Work-life balance practices and turnover intention: The mediating role of person-organization fit

The higher educational institutions in Pakistan are facing challenges to reduce the turnover of their faculty members. To date, research has failed to explore the antecedents of faculty members’ turnover intention, especially in the context of Pakistan. The purpose of this study was to investigate work-life balance practices and person-organization fit as antecedents of faculty members’ turnover intention. This study hypothesized that work-life balance practices and person-organization fit are negatively related to intention to leave the job. Furthermore, the study proposed that person-organization fit mediates the negative relationship between work-life balance practices and intention to leave the job. This study adopted a cross-sectional design and a quantitative approach. The data was collected from a total of 250 faculty members from the public sector colleges in Baluchistan, Pakistan, through self-administered questionnaires. This study used partial least square structural equation modeling (PLS-SEM) for the analysis of measurement and structural model. The results revealed that work-life balance practices and person-organization fit had a direct negative and significant effect on employees' intention to leave the job. Furthermore, the results also showed that person-organization fit mediated the relationship between work-life balance practices and intention to leave the job. This study contributes to work-life balance practices and person-organization fit literature by examining the mediating processes through which work-life balance practices and intention to leave are linked. These findings provide important implications for human resource professionals to reduce the turnover of faculty members. Limitations and directions for future research are discussed.


Introduction
*The academic work environment is characterized by poor working condition, job dissatisfaction (Kinman and Jones, 2008), workload (Almer et al., 2016), long working hours, high job demands (Watanabe and Falci, 2016) and occupational stress (Opstrup and Pihl-Thingvad, 2016). Such an environment often creates an imbalance between work and family lives and may result in undesirable outcomes such as stress (Opstrup and Pihl-Thingvad, 2016), and turnover intention (Barnes et al., 1998;Kinman and Jones, 2008). Turnover intention or intention to leave refers to the subjective probability of leaving the job in the near future. Prior empirical research and theories posit that employees' intention to leave the job is the significant predictor of actual turnover (Rubenstein et al., 2018), hence, the present study focuses on intention to leave rather than actual turnover.
The turnover of the academicians has become a global issue (Vekeman et al., 2017) because the turnover of academics not only affects the productivity of the organisations but also results in the disruption of the teaching and research activities and students learning. Owing to such grave implications, understanding the factors that may contribute to the academics' turnover intention would benefit organisations (Johnsrud and Rosser, 2002).
One particular factor that is associated with improvement in the academic work environment and academics' retention is Work-Life Balance Practices (WLBP). Work-life balance practices, which include part-time work, flexible-timing, telecommuting, paid parental leave and job sharing (De Cieri et al., 2005) are envisioned to enhance the autonomy of the workers in maintaining balance between their work and family lives (Felstead et al., 2002). Numerous scholars have shown that WLBP are significant predictors of employees' performance (Cegarra-Leiva et al, 2012), job embeddedness (Thakur and Bhatnagar, 2017), and career advancement (Straub, 2007). Furthermore, research in a non-academic work setting shows that WLBP is significantly related to employees' turnover intention (Allen, 2001;Thakur and Bhatnagar, 2017).
However, the mechanism linking WLBP to intention to leave remains unclear and underdeveloped (Allen, 2001). Most of the existing research demonstrates the direct relationship between WLBP and turnover intention (Chang et al., 2017). A few of the studies focusing on intervening mechanism have shown that work-life balance supportive culture (Cegarra-Leiva et al., 2012) and job embeddedness (Thakur and Bhatnagar, 2017) explain the intervening mechanism between WLBP and turnover intention. Thus, we know little how this relationship works. As a result, subsequent researchers have called for studies to investigate the intervening variables in the relationship between organizational practices and organizational outcomes (Banks and Kepes, 2015).
The goal of this paper is, thus, to examine the additional intervening factors in WLBP-intention to leave relationship. Recently, Person-Organization Fit (POF) has received considerable attention as a mediating variable between different concepts. For example, researchers have found that POF mediated the relationship among employees attitude, behaviour and their respective antecedents such as human resource management practices (Kooij and Boon, 2018) and person-vocation fit (Vogel and Feldman, 2009). However, to the best of researchers' knowledge, so far none of the studies have investigated the mediating role of POF between WLBP and intention to quit the job. To do so, we rely on Social Exchange Theory (Blau, 1964) that has been used to investigate WLBP effects on employee's attitude and behaviour. Thus, drawing on social exchange theory and previous empirical evidence, we suggest POF as a mediator to further explore the underlying mechanism between WLBP and turnover intention.
In doing so, this paper advances our understanding of WLBP and turnover literature in the following ways. First, by considering the impact of WLBP on faculty members' intention to quit the job, this research contributes to the amount of research conducted in the academic work environment, which is a promising area of inquiry still underdeveloped. Second, by examining the mediating mechanism in the relationship between WLBP and intention to leave, this study addresses the call of the researchers to investigate the influence of mediating variables on the relationship between organizational practices and their respective outcomes (Banks and Kepes, 2015). Our study is also important as it examines novel antecedents, such as WLPB, person-organization fit and well-studied employees' attitude intention to quit the job. The study also has methodological significance by applying PLS_SEM for hypotheses testing.

Theory of social exchange
The present study is hypothetico-deductive, and used the Social Exchange Theory (SET) by Blau (1964) to develop the hypotheses. "Social exchange theory is among the most influential conceptual paradigms for understanding workplace behaviour" (Cropanzano and Mitchell, 2005). This theory elucidates that the relationship between two parties depends on social and material exchanges. In an employment relationship, the social and material exchanges in organizational practices are the provisions of WLBP that the employees expect from the organization. On the other hand, the social exchanges that the organization expects from employees are their affirmations to the norms, culture, mission, and values of the organization and intention to stay. From a social exchange perspective, when parties privy to the relationship are mutually satisfied with exchanges, the relationship will continue and vice versa (Gutierrez et al., 2012). To illustrate, when faculty members perceive WLBP positively, they will feel that they are being valued by the organization and, in return, the faculty members will reciprocate organization with positive work attitude such as exhibiting fit with the organization and reducing their intention to leave the job.

Work-life balance practices (WLBP) and the intention to leave
WLBP refers to the organization's effort to reduce work-life conflict among employees and enable them to be more productive in multiple roles (Lazar et al., 2010) and enhance their self-sufficiency in integrating their work and non-work lives (Felstead et al., 2002). Increasingly, organisations have created WLBP as part of their effort to enhance employee's commitment and job satisfaction (Parakandi and Behery, 2016) and reduce their turnover (Allen, 2001;Thakur and Bhatnagar, 2017). WLBP also act as a coping mechanism to maintain a balance between work and non-work demands (Chiang et al., 2010). As mentioned above, the academic work-life is characterized by high job demands, long working hours, stress, and resource constraints (Kinman and Jones, 2008;Opstrup and Pihl-Thingvad, 2016), which results in negative outcomes such as burnout (Padilla and Thompson, 2016) and job-related stress (Opstrup and Pihl-Thingvad, 2016). Prior research has demonstrated that organization can reduce faculty members burnout (Padilla and Thompson, 2016) and turnover (Allen, 2001) through provision of WLBP (e.g., flexi-time, telecommuting) because such practices provide employees with the flexibility and autonomy in balancing their work and non-work lives (Onken-Menke et al., 2018). In contrast, lack of WLBP such as long working hours reduces employees' time to spend on social and family activities (Chiang et al., 2010), negatively affect WLB (Adisa et al., 2017) and result in stress and turnover intention (Allen, 2001). In sum, a review of the literature and previous studies demonstrate that WLBP is an important element of employees' turnover intention. To further explore the relationship between WLBP and intention to leave in academic work settings and Pakistan's context, the following hypothesis is proposed:

H1:
There is a negative relationship between WLBP and faculty members' intention to leave the job.

Work-life balance practices and personorganization fit (POF)
POF refers to the congruence between people and organization (Andela and van der Doef, 2018), that is the extent to which people share similar values (Chatman, 1991) and goals (Schneider, 1987) with the organization. A core assumption of personorganization fit is that employees tend to stay with an organization that shares similar values and goals (Schneider, 1987;Vogel and Feldman, 2009). Values are the common characteristics between organization and person (Finegan, 2000). "Individual values are enduring beliefs that a specific mode of conduct or end-state is preferable to its opposite" (O'Reilly et al., 1991). Organizational values "are evaluative standards relating to work or the work environment by which individuals discern what is 'right' or assess the importance of preferences" (Dose, 1997). Furthermore, organizational values are not merely slogans; they can be inculcated among employees (Chen and Chang, 2010). An organization may inculcate organizational values and goals among employees through the provision of organizational practices. A number of studies have shown that organizational practices such as training and development opportunities (Boon et al., 2011;Luu, 2018) selection and socialization (Chatman, 1991) and human resources management practices (Mostafa and Gould-Williams, 2014) are the means through which organization can align its norms and values with employees. Furthermore, prior research has demonstrated that employees' positive perception of human resource management practices enhances the fit between person and organization (Boon et al., 2011;Kooij and Boon, 2018). However, in general, these studies have not taken into account academic's perception of WLBP and how it affects their fit with the organization.
Following this, we argue that WLBP is an important determinant of academic members' POF. To support this argument, we draw on social exchange theory (SET) (Blau, 1964) and the norms of reciprocity (Gouldner, 1960).
The theory of social exchange posits that employees give great value to the favorable treatments from the organization. Provision of work life balance practices are indicative of the organization's favorable treatments of their employees. Thus, when individual receive favorable treatments from the organization, they are likely to reciprocate with favorable work outcomes (Gouldner, 1960) such high job engagement (Markos and Sandhya, 2010), and lower intention to leave the job (Samad and Saufi, 2017;Kakar et al., 2018). Thus, we expect that academics' positive perception of WLBP will develop a sense of feeling among academics that they are being favorably treated by the organization. In return, the academics will reciprocate the good deeds of the organization with positive work outcome such as affirming to the organization's culture, norms, and values: H2: There is a positive relationship between worklife balance practices and person-organization fit.

Person-organization fit and the intention to leave
The Schneider's (1987) Attraction-Selection-Attrition (ASA) model forms the basis of personorganization fit. The ASA model posits that individuals are attracted to, selected by and stay with an organization that matches their characteristics. In the attraction and selection stage of the ASA model, people who share similar values and goals are recruited and selected by the organization. Besides, new recruits with similar values to those of an organization are more satisfied with the job and more likely to stay with the organization longer (Chatman, 1991). More recent studies have further substantiated the benefits of high of level of POF for employees performance (Hamstra et al., 2018), organization commitment (Kooij and Boon, 2018), job satisfaction (Jin et al., 2018), and strengthening the membership of the employees with the organization (Boon and Biron, 2016). Finally, the attrition stage of the ASA model suggests that employees who do not fit the organisations are more likely to think of leaving the job.
Prior research has demonstrated that the perception of the low level of person-organization fit has negative consequences on employees attitudes and behaviors, including job dissatisfaction and turnover intention (Vogel and Feldman, 2009;Boon and Biron, 2016;Chhabra, 2016;Andela and van der Doef, 2018). Thus, drawing on the attractionselection-attrition model, we posit that academics who share similar values and goals with the organization are less likely to leave the job: H3: There is a negative relationship between personorganization fit and intention to leave the job.

The mediating role of person-organization fit (POF)
In addition to the direct linkage between WLBP and turnover intention, we expect an indirect WLBPintention to leave relationship via personorganization fit. As mentioned above, WLBP relates positively to person-organization fit, and, POF relates negatively to intention to leave the job. Thus, it is logical to predict that person-organization fit mediates the influence of WLBP on intention to leave the job. Presbitero et al. (2016) asserted that organizational practices not only reflect the values of the organization but also tend to reinforce these values. Furthermore, organizational practices are the means through which organization's norms and values are communicated to the employees (Farndale and Sanders, 2017). From social exchange perspective, when employees perceive organizational practices to be present, they are more likely to understand the norms and values of the organization (Boon et al., 2011;Kooij and Boon, 2018), which, in turn increases their fit and commitment with organization (Kooij and Boon, 2018) and reduces their intention to leave the job (Boon et al., 2011;Andela and van der Doef, 2018). Thus, we predict that POF acts as a mediating mechanism in the negative relationship between WLBP and turnover intention. Although several studies have found that POF was mediating the relations among intention to leave and their respective antecedents (Vogel and Feldman, 2009). However, to the best of present researchers' knowledge, no study has so far examined the mediating role of POF between WLBP and academic members' intention to leave the job. Thus, to fill this gap in the literature, we propose that:

H4:
The negative relationship between WLBP and intention to leave is mediated by personorganization fit.

Sample and data collection
The sample for this study was derived from fulltime employees of public sector colleges in Baluchistan, Pakistan. The respondents include faculty members only. Nineteen (19) colleges were selected based on convenience sampling. Within an organization, convenience sampling was used to select the respondents. A total of 450 paper-based questionnaires were self-administered among faculty members out of which 250 (55%) cases were returned (excluding cases with suspicious responses and missing values). Out of 250 respondents, 60.8% were male and 39.2% were female, including 77.2% lecturers, 12.8% assistant professors, 8 % associate professors, and 2% professors respectively. Demographic profile is shown in Table 1.

Research instruments
All the instruments including WLBP, personorganization fit, and intention to leave were measured on five-point Likert scale ranging from 1 ("strongly disagree") to 5 ("strongly agree"). For the measurement of work-life balance practices, 4 items were adopted from (Zheng et al., 2016), an example of the items is ''My organization provides flexible working arrangement''.
To measure person-organization fit, five items were adapted from (Vogel and Feldman, 2009) and the sample items include "The things that I value in life are very similar to the things that my institution values''. Five items for the measurement of intention to leave were adopted from (Cennamo and Gardner, 2008;O'Reilly et al., 1991). An example of items is "Thoughts about quitting this organization cross my mind".

Descriptive statistics
The mean, standard deviation, and correlations among variables are provided in Table 2. The correlations between the variables of the study are significant and range between -0.477 and 0.464. The correlation is significant and negative between WLBP and intention to leave (r= -0.226**), POF and intention to leave (r= -0.268**), and positive between WLBP and POF (r= 0.464**).

Measurement model analysis
For the analysis of the measurement model (the relationship between the latent constructs and its observable indicators), PLS-SEM was used. PLS-SEM is appropriate for studies that are predictions oriented. In the present study, the latent variables and their indicators or items were assessed for reliability and validity. The common measures of indicators and constructs reliability are Factor Loadings (FLs), Cronbach Alpha (CA) and Composite Reliability (CR). The results revealed that the FLs of all indicators (Table 3) were above 0.70 thresholds (Hair et al., 2011;2016). Thus, all the indicators were accepted. The results of CA and CR, which measure the internal consistency reliability of the constructs revealed that all the constructs exhibited internal consistency reliability and the values of CA and CR were beyond 0.70 thresholds (Cohen, 1988) (Table 3). The constructs of the study were also subject to validity assessment, and the common measure of constructs validity are Average Variance Extracted (AVE), Fornell-Larcker Criterion and heterotraitmonotrait ratio (HTMT). AVE, a measure of convergent validity, explains the amount of variance in latent constructs (Chin and Dibbern, 2010). All indicators (Table 3) explained more than 50% variance in the constructs (Chin, 2010). Furthermore, Fornell-Larcker Criterion (Fornell and Larcker, 1981) and HTMT were used to verify the discriminate validity of the constructs. The results of Fornell-larcker criterion (Table 4) revealed that square root of AVE of the constructs exceeded the value of estimated correlations of a construct with other latent variables of the study (Chin, 2010;Hair et al., 2011), thus confirming the discriminate validity of the constructs. In addition, an advance measure of discriminate validity (HTMT) was used (Henseler et al., 2015). The value of HTMT was less than the threshold value of 0.85 (Hair et al., 2016), thus affirming the discriminate validity of the constructs (Table 5).

Structural model assessment
The structural model evaluates the hypothesized relationship among latent constructs (Hair et al., 2016). The important criteria for the evaluation of the structural model are coefficient of determination (R 2 ), predictive relevance (Q 2 ), effect size (f 2 ) and the path coefficient (Hair et al., 2016). R 2 reflects the amount of variance that a model can explain in the endogenous constructs. The value of R 2 equal to 0.67, 0.33 and 0.19 are considered substantial, moderate and weak, respectively (Chin, 1998). The findings revealed that the model explained 37.60% (R 2 = 0.376) variance in intention to leave and 20.90% (R 2 = 0.209) in POF. The model was also assessed for predictive relevance (Q 2 ). The predictive relevance Q 2 of the model was 0.214 for intention to leave and 0.118 for POF, indicating that the model has sufficient predictive relevance. The effect size (f 2 ) of WLBP and POF on intention to leave to was moderate and ranged between 0.017 and 0.076. However, WLBP had a substantial effect on POF (f 2 = 0.265). Furthermore, the significance of path coefficient (β-values and t-values) was assessed using bootstrapping (5000 subsamples and two tailtest). Fig. 1 describes the total direct impact of WLBP on turnover intention. The findings of the study showed that WLBP is negatively related to intention to leave the job (c= -0.272**, t= 5.156), providing support for H1. Model 2 (Fig. 2) was run with a mediator (5000 bootstrap-re-sampling) and significance of the path coefficient of the mediating effect was estimated using percentile bootstrap, which generated 95% Confidence Interval (CI) for indirect paths (Williams and MacKinnon, 2008). The path "a", which shows a positive relationship between WLBP and POF was significant (a= 0.457**, t= 8.214), thus, H2 is supported. Similarly, path "b1", shows a negative and significant association between POF and ITL (b1= -0.219**, t= 2.826), providing support for H3. Therefore, the significance of the path "a" and path "b1" and "c" suggest the potential indirect effect of WLBP on ITL via POF. With the introduction of the mediator, the direct effect of WLBP on intention to leave loses its importance and become non-significant (c'= -0.143, t = 1.692). Thus, H4, which shows the mediating role of POF is supported (a×b1= -0.100**, t= 2.627). Consequently, these results show full mediation, since the direct relationship between WLBP and intention to leave (path "c") become non-significant (Baron and Kenny, 1986).

Discussion
The main objective of this research was to analyze the relationship between faculty members' perception of work-life balance practices and intention to leave the job via person-organization fit. Such an objective offers practitioners an improved understanding of the factors that may lead to the academic's turnover and provides some insights to the organisations on how to improve the compatibility of faculty with the organisations and reduce their turnover intention. This is important because fit with the organization was found to be negatively related to the employees' willingness to leave the schools (Vekeman et al., 2017).
Drawing on the social exchange theory, we demonstrated that faculty members' perception of WLBP were significantly related to their intention to leave. This result is in line with the studies conducted in different work setting (Nie et al., 2018;Thakur and Bhatnagar, 2017), showing that WLBP is a significant predictor of turnover intention. The results also show a positive and significant relationship between WLBP and POF and this result complements (Boon et al., 2011;Kooij and Boon, 2018) findings that organization practices lead to the fit between person and organization. In addition, the findings revealed that academic members' fit to the values and goals of the organisations is negatively related to their turnover intention. This result affirms that the compatibility of an individual with the organization goals and values is a significant predictor of their intention to leave the job (Kristofbrown et al., 2005;Vogel and Feldman, 2009;Andela and van der Doef, 2018). Furthermore, the results also revealed that POF fully mediated the relationship between WLBP and intention to leave. This result indicates that academics' perception of WLBP affects POF, which, in turn, negatively influences their turnover intention and reaffirm that fit is a significant mediator between organizational practices and intention to leave (Boon et al., 2011;Mostafa and Gould-Williams, 2014;Presbitero et al., 2016).
This study added several new contributions to the existing WLBP and fit literature. First, this study extends our understanding of the effect of WLBP on person-organization fit and turnover intention, which, thus far has rarely been investigated, specifically in academic work settings. Furthermore, this study added a completely novel insight to the existing literature by exploring the mediating role of POF in the relation between work-life balance practices and intention to leave. In doing so, we complement human resource literature by examining that the perception of WLBP can facilitate POF and reduces their intention to leave the job. We also respond to (Banks and Kepes, 2015) suggestions that organizational practices and organizational outcomes relationship are mediated by employees attitude.
The findings of this study provide several important practical implications for the manager. In order to reduce the turnover of the academics, public sectors organization should formulate and implement work-life balance practices that build the alignment between academics and organization's values and goals, thereby leveraging personorganization fit among academics. WLBP should not only purely enhance academic ability in managing multiple roles, but also instill organizational values and goals in their mindsets. Provisions of WLBP, which increases the academics' compatibility with the organization, may further mitigate the negative feeling of leaving the job. If the WLBP are not effective enough to instill organizational norms and goals and enhance the fit between academic and organization, the potential benefit of WLBP may not be fully utilized since their effect may not be fully translated into intention to leave the job without enhancing POF. Thus, the management should not only recruit an individual that share similar values and goals with the organization but also ensure provisions of WLBP that infuse organizational values among academics that would reduce their intention to leave the job.

Limitations and future directions
Despite a thorough investigation of previous literature and theories on turnover, the findings of the present study are subject to some limitations. First, the conceptual framework of the present study was examined with cross-sectional data, which may limit the generalizability of the study. Future researchers are encouraged to test the model with longitudinal data. Second, the study is limited to organizational practices that address the work and family life, thus, we encourage future researchers to broaden the model by incorporating a comprehensive set of organizational practices such as training, development, compensation, and performance related HRM practices. Furthermore, we investigated POF as a mediating variable. We encourage future researchers to investigate other organizational factors such as person-job fit, organizational support, and work-life balance.