Higher education institutions (HEIs) in South Africa (SA) are experiencing immense challenges in retaining their skilled and valuable employees (Mukwawaya, 2022; Omodan, 2022; Takawira et al., 2014; Theron et al., 2014). Empirical evidence suggests that employees’ satisfaction in terms of their psychological contracts has a significant influence on employee turnover and staff retention (Agarwal & Bhargava, 2013; Kumar & Santhosh, 2014). In SA, HEIs operate in a complex environment and have a diverse workforce in terms of socio-demographic groups (Mampane, 2019). However, there is limited research on the influence of socio-demographics pertaining to psychological contracts and retention of staff at HEIs. Thus, the present study investigates the influence of socio-demographic differences among higher education (HE) employees, in relation to psychological contract preferences and staff retention.

While research confirms the association between the psychological contract and staff retention, as well as the importance of psychological contract preferences in strengthening retention and lowering turnover (Deas, 2017; Kraak et al., 2017; Peirce et al., 2012; Snyman, 2021; Van der Vaart, et al., 2013; Van Stormbroek & Blomme, 2017), less attention has been given to the influence of socio-demographical differences on the relationship between these two concepts (Deas, 2017; Rafiee et al., 2015). Furthermore, there is a shortage of research on socio-demographic differences among employees in HEIs – especially in open-distance learning (ODL) institutions – in terms of how employees’ psychological contracts impact staff retention (Deas, 2018; Peltokorpi et al., 2015; Rafiee et al., 2015). Based on the high staff turnover rates at HEIs in SA (Barkhuizen et al., 2020), as well as the diversity of the workforce at these institutions (Setati et al., 2019), it is imperative that these institutions develop and implement retention strategies aimed at strengthening the retention of their valuable and diverse employees. Therefore, the present study aims to address this gap by specifically exploring the socio-demographical differences among employees in HEIs, and the role these differences play in the relationship between their psychological contract-related perceptions and the retention of staff. The findings of the study can therefore potentially add valuable new insights that could inform retention practices for diverse employees employed in the HE environment.

Higher education institutions in SA face immense difficulties in terms of skilled human resources and retention of their valuable and skilled employees, which has a devastating effect on the successful functioning of these institutions (Abugre, 2018; Barkhuizen et al., 2020; Deas, 2018; Erasmus et al., 2015; Gerstein & Friedman, 2016; Mukwawaya, 2022; Musakuro, 2022; Robyn, 2012; Tettey, 2006). Previous research suggests that there may be as much as a 13% shortage of academic and support staff at HEIs (Abugre, 2018; Deas & Coetzee, 2020; Dewhurst et al., 2013). Research by Theron et al. (2014) determined that 33.8% of employees in South African HEIs showed a strong intention to leave their institutions. Higher Education of South Africa (HESA, 2011), Lindathaba-Nkadimene (2020), as well as Omodan (2022) likewise, concluded that HEIs are battling with poor levels of staff retention and high labour turnover.

If HEIs are not able to retain their key employees, they will not be able to remain competitive and offer quality services (Hailu et al., 2013). The high staff turnover in HEIs may be resulting from several challenges experienced in the HE sector, such as financial constraints, uncompetitive remuneration packages, mergers, acquisitions, job insecurity, lack of resources, and an overload of demands placed on employees (Balakrishnan & Vijayalakshmi, 2014; Ngobeni & Bezuidenhout, 2011; Robyn & Du Preez, 2013). Furthermore, South African HEIs operate in a multifaceted milieu with a socio-demographically diverse labour force.

In a South African study, Döckel (2003) identified six crucial retention factors (also referred to as retention practices) that organisations should consider when developing and implementing retention strategies (Döckel et al., 2006; Van Dyk & Coetzee, 2012). These factors comprise: compensation, job characteristics, opportunities for training and development, supervisor support, career opportunities, and work-life balance policies.

Thus, effective capacity building, and staff retention along with the type of relationship which progresses between diverse groups of employees and their employers, will ultimately determine the success of HEIs (Armstrong & Taylor, 2014; Festing & Schäfer, 2014; Guo, 2017; Mukwawaya, 2022; Snyman, 2022). Employees are one of the most important assets of any organisation, and in order to ensure the success and efficient functioning of these institutions, it is imperative for HEIs to develop and implement retention practices which take the needs of various socio-demographic groups into account. Higher education institutions can only diagnose and prevent turnover of their employees when there is a fruitful employment relationship and the diverse needs of their employees are appreciated and incorporated into retention practices (Grobler & Jansen van Rensburg, 2019; Ng’ethe et al., 2012).

For HEIs to develop and implement retention strategies and practices aimed at addressing the high turnover levels, it is important to determine the factors that may have an impact on employees’ decision to stay with or leave their organisation. Empirical evidence shows that the type of relationship that exists between employees and their employer, and the extent to which employees perceive their employer to adhere to commitments made within the relationship, strongly impact retention (Guest, 2004; Le Roux & Rothman, 2013; Van Stormbroek & Blomme, 2017). This may be referred to as the psychological contract between employees and employers (Bal & Kooij, 2011; Rousseau, 1989). The psychological contract is a subjective, unwritten, open-ended contract based on the reciprocal expectations of both parties to the employment relationship (Eds. Guest et al., 2010; Kraak et al., 2017; Rousseau, 1989; 1990; 1995).

The state of employees’ psychological contracts is largely determined by the extent to which perceived promises made to them have been kept and obligations adhered to (Van der Vaart et al., 2013; Van Stormbroek & Blomme, 2017). The psychological contract is the basis of the employment relationship and has an enormous impact on employee retention (Guest, 1998; Kraak et al., 2017; Van der Vaart et al., 2015). When employees have a strong psychological contract with their employer, they are less likely to leave their organisation and more likely to be committed to their employer (Chin & Hung, 2013; Deas, 2017; Ngakantsi, 2022). Previous studies concur that perceived psychological contract breach negatively affects commitment and retention, and increases both planned and actual turnover (Deas, 2017; Peirce et al., 2012; Snyman, 2021; Van Dijk & Ramatswi, 2016).

A study by Deas (2017) showed that positive psychological contract-related perceptions are associated with higher satisfaction with the human resource factors that influence retention, namely, compensation, job characteristics, training and development opportunities, supervisor support, career opportunities, and work-life balance policies (Deas, 2017; Döckel; 2003; Döckel et al., 2006).

Several empirical studies have indicated that socio-demographic differences between diverse employees are among the most important factors to be considered when retention practices and strategies are developed and implemented by organisations (Peltokorpi et al., 2015; Potgieter & Mathonsi, 2021; Rafiee et al., 2015; Randmann, 2013). Socio-demographic variables such as race, gender, age, job level and tenure may also assist HEIs in predicting organisational commitment and strengthening retention (Rafiee et al., 2015; Randmann, 2013). Thus, considering the needs of diverse racial and age groups, and the predilections of different job levels, tenure groups and genders within an organisation can assist HEIs in the reinforcement of the retention of their skilled and valuable staff (Peltokorpi et al., 2015).

Various empirical studies, especially in the HE environment, have indicated that the extent to which employees value the retention practices offered by an organisation, for instance, their compensation packages, their work-life balance opportunities or their career development prospects, may be different among socio-demographic groups (Deas, 2017; Ng’ethe et al., 2012; Snyman et al., 2015). This implies that employees from dissimilar socio-demographic groups may not consider the same factors as equally important when deciding whether to stay or leave their organisation (Chin & Hung, 2013; Ryan et al., 2012).

In terms of employees’ psychological contracts, research has correspondingly shown that diverse socio-demographic groups do necessarily not share the same expectations within their employment relationships (Kraak et al., 2017; Peltokorpi et al., 2015). Different socio-demographic groups may not have the same perceptions in terms of what is regarded as detrimental to the continuation of a fruitful employment relationship, and what would result in psychological contract fulfilment and/or breach (Blomme et al., 2010; Chin & Hung, 2013). In the HEIs specifically, studies concluded that employees’ preferences related to their psychological contracts, that is, their perceptions of the necessary components for a healthy employment relationship with their employer, vary among different socio-demographic groups (Deas, 2017; Snyman, 2015).

The preceding discussion leads to the main research question related to the present study, namely, ‘What is the influence of socio-demographic differences on the relationship between psychological contract preferences and staff retention among South African HE employees?’.

This study aimed to extend the extant body of knowledge on the relationship between the psychological contract and staff retention of South African HE employees by their socio-demographic information. The present study aimed to explore: (1) the socio-demographical differences among employees in HEIs, in terms of their psychological contract preferences in relation to staff retention, and (2) the influence of socio-demographic differences on the relationship between psychological contract preferences and staff retention among South African HE employees. Exploring the relationship between the psychological contract and satisfaction with retention practices could aid in devising successful retention strategies, particularly for the varied and multicultural workforce in South African HEIs.

A cross-sectional quantitative research approach was followed in order to achieve the aim of the study. Empirical data was collected using an electronic survey, from full-time employees in a single ODL HEI in SA. A cross-sectional research design is useful in the collection of large-scale data from a target population, and this design has the possibility to contain multiple variables at a single time (Spector, 2019).

The socio-demographic characteristics of the participants are outlined in Table 1.

The participants were employees (N = 493) of a South African (SA) ODL HEI. As specified in Table 1, the sample comprised mainly black (48%), permanently employed (99%) females (63%), and aged between 26 years and 35 years. Table 1 also shows that most of the participants (22%) were employed as administrative officers or senior lecturers (14%). Furthermore, most of the participants (35%) had worked for the institution for longer than 15 years or between 6 years and 10 years (34%), and had a Master’s degree (32%) or a Doctorate degree (21%).

Participants completed the Psycones Questionnaire (PQ) (Guest et al., 2010; Psycones, 2006) and the Retention Factor Measurement Scale (RFMS) (Döckel, 2003; Döckel et al., 2006). Participants also self-reported their socio-demographic information, including their race, gender, age, job level, and tenure.

Participants’ psychological contract preferences were measured using the PQ (Guest et al., 2010; Psycones, 2006), developed by the Psycones project (Psycones, 2006). The PQ consists of 43 items and 4 subscales (employer obligations, employee obligations, state of the psychological contract and job satisfaction), and items are scored on a five-point Likert-type scale where 1 is ‘strongly disagree’, and 5 is ‘strongly agree’.

The employer obligations subscale relates to a person’s perception of promises made by the organisation and encompasses questions like: ‘Has your organisation promised or committed itself to providing you with a job that is challenging?’ and ‘Has your organisation promised or committed itself to allowing you to participate in decision-making?’. The employee obligations subscale pertains to an individual’s perception of his or her promises made to the organisation and entails questions such as: ‘Have you promised or committed yourself to showing loyalty to your organisation?’ and ‘Have you promised or committed yourself to being a good team player?’. The job satisfaction subscale includes six statements to establish participants’ emotional feelings related to the psychological contract, for example, statements such as: ‘I feel happy’, ‘I feel sad’, ‘I feel pleased’. Lastly, the fourth subscale concerns the overall state of the psychological contract and contains statements like: ‘Do you feel that organisational changes are implemented fairly in your organisation?’ and ‘Do you feel fairly treated by managers and supervisors?’.

The Cronbach’s alpha coefficients for scores from the PQ range from 0.70 to 0.95 (Psycones, 2006). In the current study, the reliability of scores from the PQ was 0.94 for employer obligations, 0.90 for employee obligations, 0.90 for job satisfaction, and 0.87 for the overall state of the psychological contract.

The RFMS comprises 35 items and 6 subscales (compensation, job characteristics, training and development opportunities, supervisor support, career opportunities and work/life balance) and are scored on a six-point Likert-type scale, where participants had to indicate how satisfied or dissatisfied they feel about their organisation regarding certain statements. The scale ranges from 1 being ‘strongly disagree’ to 5 being ‘strongly agree’.

The first subscale measures participants’ opinions about the importance of compensation and includes statements regarding, for example: ‘My benefits package’ and ‘My most recent raise’. In the second subscale, participants’ views about the importance of job satisfaction are determined and comprises statements, for instance: ‘The job requires me to use a number of complex or high-level skills’ and ‘The job is quite simple and repetitive’. The third subscale relates to training and measures participants’ perceptions of the importance of training, including for example: ‘This company provides me with job-specific training’ and ‘Sufficient time is allocated for training’.

The fourth subscale relates to participants’ views regarding their supervisor support, and includes statements such as: ‘I feel undervalued by my supervisor’ and ‘My supervisor seldom recognises an employee for work well done’. The fifth subscale measures participants’ perceptions regarding their career opportunities, for example: ‘My chances for being promoted are good’ and ‘It would be easy to find a job in another department’. In the sixth and final subscale, participants’ views on work/life balance were assessed, with statements such as: ‘I often feel like there is too much work to do’ and ‘My work schedule is often in conflict with my personal life’.

Previous empirical research reported internal consistency reliabilities of 0.80 to 0.90 for scores from the RFMS (Döckel, 2003). In the present study, the reliability of scores from the RFMS was 0.94 for compensation, 0.64 for job characteristics, 0.90 for training and development opportunities, 0.85 for supervisor support, 0.80 for career opportunities, and 0.89 for work-life balance.

In all, 4 882 questionnaires were distributed, with a total of 493 usable questionnaires returned (N = 493), yielding a response rate of 10.1%. The participants were invited to voluntarily participate in the study. The questionnaires were distributed electronically through an e-mail link. Each questionnaire encompassed a cover letter inviting respondents to participate voluntarily in the research study and guaranteeing them the anonymity and confidentiality of their individual responses. The cover letter also stated that by completing the questionnaires constituted informed consent and agreement for the results to be used for research purposes only.

The first stage of the data analysis process involved determining the means, standard deviations and Cronbach’s alpha coefficients. In the second stage, to test the strength and direction of the relationship between the PQ and RFMS variables, Pearson’s product-moment correlation coefficient (r) was applied. Also, to determine the relationship between the socio-demographic variables (race, gender, age, job level and tenure), and the PQ and RFMS variables, Spearman’s rank correlation coefficient (p) was employed. The practical significance of correlation coefficients was determined using the following cut-off points: r ≥ 0.30 (medium effect) and p ≤ 0.05 (Humphreys et al., 2019; Liu, 2019).

The third stage of the data analysis involved testing for significant differences in psychological contract-related preferences and satisfaction with retention factors, by the socio-demographic factors (race, gender, age, job level and tenure). The Shapiro-Wilk, Kolmogorov-Smirnov, Cramer-von Mises and Anderson-Darling tests showed that the data were normally distributed, and thus, parametric tests were applied. In order to measure the differences between the socio-demographic variables of race, gender, age, job level and tenure, ANOVAs (analysis of variance) and post-hoc tests were conducted. Also, to assess the differences between the genders, a t-test and Tukey’s studentised range tests were used (Lee & Lee, 2018). Finally, Cohen’s d test was utilised to establish the practical effect size regarding the differences between the relevant groups (Cohen & Cohen, 2014; Cohen et al., 2013).

The descriptive statistics and the construct validity and reliability statistics for the study variables are presented in Table 2.

As shown in Table 2, Cronbach’s alpha coefficients and composite reliability (CR) values were all above > 0.85, which displays high internal consistency reliability of the PQ. Table 2 also indicates the convergent validity of the PQ, with the CR values being more significant than the average variance extracted (AVE) values and the AVE values of employer obligations (0.50) and job satisfaction (0.58) being ≥ 0.50, which is acceptable. The AVE values of employee obligations (0.41) and the state of the psychological contract (0.46) were just below the threshold of ≥ 0.50.

Furthermore, Table 2 specifies that the Cronbach alpha coefficients and CR values were above > 0.70 (except for the job characteristics subscale), indicating good internal consistency reliability of the RFMS. The reliability coefficients for the job characteristics subscale (α = 0.64; CR = 0.60) were acceptable for large-group analysis purposes. However, the lower reliability of the subscale was considered in the interpretation of the findings. The convergent validity of the RFMS with the CR values being larger than the AVE values, and the AVE values of compensation (0.56), training and development opportunities (0.61) and work-life balance (0.70) being ≥ 0.50, is acceptable.

Table 3 summarises the results of the correlations between the psychological contract (PQ) and retention factors (RFMS).

As shown in Table 3, the relations were all significant and positive, ranging between r ≥ 0.15 ≤ 0.90 (small to sizeable practical effect size; p ≤ 0.05). The four subscale dimensions of the PQ also had significant and positive correlations with the overall RFMS (r ≥ 0.45 ≤ 0.67; large practical effect size; p ≤ 0.001) signifying the construct validity of the general construct of the psychological contract.

In addition, as indicated in Table 3, the results showed a significant correlation between the six subscale dimensions of the RFMS, in the range of r ≥ 0.14 ≤ 0.48 (small to sizeable practical effect size; p ≤ 0.05). The six subscale dimensions of the RFMS also had significant and positive correlations with the overall PQ scale (r ≥ 0.64 ≤ 0.73; large practical effect size; p ≤ 0.001), indicating the construct validity of the general construct of satisfaction with human resource retention factors.

Table 4 summarises the results of the correlations between the socio-demographic variables, the psychological contract (PQ), and retention factors (RFMS).

As specified in Table 4, the results displayed a significant negative bivariate correlation between race and employee obligations (PQ scale) (r = –0.14; small practical effect size; p ≤ 0.001); job satisfaction (PQ scale) (r = –0.17; small practical effect size; p ≤ 0.001); and state of the psychological contract (PQ scale) (r = –0.16; small practical effect size; p ≤ 0.001). The results further showed a significant negative bivariate correlation with the overall PQ (r = –0.11; small practical effect size; p ≤ 0.01) career opportunities (RFMS) (r = –0.27; small practical effect size; p ≤ 0.001).

In terms of gender, Table 4 shows no significant bivariate correlations between gender and any of the subscales and overall scales of the PQ and RFMS. In terms of age, Table 4 indicates no significant bivariate correlations between age and any of the subscales and overall scales of the PQ and RFMS. Furthermore, the results showed a significant positive bivariate correlation between job level and job characteristics (RFMS) (r = 0.20; small practical effect size; p ≤ 0.001); supervisor support (RFMS) (r = 0.11; small practical effect size; p ≤ 0.01); work-life balance (RFMS) (r = 0.15; small practical effect size; p ≤ 0.001); and employer obligations (PQ scale) (r = 0.09; small practical effect size; p ≤ 0.05).

Lastly, the results showed a significant negative bivariate correlation between tenure and job satisfaction (PQ scale) (r = –0.17; small practical effect size; p ≤ 0.001). A significant negative bivariate correlation was found between career opportunities (r = –0.30; moderate practical effect size; p ≤ 0.001) and the overall RFMS (r = –0.11; small practical effect size; p ≤ 0.01).

The results from the tests for significant differences between the socio-demographic groups in terms of their psychological contract-related preferences, as well as their satisfaction with retention factors, are outlined and discussed below.

This research has provided evidence that the psychological contract-related preferences and perceptions of HE employees differ among various socio-demographic groups. The study has also highlighted the crucial role played by employees’ psychological contracts in influencing their satisfaction with their organisation’s retention practices and their decision to stay or leave. Although previous research has established a connection between psychological contract fulfilment and staff retention, this paper’s unique contribution is its novel perspective on the psychological contract and retention-related predispositions of the diverse workforce in the South African HE sector. Hence, the variations in socio-demographic factors among HE employees regarding their psychological contract preferences in relation to staff retention should inform the customisation of retention interventions and strategies in HEIs.

The findings of this study add to the existing body of knowledge on human resource management and staff retention strategies in the South African HE context. Understanding how socio-demographic factors influence employees’ psychological contract fulfilment and satisfaction with organisational retention practices, can assist HEIs in retaining their diverse and valuable workforce.

The present study’s sample comprised predominantly permanently employed, married, black African and white females, between the ages of 46 years and 55 years, in a single ODL, HEI. Thus, the research findings are not generalisable to employees of other sociodemographic groups or different industries. Furthermore, the participants in the sample were randomly sampled and consisted of only 493 participants; therefore, the sample was not big enough to adequately embody the entire population. Finally, the study did not assess the bi-directionality of the employment relationship, as only the views and perceptions of employees were considered. Subsequent studies should consider the viewpoints of both employees and employers, pertaining to the psychological contract and retention. Additionally, it is important to replicate the results of this study in the wider South African HE landscape.