The key research question the study aimed to answer was: What are the engagement levels of personnel working from home in a private higher education institution in South Africa in the context of the COVID-19 pandemic, and how are they affected by socio-demographic variables? Therefore, the objective of this study was firstly, to assess the levels of engagement of personnel working from home in a private higher education institution in South Africa in the context of the COVID-19 pandemic; secondly, to explore the association between selected socio-demographic variables and employee engagement; and thirdly, to suggest recommendations, based on the literature review and empirical results, to assist higher education institutions, generally, and specifically, in the private space, to improve the engagement levels of personnel.

Employee engagement is a term that has attracted considerable debate regarding its conceptual definition. The concept has evolved, providing more insight into contemporary thinking around defining employee engagement in the workplace (Shuck & Wollard, 2009). Because of the recent inclusion of positive psychology and a focus on work engagement in recent management practice, there has been a greater focus on rethinking what employee engagement refers to (Sun & Bunchapattanasakda, 2019). The concept of employee engagement was first introduced by Kahn (1990), who saw engagement as harnessing workers’ selves towards their duties and cognitively investing in their work as a form of dedication. However, over the years, many researchers believed there is more to employee engagement than harnessing oneself towards work. As a result, the contributions made towards the understanding of employee engagement are discussed further in the text.

Two influential frameworks have been used to understand employee engagement: the social exchange theory and the job resources demand model. These theories offer insights into the factors affecting employee engagement and how it can be enhanced during WFH.

Because of the rapid implementation of the WFH model during the COVID-19 pandemic, organisations have begun to provide insight into how well the model works. A key variable within this success was employee engagement. The drivers of engagement that were identified are discussed further in the text.

Many organisations, given the importance of engagement and its subsequent impact on employee and organisational productivity, have devised innovative ways to counteract the effect of WFH on engagement. This section briefly elaborates on the practices used.

Pre- and during the COVID-19 pandemic, research suggests that socio-demographic variables such as age, gender, marital status, level of education and tenure may influence the levels of employees’ engagement (Botha & Coetzee, 2022; Botha et al., 2023; Chaudhary & Rangnekar, 2017; Mvana & Louw, 2020; Romero-Martín et al., 2022).

Cascio et al. (2014), Botha and Coetzee (2022), Botha et al. (2023), Mvana and Louw (2020), and Romero-Martín et al. (2022) in their respective studies found that age influences the engagement levels of employees. Mvana and Louw (2020) found that younger teachers are less engaged in their work than older teachers. Cascio et al. (2014) mentioned that when educators reach the age of 50, they can become disengaged from their work as they are less likely or capable of dealing with stressful circumstances. In the WFH context, Romero-Martín et al. (2022) found that younger employees displayed greater levels of engagement compared to older employees. Romero-Martín et al. (2022) conducted their study among active workers in the United Kingdom during the COVID-19 pandemic.

Interestingly, Botha et al. (2023) found in their study conducted among employees within higher education in South Africa, that older employees showed more dedication, vigour and absorption when working from home during the pandemic, than the younger employees. A study conducted by Botha and Coetzee (2022) also revealed that older employees displayed higher levels of absorption than the younger employees. The authors conducted their study during the COVID-19 pandemic among employees working in a debt-collection organisation in South Africa.

The study conducted by Botha et al. (2023) revealed that gender affects the engagement levels of employees; the results showed that the male employees displayed higher levels of engagement than the female employees during the WFH period. The study conducted by Romero-Martín et al. (2022) confirmed the results. However, Botha and Coetzee (2022) found no significant differences between the genders in employee engagement during WFH.

Earlier studies suggested that married people are more engaged in the workplace (Rigg et al., 2014; Zeng et al., 2009). Romero-Martín et al. (2022), whose study was conducted during the pandemic, showed that employees within the UK who were married or lived with a life partner showed greater engagement in their work during the WFH period. However, Sungmala and Verawat (2021) and Botha et al. (2023) did not find that marital status does affect employee engagement.

Mvana and Louw (2020) and Sungmala and Verawat (2021) found that the level of qualification influences employees’ engagement levels, where employees with higher qualifications were less engaged in their work than their counterparts with lower qualifications. However, the studies conducted by Botha and Coetzee (2022) and Botha et al. (2023) revealed no association between the level of qualification and the engagement levels of employees during WFH.

Mvana and Louw (2020) found that employees, particularly teachers with more experience, displayed lower levels of engagement. Klassen and Chiu (2010) explained that educators show greater engagement and psychological capital at their career’s commencement and mid-stages. This would decrease as they mature in their career, which could explain why younger teachers are more vibrant and display greater engagement.

Earlier studies (Avery et al., 2007; Coetzee & Rothmann, 2005; Montes & Irving, 2008) revealed an association between tenure and employees’ engagement levels, where engagement tends to decline the longer the employee’s tenure is within the organisation. However, in the COVID-19 context, Botha and Coetzee (2022) and Botha et al. (2023) found no significant relationship between tenure and employee engagement. According to Robinson et al. (2007) the kind of experience the organisation offers may influence employees’ engagement levels.

Despite the differences in the results of the studies mentioned earlier, it can be deduced that socio-demographic variables may influence the engagement levels of employees. This will further be explored in this study.

Four research approaches are commonly utilised in research: positivism, post-positivism, interpretivism, and critical theory. These paradigms reference different views of the world, allowing researchers to perceive, quantify and realise the unique and social experiences of individuals (Bryman & Bell, 2018). This study adopted the positivist paradigm and quantitative method of inquiry to measure employees’ engagement levels at a private higher education institution and to determine its association with selected socio-demographic variables (Bryman et al., 2021; Dawson, 2019).

The target population of this study comprised individuals from a private higher education institution in South Africa who were employed at different categories or levels of employment. Thus, the target population (133) comprised management (16), academics lecturers (27) and academic administrators (i.e. support staff) (90). The total population was targeted. Total population sampling is a form of purposive sampling which utilises or surveys a total population based on specific traits (i.e. academic personnel in this regard) (Dawson, 2019). The population received the invitation to participate in the study via email. A total of 105 responded, signifying a satisfactory response rate of 79% (N = 105).

Ethical clearance was obtained from the Economic and Management Sciences Research Ethics Committee (EMS-REC), North-West University (ethics number: NWU-01253-21-A4).

Table 1 reflects the socio-demographic information collected from the participants.

The data collected indicate that all of the respondents included in the study are South African, represent a young adult population with over 65% of respondents being between the ages of 20 years and 39 years and have been employed at the institution at varied tenures (i.e. 6 months to over 10 years) and are predominantly female (i.e. 62.9%). The data also represented all employment levels (i.e. management, academic and support staff). Support staff was expected to represent the larger population because of the South African higher education requirement of a 30:1 support staff-to-student ratio (Kosie, 2022). Most respondents have a university degree or postgraduate degree (72%), and a large proportion (47%) indicated they were married. All respondents mentioned that they pursued WFH for specific periods during the COVID-19 pandemic, inferring they can offer significant insight into this study.

The confirmatory factor analyses (CFA) conducted on the UWES found that the three-factor structure (i.e. vigour, dedication and absorption) has a good fit to the data derived from the respondents (see Figure 1). Six items loaded on vigour (ranging from 0.465 to 0.726), five on dedication (ranging from 0.293 to 0.858), and six on absorption (ranging from 0.546 to 0.697). All factor loadings were found to be statistically significant at the 0.05 level. Arguably, item QB11 (with a loading of 0.293) did not load adequately on dedication. Field (2009) suggested that a factor loading of 0.3 is regarded meaningful and that for a sample of 100, the factor loadings should be preferably greater than 0.5. Botha and Coetzee (2022) also suggested omitting this item from future studies.

Cronbach’s coefficient alpha determined the reliability and internal consistency of the three factors. Cronbach’s alpha coefficients were satisfactory (α ≥ 0.70) for vigour (0.786), dedication (0.797) and absorption (0.812) (Field, 2009). An alpha coefficient should ideally exceed the value of 0.7 to achieve reliability and internal consistency (Field, 2009). Thus, the data and the questionnaire scales are considered reliable and internally consistent.

The questionnaire captured data on a five-point Likert-type scale (with responses rated between 1 and 5, where 1 = strongly disagree and 5 = strongly agree). The responses indicated that the mean scores for all the factors were 3.7 and above. The highest mean score was achieved for dedication (M = 4.1), followed by vigour (M = 3.71) and absorption (M = 3.69) on a five-point Likert scale. Thus, the results showed that the respondents displayed acceptable levels of engagement during WFH, although there is room for improvement, in particular, to increase their vigour and absorption levels.

Hancock and Mueller (2010) suggested that it is good practice to report fit indices from three broad classes. Therefore, the following three goodness-of-model-fit indices were used to evaluate the model-data fit: the Chi-square statistic divided by degrees of freedom (CMIN/DF), the comparative fit index (CFI), and the root mean square error of approximation (RMSEA). The CMIN/DF ratio should be close to 1 for correct models (Bollen & Jackman, 1993). Scores below 1 indicate a poor fit between the data sampled and the hypothetical model utilised (Shadfar & Malekmohammadi, 2013). According to Kline (1998), a CMIN/DF ratio smaller than 3 indicates an acceptable model-data fit. According to Shadfar and Malekmohammadi (2013), a ratio closer to 2 is considered a strong fit, while ratios between 2 and 5 are considered appropriate. A CFI index of 1 indicates a perfect fit, while an index below 0.95 generally indicates a poor fit (Bentler, 1990). However, according to Hair et al. (2010) and Bentler (1990), an index greater than 0.9 for CFI indicates a good fit. For the RMSEA, an index of 0 indicates a perfect fit, while indices below 0.05 demonstrate a good fit. Indices between 0.05 and 0.08 are acceptable, while indices above 0.10 indicate a poor fit (Steiger, 1990). Table 2 reveals the model’s ratios and indices concerning the CMIN/DF, CFI and RMSEA indexes.

The three goodness-of-model-fit indices indicated an acceptable fit between the measurement model and the sampled data (see Table 2). The measurement model obtained a CMIN/DF ratio of 1.76, a CFI index of 0.864 and the following RMSEA values: 0.087; 0.068 (low); 0.106 (high).

A measurement model needs construct validity to be fit for use. Construct validity can be achieved if the model possesses both discriminant and convergent validity. The model has convergent validity if the average variance explained (AVE) value exceeds 0.5 (Fornell & Larcker, 1981). The results showed that the AVE of vigour = 0.466, dedication = 0.614 and absorption = 0.419. Thus, the AVE values of two of the factors were marginally below the recommended 0.5 AVE cutoff value. The dedication factor shows acceptable convergent validity, the vigour factor shows almost acceptable convergent validity, while the absorption factor lacks convergent validity. For a model to achieve discriminant validity, the square root of the AVE (√AVE) should be higher than the correlation coefficient between the two factors under scrutiny (Fornell & Larcker, 1981). The results show the √AVE for vigour = 0.683, dedication = 0.784 and absorption = 0.647 (see Table 3). Table 3 shows that the criterion (√AVE > r) was met, and there is discriminant validity between vigour and dedication (√AVE = 0.784; r = 0.710), but not concerning absorption, vigour (√AVE = 0.647; r = 0.677) and dedication (√AVE = 0.647; r = 0.671). Thus, it is concluded that the model possesses some evidence of construct validity. Campbell and Fiske (1959) seminally warned that these findings are not uncommon because partial convergent and discriminant validity is common in exploratory models (such as this one). Likewise, an exploratory model could have good discriminant validity but relatively poor convergent validity (or vice versa) (Hill & Hughes, 2007). In such cases, emphasis should be on the confirmatory results showing consistency between the identified factors and the data (Cole, 1987).

Independent sample t-tests determined the association between gender and employee engagement (Table 4).

The independent sample t-test results revealed significant differences in the mean scores of male and female respondents for vigour (p = 0.046; d = 0.41) and dedication (p = 0.048; d = 0.37); the effect sizes showed a small to medium effect. The male respondents displayed more vigour (Male: M = 3.87; Female: M = 3.63) and dedication (Male: M = 4.25; Female: M = 4.04) during WFH than the female respondents.

The ANOVA revealed the associations between the socio-demographic variables marital status, nature of employment, WFH situation and employee engagement (see Table 5).

The ANOVA indicated significant differences between the mean scores of the different marital categories for vigour (p = 0.013) and dedication (p = 0.010). The respondents who were single scored lower on vigour (M = 3.40) and dedication (M = 3.78) than the respondents who were in a relationship, widowed, divorced and separated (vigour: M = 3.76, d = 0.62; dedication: M = 4.19, d = 0.52), and were married (vigour: M = 3.84, d = 0.71; dedication: M = 4.22, d = 0.56); the effect sizes showed medium to large effects.

The effect sizes further showed that the respondents who belonged to the management category scored higher on vigour (M = 3.83) and absorption (M = 3.98) than the respondents who belonged to the academic category (vigour: M = 3.66, d = 0.29; absorption: M = 3.69, d = 0.46) and support (vigour: M = 3.68, d = 0.24; absorption: M = 3.63, d = 0.46) categories; the effect sizes indicated small to medium effects.

Furthermore, the effect sizes showed, regarding WFH situation, that respondents who were working from home but came to the office occasionally or a few days a week scored higher on vigour (M = 3.81) and absorption (M = 3.78) than respondents who were working from the office most of the time (vigour: M = 3.64, d = 0.26; absorption: M = 3.53, d = 0.34). Interestingly, the effect sizes showed that the respondents who were working from the office all the time (M = 3.82: d = 0.38) were more absorbed in their work than the respondents who were working from the office most of the time (M = 3.53); however, the effect was small.

Spearman’s rank-order correlation was performed to establish the linear relationship between age, highest qualification, years working at the university and period working from home with employee engagement (see Table 6).

Spearman’s rank-order correlation established medium positive correlations between age, vigour (p = 0.00, r = 0.296) and absorption (p = 0.00, r = 0.278). This indicates that the older employees were, the more vigour and absorption they displayed. In addition, a small positive correlation was found between the years working at the university and absorption (p = 0.037, r = 0.205), signifying that the longer employees were in the institution, the more they found themselves absorbed in their work.

Spearman’s rank-order correlation test determined the linear relationship between vigour, dedication and absorption (see Table 7).

Large positive correlations exist between vigour and dedication (p = 0.00, r = 0.710), and vigour and absorption (p = 0.00, r = 0.677). This indicates that the more vigour employees expressed in their jobs, the more dedicated and absorbed they are. In addition, a large positive correlation was found between dedication and absorption (p = 0.00, r = 0.671), signifying that the more dedicated employees are to their jobs, the more they find themselves absorbed in their work.

This study contributes to the body of knowledge surrounding employee engagement in the WFH context within the private higher education sphere. The study’s results could be used to address the issue of employee engagement during WFH within private higher education in South Africa. As the WFH model has been identified as a model that would be viable for the foreseeable future, the results of this study would be beneficial to private higher education even post the COVID-19 pandemic as many institutions continue remote teaching, learning and assessment. Higher education leaders should teach and promote healthy and engaged workplace cultures in the new world of work through the appropriate adaptations of the recommendations made below.

The study focussed on one private higher education institution in South Africa. Therefore, the results cannot be generalised to other private or public higher education institutions worldwide.

The following recommendations are suggested to ensure that the engagement of personnel in higher education is improved and sustained.

The most important engagement strategy is organisational support, given the complexities of WFH. Organisational support allows employees to successfully WFH, and is the most viable engagement strategy in the digital age. Management of organisations should ensure that employees working from home are provided with appropriate guidance on their roles, constant communication, appropriate team building, required ICT infrastructure, and meaningful deliberation is conducted to ensure that employees have conducive spaces to WFH.

Given the anxieties and strain placed on employees working from home, a meaningful focus on employee wellbeing was noted as another viable solution to engagement. Organisations could assist employees by ensuring they are equipped with healthy workspaces, work routines, socialisation opportunities, wellbeing physical and mental practices, and investment in technologies that promote wellbeing by allowing employees to communicate and keep abreast with one another and to recognise employee efforts, for example.

Although mentioned above as communication, the emphasis on developing a social union cannot be overstated. Organisations should make concerted efforts to ensure that employees experience a sense of belonging to the organisation, are communicated with on shared visions, are confident that their input is trusted and provide them with recognition of their work.

Furthermore, as only a few studies in South Africa reported on employee engagement during the WFH context, and because it is regarded as a viable workplace model for the foreseeable future, it is suggested that the topic be further explored, not only in the private and public higher education sector but also in organisations in the private and public sector. A qualitative study could also provide a deeper understanding of the phenomenon.