Determinants of Intention to Stay of Higher Learning Institutions Administrators: A Review of Theory and Evidence

Today’s business challenges are no longer only about facing the economic uncertainty but also the ability to strategise and reconfigure the business operations in facing the changing world business landscape, which is highly dependent on the efficiency of organisations in managing their employees. This is also a growing concern shared by the Malaysian Private Higher Learning Institutions (PHLIs). Thus, this paper intends to fill the gap in the body of knowledge of HRM practices with specific emphasis on identifying determinants of Malaysia PHLIs’ administrators’ intention to stay. Grounded by several turnover models, this study proposes a framework linking Human Resources Management (HRM) practices and intention to stay. The findings revealed that from the four HRM practice variables used in the study, only training and performance appraisal were found to have a direct influence on the administrators’ intention to stay. On the evaluation of PHLIs ’ administrators on HRM practices (performance appraisal, training, and selection), the regression results showed a significant effect, F (3, 268) = 25.95, p < 0.000. This indicates that the inclusion of independent variables improved the model significantly. Looking at the individual interactions, the results indicated that Training ( β = 0.30, p < 0.001) had the strongest influence on PHLIs ’ administrator’s intention to stay, followed by Performance Appraisal ( β = 0.21, p < 0.001). Selection was not significant in predicting PHLIs ’ administrator’s intention to stay.


Introduction
The COVID-19 pandemic did not only cause serious implications to people's health but has also been significantly impacting businesses and the economy. The setback from this impact had caused businesses to put limited focus on employees' psychological and safety considerations, which had adversely affected business' resilience and staff retention (Rangachari & Woods, 2020). This has resulted in an increasing challenge for business organisations to develop sustainable Human Resource Management (HRM) practices that can enhance human sustainability and support business corporate sustainability goals. It was also reported that the COVID-19 pandemic has also caused chronic emotional distress to the employees. This has led to high staff turnover and low retention, which will be a challenge during COVID-19 and beyond the pandemic (Rangachari & Woods, 2020).
Evidently, HRM practices have been proven by previous studies that they affect not only the employees but also other contexts of the organisation, such as the human, social and environment. Therefore, the ability to attract and retain performing employees is one of the vital HR functions besides strategy execution, managerial credibility, management experiences, and compensation strategy, which link individuals' experiences to the organisation's critical success factors (Aust et al., 2020). Hence, understanding the elements that will positively influence the employees' intention to stay will allow a company to understand better the behaviour of its employees and their motivational factors. This aligns with the Job Embeddedness Model (Lee & Mitchell, 1994) contentions that employees' intention to stay is the best predictor in predicting the employees' behavioural outcome.
Specifically, this study focuses on the case of Higher Education Institutions (HEIs) in Malaysia because education is a significant industry in developing the nation's economy. Furthermore, with the current landscape of the education industry, where the ranking and status of a university are predominant factors that attract students and sponsors, the HEIs need to give extra attention to their staff retention strategy. The ability of HEIs to retain their competent employees will assist the university to stay focused on achieving its goals. Therefore, this research is expected to provide significant theoretical and practical contributions by providing insights into the factors that can influence the intention to stay of the HEIs workforce. Practically, this study is critical because the findings would provide a better understanding to the management on what they should do to retain the administrative group employees. This is because they are part of the vital components in HEIs and the engine of their institutions who perform significant, fast-paced, stimulating, stressful and demanding responsibilities such as coordinating and interacting with faculty members, parents, students, community members, business leaders and policymakers (Shariff et al., 2017). Undeniably, any weak link in the running of a university will affect not only the institutional rating but also its performance.
In terms of theoretical significance, this study is expected to fill the gap in the body of knowledge of Human Resource Management (HRM) practices, emphasising Malaysian Private Higher Learning Institutions (PHLIs). Specifically, this study intends to identify the determinants of intention to stay of PHLIs' administrators and the influence of these determinants on the intention to stay of PHLIs' administrators.

The conceptualisation of HRM Practices
An organisation's ability to formulate suitable HRM practices that can contribute toward the development of the positive behavioural intentions of their employees is very important. This is because HRM practices have been repeatedly reported by many scholars as having significant effects in influencing employees' intention to stay. Their importance to the success of an organisation has grown significantly, especially in today's challenging Covid-19 pandemic period. According to Boselie et al (2005), HRM, in its broadest sense, is all about carefully selecting the best practices in managing the employees towards enhancing the organisational effectiveness to produce better organisational outcomes.
The Strategic Human Resources Management (SHRM) researchers who studied the broader focus of HRM, from the microanalytic research to a more macro or strategic perspective, divided the practices into three different approaches -universalistic, contingency and configurational (Delery & Doty, 1996). Under the SHRM universalistic perspective, some HRM practices are considered universally effective across the board; therefore, companies that adopt these best practices will experience a competitive advantage. However, the contingency perspective suggests that each HRM practice is unique, and its relationship with the organisation's strategy will strongly influence its performance. Therefore, the correct selection of practices will assist a company in achieving its target effectively. On the other hand, the configurational perspective highlights the synergistic effects among the HRM practices. The universal approach has received empirical support as the best approach compared to the contingency and configurational approach (Boselie et al., 2005;Huselid, 1995;Meijerink et al., 2016). The researchers also reported that companies adopting the universal or best practices of high-performance HRM practices would experience a significant improvement in their organisational performance and productivity and lower turnover intention.
Unfortunately, there is a lack of consensus regarding the use of a universal set of HRM practices. Therefore, to understand the causal factors that can influence employees' positive behavioural intentions or specifically the intention to stay, a detailed review of behavioural intentions theories was conducted. This is important because understanding and effectively managing employee turnover will require the researcher to understand both the positive and negative potential effects of turnover because there are multiple causes and consequences towards it. Furthermore, it is not a static event but an interrelated and ongoing process (Mobley, 1982). However, due to the enormous volumes of documented employee retention and turnover theories, this study would not be able to cover all available theories exhaustively. Therefore, the discussion on the behavioural intentions theories in this study would only be limited to seven main employee's behavioural intention theories that have been regarded as the main reference of the causes and correlate of behavioural intention studies. Table 1 summarises the turnover models as delineated by their proponents.  (March & Simon, 1958) If employees expected inducements (such as pay) that influence their job satisfaction match or exceed their perceived contribution values, the propensity to leave would be reduced. Otherwise, they will leave if there are other suitable job alternatives.
Met Expectations Model (Porter & Steers, 1973) If an organisation fails to fulfil employees' set of expectations such as pay, promotions, supervisory or peer group relationship (which varies from one individual to another), these unmet expectations will lead to job dissatisfaction, and the probability of withdrawal will result.

Mobley's Intermediate
Linkages Model of Turnover (Mobley, 1977) Job dissatisfaction is highlighted as the primary determinant of quitting intention, and the intention to quit can be prevented if employees feel committed and satisfied with the organisation.

The Expanded Mobley et al.'s Model of
Turnover (Mobley et al., 1979) The four main determinants of intention to quit that could cause turnover are job satisfaction -dissatisfaction, the expected utility of alternative internal work roles, external work roles, and non-work values and roles.
CUSP Catastrophe Model of Turnover (Sheridan & Abelson, 1983) There are interrelated relationships between organisational commitment, job tension, and withdrawal behaviour. Employee withdrawal behavioural is influenced by job tension-related factors such as role conflict, role ambiguity and role accuracy. Organisational commitment, on the other hand, will reduce employees' withdrawal behaviour.
Price and Muller's Model of Turnover (Price & Mueller, 1986) Job dissatisfaction is identified as the core influencer of turnover decisions, while organisational commitment was identified as the mediator that might influence the decision between dissatisfaction and turnover. Role overload, nature of the job (routinisation), centralisation, instrumental communication, integration, pay, distributive justice, and promotional opportunities are factors that influence job dissatisfaction. Professionalism, general training, kinship responsibility, and company work units are factors that influence work commitment.

Job
Embeddedness Model (Lee & Mitchell, 1994) The whole screening and decision-making processes generally begin due to some distinguishable events experienced by the employees. These events trigger the employees to make suitable judgments that affect their turnover. However, employees who are embedded with the organisation's on and off job factors such as links, fits, and sacrifices may not be affected and may increase their propensity to stay.
The integrative theoretical turnover models discussed showed that there are various aspects of the job and the individual factors that potentially affect and trigger an employee's intention to stay. Nonetheless, to influence employees' intention to stay, the Theory of Organisational Equilibrium (March & Simon, 1958) suggested that efforts should focus on the effect of job satisfaction and fulfilling employees' expected inducements. Next, the Met-Expectations Model (Porter & Steers, 1973), the Mobley's Intermediate Linkages Model of Turnover (Mobley, 1977) and the Expanded Mobley et al.'s Model of Turnover (Mobley et al., 1979), emphasised more on fulfilling the employees' expectations on the set of job satisfaction and dissatisfaction about pay, promotions, supervisory or peer group relationship, the expected utility of alternative internal work roles, the expected utility of external work roles, and non-work values and roles. Subsequently, the CUSP Catastrophe Model of Turnover (Sheridan & Abelson, 1983), and the Price and Muller's Model of Turnover (Price & Mueller, 1986) had then focused on organisational commitment, job tension, role overload, nature of the job, centralisation, instrumental communication, integration, pay, distributive justice, promotional opportunities, professionalism, general training, kinship responsibility, and company work units. Finally, the Job Embeddedness Model (Lee & Mitchell, 1994) proposed that efforts should focus towards on and off the job factors such as links, fits, and sacrifice.
The findings uncovered several underlying HRM practice constructs that can help companies increase employees' commitment and reduce turnover intention. Essentially, the practices highlighted are part of the overall HRM processes, which include human resource planning, recruitment, selection, socialisation (orientation), training, development, performance appraisal, promotion, transfer, demotion, and separation. Subsequently, based on the varieties of effective HRM practices reported by the theories, the general practices were re-categorised into four sets of practice-level categories: recruitment and selection, training and development, performance appraisal, and compensation and benefits. Therefore, in the context of this study and consistent with the findings of past researchers, the above HRM practice-level categories are outlined as the HRM practice variables for this study.
In this study, recruitment is defined as the process of attracting suitable candidates to apply for the available vacancies in the organisation. While selection is the process of selecting the most qualified candidates from a group of selected applicants. The main objective of this activity is to ensure that only suitable candidates are recruited for the offered position (Aminuddin, 2018). Subsequently, training in this study is indicated as the use of systematic and planned instructional activities that can help the employees acquire the skills and knowledge necessary for them to perform their job satisfactorily and development is programmes or activities conducted to prepare the employees for their career promotion (Armstrong & Taylor, 2020). Both training and development are considered the most significant investment that an organisation should contribute to and focus on. This is because there are direct and indirect benefits that the company might obtain for the betterment of the organisation, the employees, and the employer.
Performance appraisal, in this study, is defined as the process of evaluating an employee's job performance and comparing the achievements with the standards and objectives which have been preassigned based on individual job roles and positions (Armstrong & Taylor, 2020). It is a formal, planned, and organised process which refers to the employees' agreed physical, technical, behavioural and physiological terms and conditions. Finally, compensation refers to all forms of direct financial payments such as wages, salaries, incentives, commissions, and bonuses paid to the employees in return for their employment with the organisation. On the other hand, benefits are indirect financial and nonfinancial payments such as company-paid medical coverage, vacations, and work insurance provided to the employees for continuing their employment with the organisation (Gary, 2019). Compensation and benefits were both highlighted as part of the elements in the organisational reward system that involves people, processes, rules and procedures, and decision-making activities in exchange for the contributions of employees to the organisation.
Apart from that, intention to stay is selected as the dependent variable for this study. This is because focusing on employees' intention to stay rather than on other turnover cognitions will be more beneficial to the organisation because it instils employee loyalty and promotes the spirit of togetherness among employees towards their organisation. Furthermore, focusing on the intention to stay will also allow the researcher to have access to first-hand data because the employees are still around to give appropriate feedback and responses concerning their intention to stay. This will also allow the study to be completed on time because the researcher can always cross-check with the participants on certain uncertainties that may arise from the study. The highlighted variables are used and adopted into this study and outline the underlying conceptual framework as described in Figure 1.

Methodology
This study utilises the descriptive research design, specifically, the cross-sectional approach. Through this approach, various segments of the population were sampled, and all data on the variables were collected simultaneously to investigate the relationships among the variables. Specifically, this study examines the perceptions of Malaysia's PHLIs administrators within the position of officers and above. To eliminate any bias and possible differences in terms of the needs and expectations of the administrators, only full-time staff who have been employed for a minimum of one year were chosen as participants for this study.
The purposive sampling technique was utilised as the selection method because it is impossible to obtain a comprehensive list of the population. According to Tongco (2007), purposive sampling is a non-random technique that has no specific underlying theory requirements or a minimal number of participants. Therefore, the researcher only needs to determine the research objectives and decide which sets of individuals are willing to share their knowledge or experience in resolving the research issues. For this study, a total of 279 responses were received; however, seven responses were excluded from further analyses because they were found to be incomplete or completed by participants who did not fit the participants' criteria. Therefore, only 272 responses were accepted, coded, and analysed accordingly.
The questions asked in the questionnaire were taken and adapted from established instruments developed by previous researchers. Four dimensions of HRM practices using 20 items to measure the HRM practices comprising recruitment and selection, training and development, performance appraisal, and compensation and benefits were measured using the instruments developed by Delery and Doty (1996), Koch andMcGrath (1996), andMacDuffie (1995). Seven items measuring scale developed by Mowday et al. (1984), and Walsh et al. (1985) were used to measure intention to stay. A 5-point Likert Scale, ranging from 1 = strongly disagree to 5 = strongly agree, was applied to measure the participants' feelings about certain issues under study.

Results
In determining the validity of the HRM practices scale, the principal component factor analysis with varimax rotation was used. Initially, 20 items were included, however, the initial results did not produce a clean and meaningful structure. The results revealed that ten items Recruitment and Selection Intention to Stay Training and Development

Performance Appraisal
Compensation and Benefits had either high cross-loadings (an item that loads at 0.35 or higher on more than one component) or low factor loadings (0.50 or below). Therefore, they were eliminated from further analyses, and the factor analysis was repeated on the remaining ten HRM practices items.
The results of factor analysis on HRM practices indicated the existence of three significant factors. The results of a scree test also provided support for a three-factor solution. The first factor is comprised of five items relating to the performance evaluation. Therefore, factor 1 was categorised as Performance Appraisal. The second factor comprises three items, which reflect the participants' perceptions of the training programmes. Thus, the name Training was retained. Finally, factor 3, initially named recruitment and selection, was changed to selection because both items reflect the participants' perceptions of the selection process. Table 2 shows the results of factor analyses on HRM practices.  Table 3 highlights the results of factor analyses on intention to stay. The principal component factor analysis conducted on the seven items of intention to stay resulted in a single factor. The results of a scree plot test also supported the one-factor solution. Since only one component was extracted, the solution could not be rotated. Therefore, the seven items measuring intention to stay were retained as per the original construct because the items were interrelated, and they shared common factors. 0.000 The outcome from the factor analyses resulted in some of the earlier identified variables being deleted, re-categorised and renamed. Therefore, the research framework developed earlier was revised to incorporate the factor analysis results. The revised research framework is as per Figure 2 below: Figure 2 Revised research framework.
Subsequently, the below hypotheses were reformulated to reflect the findings of factor analyses:

H1
Performance appraisal significantly influences the intention to stay of PHLIs' administrators.

H2
Training significantly influences the intention to stay of PHLIs' administrators.

H3
Selection significantly influences the intention to stay of PHLIs' administrators.
Performance Appraisal Table 4 shows the reliability of independent and dependent variables. In this study, the indicator suggested by Zikmund et al. (2013) was adopted, where very good reliability is when 0.80 < α < 0.95; good reliability is when 0.70 < α < 0.80; fair reliability is when 0.60 < α < 0.70; and poor reliability is when α < 0.60. Hence, the measures have high internal consistency and stability.  Table 5 lists descriptive statistics for each of the variables in the study. This study used the following criteria: mean scores of 2.99 or less are considered "low", mean scores between 3.00 and 4.99 are categorised as "moderate", and mean scores of 5.00 and above are categorised as "high". Findings reported in the table below indicate that the PHLIs' administrators had a moderate perception regarding the existence of a formal performance appraisal process, training programme, and selection mechanism at their respective universities. The "Intention to stay" felt by the PHLIs' administrators was also reported to be at a moderate level.  Table 6 shows the regression analysis results, which were conducted to analyse the influence of the participants' perceptions about their organisation's HRM practices (selection, training, and performance appraisal) on their behavioural intention (intention to stay). On the evaluation of PHLIs' administrators on HRM practices (performance appraisal, training, and selection), the regression results showed a significant effect, F (3, 268) = 25.95, p< 0.000. This indicates that the inclusion of independent variables improved the model significantly. Looking at the individual interactions, the results indicated that Training (β = 0.30, p< 0.001) had the strongest influence on PHLIs' administrator's intention to stay, followed by Performance Appraisal (β = 0.21, p< 0.001). Selection was not significant in predicting PHLIs' administrator's intention to stay.

Conclusion
The results revealed that only two dimensions of HRM practices, performance appraisal and training, significantly and positively influenced the PHLIs' administrators' intention to stay. Training was the strongest predictor of the intention to stay of the PHLIs' administrators, followed by performance appraisal. Selection was not a significant predictor of intention to stay. This study has empirically provided evidence on the importance of HRM practices in fostering positive attitudes among the administrators. Most important is that this study has provided evidence to the independent effects of HRM practices -training and performance appraisal on behavioural intention.
These findings indicated that when PHLIs' administrators perceive that they are provided with relevant training related to their future career development, conducted formally to prepare them for the job and coupled with a very extensive training programme, they tend to experience a higher sense of belongingness. This will create loyalty and motivate them to stay longer with their respective PHLIs. This finding is consistent with the results of previous researchers, whereby training has a significant positive association with the intention to stay (Harun et al., 2021;Shahid et al., 2018;Othman & Lembang, 2017). One possible reason is that PHLIs' administrator regards training as one of the ways for them to upgrade their skills and knowledge in preparation for their future career advancement.
Performance appraisal was also found to directly affect the intention to stay of the PHLIs' administrators. When the PHLIs' administrators perceive that the organisation's performance appraisal process is reliable and objective, supported by a transparent promotion process, and the staff are aware of the availability of positions for promotion, it will positively influence the intention to stay of the PHLIs' administrators. This is true especially when the heads of departments are supportive and aware of the career aspirations of their staff, and they support it through a clear and objective performance appraisal process. This finding is somewhat in line with the recommendation made by previous researchers that the staff perceptions of supervisor's support will significantly reduce the employees' turnover intentions (Shahid et al., 2018). This finding also concurs with Daoanis (2012) findings that a sound performance appraisal system plays a crucial role in reinforcing commitment and loyalty.
The findings from this study also serve as an eye-opener to businesses that have been relying primarily on providing rewards and remunerations packages as the main attraction in promoting staff retention. Undeniably, it may significantly affect staff retention, but it requires the business to spend a considerable amount of money to implement this solution. The evidence from this study suggests that PHLIs should focus more on their training programmes and performance appraisal processes. This is because the results showed that those are the main contributors to the positive behavioural intentions of the PHLIs administrators. This approach may also require the PHLIs to spend less and obtain better outcomes. Concerning this, the findings of this study suggest that there is a possibility that when PHLIs show their commitments and willingness to invest in the administrators' wellbeing through good HRM practices, the administrators will feel more committed to their university. This argument is in line with Renaud et al (2015) argument that organisations can foster functional retention of their experts through HRM practices.
Similar to other studies, several limitations of the study have been identified. First, the scope of the HRM practices investigated in this study was limited. HRM practices utilised in this study focused only on training and development, compensation and benefits, performance appraisal, and recruitment and selection. More research is needed to better distinguish between the general HRM practices that can suit the general groups of employees and the special HRM practices designed for a specific group of employees, thought to have specific needs. Second, since the focus of this study was on the Malaysian PHLIs, future studies should expand their scope to cover Malaysian public universities so that the findings can be generalised to all Malaysian universities regardless of their status, whether they are private or public universities.