 (M = 4.71, SD = 1.10).

Independent Variables

Several variables related to job demands and resources as well as employees’ ideological orientation to work were used as explanatory variables. Most of them were adopted from the Finnish Quality of Work Life Survey (QWLS) and the general Nordic questionnaire for psychological and social factors at work (QPSnordic). The questions were originally given in Finnish.

Job demands include three variables: time demands, unpredictability of work, and employment insecurity. Time demands were measured by three questions: “To what extent does your work involve tight schedules?”, “How often do you need to extend your working day in order to get all the work done?”, and “How often do you feel that you do not have enough time to do your work as well and as thoroughly as you would like to?” (response options from 1 = totally disagree to 4 = totally agree). The mean of these variables forms a sum scale (M = 2.69, SD = 0.71, Cronbach α = 0.75). The second factor of demands is the unpredictability of work, which was measured by two variables (response options: 0 = no, 1 = yes): the threat of “unforeseen changes” and “intolerable increase in workload”. Together, these form the unpredictability index, which can have values from 0 to 2 (M = 1.04, SD = 0.82, Cronbach α = 0.54). The third job-related demand is employment insecurity. It consists of three threats: “temporary dismissal”, “dismissal”, and “unemployment” (0 = no, 1 = yes). Together, these form the employment insecurity index, which can have values from 0 to 3 (M = 0.51, SD = 0.90, Cronbach α = 0.75).

Job control, social support from supervisors, and social support from colleagues were identified as job resources. The first, job control, consists of six questions. Respondents were asked whether they are able to “learn new things”, “apply their own ideas at work”, whether “their job requires creativity”, and are they able to influence “…the content of their tasks”, “…working methods”, or “…working hours” (response options from 1 = never to 4 = always). The mean of these scores forms a sum variable (M = 2.78, SD = 0.54, Cronbach α = 0.81). The second job resource is social support from supervisors. It consists of seven statements concerning the respondent’s immediate superior: my superior “supports and encourages me”, “rewards good work performance”, “discusses with us often”, “speaks openly about everything concerning the workplace”, “encourages his/her subordinates to study and develop in their work”, “gives sufficient feedback about how well I have succeeded in my work”, and “delegates responsibility sensibly to his or her subordinates” (response options from 1 = totally disagree to 5 = totally agree). The mean of these seven items forms a sum scale (M = 3.44, SD = 1.03, Cronbach α = 0.93). In addition to support from supervisors, social support from colleagues was also included as a job resource. Social support from colleagues was queried through two questions: “When your work seems difficult, do you receive support and encouragement from your co-workers?”, and “If needed, are your co-workers willing to listen to your work-related problems?” (response options from 1 = newer to 4 = always). Together, these form a mean sum variable (M = 3.08, SD = 0.76, Cronbach α = 0.88).

Employees’ ideological orientation to work was analyzed based on public service motivation and value congruence. Public service motivation has been most often measured with James Perry’s scale (1996). Because of data limitations, however, the present study was not able to use this scale. We recognize that the use of some other scale is not ideal because it undermines the comparability of results with previous studies. The scale used in this study, however, mostly resembles Perry’s subscale commitment to public interest and thus gives insight into the relationship between work engagement and third-sector employees’ desire to serve the community. It consisted of three questions: How important do you consider that your work “is socially beneficial?”, “gives you a sense of accomplishing something worthwhile?”, and “enables you to help and take care of other people?” (1 = not important, 5 = very important). The mean of these items forms a sum scale (M = 4.24, SD = 0.68, Cronbach α = 0.70). Second, the study examined value congruence through one statement: “I feel that my values and the organization’s values are very similar” (1 = totally disagree, 4 = totally agree, M = 3.83, SD = 0.99).

Background Variables

Several background variables were used in addition to the variables described above. These include gender, age, occupational class, field of employer activity, workplace size, and employment contract. Information of demographic characteristics of the sample is presented in Table 1. It shows that the third-sector workforce is female-dominated and concentrated on expert positions. Most of employees are working in small, social, and health care organizations with permanent work contract.

Table 1 Characteristics of third-sector workforce and mean of work engagements in each group
Full size table

Statistical Methods

The article uses mean comparisons and variance analysis to reveal whether third-sector employees report higher work engagement than employees generally in the work engagement studies. These will answer hypothesis 1. This is followed by hierarchical multiple regression analysis, which examines the connections between background variables, job demands, job resources, and employees’ ideological orientation to work for work engagement. Thus, it will provide answers to the rest of the hypotheses (2–4). Hierarchical regression analysis was also performed by using natural logarithmic transformation of the work engagement scale. This was done because work engagement scale is skew (−1.33). The results, however, remained the same, which authorizes the use of the original scale. These, therefore, will be reported.

Level of Work Engagement

Mean of work engagement in the Finnish third sector is 4.71 (95 % confidence interval for mean 4.65–4.77, N = 1,253). This illustrates high work engagement if compared with average value in the Finnish UWES-9 manual (M = 4.26, 95 % confidence interval for mean 4.24–4.28, N = 16,335).Footnote 2 For reference, the UWES manual was chosen, since in Finland there is no representative national sample of work engagement. Thus, the UWES manual, that has a combined number of datasets from different professional groups, provides the best approximate of the work engagement level in Finland.

Based on this comparison, third-sector employees seem to experience higher levels of work engagement than Finnish employees in the work engagement studies generally. Some of these differences, however, can be related to third sector special labor market structure. Based on previous study by Ruuskanen et al. (2012), the third-sector workforce is female dominated (85 %) compared with public (71 %) and private (44 %) sectors. On average, they are slightly older (M = 45) than employees in the private sector (M = 40). There are more employees working in expert or managerial positions (70 %) than in the public (63 %) or private (45 %) sectors. Furthermore, based on international classification of nonprofit organizations (ICNPO, Salamon and Anheier 1996), the majority of Finnish third-sector employees are concentrated in social and health care (47 %), followed by culture, sport, and recreation (14 %) and development and housing (10 %) (see Table 1). Most of employees work in a small workplace with less than 10 employees (40 %), whereas in the public (20 %) or private (26 %) sector small workplaces are not that common. Furthermore, in the third sector there are more employees who are working on a fixed term contract (21 %) than in the public (19 %) or private (8 %) sector (See Ruuskanen et al. 2012).

Because the UWES manual (Hakanen 2009) does not provide mean and standard deviations for each of these groups, the level of work engagement was compared inside the third sector. The results of the variance analysis show that the third-sector workforce is quite homogenous in terms of work engagement (see Table 1). Level of work engagement is not related to gender, field of employment activity, size of the workplace or type of employment contract. Age, however, is weakly correlated with work engagement (r = 0.059, p = 0.042). Older employees are more engaged in their work than younger ones. This connection, however, is so weak that it does not have any practical significance. The only variable which has strong connection to work engagement is occupational class (see Table 1). As the occupational class rises, employees’ experience of work engagement also increases. This is consistent with previous studies which observed that executives report higher work engagement than blue-collar workers (e.g., Mauno et al. 2007; Schaufeli et al. 2006). Thus, it can partly explain mean differences observed between third-sector employees and mean value of the Finnish UWES-9 instrument.

Results from the third sector confirm hypothesis 1. Third-sector employees report higher work engagement than employees on average in the work engagement studies. This is not related to third-sector female domination, bigger share of small workplaces or concentration in social and health care organizations. Nevertheless, it can be partly related to domination of expert and managerial positions in the third-sector workforce. In order to have more information as to why third-sector employees report high work engagement, hierarchical multiple regression analysis is next presented.

Work Engagement and Its Antecedents in the Third Sector

Table 2 presents results of a hierarchical multiple regression analysis. This statistical method was chosen for two reasons. First, it allows determining how much added variables increase a model’s predictive power. Thus, it will answer whether employees’ ideological orientation has predictive power beyond job demands and resources. Second, multiple regression analysis can evaluate simultaneous contributions of multiple predictors to an observed response. This eliminates the effect of the third variable in the analysis. In the first model background variables previously observed as statistically significant were entered (see Table 1). These include age and occupational class. In the second model job demands and resources were included. These are the most commonly examined variables in the JD-R model. Time demands, unpredictability of work, and employment insecurity were studied as job demands. Job control and social support from the supervisor and colleagues were examined as job resources. In the last model the previous conception of the JD-R model was extended by including employees’ ideological orientation to work. These were public service motivation and value congruence.

Table 2 Work engagement and its antecedents among third-sector employees
Full size table

Model 1 shows that age and occupational class account for 1 % of the variance in work engagement and form a statistically significant regression model (see Table 2). As previously, in the variance analysis (see Table 1) age and occupational class was observed to increase work engagement. Older employees and those in expert and managerial positions report higher work engagement than younger ones or workers. The addition of job demands and resources in model 2, however, changes the connection between occupational class and work engagement. After job demands and resources are controlled, managers and experts report less work engagement than workers, but only the difference between managers and workers is statistically significant (Table 2). This observation suggests that third-sector managers have energetic and dedicated attitude toward their work, but they do not have enough resources to meet the job demands they are set (see Karasek 1979). Therefore, work engagement experiences disappear after job demands and resources are taken into account.

The addition of job demands and resources in model 2 significantly improves model prediction of work engagement beyond background variables (F change = 74.09, df1 = 7, df2 = 1,142, p value < 0.001). Increase in the explanation power is 31 % points. Specifically, as regards job demands, time demands, unpredictability of work, and employment insecurity were analyzed. Nearly all of these are negatively associated with work engagement thus supporting hypothesis 2. The only exception to this can be found from time demands. Initially, as the linear connection was analyzed, there was positive connection between work engagement and time demands. This has been explained by Mauno et al. (2007), who speculated that in small amounts time demands may increase work engagement until they become too excessive and the link becomes negative. Hence, the link between work engagement and time demands may appear as an inverted j-curve. Therefore, in the model the second power of time demands was included. Use of the squared variable reveals whether the link with work engagement in high values of time demands is negative. As can be seen in Table 2, excessive time demands decrease work engagement as predicted by the JD-R model. Thus, it appears that in small amounts the high pace of work promotes vigorous, energetic, and dedicated attitudes to work, but when the work becomes too hectic, work engagement starts to decline.

The remaining job demands have weaker link to work engagement than time demands. Insecurity in the form of unpredictability reduces work engagement, whereas employment insecurity is not linked to work engagement (see Table 2). Thus, for third-sector employees unpredictability of work is a more serious threat than fear of losing a job. Since most previous studies on work engagement have focused only on employment insecurity (e.g., Bosman et al. 2005), this result suggests that the other form of insecurity should also be taken into account in the future.

As regards job resources, job control and social support from supervisors and colleagues were considered. All of them are statistically significant and positively related to work engagement (see Table 2). This provides support for hypothesis 3. Employees who have more opportunities to use their skills and influence the content of their duties, working methods, and working times are more engaged in their work. The same is observed in terms of social support from supervisors. The more supervisors support and encourage employees, reward good performance, have discussions with them, and delegate responsibility sensibly, the higher the level of work engagement among employees. Furthermore, social support and encouragement from colleagues promote work engagement, but the link is barely statistically significant and eventually it disappears in model 3 (See table 2). Thus, it seems that social support received from supervisors is more important than social support from colleagues.

In model 3, the JD-R model was extended by including employees’ ideological orientation to work. This included public service motivation and value congruence. The addition of ideological orientation to work significantly improved model prediction beyond job demands and resources (F change = 38.75, df1 = 2, df2 = 1,140, p value < 0.001) explaining an additional 4 % points variance in work engagement. Therefore, the results indicate that the JD-R model can be extended beyond job demands and resources. Both variables—public service motivation and value congruence—turned out to be statistically significant and positively related on work engagement. This confirms hypothesis 4. Employees who are motivated by public interest and share employer organization’s values are more engaged in their work despite job demands and resources. The results suggest that employees’ ideological orientation to work form an important part of third-sector employees’ work engagement.

Discussion

Previous work engagement models, such as the JD-R model (e.g., Bakker and Demerouti 2007) or Karasek’s (1979) and Siegrist’s (1996) balance models, are unable to fully understand third-sector employees’ work engagement. They concentrate on job demands and resources, and thus ignore the assumption that third-sector organizations are mission-driven entities that attract ideologically oriented employees (Mirvis and Hackett 1983; Rose-Ackerman 1996). Based on this study, third-sector employees’ work engagement stems not only from job demands and resources as suggested in the previous models, but also from their ideological orientation to work.

The third-sector operating environment further seems to support employees’ high work engagement. Compared to previous work engagement studies, third-sector employees were observed to report higher work engagement than employees in general. This is consistent with the previous findings, which have shown that concepts that are closely related to work engagement, such as intrinsic motivation and commitment, are more prevalent among third-sector employees than their counterparts in either the public or the private sector (Borzaga and Tortia 2006; De Cooman et al. 2011). According to this study, high work engagement can be related to the fact that third-sector organizations select ideologically oriented employees and use managerial tools that are designed to support employees’ involvement (e.g., Mirvis and Hackett 1983). These tools include, for example, better control over work (e.g., Borzaga and Tortia 2006; Ruuskanen et al. 2012), which enables employees to express themselves freely on the job.

Nevertheless, the third-sector operating environment also includes aspects that limit employees’ work engagement. These include inadequate social support, high intensity, and unpredictability of work (Cunningham 2001; Ruuskanen et al. 2012). In particular, social support from supervisors is important for higher work engagement, but previous studies suggest that third-sector employees do not get enough feedback regarding their job performance (Mirvis and Hackett 1983). Therefore, third-sector organizations should develop their managerial tools in a way as to be able to provide employees with more social support and to relieve job demands. A practical tool for achieving this could be mentoring, that is, offering guidance and support from more experienced employees. Mentoring establishes dialog between parties, and therefore enables more experienced employees to provide social support and to give advice on how to handle the intensity and unpredictability of work. This would be especially useful for third-sector managers, since they were observed to show less work engagement than workers because of the problems associated with job demands and resources. Thus, mentoring could be used to detect which job demands and resources cause the most problems for managers, and could assist in trying to resolve them.

Furthermore, since ideological orientation was observed to increase work engagement, it should be encouraged by mentoring. Mentors could present the organization’s goals and values to new employees and thus help them to understand why the organization is doing the job. They could also explain how employees’ own work promotes these goals. This would help employees to identify with the organization, and subsequently create and strengthen public service motivation and value congruence (See Leiter and Maslach 2010).

Nevertheless, since there are certain limitations related to the present study, these suggestions should be interpreted with caution. First, this article was based on a cross-sectional dataset and, therefore, conclusions of causality cannot be made. Also, data collection from four trade organizations raises the question whether the results are generalizable to non-unionized third-sector employees even though it was shown that the data are fairly representative. Therefore, future analysis should be repeated by using a longitudinal dataset that extends beyond trade unions.

Despite the limitations, however, this article is one of a few studies that have provided information on third-sector employees’ work engagement. It has shown that public service motivation theory gives more insight into third-sector employees’ work engagement than does the conventionally used JD-R model. This analysis should be continued in future studies. For example, since public service motivation theory assumes that employees with strong public service motivation are willing to sacrifice their own well-being for the sake of others (Perry and Wise 1990), forthcoming studies could use mediation analysis to test whether employees with strong public service motivation are willing to withstand more job demands and lower job resources.