EXAMINING STRUCTURE RELATIONSHIPS AMONG PROCEDURAL JUSTICE, PERCEIVED ORGANIZATIONAL SUPPORT, EMPLOYEE ENGAGEMENT, AND TURNOVER INTENTION IN LAO PUBLIC ORGANIZATIONS
The purpose of this research was to examine practical issues that predict and prevent turnover intention in the context of Lao public organizations, focusing on the relationship between procedural justice, perceived organizational support, and employee engagement. Data for this study were collected from 331 public officials. In this study, structural equation modeling was used to analyze the collected data and test the hypothesized relationships. The results indicated that procedural justice had a significant effect on perceived organizational support, employee engagement, and turnover intention, whereas perceived organizational support and employee engagement were not significantly related to turnover intention. In addition, the mediating effect of perceived organizational support in the relationship between procedural justice and employee engagement was statistically significant. The findings suggest that organizations should contemplate how to improve procedural justice and promote their level of perceived organizational support. On the basis of the findings, human resource development implications and recommendations are discussed and suggested.
INTRODUCTION
An organization's human resources (HR) orientation is vital to strategic management (Gabčanová, 2011) given that an organization's most important resource and success factor is its employees (A. Gupta, 2013). Without them, it would be challenging for an organization to be successful. Thus, employees' loyalty and retention have a significant effect on the organization (Kabinritthiwat, 2014). However, there has been an increasing tendency for employees to develop job performance competencies for their own careers and to work in multiple workplaces over their career. Working for one company is no longer an absolute for employees. They are likely to seek more suitable conditions for themselves and will continue to seek new job opportunities, meaning that turnover will occur more frequently (Huang et al., 2022). From an organizational perspective, turnover can cause great disruption to the organization due to the high cost of hiring and training new employees (Holtom et al., 2008). From an employee's perspective, giving up on their previous job and the interpersonal relationships when they change organizations can be very stressful (Boswell et al., 2005) and can negatively affect both individuals and organizations. Thus, efforts are needed to lower turnover rates and turnover intention.
For public organizations, there has been an ongoing concern about the lack of qualified candidates, even though there is ongoing demand for competent public officials (Sun & Wang, 2017). In developing countries, the public sector workforce could be more important because a competent workforce directly affects political, economic, and social development (Bui & Chang, 2018). Even in the context of Lao public organizations, the number of civil servants leaving work and moving to other government organizations increased by 51% in 2020 compared with 2019, according to the Ministry of Home Affairs' (2019, 2020) statistical reports. Although the Lao government has introduced and enforced legislation to deal with turnover (e.g., pay systems based on high-ranking government officers, seniority, and position, public property ownership law for civil servants, and family health insurance), it is still an unresolved challenge. Thus, public organizations should create an environment where public officials can engage in their work without the threat of turnover. It is also necessary to identify the crucial elements and ways to reduce turnover intention.
Procedural justice (PJ) is a key factor that significantly affects turnover intention (TI; Saba et al., 2015). PJ has commonly been defined as employees' perceptions of the fairness of procedures in the workplace (W. Kim & Park, 2017). Lim and Saraih (2020) examined the relationships between PJ and TI and showed that fair or unfair treatment from the organization could influence the level of TI among employees. PJ also has a significant effect on perceived organizational support (POS) and employee engagement (EE). Sarianti and Armida (2021) demonstrated that PJ had a positive and significant effect on POS. PJ is also very important to employees because it increases the sense that they are supported by the organization. Thus, greater PJ increases their POS in the organization. M. Gupta and Kumar (2015) tested the effect of PJ on EE and found that greater EE can be achieved by attaching importance to PJ. On the basis of these studies, it appears that PJ not only lowers TI within the organization, but also has a positive effect on POS and EE, so it should be considered an important factor.
POS often refers to employees' perceptions that they are supported and valued by their organizations (Caesens et al., 2016), which means a supportive work environment with assistance and consideration from the organization (Sulea et al., 2012). Wang and Wang (2020) examined the effect of POS on TI and found that POS plays a role in reducing employee turnover. Previous studies have found that POS also has a positive relationship with EE (Biswas et al., 2013; Caesens et al., 2016; Nadeem et al., 2019). For example, W. Kim et al. (2018) and Y. Park et al. (2020) have demonstrated that POS could positively contribute to EE. As a result, when employees feel that their organizations support their work and well-being, employees have a higher perception of organizational support and a greater willingness to improve their work capabilities (Shore & Wayne, 1993). Previous studies have also investigated how PJ affects EE through POS as a mediator (Javed & Tariq, 2015). In other words, POS can be an important factor to lower TI and increase EE in organizations considering its role as a mediator.
In addition, for optimal performance of the organization, it is necessary for employees to be engaged in their work. Thus, EE is important for them to have a positive mindset toward the organization (J.C. Lee, 2014). Depending on employees' psychological state, they may put additional efforts into work with a sense of reciprocity in the organization, or their psychological state may adversely affect the organization through unproductive work behaviors such as absences and slowdowns (Eisenberger et al., 1986). Other studies have also found that employees' work engagement had a negative relationship with TI (W. Kim, 2017; W. Kim et al., 2019), implying that the more employees are engaged in their work, the less likely they are to leave their organization. In addition, Saks (2006) showed that EE mediated the link between the antecedents (e.g., POS and PJ) and TI. Thus, it is necessary to confirm the role of EE as a mediator in different contexts.
In evaluating and compensating for employee performance, organizations can promote a positive or negative psychological state depending on the procedures and how they are distributed (H. Kim & Ham, 2013). In addition, to improve organizational performance, relevant factors that can promote employees' positive perceptions, emotions, and attitudes need to be considered in the HR aspects of the organization (J.W. Lee, 2010). Reflecting this practical interest, numerous studies have focused on factors such as PJ, POS, EE, and TI that are directly related to employees' psychological state toward the organization. These studies have confirmed that PJ is strongly related to POS, EE, and TI in organizations. Although many studies have been conducted on each major factor, insufficient studies have examined the psychological relationship of beliefs about the organization and how individuals' positive mental states can affect TI.
Furthermore, little empirical research has investigated how PJ could affect EE and TI through POS, especially in public organizations in the context of developing countries. Although this study focuses on the specific context in Laos, considering the recent increasing tendency of a higher turnover rate of government officials in Asian countries (Bui & Chang, 2018), it seems timely to reveal broader implications from the case of Laos. More research is also needed to reveal the psychological mechanism through which PJ of employees affects their TI with POS and EE. Thus, this study aims to investigate the structural relationships among PJ, POS, EE, and employees' TI in the context of Lao public organizations.
THEORETICAL FRAMEWORK
In the study, social exchange theory (SET) elucidates the relationships among PJ, POS, EE, and TI, which is consistent with previous studies (Hussain & Khan, 2019; Joo et al., 2015; W. Kim & Park, 2017; Memon et al., 2017; Y. Park et al., 2020). Typically, SET refers to a series of interdependent variables and is contingent on the actions of another party that can lead to a high-quality relationship (Blau, 1964; Cropanzano & Mitchell, 2005).
To extend previous studies from a SET perspective, PJ is related to the quality of social exchange between employees and their organization (Flint et al., 2013; Hussain & Khan, 2019). In particular, studies have shown that individuals with firm exchange tendencies are more likely to repay employees' good deeds (Cropanzano & Mitchell, 2005). Therefore, when employees have high perceptions of justice in the workplace, they feel supported by their organization, which motivates and obligates them to reciprocate the positive treatment they have received with favorable attitudes and behaviors toward their organization (Caesens & Stinglhamber, 2014; Rhoades & Eisenberger, 2002). As a result, they engage more in their work and their organization (Biswas et al., 2013; Memon et al., 2017) and reduce TI (Flint et al., 2013; Hussain & Khan, 2019). This study is interested in determining whether these influences contribute to or reduce TI in Lao public organizations given they are experiencing high TI rates.
LITERATURE REVIEW
To conduct the literature review, we searched previous studies through Google Scholar, ProQuest Central, and RISS (a representative comprehensive database in Korea). The search keywords were procedural justice, perceived organization support, employee engagement, and turnover intention. The results showing the relationships among these variables are reviewed throughout this section. Meta-analysis and systematic literature review papers were also referenced for a comprehensive review of the main variables.
Effects of Procedural Justice
PJ refers to employees' perceptions of the fairness of decision-making procedures at work (Biswas et al., 2013; W. Kim & Park, 2017). Procedures are considered fair if employees can voice their opinions in the decision-making process (Thibaut & Walker, 1975) and the organization shows efforts to eliminate bias, allocate resources consistently, use accurate information, express concerns about poor decision-making, and reflect prevailing ethics norms (Leventhal, 1980). PJ also refers to employees' awareness of fair policies and procedures related to decisions in terms of the allocation of organizational resources (He et al., 2014; Nazir et al., 2019). Hussain and Khan (2019) described PJ as the fairness of the procedures used to determine such outcomes (Moorman, 1991). Employees generally believe that PJ is a driving force for performance, which is evaluated as the most appropriate type of organizational justice (Novianjani et al., 2019).
Research has shown that PJ is an antecedent of POS (Ahmed & Nawaz, 2015; Rhoades & Eisenberger, 2002). In several empirical studies, PJ had a significant positive relationship with POS (Biswas et al., 2013; Mukherjee, 2010; Nazir et al., 2019; Sarianti & Armida, 2021). DeConinck and Johnson (2009) also demonstrated that PJ had a positive relationship with POS and that it plays an important role in influencing TI through other variables including POS. To increase the level of POS, organizations should further improve PJ (Babic et al., 2015). Thus, we present the following hypothesis:
Hypothesis 1: Procedural justice is positively related to perceived organizational support
Cropanzano and Mitchell (2005) confirmed that SET is a strong theoretical foundation to explain why employees choose to be more engaged or less engaged in their work and their organization. Hence, it is a well-established theoretical framework to explain how individuals' perceptions of PJ may influence their engagement with their work and organization. Previous literature has also shown that PJ has a strong positive influence on EE both directly and indirectly (Biswas et al., 2013; He et al., 2014; Karatepe, 2011; W. Kim & Park, 2017; S. Li et al., 2010; Saks, 2006). Al-Shbiel et al. (2018) and M. Gupta and Kumar (2015) also found that PJ and EE were significantly correlated. Researchers have also argued that when employees are treated procedurally, they recognize the workplace as a desirable environment and that the organizations show them respect and dignity (Lind & Tyler, 1988). On the basis of these studies, we present the second hypothesis:
Hypothesis 2: Procedural justice is positively related to employee engagement
In contrast, when employees are treated poorly, they may feel that they are not valued or respected and may seek alternative employment in search of fairer treatment. This argument is in line with SET, which offers the theoretical framework for explaining how PJ affects TI (Hussain & Khan, 2019). Employees are more likely to maintain a relationship with the exchange organization if they perceive benefits from the exchange. If not, they could stop participating in future interactions and may even quit the organization (Flint et al., 2013). Several studies have also reported a significant negative relationship between PJ and TI (Al-Shbiel et al., 2018; Flint et al., 2013; Hussain & Khan, 2019; A. Li & Bagger, 2012; Lim & Saraih, 2020; Ponnu & Chuah, 2010). In addition, Saba et al. (2015) demonstrated that procedural injustice had a significant effect on TI. Thus, it can be minimized by enhancing fairness in the procedures of an organization. The following hypothesis predicts this relationship:
Hypothesis 3: Procedural justice is negatively related to turnover intention
Effects of Perceived Organizational Support
POS is defined as workers' perceptions that the “organization values their contributions and cares about their well-being” (Eisenberger et al., 1986, p. 501). The term refers to the general perception that employees are supported and valued by the organization (Caesens et al., 2016). POS also represents a supportive work environment that includes assistance and consideration for employees' purpose and worth (Sulea et al., 2012). Therefore, POS imposes an obligation on both the organizations and employees to help each other reach their goals based on SET (Rhoades et al., 2001; Saks, 2006). Employees who perceive organizational support are more likely to remain attached to the organization than those without POS (Joo et al., 2015). From a social exchange perspective, when employees have a high degree of POS, it may be expressed as improved work performance (Shore & Wayne, 1993). In other words, if employees within the organization perceive that they are receiving additional incentives from the organization, they form a reciprocal relationship by striving to achieve organizational goals.
Current studies have suggested that POS also has a positive relationship with EE (Biswas et al., 2013; Caesens et al., 2016; Nadeem et al., 2019; Sulea et al., 2012; Thirapatsakun et al., 2014). Karatepe and Aga (2016) found that employees are engaged in their work with vitality and dedication when they recognize that management is trying to achieve organizational goals through interest and investment in employees and in technology. Some studies have also noted that the relationship between POS and TI is negative (e.g., Arshadi, 2011; Joo et al., 2015; Nadeem et al., 2019; Rhoades & Eisenberger, 2002). Perryer et al. (2010) also demonstrated that POS was an important predictor of TI. Specifically, employees who perceive a high level of organizational support are less likely to leave their organization. Thus, we present the following hypotheses:
Hypothesis 4: Perceived organizational support is positively related to employee engagement.
Hypothesis 5: Perceived organizational support is negatively related to turnover intention
Effects of Employee Engagement on Turnover Intention
EE is “an active, work-related positive psychological state operationalized by cognitive, emotional, and behavioral energy” (Shuck et al., 2017, p. 959). Since the concept of engagement was initially established in the '90s (Kahn, 1990), considerable research has been conducted on engagement related to employees' work, job, and roles. Saks (2006) noted that engagement has been defined academically as a unique structure including “cognitive, emotional, and behavioral components that are associated with individual role performance” (Saks, 2006, p. 602). This concept of engagement is slightly different but is generally used interchangeably in the literature (Carasco-Saul et al., 2015). This study uses the term employee engagement on the basis of Shuck and colleagues' (2017) study. Employees who are highly engaged feel that work is fun, energetic, and meaningful; thus, they feel positive emotions such as enthusiasm, joy, and happiness (Bakker & Demerouti, 2008). Therefore, EE could be perceived as a state in which employees are positively affected by their work and their TI is weakened (Al-Shbiel et al., 2018).
Previous studies have found that EE has a significantly negative relationship with TI (Al-Shbiel et al., 2018; Bakker & Demerouti, 2008; Kang et al., 2022; W. Kim, 2017; W. Kim & Hyun, 2017; Salahudin et al., 2019). Saks (2006) showed that employee engagement (job engagement and organizational engagement) had a negative correlation with the intention to quit. Similarly, Azanza et al. (2015), Schaufeli and Bakker (2004), and Thirapatsakun et al. (2014) found a direct negative relationship between EE and TI. Several researchers have also demonstrated the same results in their research where EE has a negative effect on TI and can reduce TI (e.g., Agarwal & Gupta, 2018; W. Kim et al., 2019; Memon et al., 2017; Shuck et al., 2014; Zhang et al., 2018). Based on Kang and colleagues' (2022) recent meta-analysis, the relationship between EE and TI can be negative. These findings imply that in order to lower TI, EE, which has a negative relationship with TI, must be managed above a certain level. Given that numerous studies have reported a negative relationship between EE and TI, the following hypothesis is presented:
Hypothesis 6: Employee engagement is negatively related to turnover intention
Indirect Effects of Perceived Organizational Support and Employee Engagement
Studies on antecedents and consequences (e.g., employee engagement) in the context of an organization emphasize social exchange theory as one of the core theories (Biswas et al., 2013). On the basis of SET, the relevant literature has shown that the mediating effect of POS and EE is significant. The results of a meta-analysis indicated that POS plays an important role in the relationships between antecedents and consequences (Ahmed & Nawaz, 2015; Kurtessis et al., 2017; Rhoades & Eisenberger, 2002). Specifically, POS plays a vital role as a mediator between PJ and EE (Biswas et al., 2013; Javed & Tariq, 2015; Loi et al., 2006) and between PJ and TI (Loi et al., 2006). PJ is also an indirect predictor of TI through EE (DeConinck & Johnson, 2009). In addition, several studies have demonstrated the dominant role of EE as a mediating attribute that plays a significant role between PJ and TI (Al-Shbiel et al., 2018; Saks, 2006), and EE mediates the relationship between POS and TI (Bas & Çınar, 2021; Nadeem et al., 2019; Saks, 2006; Thirapatsakun et al., 2014). Assuming that the antecedent variable predicts the mediator and the mediator predicts the outcome variable, the mediators, (i.e., EE and POS) can mediate the relationship between PJ and TI. Therefore, the following hypotheses are presented:
Hypothesis 7: The relationship between procedural justice and employee engagement is mediated by perceived organizational support.
Hypothesis 8: The relationship between procedural justice and turnover intention is mediated by perceived organizational support.
Hypothesis 9: The relationship between procedural justice and turnover intention is mediated by perceived organizational support and employee engagement.
METHODS
The main purpose of this study is to examine the structural relationships among PJ, POS, EE, and TI in Lao public organizations. Figure 1 shows the hypothesized research model based on the theoretical framework and extant empirical literature to test the research hypotheses using structural equation modeling (SEM).
Citation: Performance Improvement Quarterly 36, 3; 10.56811/PIQ-22-0016
Sample and Data Collection
The study employed a quantitative approach with a self-report questionnaire form using a convenience sampling technique with self-selected participants. Questionnaires were sent to government officials ranging from central to local levels in Laos. First surveys were distributed through email and social network service messenger; 350 were returned. After excluding outliers and incomplete questionnaires, 331 were used for SEM analysis. Table 1 shows the demographic characteristics including gender, age, status, education, organizational level, position, working sector, and job tenure in the current organization. Of the total respondents, 56.5% were men, almost 80% were in their 30s and older (31–39 years: 40.1%; 40 years and older: 36.9%), 68.9% were married, and 57.4% had graduated with a bachelor's degree. In addition, 58.6% of the employees worked for local authorities, and of those, 75.8% of the positions were nonmanager positions. In terms of working sectors, 63.7% worked in three sectors: public health (22.1%), economics and finance (21.1%), and education (20.5%). In terms of job tenure in the current organization, 65.5% had worked in their current organization for more than 7 years, 26.3% had worked 16 years or more, 20.2% had worked 8–10 years, and 19% had worked 11–15 years.
Measurement
Reliability and validity of all measurement were secured in previous studies.
Procedural Justice
To measure PJ as a type of organizational justice, this study used the seven-item instrument by Moorman (1991) measured on a 5-point Likert scale. Cronbach's alpha (α) was from .924–.926 (Kang et al., 2012; W. Kim & Park, 2017).
Perceived Organizational Support
The study employed a widely adapted instrument by Eisenberger et al. (1986). The instrument includes eight items measured on a 7-point Likert scale. Cronbach's alpha was from .721–.956 (Bano et al., 2015; W. Kim et al., 2018).
Employee Engagement
EE was measured by three subscales (cognitive, emotional, and behavioral engagement), and each subscale included four items measured on the 5-point Likert scale by Shuck et al. (2017). Cronbach's alpha was .891 (J. Park et al., 2021).
Turnover Intention
TI was measured with three-items with the 5-point Likert scale adopted from Colarelli (1982). Cronbach's alpha was .713 (W. Kim & Hyun, 2017).
Data Analysis
This study employed SEM to analyze the collected data. Preliminary analysis was performed for SEM, and then the measurement model and structural model were analyzed in order according to Anderson and Gerbing's (1988) two-step approach. First, preliminary analysis checked common method bias (CMB), normality, reliability, and correlations. Second, the overall fit indices of the research model and improper solutions were examined. This study used Satorra-Bentler (SB) scaled chi-square to address the potential nonnormality of the data with other absolute and incremental fit indices including the Tucker-Lewis index (TLI), comparative fit index (CFI), standardized root mean residual (SRMR), and root mean square error of approximation (RMSEA; CFI ≥ .90, TLI ≥ .90, RMSEA ≤ .08, SRMR ≤ .08; Bae, 2014). To check for improper solutions, reasonable signs and magnitude, negative error variances, and significant paths of individual parameter estimates were examined (Lei & Wu, 2007). Third, to test the research hypotheses, this study mainly used standardized path coefficients of the direct effects and the bias-corrected (BC) bootstrap estimates of the mediating effects with a 99% confidence interval (Preacher & Hayes, 2008). Fourth, to measure EE, this study used a facet-representative parceling approach for multidimensional constructs to optimally represent the latent construct as this measurement has been widely validated in the literature (Little et al., 2013).
RESULTS
Preliminary Analysis
CMB, correlations, reliability, and normality were examined. Given that CMB could significantly influence the findings, confirmative factor analysis for the one-factor model was used to check for CMB (Podsakoff et al., 2003). The results demonstrated that the one-factor model did not have a good fit with the data, χ2(464) = 10461.210, p < .001; RMSEA = .255; SRMR = .158; CFI = .332; TLI = .286. Given that the one-factor model did not account for the major variance in the collected data, it would be reasonable to conclude that CMB is not a major concern for this study.
As shown in Table 2, the results of the Cronbach's alphas of all measurements were acceptable (α ranged from .879–.956; Urdan, 2010). In addition, all correlations among the latent research constructs were statistically significant and indicated no critical issue of multicollinearity (Lei & Wu, 2007). Furthermore, although the results of the univariate test (|skewness| < 2, |kurtosis| < 7]) and multivariate normality (p < .05) demonstrated that the collected data set had a moderate nonnormality, it could be handled by using the robust maximum likelihood (ML) approach (Finney & DiStefano, 2013; Pellegrini & Scandura, 2005).
Model Evaluation
The overall fit indices and any improper solutions of the measurement model and the full model were assessed. According to the overall fit statistics of the measurement models presented in Table 3, the SB chi-square of the measurement model (MM1) was statistically significant, χ2(183) = 624.254, p < .001. Other fit statistics except CFI did not meet the cutoff criteria, RSMEA = .085 (≤.08), SRMR = .091 (≤.08), CFI = .904 (≥.90), TLI = .890 (≥.90). Regarding improper solutions, although there was no identified issue when checking negative error variances, out-of-range of r, and all signs and magnitudes of parameters, one of the factor loadings (POS6 = .416) was less than .5 (Hair et al., 2019). On the basis of these results, we considered and reevaluated the modified measurement model (MM2) by removing the observed item (POS6). The results of MM2 demonstrated that although SB chi-square of MM2 was statistically significant, χ2(164) = 459.323, p < .001, other fit statistics were satisfied, RSMEA = .074 (≤.08), SRMR = .073 (≤.08), CFI = .932 (≥.90), TLI = .922 (≥.90). Moreover, no issue emerged in checking improper solutions. Based on these results, MM2 had an acceptable fit with the current data set.
Given that MM2 appeared to be valid and feasible, the full model (FM1) was assessed. Because MM2 and FM1 in this study were equivalent models, the overall fit statistics of FM1 were the same as those of MM2, indicating that despite the statistical significance of the SB chi-square of FM1, other fit statistics satisfied the criteria (see Tables 3 and 4). With regard to possible improper solutions, although the signs and magnitudes of the parameter estimates in FM1 made sense with no negative variances and out-of-range of r, one structural path (e.g., from EE to TI) was not statistically significant (see Figure 2). Considering the insignificant path, we needed to improve the FM1.
Citation: Performance Improvement Quarterly 36, 3; 10.56811/PIQ-22-0016
Model Modification and Its Evaluation
The FM1 was modified by fixing the path from EE to TI (t = 1.750, p > .05) as the modified full model 1 (FM2). FM1 and FM2 had a nested model relationship and the current study used robust ML, so the nested model comparison was conducted using the SB chi-square difference test. According to the results presented in Tables 4 and 5, although the SB chi-square difference test was not statistically significant (Δ df = 1, p > .05), FM2 was selected because it was more parsimonious. However, in terms of improper solutions, the results demonstrated that another structural path (i.e., from POS to TI) became statistically insignificant (see Figure 3). Thus, FM2 also needed to be modified. The FM2 was respecified by fixing the path from POS to TI (t = −1.769, p > .05) as the modified full model 2 (FM3). Through the process of model modification, the results of the SB chi-square difference test and the nested model comparison revealed that even though the SB chi-square difference test was not statistically significant (Δ df = 1, p > .05), FM3 was chosen because it was more parsimonious (see Tables 4 and 5).
Citation: Performance Improvement Quarterly 36, 3; 10.56811/PIQ-22-0016
With regard to evaluating FM3, the SB chi-square for FM3 was statistically significant, χ2(166) = 464.071, p < .001, and other overall fit indices satisfied the cutoff criteria, RMSEA = .074, SRMR = .074, TLI = .932, and CFI = .922 (see Table 5). With regard to improper solutions, all structural path coefficients were statistically significant (p < .05, see Figure 3). There was no issue found from error variances, range of r, and signs and magnitudes of parameters. Furthermore, according to the results of the nested model comparison between FM1 and FM3, although the SB chi-square difference was not statistically significant, FM3 was selected due to parsimony (Δ df = 2, p > .05). Taken together, FM3 had a more adequate fit with the collected data than FM1 and FM2.
Hypothesis Testing
On the basis of the results of this study, all proposed research hypotheses were tested. Standardized path coefficient estimates were the main focus to examine the magnitudes of the paths among the four proposed research variables of FM3 (see Figures 2 and 3).
Regarding Hypotheses 1 and 2, the results of FM3 showed that the positive and direct effects of PJ on POS (γ11 = .739, p < .001) and EE (γ21 = .440, p < .001) were statistically significant. Thus, Hypotheses 1 and 2 were supported. For Hypotheses 3 and 4, the negative and direct effect of PJ on TI (γ31 = −.415, p < .001) and the direct and positive effect of POS on EE (β21 = .300, p < .001) were both statistically significant. Thus, Hypotheses 3 and 4 were supported. For Hypotheses 5 and 6, the negative and direct effect of POS on TI (β31 = −.129, p > .05) as well as the negative and direct effect of EE on TI (β32 = .129, p > .05) were not statistically significant. Thus, Hypotheses 5 and 6 were not supported.
To investigate the mediating effects of POS and EE in FM3 for Hypotheses 7–9, we examined the BC bootstrap estimates, as strongly recommended by Preacher and Hayes (2008). This study tested the mediating effects using the BC bootstrapping procedure with 1,000 bootstrap samples. For Hypothesis 7, the results presented in Table 6 demonstrate that the mediating effect of POS in the relationship between PJ and EE (ab = .222, 99% CI [.027, .442]) was statistically significant in that the BC 99% CI excluded zero. Therefore, Hypothesis 7 was supported. Regarding Hypotheses 8 and 9, because the direct relationship between POS and TI as well as the relationship between EE and TI were not statistically significant, the mediating effects of POS in the relationship between PJ and EE and the multimediating effect of POS and EE in the relationship between PJ and TI were neither meaningful nor significant (see Figures 2 and 3). Thus, Hypotheses 8 and 9 were not supported.
DISCUSSION AND CONCLUSIONS
This study examined the effects of PJ, POS, and EE on TI in the context of Lao public organizations. First, PJ positively influenced POS and EE and negatively influenced TI. This result supports the findings of numerous studies (e.g., Al-Shbiel et al., 2018; Biswas et al., 2013; Flint et al., 2013; He et al., 2014; Hussain & Khan, 2019; Karatepe, 2011; W. Kim & Park, 2017; Lim & Saraih, 2020; Nazir et al., 2019; Sarianti & Armida, 2021). Furthermore, POS positively influenced EE, which confirms previous findings (e.g., Caesens et al., 2016; Nadeem et al., 2019; Sulea et al., 2012; Thirapatsakun et al., 2014), but POS did not influence TI. This finding contradicts the results of previous studies that found a negative effect between POS and TI (e.g., Joo et al., 2015; Nadeem et al., 2019; Perryer et al., 2010). The results of Ahmed and colleagues' (2015) meta-analysis related to POS and the outcome variables including TI may be different on the basis of the type of organization even though POS and TI had a strong relationship. They found that there was a strong relationship in the manufacturing industry, whereas educational institutions and public institutions had a relatively moderate relationship (Ahmed et al., 2015). Thus, the results may vary depending on the support environment in organizations, such as regular feedback, training, and compensation for performance. For public organizations, leaders often change suddenly because they are elected or appointed politically, making it highly likely that the influence of POS that values social exchange would not be expressed (Wayne et al., 1997). Even worse, there are few incentives such as high salaries, wage hikes, and bonuses, unlike in the private sector (Jin & McDonald, 2017). Officers in public organizations are strongly aware of these characteristics; therefore, POS may not have a direct affect on TI.
It is also noteworthy that EE did not influence TI, which is in contrast to the results of Kang et al. (2022), W. Kim (2017), W. Kim et al. (2019), and Salahudin et al. (2019). Our findings could be explained by SET in that the link between engagement and outcomes (e.g., job performance) greatly depends on how satisfied employees are with their relationship with their managers (Harter et al., 2002; Judge et al., 2001). Therefore, this result is likely to stem from the lack of active interaction among individuals, groups, or organizations due to the typically rigid culture in public sectors. Future studies could explore additional variables (e.g., psychological variables) in certain situations such as public organizations in Laos, since variables such as burnout and stress of public officials have been the main focus in previous studies (e.g., J. Kim, 2015; Siu et al., 2015).
POS also mediated the relationship between PJ and EE. These results confirm the findings of Loi et al. (2006) and Javed and Tariq (2015) on the mediating role of POS. However, POS did not mediate the relationship between PJ and TI, and POS and EE did not mediate the relationship between PJ and TI. These findings are not consistent with Loi et al. (2006), who identified a mediating role of POS between PJ and TI. They also do not support Al-Shbiel and colleagues' (2018) and Saks's (2006) findings, which emphasized the dominant role of EE as a mediator. They assumed that there was no conclusive link between EE and TI because employees in public sectors might leave the organizations for numerous reasons such as the office environment, supervisors, colleagues, family, finances, and health conditions. Therefore, it is necessary to find a direct way to lower TI such as increasing job satisfaction or promoting stability.
Implications
Our findings have several implications for HR development professionals and officers in Lao public organizations. The research results indicate that public organizations need to increase PJ and POS at the organizational level to enhance EE and lower TI. To increase PJ, public officials must recognize the organizational fairness within the organization. In particular, officers have to receive reasonable compensation for their efforts and contributions to encourage a stable workforce. The organizations need to strive to ensure that officers feel that there is fairness within the organization, and officers should faithfully perform their duties and roles to enhance organizational performance. In addition, even if the results or procedures of compensation are somewhat unsatisfactory due to the nature of public organizations, sufficient pre- and postinformation needs to be provided with emotional exchange or considerate communication to increase EE and lower TI. More specifically, prior explanations, such as promotions, educational opportunities, work evaluations, and consideration and respect, including organizational and emotional support will lead to greater satisfaction for public officials and promote loyalty to the organization.
Limitations and Future Research
There are several limitations and suggestions for future research. First, the data were collected from public organizations in Laos using a convenience sampling approach, which restricts the generalizability of the results. Future research should invite a larger number of study targets in more diverse contexts. Second, although the sample size was relatively large, this study did not consider the proportions of the statistical backgrounds of the sample. Among the research sample, 77.1% of the participants were older than age 30, 68.9% were married, 58.6% worked for local authorities, and 63.7% worked in three sectors—public health, education, economics, and finance—which may limit the generalizability of the results. To obtain more precise results and increase the generalizability of the study findings, future research could consider a stratified sampling approach. Third, this study analyzed quantitative data using one data collection method, a self-report questionnaire. Future studies should employ various data collection methods using a mixed-method approach to increase the validity of the research findings. Another limitation of this study is the noninclusion of other types of justice such as distributive justice and interactional justice, which could also be relevant factors that affect POS, EE, and TI. These types of justice need to be considered in future work. Finally, many previous studies have reported that POS affects TI, and EE affects TI, but no statistically significant relationship was derived in this study. Future studies should search for mediating variables or moderating variables that can explain the relationship between POS and TI and between EE and TI to identify the mechanism(s) to lower TI. For example, leader member exchange could be tested as a mediating variable, and age, gender, and tenure could be moderating variables of the relationships in future studies. Despite these limitations, our findings add an important nuance to previous studies.
Hypothesized Research Model
Full Model With Standardized Path Coefficient Estimates
Modified Full Models With Standardized Path Coefficient Estimates
Contributor Notes
DR. THIPPHAVANH AROUNLEUTH is a researcher. She has a Ph.D. in human resource development (HRD) from the Graduate School of HRD at the Korea University of Technology and Education. Her research interests include employee engagement, organizational justice, and organizational performance. Email: tingnoy2@hotmail.com
HYUNJEONG JO is a doctoral student in the Graduate School of Human Resource Development at the Korea University of Technology and Education. Her research interests include employee engagement, leadership, perceived organizational support, and career development. Email: akrinan@naver.com
DR. WOOCHEOL KIM is an associate professor in the Department of Human Resource Development (HRD) and the Graduate School of HRD at Korea University of Technology and Education. Dr. Kim has a Ph.D. in workforce education and development with an emphasis on HRD & organization development at Pennsylvania State University. His research interests include employee engagement, leadership, technical and vocation education and training, training performance/transfer, and career development in organizations. Email: kwccwk97@koreatech.ac.kr
DR. JUNGWON KIM (corresponding author) is a research professor at the Research Institute of Human Resource Development Policy, Korea University. She has a Ph.D. in Adult Continuing Education at Korea University. Her research interests include competencies, career development, and lifelong education. Email: jkim829@naver.com
This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2020S1A3A2A02091529).
