Cocker, Martin, Scott, Venn and Sanderson (2013) note that there is scant SME-specific research focused on the mental health of SME employees. As a result, available resources are often informed by research carried out in large organisations and developed without evidence that mental health factors in large organisations mirror those of SMEs (Cocker et al., 2013). Martin et al. (2009) point out that SMEs provide an outstanding environment for emotional infection research, as staff are often closer to the feelings of their executives. Some PWB studies were among entrepreneurial ventures (Parasuraman, Purohit, Godshalk, & Beutell, 1996; Uy, Foo, & Song, 2013), but a search of more than 700 studies using Ryff’s (1989) PWB scale found no reference to small business studies (Ryff’s Scale Information Document, 2016). It can be observed from the aforementioned interpretations that SME employees are academically relevant, especially in terms of mental health research. The following section focuses on the study’s theoretical framework.

Social Exchange Theory is based on the relational process and the principle of voluntary value exchange between individuals (employer-to-employee) and organisations (business-to-business [B2B]). According to the SET, ‘relationship building is based on a subjective cost-benefit analysis and options assessment’ (Liu, Min, Zhai, & Smyth, 2016, p. 54; Roberts-Lombard et al., 2017, p. 2). This means that the partners in a relationship will assess the future of a relationship on the value still to be accrued from a partner (e.g. financial rewards, a partner’s trustworthiness and the level of future relationship satisfaction) (Sierra & Quitty, 2005:393). In addition, Ward and Berno (2011) opine that:

In this regard, the SET principle of reciprocity explains the positive effects of POS, which is one of PWB’s precedents. If employees feel supported, they are likely to return the appreciation through good performance, which, in turn, raises expectations for reward (Kurtessis et al., 2017). The principle of reciprocity is also applied when it is evident that an organisation lacks support or displays abusive behaviour. Poor employee behaviour is likely to result from unfavourable treatment (Cohen-Charash & Mueller, 2007).

This study postulates that the SET is associated with POS, CGD, EL and PWB. The SET confirms that when an organisation offers tangible and intangible resources, employees will respond in kind (Blau, 1964; Saks, 2006). Perceived organisational support has conventionally been considered as a social exchange process, as it forms the basis of exchange relationships (Blau, 1964; Eisenberger, Huntington, Hutchison, & Sowa, 1986; Shantz, Alfes, & Latham, 2016). Scholars have posited that POS acts as a beneficial resource for employees (Ensher, Thomas, & Murphy, 2001), and some consider career development as a part of the SET (Shantz et al., 2016). According to Su and Swanson (2019), the SET allows for greater levels of organisational trust and identification, which can subsequently enhance employee well-being. Perceived organisational support is a central construct in the SET, and the present study aligns with the limited research available that considers conceptualised POS as an essential resource that can alleviate the negative consequences of work-related attitudes and states (Hochwarter, Witt, Treadway, & Ferris, 2006; Panaccio & Vandenberghe, 2009; Witt & Carlson, 2006).

According to Kay and Heckhausen (2015), the Motivational Theory of Lifespan Development (MTD) explains the processes through which humans optimally pursue their development goals. In addition, MTD suggests that people select targets based on opportunities and constraints to promote development across multiple domains, and preserve various objectives (Heckhausen, 1997; Heckhausen & Schulz, 1993; Heckhausen, Schulz, & Wrosch, 1998; Schulz & Heckhausen, 1996). Development is most successful when people pursue suitable goals, engage in goal-oriented efforts to pursue their goals and disengage when goals have been achieved or are no longer achievable (Kay & Heckhausen, 2015). The MTD has consequences for lifelong achievement of goals, health and well-being (Kay & Heckhausen, 2015). Shane and Heckhausen (2019), however, claim that the MTD can be applied to a work environment also. The MTD questions aspects of the working life of an individual, including how work and career goals function as development goals that structure and direct the development of individuals; how control capacities and trajectories differ across careers and career stages; and how individuals organise their control efforts to capitalise on career development. As this study investigates the impact of CGD on PWB, the MTD will serve as the theoretical lens.

Budeli (2012) states that one of the best ways to support people’s interests is by applying the empowerment theory. Robbins, Chatterjee and Canda (1998) posit that empowerment aims to provide conceptualisations of social stratification and oppression to identify personal and political barriers and dynamics that maintain oppression. In addition, empowerment provides value frameworks to promote human empowerment and liberation and identifies practical strategies to overcome oppression and achieve social justice. The empowerment theory recognises conditions of inequality and injustice (Budeli, 2012) and aims to help marginalised people. Small and medium-sized enterprise employees have the power to ensure effectiveness and social justice in their lives and work environment. Applying the empowerment theory can increase employees’ PWB and stimulate SME performance. If SME leaders or managers empower their employees, their employees will feel more active in their organisations and this will help improve the quality of the organisation’s output.

Perceived organisational support refers to the perception of employees regarding the organisation’s appreciation of their efforts and care about their well-being (Eisenberger et al., 1986), which may play a crucial role in employee readiness (Gigliotti, Vardaman, Marshall, & Gonzalez, 2019), especially during organisational change (Gigliotti et al., 2019; Vardaman, Amis, Dyson, Wright, & Randolph, 2012). Tremblay, Gaudet and Vandenberghe (2019) consider POS as a collective, commonly accepted view that the organisation supports its members. For employees, POS is expressed through payment, job enrichment, recognition and other benefits. Concern about the well-being of employees is also reflected in the offer of organisational policies and human resources practices that facilitate meeting and balancing job and external demands (Worley, Fuqua, & Hellman, 2009). Perceived organisational support meets the socio-emotional needs of the employee, contributing to organisational identification, emotional commitment (Kurtessis et al., 2017) and positive impact (Caesens, Stinglhamber, & Ohana, 2016). Prior literature on POS (Gigliotti et al., 2019; Tremblay et al., 2019) concurs that it epitomises organisations that are concerned for their employees. Other scholars (Arogundade, Arogundade, & Adebajo, 2015) posit that the personification of organisations is the result of POS. The aforementioned studies did not, however, empirically examine the impact of POS on the PWB of employees.

Career goal development refers to the objectives that people pursue at work. These goals include promoting, increasing pay and developing skills (Rahim & Siti-Rohaida, 2015; Zikic & Klehe, 2006). Koekemoer (2014), however, argues that a successful career development by the employee implies that certain organisational objectives have been achieved. Career development takes the form of a lifelong career in one organisation, or a steady climb up the corporate ladder for the individual (Koekemoer, 2014). Career goal development can therefore be defined as individuals’ objective or subjective perception of achievements during their work experiences, suggesting an objective and subjective dimension (Koekemoer, 2014). Some researchers postulate that traditional career development models are only marginally effective in the modern global economic context (Dean, Shogren, Wehmeyer, Almire, & Mellenbruch, 2019). Hu, Hood and Creed (2018) investigated career feedback and stress, while Dickmann and Cerdin (2018) explored career development and transfer of career capital, such as skills and connections. These studies did not, however, assess the relationship between CGD and mental health.

Empowering leadership is defined as behaviours that share power with team members, and, in turn, raise the intrinsic motivation levels of all members (Gyu-Park, Sik-Kim, Yoon, & Joo, 2017). Empowering leaders show four types of behaviours, namely, emphasising the importance of work, providing decision-making participation, fostering confidence that performance will be excellent and removing any bureaucratic constraints (Ahearne, Mathieu, & Rapp, 2005). These leadership empowerment behaviours are based on Conger and Kanungo’s (1988) broad conceptualisation of empowerment. They indicate that empowerment is a process of motivation, rather than a simple delegation of power to employees. A team leader needs to help team members understand the importance of their roles, involve them in the decision-making process, believe in their ability to achieve high performance and simplify administrative rules and procedures in order to be empowered and motivated (Ahearne et al., 2005).

Previous empirical studies have shown substantial evidence of the positive relationship between EL and different team and organisational outcomes, including task performance (Vecchio, Justin, & Pearce, 2010), job satisfaction (Konczak, Stelly, & Trusty, 2000; Vecchio et al., 2010), engagement (Konczak et al., 2000) and organisational citizenship behaviour (Yun et al., 2007). Gyu-Park et al. (2017) explain that only a few studies have looked at the influence of EL on employee psychology. It is, therefore, important to understand how leaders influence the mindset and organisational behaviour of employees. Kearney, Shemla, Van Knippenberg and Scholz (2019) posited that EL leads to goal clarity in organisations but failed to address employee mental well-being, while Han, Harold and Cheong (2019) examined empowerment leadership and trust in organisations. The above-mentioned research however did not examine the influence of EL on an employee’s PWB.

Psychological well-being refers to positive mental health (Edwards, 2005; Ismail & Desmukh, 2012). Farrington (2017:47) describes PWB as the ‘perception of engaging with life’s existential challenges’, whereas Winefield, Gill, Taylor and Pilkington (2012:2) describe it as ‘a combination of positive affective states such as happiness and functioning with optimal efficacy in individual and social life’. Optimum PWB levels are likely to depend on the external circumstances of an individual, as well as the resources and challenges they face (Farrington, 2017). Psychological well-being has been investigated in various organisational contexts, including employee well-being (Žižek, Treven, & Čančer, 2015), commitment (Kanten & Yesıltas, 2015) and job satisfaction (Jones, Hill, & Henn, 2015), as well as relationships between family and work (Grzywacz & Butler, 2005). It has been found that PWB is associated with positive results in terms of work performance of an employee. For instance, Wright and Cropanzano (2000) point out that PWB benefits include an increase in an individuals’ cognitive functioning and health, and ultimately the happy functioning of society.

Psychological well-being is generally considered necessary for effective human functioning (Ryan & Deci, 2001). Previous literature on PWB explored workplace collegiality, such as the relationship between employees and PWB (López, Sanderman, Ranchor, & Schroevers, 2018). Hernandez et al. (2018) assessed PWB as physical health, reinforcing the notion that PWB is essential in organisations. However, these studies did not investigate the impact of POS, CGD and EL on PWB. Robertson and Cooper (2010) found that individuals with higher levels of PWB behave in ways that may lead to higher levels of workplace engagement, which, in turn, affects business success. In this context, it may be concluded that SME employees who display positive PWB are more likely to contribute to positive outcomes, like successful businesses.

It is imperative to elucidate on the nexus between POS and PWB. Caesens et al. (2016) revealed that PWB is positively linked with overall well-being. In their study that determines the influence of POS and PWB on organisational commitment, Aggarwal-Gupta, Vohra and Bhatnagor (2010) revealed that POS significantly influences PWB and all components of organisational commitment. Roemer and Harris (2018) also pointed out that, in terms of POS, employees with perceived higher levels of organisational support are more likely to experience higher levels of well-being. Perceptions of support seem to foster positive emotions in employees and may be strongly associated with their PWB (Ryff & Singer, 1998). Therefore, the following hypothesis has been formulated:

The more focused employees’ career goals are, the more likely they are to engage in behaviours that will help them reach their goals, and the greater their motivation to participate in development activities (Rahim & Siti-Rohaida, 2015). Evidence shows that individuals who actively reflect on career goals and have stronger vision in what they want to attain during their career report a higher level of career success (De Vos & Soens, 2008). In addition, self-managing individuals are more likely to strive to obtain their desired career goals, which in turn will make them feel more successful in their career (Arthur et al., 2005). In the career literature, goal focus has been shown to be an important determinant of career goal attainment and satisfaction with career progress (Sugalski & Greenhaus, 1986). Therefore, career goals are expected to influence the level of PWB of individuals (Rahim & Siti-Rohaida, 2015). However, very few studies have been conducted on the theoretical relationship between CGD and PWB (Rahim & Siti-Rohaida, 2015). Based on this assumption, a second hypothesis is formulated as follows:

Gyu-Park et al. (2017) explained that empowerment can affect job-related characteristics of PWB, such as skill utilisation and workload. In this vein, Heller and Wilpert (1981) demonstrated that employees’ skill utilisation is associated with participatory decision-making, which is a characteristic of empowering leaders (Arnold, Arad, Rhoades, & Drasgow, 2000). Thus, empowerment can be an effective means to maintain job satisfaction and reduce burnout when an employee’s workload is increased (Andrews, Wilmington, & Kacmar, 2014). Empowering leadership focuses on the relationship between leaders and subordinates, while attending to individual psychological reasons for motivation (Thomas & Velthouse, 1990). In another study, Fitzsimons and Fuller (2002) showed that having a sense of empowerment is associated with reduced psychological distress. They also reported a positive association between psychological empowerment and self-esteem. Molix and Bettencourt (2010) posited that psychological empowerment partially mediates the relationship between group identity and PWB. Along these lines, deriving from the literature and the empirical proof above, it is hypothesised that:

For this study, a self-administered questionnaire was used to collect the necessary data. Leedy and Ormrod (2010:197) describe a questionnaire as a research instrument in which a researcher poses a series of questions to willing respondents and summarises their responses with percentages, frequency counts or more sophisticated statistical indexes, upon which inferences are made about a population. Because of the nature of this research, a questionnaire was deemed the most appropriate data collection method. The questionnaire was categorised into five sections, all of which are described in the following.

Section A of the questionnaire comprises demographic questions, such as gender, age, level of education, number of employees (full-time) in the business and age (years) and location of the business. Section B assesses POS and comprises eight items adapted from Eisenberger et al.’s (1986) study, with a Cronbach’s alpha of 0.93. In Section C, CGD was measured using four items adapted from Rahim and Siti-Rohaida’s (2015) work, which had a composite reliability value of 0.901. Section D of the questionnaire comprises questions on EL that were assessed using a 12-item scale adapted from Gyu-Park et al. (2017), with a Cronbach’s alpha of 0.97. Section E measures PWB and adapted five items from the scale used by Zeng et al. (2014), which met the reliability criteria with a Cronbach’s alpha of 0.61. Responses were recorded using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) for all the variables in this study.

In terms of reliability, all Cronbach’s alpha coefficients were above 0.8, exceeding the expected value of 0.7 (Nunnally & Bernstein, 1994). The Cronbach’s aplha values are as follows: POS (0.795), CGD (0.782), EL (0.818) and PWB 0.742, respectively. A more detailed presentation of the reliability and validity results is given in Table 2.

The data gathered from SMEs were recorded on an Excel spreadsheet after screening returned questionnaires. The data were analysed using a software for descriptive statistics, Cronbach’s alpha values and correlations, and the Statistical Package for Social Sciences (SPSS version 25.0). To test the psychometric properties of the measurement scales, and the hypotheses, the Analysis of Moment Structures (AMOS version 25.0) statistical software was used.

In the wake of checking for missing values and anomalies, it was found that 250 questionnaires out of the sample of 321 were completed (a 77.8% response rate). Of these, 71 were unusable because a few items on the questionnaires were not answered. Table 1 presents the classification of information related to gender. The largest proportion of the sample indicated that they were men (62.4%; n = 156), followed by those who revealed that they were women (33.2%; n = 83). The remaining 4.4% (n = 11) preferred not to mention their gender.

Most of the respondents (32.4%; n = 81) were between 50 and 59 years old, 30.0% (n = 75) were between 31 and 39 years old, 18.4% (n = 46) were between 40 and 49 years old, 12.4% (n = 31) were between 18 and 30 years old, and the remaining 6.8% (n = 17) revealed that they were 60 years and above.

In terms of the level of education, 37.6% of respondents had a diploma, 30.0% (n = 75) had degrees, 20.0% (n = 50) had no formal education and the remaining 12.4% (n = 31) had only obtained some basic education.

The employment figures for this specific sample profile (Table 1) indicate that most of the SME businesses employ fewer than 50 employees, 44% (n = 110) employ 10–50 employees, 32% employ 100–200 employees, 21.2% (n = 53) employ 50–100 employees, with the remaining 2.4% (n = 6) employing fewer than 10 employees. These findings coincide with the findings of Moodley (2002:37), who asserts that ‘SMEs are important, because, although recruiting less per entity, their potential for job creation is in numbers’. Table 1 provides the profile of the surveyed SMEs in terms of their location. Most of the respondents (74%; n = 185) indicated that their business was local (in the industrial areas of the Gauteng province of South Africa). The remaining 26% (n = 65) indicated that they are located in the various Central Business Districts (CBDs) of the Gauteng province.

The assessment of the measurement scales’ psychometric characteristics was performed through a CFA to determine the construct’s reliability, validity and model fit. Table 2 presents the outcomes of the CFA assessment.

The means, Standard Deviations (SD), item to total correlation values, Cronbach’s alpha values, composite reliability values, average variance extracted values and standardised factor loadings are reported in Table 2.

As shown in Table 2, mean scores ranging between 3.58 and 4.21 (out of 5.0) were computed for all the constructs examined in the study. These scores depict a collective inclination towards both the ‘agree’ and ‘strongly agree’ positions on the Likert scales. As posited by Hair, Babin, Anderson and Tatham (2010), the relationship between the mean and the SD is that a small estimated SD denotes that respondents’ responses were consistent and that the response distributions lie close to the mean. Conversely, a large SD indicates that the responses are varying, making the response distribution values fall away from the mean of the distribution (Hair et al., 2010). Moreover, the SD value should be less than 1, but it is recommended to at least include a value of less than 2 to ensure that there is no issue of outliers (Drost, 2011). Table 2 reveals that the highest SD value was reported at 1.266 and the lowest SD value was 0.881. This information indicates that the data points are clustered around the mean. The SD values are below 2; hence, it indicates that there was no presence of outliers.

A further analysis of Table 2 reveals that, on POS, three items (POS6, POS7 and POS8) were deleted, as the item to total correlation values were less than 0.5. On EL, four items (EL9, EL10, EL11 and EL12) were deleted, as the POS8 item to total correlation values were less than 0.5. On PWB, two items were deleted, namely PWB1 and PWB2, as the item to total correlation values were less than 0.5. It is worth mentioning that all these items were deleted because they did not meet convergent validity, they did not measure at least 50% of what they were supposed to measure. As such, these items did not require further analysis.

According to Nunnally (1978), the reliability of a measure is supported if Cronbach’s alpha is 0.7 or higher. The results provided in Table 2 range from the lowest Cronbach’s alpha (0.742) to the highest (0.818). Cronbach’s alpha scores indicate that each construct exhibits strong internal reliability (Lee, 2009). Because the Cronbach’s alpha values of the constructs exceeded the recommended value of 0.70, they met the required threshold, demonstrating that the constructs used to measure variables are very reliable for all the variables.

Table 2 shows the loading of each item on their construct. The lowest value for each respective item loading for the research constructs is 0.514; all the individual item loadings exceed the recommended value of 0.5 (Anderson & Gerbing, 1988). This indicates that all the measurement instruments are acceptable and reliable, because all the individual items converged well and with more than 50% of each item’s variance shared with its respective construct (Fraering & Minor, 2006).

Composite reliability and AVE for each construct were also computed and assessed to determine if they met the required thresholds for reliability and validity. As per the results shown in Table 2, the lowest CR value (0.740) is well above the recommended value of 0.6 (Hulland, 1999), while the lowest obtained AVE value (0.420) is above the recommended value of 0.4 (Anderson & Gerbing, 1988). This indicates that convergent validity was achieved, further confirming excellent internal consistency and reliability of the measurement instruments used. As such, all pairs of constructs revealed an adequate level of discriminant validity (see Table 2). By and large, these results provided evidence for acceptable levels of research scale reliability (Chinomona & Chinomona, 2013:20; Chinomona & Mofokeng, 2016).

In this study, hypotheses testing was determined by path coefficient values as well as the p-values for the structural model. In the model, relationships of constructs proposed in this study generate the coefficients of paths. These coefficients form the basis of hypothesis examination.

Permission was granted by the management of the Small Enterprise Development Agency (SEDA). The researchers obtained the permission letter which allowed them to collect data from SMEs registered with SEDA before questionnaires were given out to respondents. Ethical consideration was approved unconditionally and this research study acted in accordance with the ethical standards of academic research, which include, among other things, protecting the identities and interest of respondents and assuring confidentiality of information provided by the respondents. Respondents gave their informed consent to participate in this research and were informed beforehand about the purpose of the investigation to ensure that they were not misled. Despite all the above-mentioned precautions, it was made clear to the respondents that the research was only for academic research purpose and their participation was absolutely voluntary. No one was forced to participate.