1. Introduction
In the rapidly changing business world, organizational sustainability is a vital goal for long-term success [
1]. To achieve and maintain sustainability, employees’ job engagement can serve as a crucial driver of organizational performance and sustainable success [
2]. Engaged employees not only tend to maintain enthusiasm and vigor toward their job, but also often demonstrate a strong commitment to their organization [
3,
4]. For this reason, they are expected to successfully perform their tasks and often go beyond the call of their duties [
3]. As such, job engagement may help organizations to improve or sustain their competitive advantage [
3]. Furthermore, job engagement is directly linked to work-related wellbeing (e.g., Schaufeli et al., Caesens et al.) [
5,
6]. Given that organizational sustainability concerns social performance, which is deeply related to the process of improving employees’ wellbeing, job engagement may be the key aspect of social performance that can lead to the achievement of organizational sustainability [
1,
7].
Although the definition of job engagement may differ depending on the researcher, it seems to converge to a general definition as a high level of effort put into the tasks of a job by an individual [
8,
9]. Many researchers have suggested that job engagement is positively related to various work outcomes. That is, it leads to high performance levels of in-role behaviors [
3], and is also highly likely to motivate employees to act toward a direction that benefits the organization. According to empirical studies, job engagement can lead to organizational citizenship behaviors (OCB) [
9], proactive behaviors [
10], contextual performance [
3], and innovative work behaviors [
11]. Given the benefits that job engagement provides, there are both theoretical and practical implications of exploring what encourages employees to engage at work and how engagement affects meaningful behavioral outcomes. As a result, this study examines the antecedents of job engagement and their impact on employees’ voice behavior based on Kahn’s [
8] study, which is recognized as the first research that theorized job engagement.
Kahn [
8] maintained that when individuals have a psychological experience of meaningfulness (i.e., Q1: How meaningful is it to me to engage in this work?), safety (i.e., Q2: How safe is it to do that?), and availability (i.e., Q3: Do I have the ability to do that?), they will fully engage in their jobs. May et al. [
4] carried out an empirical analysis of Kahn’s [
8] study and established empirical grounds for a theory that demonstrated that these three psychological conditions (i.e., meaningfulness, safety, and availability) had an important relationship with job engagement. Based on Kahn’s [
8] three dimensions of psychological conditions (i.e., work, social system, individual elements), this study adopts person–job fit (i.e., work characteristics), psychological contract fulfillment (i.e., social context), and self-efficacy (i.e., personal trait) as the antecedents of job engagement, that is, what might drive employees’ attitude and behaviors. Therefore, the first objective of this study is to explore the relationships of the antecedents with job engagement.
The second objective is to investigate the impact of job engagement on employees’ voice behavior. Voice behavior is a type of extra-role behavior challenging the status quo with the intent of making situations better [
12]. Exploring voice behavior is especially meaningful given the importance of rapid innovations today [
12]. Voice behavior includes making constructive suggestions, expressing ideas, and persuading other people to consider those ideas or directions [
13]. Since expressions of, or suggestions for, altering status quo may hurt relations with others, employees may feel that it is risky to do so and, thus, be reluctant to speak up in some situations [
13,
14]. Even so, improving voice behavior is a practical way for engaged employees not to rest on their personal performance but to contribute to constructive changes in an organization. Thus, it is worth exploring how to enable engaged employees to be willing to speak up by taking such risks. Nevertheless, so far, there are limited studies on the direct effect of job engagement on voice behavior [
15]. Thus, in this study, we verify the relationship between job engagement and employees’ voice behavior.
Furthermore, we consider the moderating role of perceived coworker support by focusing on an environment that can encourage employees to increase their contribution to the organization through job engagement. Nowadays, due to emphasis on team-based work and a flat organizational structure, employees tend to interact more frequently with coworkers than with their supervisors [
16]. Thus, the support from coworkers may have a significant impact on the employees’ attitude and behaviors. While voice behavior may require sufficient resources (i.e., job engagement) to enhance speaking up and the expression of suggestions [
13,
17], some individuals may still suffer from stress and anxiety due to lack of job engagement. In this regard, by applying the Conservation of Resources theory (COR theory; Hobfoll) [
18], this study makes the assumption that coworker support may compensate for the resource deficiency. Therefore, the third objective of this study is to investigate the moderating effect of perceived coworker support in complementing job engagement.
In summary, this study contributes to the literature on job engagement and voice behavior in at least four ways. First, we add to the literature by suggesting the antecedents of job engagement in terms of work characteristics, social context, and personal traits based on the three dimensions of psychological experiences that Kahn [
8] suggested. Specifically, we shed light on the role of person–job fit (P–J fit), psychological contract fulfillment, and self-efficacy. As for psychological contract fulfillment, it is worth noting that our study deals with how “employees’ expectations” affect their job engagement; however, so far, this has not been widely explored. Second, we investigate the relationship between job engagement and voice behavior. This is meaningful in that there are few studies on the direct effect of job engagement on voice behavior within the scope of this research. Third, this study aims to investigate the mechanism that explains how the antecedents influence employee voice behavior via job engagement. Thus, our study notes the mediating role of job engagement in the relationship between the antecedents and voice behavior. Fourth, this study highlights the role of perceived coworker support as a complementary resource for when the level of job engagement is low.
Figure 1 presents the conceptual model of this study.
4. Results
In order to verify the construct validity of the hypothesized variables, we conducted a confirmatory factor analysis. The result of the analysis showed that the hypothesized model (i.e. six-factor model) fit well to the data:
χ2(215) = 436.18,
p < 0.01, CFI = 0.93, TLI = 0.92, and RMSEA = 0.07. In addition, the hypothesized model was compared with alternative models. As presented in
Table 1, the results suggested that the hypothesized model is superior to all other alternative models. For instance, the hypothesized model fit better than five-factor models in which two of the three antecedents loaded on the same factor (
χ2(220) =688.16,
p < 0.01, CFI = 0.85, TLI = 0.83, RMSEA = 0.11). Therefore, the results indicate that the six constructs are distinct from each other.
The means, standard deviations, reliabilities, and correlations among the main study variables are presented in
Table 2. All scales had very good internal consistency (alpha values above 0.90).
Hypothesis 1-1 proposed that P–J fit is positively related to job engagement. As illustrated in Model 2 of
Table 3, P–J fit was indeed positively associated with employees’ job engagement (
β = 0.70,
p < 0.001). Therefore, Hypothesis 1-1 was supported. In addition, Hypothesis 1-2 suggested that psychological contract fulfillment is positively associated with job engagement. The results of Model 2 in
Table 4 show that psychological contract fulfillment is significantly related to job engagement (
β = 0.24,
p < 0.01), supporting Hypothesis 1-2. In the same way, we tested Hypothesis 1-3, which suggested that self-efficacy is positively related to employees’ job engagement. As illustrated in Model 2 of
Table 5, for job engagement, the coefficient associated with self-efficacy was statistically significant (
β = 0.66,
p < 0.001). Therefore, Hypothesis 1-3 was also supported.
Hypothesis 2 proposed that job engagement has a positive relationship with employees’ voice behavior. The results, as illustrated in Model 5 of
Table 3, indicated that job engagement has a positive effect on employees’ voice behavior (
β = 0.26,
p < 0.001). Therefore, Hypothesis 2 was supported.
Hypotheses 3-1, 3-2, and 3-3 posited that job engagement mediates the relationships between the antecedents of this study (i.e., P–J fit, psychological contract fulfillment, and self-efficacy) and voice behavior. To test our hypotheses regarding the mediating effect of job engagement, Baron and Kenny’s procedure [
51] was adopted. Moreover, following the approach suggested by Hayes and Preacher [
52], both the Sobel test and bootstrapping were used to assess the significance of the indirect effects.
Specifically, Hypothesis 3-1 proposed the mediating role of job engagement in the relationship between P–J fit and voice behavior. P–J fit was positively associated with voice behavior (
β = 0.16,
p < 0.05; Model 4 of
Table 3) and was significantly related to job engagement (
β = 0.70,
p < 0.001; Model 2 of
Table 3), thus, the first and second prerequisites for mediation were fulfilled. To test the third condition for mediation, voice behavior was regressed on job engagement, controlling for P–J fit. As shown in Model 6 of
Table 3, the beta coefficient for job engagement was statistically significant (
β = 0.29,
p < 0.01); however, the effect of P–J fit was no longer significant (
β = –0.05,
ns). Thus, the mediation analysis indicates that the effect of P–J fit on employees’ voice behavior is fully mediated by job engagement. In addition, the lower part of
Table 3 indicates the results of the Sobel test and bootstrapping. The formal two-tailed significance test (assuming that a distribution is normal) presented the significance of the indirect effect (Sobel
z = 2.890,
p = 0.004), and the result of bootstrapping reaffirmed the Sobel test. To be specific, we calculated a 95% bias-corrected confidence interval (CI) for indirect effects using 10,000 bootstrapped samples. Shrout and Bolger [
53] explained that if the CI is not zero, the researcher can have confidence that the indirect effect is different from zero. In this study, the CI was between 0.074 and 0.397, excluding zero, which indicates the statistical significance of the indirect effect in this research model. Therefore, Hypothesis 3-1 was supported.
Hypothesis 3-2 posited that job engagement mediates the relationship between psychological contract fulfillment and voice behavior. Psychological contract fulfillment was not significantly associated with voice behavior (
β = –0.01,
ns; Model 4 of
Table 4); thus, the first prerequisite for mediation was not fulfilled. However, psychological contract fulfillment had a significant relationship with job engagement (
β = 0.24,
p < 0.01; Model 2 of
Table 4), and job engagement had a significant relationship with voice behavior after including psychological contract fulfillment (
β = 0.28,
p < 0.001; Model 5 of
Table 4). Thus, to confirm the indirect effect more accurately, we examined the results of the Sobel test and bootstrapping. As illustrated in
Table 4, the indirect effect was statistically significant in our model (Sobel
z = 2.450,
p = 0.014; 95% bootstrap CI = 0.023 to 0.142), providing partial support for Hypothesis 3-2.
In the same manner, we tested Hypothesis 3-3, which predicted the mediating role of job engagement in the relationship between self-efficacy and voice behavior. As reported in
Table 5, self-efficacy was significantly related to voice behavior (
β = 0.18,
p < 0.05; Model 4 of
Table 5) and positively associated with job engagement (
β = 0.66,
p < 0.001; Model 2 of
Table 5). Additionally, job engagement had a significant relationship with voice behavior, controlling for self-efficacy (
β = 0.26,
p < 0.01; Model 5 of
Table 5); however, the effect of self-efficacy was no longer significant (
β = 0.01,
ns). These results suggested that job engagement fully mediated the relationship between self-efficacy and employees’ voice behavior. Moreover, the results of the Sobel test and bootstrapping confirmed the indirect effect (Sobel
z = 2.618,
p= 0.009; 95% bootstrap CI = 0.070 to 0.429), providing support for Hypothesis 3-3.
Hypothesis 4 predicted the moderating effect of perceived coworker support on the relationship between job engagement and voice behavior. The results of Model 5 in
Table 6 present that the coefficient for the interaction term, including job engagement and perceived coworker support, is negative and statistically significant (
β = –0.17,
p < 0.05). To further investigate this interaction effect, we plotted the results following Aiken and West’s procedure [
54]. As presented in
Figure 2, the positive relationship between job engagement and voice behavior is stronger when perceived coworker support is low rather than high. Thus, these results supported Hypothesis 4.