The Mediating Role of Person-Job Fit between Person-Organisation Fit and Intention to Leave the Job: Empirical Evidence from Pakistan

The turnover of faculty members is a pressing problem that has adversely affected the performance and productivity of higher education institutions. The purpose of this study is to examine the direct effect of person-organisation fit (POF) and person-job fit (PJF) on faculty members’ turnover intentions. Furthermore, the study explores the mediating role of PJF between POF and turnover intention. Data collected from faculty members of public sector colleges (n = 250) were analysed using partial least squares structural equation modelling. The results, contrary to our expectations, revealed that POF is not a significant predictor of faculty turnover intention. However, the results provided support for the negative relationship between PJF and turnover intention. Additionally, as predicted, the results confirmed that PJF mediates the relationship between POF and turnover intention. The findings also revealed that the reliability and validity of the adopted/adapted scales change in the context of different countries and organisational settings. The theoretical and practical implications, research limitations and directions for future research are provided.


Introduction
Over the years, the turnover of faculty members has been the focus of discussion among practitioners and researchers. This is because the turnover of faculty members is disruptive and costly to the organisation, involving both monetary and non-monetary costs. Schloss [1] reported that the cost of faculty turnover is $400,000 in the USA, which includes the cost of hiring and recruiting new faculty. In addition to monetary cost, high faculty turnover adversely affects organisation productivity [2], disrupts the quality of education, and impedes students' learning and research activities [3,4]. Moreover, the turnover of faculty members also reduces the motivation and morale of existing academics [4,5]. Thus, owing to the negative implications of academic faculty turnover, it is viable to investigate the factors that cause faculty members' to consider leaving the job.
One of the factors that have been consistently linked with employees' turnover intentions is person-organisation fit (POF). POF is defined as the extent to which employees' personal values and goals are identical with the values and goals of the organisation [6]. When employees perceive compatibility between themselves and their organisation (i.e., POF), they become more attached to the organisation and intend to remain within the organisation [7][8][9]. Employees also exhibit job satisfaction [10], organisation commitment (OC) [11], and organisational citizenship behavior (OCB) [12,13] when they perceive a strong fit between themselves and an organisation. In contrast, poor POF results in negative work outcomes such as intent to leave the organisation [14][15][16].
Considering the indispensable role of POF, many researchers have explored it in various industrial settings, such as the health [14], insurance [17], and tourism industries [7]. Nevertheless, the concept of POF has received little attention in educational institutions (for an exception see, Jin et al. [9]. Jin et al. [9] conducted an empirical study among a sample of faculty members and found POF to have a significant negative effect on turnover intention. However, Jin et al. [9] conceptualised POF with values congruence only, whereas we use values and goals congruence to conceptualise POF, thus providing a more robust conceptualisation of fit. Additionally, although the beneficial direct effect of POF cannot be undermined, researchers have called for further studies to explore the underlying mechanisms that may transform the influence of POF on turnover intention [9]. Prior scholars have explored a myriad of factors that may mediate the influence of POF on turnover intention. For instance, scholars tested whether the effects of POF on work outcomes were mediated by job satisfaction [18], job stress [15], OC [19], and work engagement [20]. Recently, person-job fit (i.e., congruence between an individual and his or her job's characteristics) has received considerable attention as a mediating variable between employee attitudes, behaviours, and their respective antecedents [17,21]. However, previous research on the mediating role of person-job fit (PJF) between POF and turnover intention is limited. Thus, despite the extensive amount of studies on POF, further research into the underlying mechanisms that link POF and essential work outcomes, such as turnover intention, is needed.
Furthermore, current understanding of the causes and consequences of POF is based on research conducted in non-academic work settings. The values and goals of academic institutions are different to those of non-academic organisations; for example, educational institutions, especially public sector institutions, are non-profit seeking, while other organisations value profit. Therefore, the findings of studies conducted in non-academic work settings cannot be generalised to academic institutions. Additionally, the literature on POF is mostly dominated by studies from Western and developed countries, whereas only a few studies are available from Asia in general and Pakistan in particular. This omission is critical as several studies have shown that the findings of studies conducted in Western culture cannot be generalised in other cultural settings [12,22]. For example, in a crosscultural empirical study on fit, Astakhova [23] found a significant and positive relationship between person-supervisor fit and OC in Japan; however, the same relationship was insignificant in the USA. Thus, it is worth exploring the role of cross-cultural differences on the implications of POF. This study is aimed to fill this gap in the literature by investigating the direct and indirect effect (through PJF) of POF on faculty members' turnover intentions in the educational institutions of Pakistan.
In doing so, the study contributes to POF, PJF, and turnover intention literature in many ways. First, the study contributes to POF and PJF literature by investigating its impact on faculty members' turnover intentions. Second, by examining PJF as a mediator, the study explores the psychological mechanisms underlying the influence of POF on turnover intention. Third, following Huang et al.'s [22] assertion that studies conducted in a Western context cannot be effective in different cultural settings, this study adds to the literature by investigating the implications of POF and PJF in the context of Pakistan. Fourth, by validating the measurement scales of POF and PJF in different cultural settings, this study also has methodological contributions.

Person-Organisation Fit and Person-Job Fit
POF has been conceptualised as the match between the individual and his or her organisation. POF can be measured using either person-organisation values fit, whereby individual values are compared to the organisation, or person-organisation goals fit, whereby employees' goals and objectives are compared to the organisation [24,25]. On the other hand, person-job fit (PJF) is defined as the match between the person and job's characteristics [26]. In other words, PJF occurs when the knowledge, skills, and abilities (KSAs) of a person are compatible with the job demands and resources (e.g., pay, fringe benefits, allowances, and working conditions), thus fulfilling the physiological and psychological needs and preferences of the employees.
Although POF and PJF are somewhat closely linked, the former puts much more emphasis on the congruence of a person with an organisation, while the latter emphasises the job's characteristics. Lauver and Kristof-Brown [27] also sought to differentiate between the two concepts by establishing the associations of POF and PJF with important work outcomes such as turnover intention and job satisfaction. They observed that POF was significantly associated with organisational level outcomes, such as organisation commitment and performance, while PJF's relationship with organisational outcomes (e.g., task performance) was weak. Additionally, the interrelationship between POF and PJF was also weak. These findings suggest that both POF and PJF are conceptually different from each other.

Turnover Intention
Turnover intention is defined as the likelihood of leaving a job in the near future. It is considered to be the last stage of withdrawal cognition and a significant predictor of actual turnover [30,31]. In their meta-analysis study, Rubenstein et al. [32] found a positive and significant correlation of 0.56 between turnover intention and actual turnover. Despite significant correlation between the two, turnover intention and actual turnover are theoretically and conceptually different from each other. For instance, actual turnover is time-specific and represents the physical separation of employees from an organisation [33], while turnover intention is subjective and reflects employees' willingness to leave the job in the near future [34,35]. Additionally, in the literature, researchers have utilised intention as a proxy of actual turnover for a number of reasons. First, turnover intention is subjective and amenable to measures [36], while actual turnover is dichotomous and requires longitudinal data. Second, theories and literature suggest that intention is the single best predictor of actual turnover [32,[37][38][39]. Hence, this study focuses on the turnover intention of faculty members.

Hypotheses Development
In this study, we hypothesise that POF relates positively to PJF and negatively to the turnover intention of academics. We draw on the Attraction-Selection-Attrition (ASA) model to justify the influence of POF on turnover intention [40]. Likewise, the relationship of POF with PJF is explained by spillover theory [41].
A core assumption of the ASA framework is that individuals are attracted to and selected by organisations whose members' attributes, such as values, personality, and goals, are identical to their characteristics [40]. However, individuals who do not share identical characteristics with the organisation tend to quit the job [40]. There is also empirical evidence from non-academic work settings that people who do not feel compatible with the values and goals of the organisation quit the job. For instance, Vogel and Feldman [7], in an empirical study on employees working in restaurants, found that POF had a negative impact on turnover intention. Rurkkhum [42] showed that employees' congruence with an organisation negatively influenced their intention to leave the job. Further, Andela and van der Doef [8], in an occupationally heterogeneous sample, found that POF was related negatively to turnover intention. Along the same lines, Ribando and Evans [43] found a significant and negative relationship between POF and turnover intention among prospective employees in Georgia. Also, Jin et al. [9], in a study among academics in a public university in the United States, found that POF had a direct negative impact on turnover intention. However, they did not operationalise POF with values and goals congruence.
Following these theoretical and empirical arguments, we propose that: Hypothesis H1. Person-organisation fit is negatively related to intention to leave the job.

Person-Job and Turnover Intention
The ASA model can also be applied to explain the influence of PJF on turnover intention. As stated above, organisations not only look for employees who match their characteristics but also who possess the capabilities to meet the job requirements [25,26]. On the other hand, employees are interested in joining organisations whose resources are sufficient enough to meet their needs and preferences [18,44,45]. In simple words, employees prefer to stay with the organisation when their KSAs are adequate to meet the job's requirements and the job supplies are sufficient to meet their needs and preferences. For example, Vogel and Feldman [7] found that restaurant workers were more willing to stay with the organisation when their KSAs were compatible with the job requirements and the job resources met their needs. Andela and van der Doef [8] also found that PJF had a negative influence on turnover intention. We expect similar results (in the context of higher education institutions (HEIs) and Pakistan) when it comes to the relationship between faculty members' perception of PJF and turnover intention (i.e., a higher PJF will lead to lower faculty turnover intention).
Hypothesis H2. PJF relates negatively to academics' intention to leave the job.

Person-Organisation Fit and Person-Job Fit
We also propose that POF is positively related to PJF fit, and this relationship is explained by spillover theory. Spillover theory suggests that fit in one aspect may influence fit in another aspect [41]. When an individual experiences fit in one aspect, then it is likely that he or she may strive to modify other aspects that are a poor fit or misfit. For example, if an individual experiences strong fit with the job's resources but poor fit with the job's requirements, over time the individual is likely to reduce the misfit by either changing his abilities, knowledge, and skills to match the job requirements, or working to change the requirements of the job. Recent fit researchers also found that there is a positive spillover of fit in one aspect (e.g., POF) on needs-supplies fit and demands-abilities fit [18]. Boon and Biron [45] reported that, on the one hand, an individual with a higher degree of POF adapts, develops and applies his/her KSAs to match the job's requirements; on the other hand, a higher degree of POF enables employees to communicate their needs, desires, preferences, and thoughts to organisations, so organisations can modify job supplies to meet employees' needs and preferences. Hence, following spillover theory and related literature, we predict that a person with a higher degree of POF is likely to experience PJF.
Hypothesis H3. POF is positively related to PJF.

Person-Job Fit as a Mediator
Thus far, we have suggested that high POF is likely to lower faculty turnover intention (Hypothesis 1) and enhance their perception of PJF (Hypothesis 3). We also predict that PJF relates negatively to turnover intention (Hypothesis 3). Given hypotheses 1-3, we posit that PJF mediates the influence of POF on intention to quit the job. In addition, scholars such as Boon and Biron [45] have observed that when employees feel compatible with the values and goals of the organisation, they tend to adapt or develop KSAs to meet the job requirements and convince the organisation to meet their needs and preferences. Accordingly, when employees believe their KSAs are compatible with job requirements, and job resources are adequate to meet their needs and preferences, they feel satisfied with the job and intend to remain within organisations [7,8,18,26]. In contrast, in the case of a poor match between an individual and job characteristics (i.e., a poorer PJF), adverse work outcomes such as stress, emotional exhaustion, and turnover intention will ensue [7,8,45]. Thus, we predict that a high perception of POF strengthens PJF and, in return, PJF reduces academics' turnover intention; therefore, we propose that (Figure 1): Hypothesis H4. The impact of POF on turnover intention is mediated by PJF.

Sample and Data Collection
The target respondents of this study were full-time academics in public sector colleges in Baluchistan, Pakistan. The sample size was calculated using Faul et al. [46] G*power 3.1.9.4 software. Using 0.15 effect size, (1-β) = 0.80 power, 0.05 significant level, and three predictors, the minimum required sample size for the study was 74. The sample from the population was drawn using convenience sampling, i.e., approaching participants that were easily accessible [47]. To elicit information from the sample, 450 questionnaires were self-administered among faculty members. Initially, 250 surveys were returned, indicating a response rate of 55%. Out of the 250 participants, 61% were male, and 39% were female. Participants were categorised into four groups according to their designation. Out of the 250 respondents, 72% (n = 180) were lecturers, 14.8 (n = 37) were assistant professors, and 8% (n = 20) were associate professors. Professors constitute only 5.2% (n = 13) of the sample. The average age of the respondents was 38 years (minimum age = 22 years, and maximum age = 67 years). The average working experience of the faculty members was 7 years.

Research Instruments
Research instruments for the present study were adopted from the literature. These adopted scales have been used in the past [7,45,48] and were found to be valid and reliable. All items of the constructs were measured on a Likert scale of 1 to 5. The scale ranged from "strongly disagree" = 1 to "strongly agree" = 5.
POF was measured with six items. These items were adopted from Cable and DeRue [26] and Vogel and Feldman [7]. An example item of the scale is: "My personal values match my organisation's values and culture." Likewise, Cable and DeRue's [26] six items scale was used for the measurement of PJF. These items represent the congruence of individual ability and needs with the demand and resources of the job. The sample item of the scale is: "My abilities and training are a good fit with the requirements of my job." Finally, scales were adopted from O'Reilly et al. [49] and Cennamo and Gardner [50] for the measurement of turnover intention. An example item of the scale includes: "Thoughts about quitting this job cross my mind." Since this study is cross-sectional, therefore, the possibility of common method variance (CMV) was assessed using Podsakoff et al.'s [51] procedural remedies and Harman's single factor test. Significant CMV would occur if one general factor accounted for more than 50% of the variance in the items used. In this study, the variance explained by the single factor was 43.37%, which is less the minimum threshold. Thus, there is no CMV issue in this study.

Data Analysis Tools
We used the Statistical Package for the Social Sciences (SPSS 24) and PLS-SEM to examine the research model. PLS-SEM was suitable for the present study since the study aimed to explain the amount of variance in the endogenous constructs (i.e., turnover intention) [52,53]. The means (M), the deviation of the data from the means (SD), and the correlations of the study variables were calculated through SPSS 24. SmartPls3.2.8 was used to assess the reliability and validity of the measurement model. The goodness of fit of the structural model and the significance of the path coefficient was evaluated using SmartPls3.2.8 software.

Results
Means, SD and bivariate correlations between the variables of interest are provided in Table 1. The results (Table 1) reveal that POF and PJF are strongly and negatively correlated with intention to quit (r = −0.27 **, p < 0.001; r = −0.40 **, p < 0.001), respectively. The correlations of POF with PJF are significant and positive (r = 0.56 **, p < 0.001).

Constructs Reliability and Validity
To assess whether the latent constructs and items used in the survey were reliable and valid, we estimated the values of factors loading (FLs), Cronbach Alpha (CAs; α), and Composite Reliability (CR). α and CR are the measures of the constructs' internal consistency reliability, while the reliability of the items are measured by FLs. We also calculated the constructs' validity using Average Variance Extract (AVE). Furthermore, Fornell-Larcker Criterion and Heterotrait-Monotrait Ratio (HTMT) were used for the assessment of discriminate validity.
Hair et al. [54] posit that an item or indicator is said to be reliable if the values of FLs are greater than 0.70 (FLs ≥ 0.70). In the present study, all items ( Figure 2) of the constructs were reliable (FL ≥ 0.70), with the exception of the item FOF6. The factor loadings of the item (FOF6) was 0.510; therefore, item FOF6 was deleted. The FLs of the remaining POF with five items range from 0.723 to 0.800. Similarly, the FLs for PJF and intention to leave range from 0.700 to 0.846. Furthermore, the results (see Table 2) show that CAs and CRs are higher than 0.70 and in the acceptable range [54]. Additionally, all the constructs of the study ensured convergent validity since the values of AVE were greater than the 0.50 threshold [55]. The results also revealed that the Variance Inflation Factors (VIF) were less than 5; thus, collinearity was not an issue among the constructs ( Table 2).  Lastly, the constructs of the study were subject to discriminate validity. Discriminate validity represents the distinctiveness of the construct from the rest of the constructs in the model. The results of Fornell-Larcker Criterion (Table 3) and HTMT (Table 4) show that all the constructs are discriminately valid.

Structural Model
Assessment of the structural model provides an estimation of the hypothesised relationship among variables of interest. In PLS-SEM, the measures of the structural model are R-square (R 2 ), path coefficients (β), and Q-square (Q 2 ), or the predictive relevancy of the model, respectively. R 2 represents the explanatory power of the overall model. In the structural model, R 2 values of 0.75 are considered substantial, while R 2 equal to 0.50 is regarded as moderate. R 2 equal to 0.25 represents the weak explanatory power of the model [54]. In the present study, the R 2 value for an endogenous latent construct such as turnover intention was weak (R 2 = 0.180). However, the R 2 value for PJF was moderate (R 2 = 0.298).
Q-square (Q 2 ), which reflects the relative predictive relevance of each predictor construct on endogenous constructs, was calculated using the blindfolding procedure. A model is said to possess predictive relevance if Q 2 values are larger than zero [54]. In this study, Q 2 values for all exogenous constructs were larger than zero, thus confirming the predictive significance of each exogenous construct.
Finally, path coefficient (β), which shows the significance and strength of the relationship among constructs, was obtained using bootstrapping (5000 subsamples, two-tail test). The results of the structural model's path-coefficient are provided in Table 5. The findings (see Table 5) revealed that the path-coefficient between POF and turnover intention was insignificant (β = −0.076, t = 1.023, p = 0.153), thus rejecting H1. However, the path-coefficient between PJF and turnover intention was negatively significant (β = −0.379, t = 5.747, p < 0.05), providing support for H2. H3, which shows the relationship between POF and PJF, was significant and positive (β = 0.546, t = 10.437, p < 0.05).
For the assessment of the mediating effect of PJF, Preacher and Hayes [56] procedure was used. The result revealed that POF had an indirect effect on turnover intention via PJF (β = −0.207, t = 4.781, p < 0.05), thus providing support for the full mediation of PJF between POF and turnover intention.

Discussion and Conclusions
In today's uncertain and competitive business world, one of the critical success factors in retaining employees is their compatibility with the attributes of the organisation [9,57]. This study investigated the direct and indirect effects of POF (via PJF) on academics' turnover intentions. Drawing on spillover theory, the study also hypothesised the positive influence of POF on PJF, and, in return, PJF's negative influence on faculty members' intention to leave their jobs in HEIs of Pakistan. Additionally, the objective of the study was to validate the POF and PJF scale in Pakistan's academic work context. Contrary to our expectations, the results of the study revealed that POF is not a significant predictor of faculty members' turnover intentions in Pakistan. This finding is in contrast to previous literature and theories [7,8,11,58], which always provide support for the negative and significant impact of POF on employees' turnover intentions. However, the finding, to some extent, corroborates the work of Astakhova [23], who reported no significant relationship between fit (i.e., personsupervisor fit) and OC in the USA. One of the reasons for this insignificance may be the different context of the study, i.e., HEIs, since the public sector's HEIs are non-profit seeking, and their values and goals are different from other profit-oriented organisation. Therefore, our findings suggest that employees working in HEIs have a different perception of organisations than employees working in non-academic organisations. This change in employees' perception may be one of the reasons for the insignificant relationship between POF and turnover intention. Additionally, this study used both values and goals congruence for the measurement of POF, while prior studies have either used goals or values congruence only. This change in the measurement method may have resulted in the insignificant path coefficient between POF and turnover intention. We also expect that a high correlation between POF and PJF (0.56 **, p < 0.001) may have contributed to this insignificance. Furthermore, the majority of earlier studies were conducted in Western countries; therefore, we expect that Pakistan's cultural values and norms may have caused this insignificance.
Another imperative finding of this study was the significant and positive influence of POF on PJF fit. This finding suggests that faculty members who perceive a high congruence with their organisation are more likely to feel compatible with job characteristics. This result is aligned with Vekeman et al. [57] contention that teachers who fit the characteristics of the schools (i.e., organisation) were more satisfied and less willing to quit the profession. This finding is also in line with spillover theory, which maintains that perception of fit in one sphere (e.g., POF) can spillover positively into another sphere (i.e., PJF) [41]. Furthermore, the study revealed that academic members who fit the demand and resources of the job (PJF) had reduced turnover intention. This finding is in line with Badger Darrow and Behrend [10] and Vogel and Feldman's [7] studies, which report that PJF is significantly related to employees' intention to quit their job.
One interesting conclusion from this study is the mediating role of PJF in POF-turnover intention relationships. This finding suggests that the influence of POF on turnover intention is not straightforward; rather, it is mediated by PJF. This result implies that when faculty members feel compatible with the values and goals of the organisation, they will exhibit a better fit with the job, which, in return, negatively influences their intention to quit the job. Previous research also substantiates the mediating role of PJF between employees' turnover intention and their respective antecedents, including perceived organisation support [17] and person-vocation fit [7]. Thus, drawing on the findings, we suggest that there is an underlying mechanism or mediation between POF and its outcomes.
Moreover, the study objective was to validate the POF and PJF scales in Pakistan's context. Both scales for POF and PJF were adopted from the literature. Prior literature has shown that POF and PJF scales are valid and reliable [7,18]. However, contrary to the literature, our results showed that the scales used for POF are not always reliable. The findings revealed that the factor loading of one item (POF6) of POF was very low (i.e., 0.510), suggesting that item POF6 is not important. Hairet al. [54] recommend that items with a factor loading less than 0.70 are not reliable and should therefore be deleted. Hence, item FOF6 was deleted from the final analysis.
Item POF6 has yielded a reliable score in various work settings, however, in the context of HEIs and Pakistan, its loading was less than the 0.70 threshold [53]. Hence, based on these results, we suggest that the validity and reliability of the measurement scales vary across countries and organisations. Thus, we conclude that country and organisational context matter in the reliability and validity of the constructs. Therefore, it is appropriate to check the validity and reliability of a construct, even if the items (indicators) of the construct are adopted or adapted from reliable and valid sources.

Implications
This study has several theoretical and practical implications. First, it extends our understanding of how academic members' compatibility with their organisation affects their fit to the job and turnover intention. Second, it provides empirical support for the mediating role of PJF between POF and turnover intention. To our understanding, this is one of the pioneering studies that has examined the indirect influence of POF on academics' turnover intention. Third, this study provides empirical support to the ASA framework [40] that people tend to stay with their organisation when their characteristics match the organisations' characteristics.
Additionally, this study has some significant practical implications. First, the study specifies that POF and PJF are essential predictors of academics' intention to leave their job. Thus, a college principal who wants to retain faculty members should emphasise the enhancement of the fit of the organisation's values and goals with those of the faculty members. In addition, compatibility between job characteristics and faculty attributes are also important elements of faculty retention. Thus, management should not only enhance the fit between the academic and the organisation by arranging orientation and training programs but also ensure that such practices enhance the fit between faculty and job. If management practices fail to enhance the fit between faculty and job, the potential benefits of organisational practices (e.g., orientation and training) will not yield the desired results of lowering turnover intention because the effect of POF on turnover intention is transferred through PJF. Thus, a way through which colleges can retain their faculty members is to provide training and job opportunities, development opportunities and enhance the monetary and nonmonetary allowances of faculty members. The provision of such practices will result in a better POF, which will positively influence PJF, and, in turn, PJF will negatively influence their intention to quit.

Research Limitation and Future Research
Although the study provides useful insight on some important issues, its results should be considered in light of its restrictions. First, the study is cross-sectional in nature; therefore, the findings may not depict the real picture of causality. Second, the data were only collected from fulltime academics in public sector HEIs. If data had been collected from different sources, the findings might have been different. Third, this study only investigated POF and PJF as antecedents of intention to leave; future researchers are encouraged to assess the impact of other fit concepts such as personvocation fit (PVF), person-supervisors fit (PSF) and person-group fit (PGF) on intention to leave. In addition, the present study tested the mediating role of PJF; we expect that PVF, PSF and PGF could be possible mediators or moderators of the POF-intention to leave relationship.
We encourage future researchers to validate the conceptual model (see Figure 1) with a longitudinal study in different work settings, such as manufacturing, textile, information technology, health, and public sector organisations.