A South African study on antecedents of intention to quit amongst employees in bed and breakfast establishments in the Free State province

Bed and breakfasts (B&Bs) in South Africa are small accommodation establishments that strive to grow, whilst competing with large accommodation establishments (Hikido 2017). They are faced with several challenges such as low-paid employees who are not motivated to work because the work environment is stressful and demanding (Hlanyan & Acheampong 2017). Small accommodation businesses, especially B&Bs in South Africa, are also faced with the challenges of low income during off-seasons (Hsieh & Lin 2010), as well as competition from large accommodation establishments (Van Schalkwyk et al. 2010). It is, therefore, highly likely that B&B employees have high intentions to quit.


Introduction
Bed and breakfasts (B&Bs) in South Africa are small accommodation establishments that strive to grow, whilst competing with large accommodation establishments (Hikido 2017). They are faced with several challenges such as low-paid employees who are not motivated to work because the work environment is stressful and demanding (Hlanyan & Acheampong 2017). Small accommodation businesses, especially B&Bs in South Africa, are also faced with the challenges of low income during off-seasons (Hsieh & Lin 2010), as well as competition from large accommodation establishments (Van Schalkwyk et al. 2010). It is, therefore, highly likely that B&B employees have high intentions to quit.
It, therefore, goes without saying that appreciating the antecedents to employees' intention to quit and the identification of strategies to retain such employees cannot be over-emphasised. Researchers (Burakova et al. 2014;Halawi 2014;Treuren & Frankish 2014) have pointed to a set of occupational, environmental and personal antecedents to employees' intention to quit. Job satisfaction and organisational commitment have been identified as the two common occupational antecedents to employees' intention to leave (Steil, Floriani & Bello 2019), and with reference to environmental antecedents, these include but are not limited to the existence of employment alternatives (Heinen et al. 2013), the economic development level, social security policy, employment policy, labour demand and supply conditions (Li & Lu 2014).
According to Steil et al. (2019), age, schooling, sex, marital status, professional experience and family or kinship responsibility are the six personal antecedents that have been researched the most. Whilst considerable research has been conducted on antecedents to employees' intention to leave, there is a dearth of research on the antecedents of employees' intention to quit amongst small businesses, especially B&B establishments in the hospitality sector in South Africa. Whilst evidence is available on how organisation-related antecedents such as low compensation and those that are individual-related, for example, job satisfaction have been linked to intention to quit in large hospitality organisations in developing economies (Mat-Nor et al. 2017), the same cannot be said of such a relationship amongst B&Bs found in either urban or rural contexts in developing economies.
With such paucity of research, the objective of this study was, therefore, to determine the influence of selected individual (job satisfaction, organisational commitment and job stress) and organisational factors (human resource practices, quality of work environment [QWE] and organisational structure) on intention to quit amongst employees in B&Bs in a particular district in the Free State province of South Africa. The choice of these factors is premised on empirical evidence showing that they have been mostly linked to several behavioural outcomes in the hospitality sector (Chen, Ayoun & Eyoun 2018;Grobelna, Sidorkiewicz & Tokarz-Kocik 2016).
It has to be noted that employee turnover is costly to both parties in the employment relationship as the establishment has to invest more effort and time in acquiring, developing and retaining its employees (Grobler & De Bruyn 2011) and to employees, as it directly interferes with their career development. High turnover amongst employees leads to direct financial costs, and also results in below average organisational performance manifested by low quality, lessened efficiency, decreases in morale and even service disruptions (Jang & George 2012). The fact that employees have a high propensity to leave and look for alternative employment because of low salaries (Omar et al. 2012) gave impetus to the current inquiry. As such, the study sought to determine the impact of selected individual and organisational factors on intention to quit amongst employees in B &Bs from a selected district in the Free State province.

Literature review Job satisfaction versus intention to quit
Job satisfaction is a general expression of workers' positive attitudes towards their jobs (Thomas 2015). It is pivotal in achieving positive organisational outcomes such as increased productivity , but can also lead to detrimental outcomes such as intention to quit (Poghosyan et al. 2017) if employees are not satisfied. Ibrar (2015) reported that job satisfaction does not only increase productivity but also decreases intention to quit amongst employees. This finding is corroborated by a study performed within the hospitality industry, which established that dissatisfied employees considered looking for alternative employment when they were not happy with working conditions, pay, relations with the supervisor and the job itself (Akgunduz & Sanli 2017). Similarly, Santerosanchez et al. (2015) posited that 24/7 working hours and high levels of job pressure in large established hotels are associated with job dissatisfaction, and ultimately intention to quit.
Contrastingly, other researchers found out that job satisfaction had no direct relationship with intention to quit but mediates the relationship between creativity and intention to quit (Zhen, Mansorzd & Chong 2019). Notwithstanding this contradiction, there is also overwhelming evidence arguing that satisfaction levels differ by employee position, and, therefore, intentions to quit are not the same across all employees in any organisation (Lu et al. 2016). Based on these findings, one can, therefore, conclude that job satisfaction has an impact on intention to quit amongst employees, even amongst B&B employees. The above evidence leads to the following hypothesis: H1: Job satisfaction has a significant positive effect on intention to quit.

Organisational commitment versus intention to quit
Organisational commitment is defined as: [A] strong belief in and acceptance of the organisation's goals and values, a willingness to exert considerable effort on behalf of the organisation and a definite desire to maintain organisational membership. (Watson 2010, p.18) The concept is conceived as associated with intention to quit in the sense that the two concepts are directly opposite to each other -one concerns the notion of being attached to the organisation and the other being detached from an organisation (Kim, Song & Lee 2016).
Organisational commitment can assume three dimensions: normative commitment, affective commitment and continuance commitment (Meyer & Allen 1997). Normative commitment reflects an employee's sense of duty to remain in employment, whilst affective commitment refers to the strength of an employee's identification with and involvement in an organisation (Porter et al. 1974). Continuance commitment deals with the cognitive attachment that exists between an employee and the organisation as the costs of quitting outweigh the benefits (Khan et al. 2018).
Employees who do not have much organisational commitment may have a high propensity to leave the organisation (Gatling, Kang & Kim 2016). In contrast, employees with stronger organisational commitment are less likely to develop intentions to quit and leave the organisation (Wijnmaalen, Heyse & Voordijk 2016). Committed employees who feel much attached to an organisation and are extremely dedicated are less likely to show intent to quit because of their high level of commitment and willingness to invest more into the organisation.
According to Kim, Im and Hwang (2015), service industry jobs are associated with high-stress levels, -a factor that works against employee organisational commitment. This is supported by Jung and Yoon (2016) who contended that the service industries' dependence on human resources makes employee commitment critical to maintain, failure of which has negative outcomes detrimental to the organisation such as intention to quit might ensue.
Related to this finding is a study which showed that whilst hospitality organisations can recruit talented and highly motivated employees, they seem to have difficulty in arousing their organisational commitment and, therefore, retain them (Chiang & Liu 2017). Because retaining employees is determined by how committed employees are (Jang & Kandampull 2018), hospitality organisations might need to ensure that there is commitment amongst their employees. As such, it is hypothesised that: H2: Organisational commitment has a significant negative effect on intention to quit.

Job stress versus intention to quit
Job stress has been defined as 'the pattern of emotional states and psychological reactions occurring in response to inability to cope with stressors from within or outside an organization' (Ekienabor 2017:124). The reactions, which lead to high job stress, are likely to result in low morale and high turnover amongst employees (Bowness 2017). Lo and Lamm (2015) established that working in the hospitality industry can be stressful because the industry is highly labour-intensive and has increasingly harsh environmental demands imposed upon it.
Similarly, Namra and Tahira (2017) found that the nature of work within hotels and B&Bs, for example, includes frequent deadlines, unexpected interactions with guests, long working hours, night and evening work, repetitive work, high emotional demands, low influence (control), shift work, extensive work space and problems with coordination of work. Consequently, such demands on employees lead to work-related stress.
Furthermore, evidence from a study by Mansor and Mohanna (2018) revealed that unpredictable and irregular working hours were stressful factors amongst hospitality organisations that led to increased intention to quit amongst employees. This was reaffirmed by Newnham (2017) who established that hospitality industry employees suffer from dissonance because they are required to display certain emotions all the time, which might not represent their actual feelings at any given time. The fact that these employees are continuously suppressing their actual emotions might lead to outcomes such as intentions to quit. The above review of literature led to the following hypothesis: H3: Job stress has a significant positive effect on intention to quit.

Human resource practices versus intention to quit
Human resource practices refer to those organisational activities which are directed at managing the pool of human resources and ensuring that the resources are employed towards fulfilment of organisational goals (Russo, Mascia & Morandi 2018). Some studies indicated that human resource activities linked to turnover intentions include the manner in which performance management is done (Nankervis & Debrah 2015) and how employee compensation is managed (Khaleefa & Al-Abdalaat 2017).
Aside from pay, promotional opportunities, effective and supportive leadership, satisfactory compensation and work-group cohesion have the potential to influence turnover (Qiu et al. 2015). Akgunduz, Gok and Alkan (2019) also concluded that good monetary incentives offered to employees often reduced employees intentions to quit. A related study by Santhanam et al. (2017) amongst hotel frontline employees established that selection, training and compensation practices had an influence on turnover intentions.
Interestingly, the study concluded that psychological contract breaches enhanced turnover intentions by employees, notwithstanding effective implementation of human resource management (HRM) practices. According to Lo and Lamm (2015) although employees in the hospitality industry are vulnerable in terms of poor working conditions and low wages, good compensation management practices might reduce their intentions to quit. However, Altarawmneh and Al-Kilan (2010) found out that despite the robust investment levels in HRM practices in Jordanian hotels, HRM practices themselves did not directly influence intentions to quit as employees' http://www.sajesbm.co.za Open Access intentions to quit may depend on other factors. Based on the discussions above, it is hypothesised that: H4: Human resource practices have a significant positive effect on intention to quit. Agbozo et al. (2017) confirmed that the QWE was a critical factor that influenced several outcomes, such as the level of satisfaction, intention to quit and motivation of the employees. The QWE is characterised by a good physical environment (e.g. in terms of heat or noise), a conducive psychological and social work environment (Agbozo et al. 2017).

Quality of work environment versus intention to quit
Bednarska (2016) posited that the hospitality environment is one of the most complicated places to work in, with pressures on meeting expectations of customers and working long hours remaining foci pressure points for possible reactions by employees in the form of counterproductive work behaviours. Robinson et al. (2016) also acknowledged that shift work and the number of working hours put pressure on hospitality employees, significantly affecting their psychological, physical and emotional well-being, leading to turnover intentions.
A small survey of employees in New Zealand by Markey, Ravenswood and Webber (2015) found that a majority of employees intending to quit viewed the QWE as poor. The results also showed that employees were likely to quit if they were stressed, if they were not a parent, if they experienced reduced job satisfaction and did not receive adequate important information, although the impact of the aforesaid factors was greater in workplaces with a good QWE. In view of this, it is hypothesised that: H5: The QWE has a significant positive effect on intention to quit.

Organisational structure versus intention to quit
Organisational structure is defined as the formal system of authority relationships and tasks that control and coordinate employee actions and behaviour to achieve organisational goals (Valentina & Cvelbar 2018). Organisational structure is related to employee attitudes and behaviour in organisations (Valentina & Cvelbar 2018). A study by Hamzat et al. (2020) amongst library and information science professionals in Nigeria's Osun State private universities revealed that the ownership structure of the organisation has a direct influence on the professionals' turnover intention.
A related study by Lensen (2016) established that although a flat or non-hierarchical organisation (characterised by few lines of command and few layers) leads to greater expectations regarding job satisfaction and is considered more attractive, but it does not necessarily lead to lower turnover intentions. Regarding a functional structure, Cummings and Worley (2015) highlighted that the functional structure's emphasis on central power discourages growth, diminishes selfconfidence amongst employees and discourages them from becoming innovative and involved in business activities. Such a structure tends to work best in small-to-mediumsized enterprises like the B&B establishments under study (Sinha 2017).
Another organisational structure, the customer-centric structure, which focuses on sub-units for the creation of solutions and satisfaction of key customers, is considered appropriate for hotels as it aims at meeting customer needs (Cummings & Worley 2015). Previous researchers (Palacios-Marques, Guijarro & Carrilero 2016) have attested that this approach can deliver significant benefits to an organisation. The benefits include: improved customer experience, reduced intentions to quit amongst employees, consistent engagement with customers and increased sales compared to other structural types like the functional structure. Based on the above discussion, the following hypothesis has been proffered: H6: Organisational structure has a significant positive effect on intention to quit.

Organisational and individual factors -More empirical evidence Human resource practices and job satisfaction
Human resource (HR) practices and job satisfaction have long been studied and are assumed to be closely associated (Ahmed, Zaman & Khattak 2017;Cortini 2016;Kampkotter 2017), because it is believed that sound HR practices result in better levels of job satisfaction, which ultimately improves organisational performance.
The aforesaid relationship has also been established within the hospitality sector in Portugal (Ferreira et al. 2017), Thailand (Ashton 2017) and Nigeria (Onyebu & Omotayo 2017). It is evident from these studies that the use of specific HR practices in organisations is associated with greater levels of job satisfaction. This evidence led to the following hypothesis: H7: HR practices have a significant positive effect on job satisfaction.

Working environment versus organisational commitment
Another area of interest is how the working environment affects organisational commitment. However, this area has not enjoyed much empirical attention in the literature (Holston-Okae 2017). Yet, workplace environmental factors are essential elements that determine the level of employees' commitment, their concentration and performance and the sustainability of a business (Funminiyi 2018). Funminiyi argued that employees are always content when they feel that their immediate environment -both physical sensations and emotional states -are in line with their obligations.
McCoy and Evans (2015) opined that the way employees connect with their organisation's immediate workplace environment, influences to a great extent their commitment, efficiency, innovativeness, collaboration with other employees, absenteeism and, ultimately, their retention. In view of this above, it can be hypothesised that: H8: The QWE has a significant positive effect on organisational commitment.

Organisational structure versus job stress
The way an organisation is structured and how it is run can have a positive significant effect on job stress amongst employees. A study by Namra and Tahira (2017) found that job stress amongst hospitality employees was associated with changes in management and lack of participation in decision-making, which ultimately caused intention to quit because of employees' inability to be involved in the decisionmaking processes.
Despite the different context, a related study by Daoli and Mohsenvand (2017) found that employees who were subjected to a centralised organisational structure experienced job stress, which then caused higher absenteeism, lower productivity, workplace aggression and increased intentions to quit. Highly-centralised organisations often have low levels of flexibility which often affect employees' stress levels as workers have limited autonomy and control over their work (Daoli & Mohsenvand 2017). Most small businesses usually follow a centralised structure (Vitez 2017); hence, one can conclude that high level of stress develops amongst B&B employees which might lead to increased intention to quit. It is, therefore, hypothesised that: H9: Organisational structure has a significant positive effect on job stress.

Conceptual framework
The conceptual framework as shown in Figure 1 depicts the researchers' understanding of how the variables in the current study connect with each other. Figure 1 summarises the hypothesised relationships discussed in the literature review section. It is hypothesised that employees' intention to quit might be influenced by the two broad categories of factors -individual and organisational. No moderation influences were investigated. The framework further assumes that organisational factors have an impact on employees' individual factors.

Research design
A research design refers to the plan for the collection, measurement and analysis of data (Blumberg, Cooper & Schindler 2014). The study used the correlational ex post facto design. A correlational design attempts to describe relationships, rather than explain them (Gravetter & Forzano 2009). Although a correlational research does not imply causality, it allows for predictions to be made even though one may not have an idea why a relationship exists. The ex post facto is pre-experimental, implying that it does not meet the scientific standards of experimental designs, nor does it involve a control group (De Vos, Strydom, Fouché & Delport 2011). The ex post facto design provides a means by which researchers may examine the degree to which an independent variable could affect the dependent variable(s) of interest. Kothari (2004) pointed out that there are two basic approaches to research -quantitative and qualitative. Researchers generally show a preference for either type of method, reflecting their research's philosophical point of view (Kothari 2004). Proponents of the positivistic school of thought commonly use quantitative methods, whereas qualitative methods tend to be chosen by researchers with an interpretivist attitude (Moon & Moon 2004). The current study adopted the quantitative approach. Nykiel (2007) posited that quantitative research methods seek to establish facts, make predictions and test hypotheses that have already been stated using a deductive approach and establishing objective knowledge.

Data collection
The Data were collected using a structured questionnaire. The questionnaire had an introductory section which summarised research objectives and addressed ethical issues. Section A contained items on demographic information. Section B had three other sub-sections, namely, B1 and B2 and B3. B1 had specific sections on selected individual factors (job satisfaction, organisational commitment and job stress).
Respondents were required to indicate their level of agreement with given statements on a Likert scale, ranging from 1-Very Dissatisfied to 5-Very Satisfied and the other scale ranged 1-Strongly Disagree to 5-Strongly Agree for items measuring organisational commitment and job stress. Section C comprised of items measuring organisational factors (human resource practices, QWE, organisational structure). A five-point Likert scale ranging from 1-Strongly Disagree to 5-Strongly was used. Lastly, Section C had intention to quit items measured using a three-point Likert scale ranging from 1-Always to 2-Mostly and 3-Never.

Data analysis
Data were captured using the Statistical Package for Social Sciences (SPSS) version 25. Descriptive statistics were used to analyse the demographic characteristics of the sample. Thereafter, data were subjected to SEM using AMOS software techniques, to ascertain model fit.

Ethical considerations
Ethical approval to conduct the study was obtained from the Faculty Research and Innovation Committee (FRIC), Central University of Technology, Free State, reference number: FMSEC05/16. One of the researchers visited each business establishment in the province and requested employees to fill consent forms before their voluntary participation. The respondents were also informed of their right to withdraw from the study at any stage should they have felt any harm or threat. In order to maintain confidentiality, data were stored in an aggregate form.

Exploratory factor analysis
An exploratory factor analysis (EFA) was performed to test the structure of the three main constructs, that is, organisational factors, individual factors and intention to quit. Such an analysis allowed for the empirical assessment of the validity of the scales used. The Kaiser-Meyer-Olkin (KMO) measure was calculated to ensure that the sample was adequate for factor analysis. Glen (2016) noted that the KMO test is a measure of how suited data are for factor analysis. The test measures sampling adequacy for each variable in the model and for the complete model (Glen 2016). The suitability of the data was supported because the KMO value (0.871) was superior to the threshold of 0.6, and Bartlett's Test of Sphericity was significant (p < 0.001) (Pallant 2010). The KMO and Bartlett's Test of Sphericity confirmed that the data were suitable for factor analysis.
After ascertaining the suitability of the data for factor analysis, a principal component analysis, using the VARIMAX method was used to extract the factors with an Eigenvalue above 1. However, the analysis showed poor results. The items were not loading well into the factors. After presenting Eigen values as indicated in the scree plot above, the models (initial and refined) were examined as indicated in the next section.

Model fit indices
Before examining the model fit indices of the final measurement model, a univariate normality test was conducted to confirm whether the model could be estimated using the maximum likelihood method (Li & Malik, 2018) The results are indicated below.

Normality test
A normality test is a statistical process used to determine if a sample or any group of data fits a standard normal distribution (Das & Imon 2016). A normality test can be performed mathematically or graphically. Generally, the values for skewness and kurtosis between -2 and +2 are considered acceptable to prove normal univariate distribution (Cain, Zhang & Yuan 2017). The results will not be affected by the non-normality distribution of data for all items used to measure the constructs, as most of their coefficients belong to the interval [-2 and +2]. Because the normality is supported, the maximum likelihood method was confidently used to assess the model fit of the initial model as indicated in Figure 2.
The model presented in Figure 2 is the measurement model before refinement. Its Chi-square (X 2 ) was equal to 568 798, its p value < 0.001 (significant) and its degrees of freedom (df) = 249. Although this initial model suggests a significant Chisquare, there was a need to further examine model fit indices before concluding on the model. The fact that the Chi-square is very sensitive to the sample size is the reason most Chisquare of large samples are often significant (Hair et al. 2014).
A close diagnosis of the modification indices and the standardised residual covariance matrix retrieved from the IBM AMOS outputs suggested that some items should be deleted to improve the model fit indices. Items whose factor loadings were below 0.5 were also deleted.
Two inter-item correlations pertaining to the same construct were added to improve the model fit using the IBM http://www.sajesbm.co.za Open Access AMOS function 'modification indices' (Hair et al. 2014). The addition of the inter-item correlations reduced the measurement error and improved the internal consistency of items (Ford, MacCallum & Tait 1986), which in turn enhanced the model fit. Following the aforementioned amendments, a final measurement model was designed (see Figure 3).
The final measurement model as indicated in Figure 3 above indicates a significant and lower Chi-square (χ 2 = 242 110; p = 000; df = 130), which implies that this later version of the model is more improved than the initial measurement model as shown in Figure 2. The model fit indices of the final model are provided in Online Appendix 1,

Reliability analysis, convergent and discriminant validity assessment
The final measurement model illustrated in Figure 4 was considered as the graphical evidence of convergent and discriminant validity. All the factor loadings were above 0.5, suggesting a convergent validity of all the items. The moderate level of correlations above 0.8 suggests a discriminant validity concern of all three latent variables. Further robust statistical evidence is provided in Online Appendix 1, Table 2-A1 to establish the validity of all the research instruments used in the study. The table indicates a good reliability for all the scales used in this study as Cronbach alphas and composite reliability coefficients were both above 0.7 (Vaske, Beaman & Sponarski 2017). Online Appendix 1, Table 2-A1 also shows that the factor loadings of all constructs were above the recommended threshold of 0.5 (Hair et al. 2014). Similarly, the average variances extracted (AVEs) of all constructs were also above the required cut-off of 0.5 (Ahmad, Zulkurnain & Khairushalimi 2016). In addition, the overall result indicates a good reliability of all the scales involved in this study (see Online Appendix 1) as Cronbach alphas and composite reliability coefficients were both above 0.7 (Bagozzi & Yi 1988).
All the estimates on Online Appendix 1, Appendix 1, Table 3-A1. Discriminant validity is assessed through a comparison between the square root of the AVE estimates and the highest inter-construct correlation of the specific construct (Malhotra, Nunan & Birks 2017). The square root of the AVE is expected to be above all the interconstruct correlation values.

Structural equation modelling
The structural model was tested using the maximum likelihood performed with AMOS 25. The structural model for the study is presented in Figure 4.

Hypotheses testing
The study tested hypotheses in relation to whether selected individual and organisational factors have an influence in determining intention to quit amongst B&B employees. Regression analysis was used to determine the impact of the independent variables of the study on the dependent variables. The results are presented in Online Appendix 1, Table 6-A1.

Regression weights
The model explains up to 92.2% of the intention to quit and 89.4% of the individual factors. The details of the impact of each independent variable are provided in Online Appendix 1, Table 4-A1. Regression weights as shown in Online Appendix 1, Table 4-A1. Showed that all of the employee individual factors did not have a significant effect on intention to quit (β = 0.495, p > 0.05). This means that improving the employee individual factors does not translate into intention to quit. The regression results also showed that organisational factors had a positive and significant effect on intention to quit (β = 0.814, p < 0.01).   This means that when any of the organisational factors goes up by 1 standard deviation (SD), there are 99% of chances that intention to quit also goes up by 0.814 of its own SD. The specific hypotheses results relating to specific factors are presented below.

Discussion of findings Job satisfaction versus intention to quit
Online Appendix 1, Table 4-A1 revealed that job satisfaction (one of the individual factors) did not have an influence on intention to quit. We, therefore, accept the null hypothesis and reject the alternative hypothesis (H1) which says, job satisfaction has a significant positive effect on intention to quit. Most studies that investigated the impact of job satisfaction on intention to quit have been conducted amongst large hospitality organisations, such as hotels.
For instance, past studies (Albattat, Som & Helalat 2013;Sangaran & Jeetesh 2015) found that intention to quit occurs when there is dissatisfaction in the job. These findings are contrary to those of the current study because given the high unemployment rate in South Africa (Wang 2019), employees could be left with but little choice. However, the current study's results corroborate those by Holston-Okae (2017), who found out that job satisfaction within the hospitality industry correlates inversely with employee intention to quit to a statistically significant degree.

Organisational commitment versus intention to quit
Results in Online Appendix 1, Table 4-A1 show that all aspects of another individual factor -organisational commitment (affective, continuance and normative) had no significant positive influence on intention to quit. We, therefore, accept the alternative hypothesis H2, which says that organisational commitment has a significant negative effect on intention to quit. This is contrary to a study conducted by Ghosh and Gurunathan (2015) that established organisational commitment as one of the major factors that influences intention to quit amongst employees in the hospitality industry. Although Johansson (2016) revealed that in most American small businesses, lack of organisational commitment is one of the factors that causes

Job stress versus intention to quit
Results in Online Appendix 1, Table 4-A1 revealed that job stress (one of the individual factors) had no significant association with B&B employees' intention to quit. This led us to accept the null hypothesis and reject the alternative hypothesis (H3) which says that job stress has a significant positive influence on intention to quit. These findings are contrary to a study by Hang-yue, Foley and Loi (2015), who found that job stress exerts a significant positive effect on intention to quit. The current study's results are also contrary to previous studies (Hwang et al. 2016;Namra & Tahira 2017), which revealed that hospitality work can be stressful leading to increased intention to quit amongst employees.

Human resource practices versus intention to quit
This study established that selected human resources practices had a significant positive influence on intention to quit amongst B&B employees. This led us to accept the alternative hypothesis (H4), which says that human resource practices have a significant positive effect on intention to quit. A study by Santhanam et al. (2015) showed that human resources practices such as compensation, performance management and recruitment can affect employee outcomes such as intention to quit. With others, satisfactory compensation was found to reduce intention to quit and led to employee retention (Pohler & Schmidt 2015). Afsar, Shahjehan and Shah (2018) observed that employees who perceive their organisation to be in poor financial condition and do not receive their desired compensation may anticipate future layoffs and may pre-emptively leave. These factors help buttress the eminence of human resource practices in small businesses.

Quality of work environment versus intention to quit
Regarding the QWE and employees' intention to quit, a statistically significant positive effect was found between the two. Therefore, the alternative hypothesis (H5) which states that the QWE has a significant positive effect on intention to quit was accepted. These findings supported the conclusions previously reported by Robinson et al. (2016) that a dynamic and interactive environment may lead to higher job satisfaction of employees and, ultimately, employee retention, amongst other organisational benefits. Similarly, Kang, Busser and Choi (2018) noted that unfavourable perceptions of work environments lead to negative workplace outcomes, including turnover, whilst meaningful working environments deter intention to quit in a variety of job contexts.

Organisational structure versus intention to quit
Organisational structure was found to have had an impact on intention to quit. This led to the acceptance of the hypothesis (H6) which states that organisational structure has a significant positive effect on intention to quit. According to Katsikea, Theodosiou and Morgan (2016), organisations that focus on a team-based approach rather than the typical hierarchical structure are in a better position to retain their employees. Afsar et al. (2018) called such organisations high performance organisations, that is, organisations that try to bring out the best in individuals and create an exceptional capability to deliver high-end results.
Cummings and Worley (2015) stated that high performance organisations include organisations with divisional and process structures. A study by Katsikea et al. (2016) found that hospitality organisations that follow the divisional and process structures have reduced rates of intention to quit than those with a functional structure. Work in the hospitality sector, especially hotels and B&Bs, follows the division structure approach, with say a frontline section, kitchen section and so on.

Association between organisational and individual factors
Online Appendix 1, Table 5-A1 also indicates that organisational factors had a positive and significant effect on employee individual factors (β = 0.946, p < 0.05). This means that when any of the organisational factors goes up by 1 SD, there are 99% chances that corresponding hypothesised individual factors also go up by 0.946 of their own standard deviation. The specific hypotheses results with corresponding factors are presented below.

HR practices versus job satisfaction
Selected human resources practices (compensation management, performance management and recruitment) were found to be associated with job satisfaction. This led to the acceptance of the alternative hypothesis (H7) which states that HR practices have a significant positive effect on job satisfaction. Haider et al. (2015) maintained that total compensation planning can help improve job satisfaction and increase employee retention. Hospitality organisations, such as B&Bs, might have to ensure that their compensation policies are satisfactory to employees.
In addition, one other HR practice, performance management, plays an important role in employee motivation and satisfaction (Ushus & Johney 2015). According to Ushus and Johney (2015), when employees receive high quality appraisal experience, they will tend to feel satisfied and motivated about their job and tasks given to them. Likewise, recruitment has been found to play a positive role in ensuring worker performance and positive job satisfaction outcomes (Agoi 2016). It is often claimed that recruitment of workers occurs not just to replace departing employees or add to a workforce but, rather, it aims to put in place workers who can perform at a high level and demonstrate commitment, thereby leading to high levels of job satisfaction (Ballantyne 2014).

Quality of work environment versus organisational commitment
Quality of work environment was found to have an influence on organisational commitment amongst B&B employees. The alternative hypothesis (H8) which says that the QWE has a significant positive effect on organisational commitment was accepted. The results are in tandem with Hanaysha (2016) who maintained that internal work events and elements of the work environment shape employees' commitment to an organisation. The present findings also support the empirical evidence that work environment has the most significant influence on organisational commitment amongst hospitality employees (Ebrahim 2014). This is corroborated by Lee, Back and Chan (2015) who reported that in the hospitality industry, employees' poor commitment to an organisation is closely linked to excess work, pressure of work and difficult customers which are all aspects of the work environment. It is, therefore, evident that the QWE can lead to organisational commitment amongst B&B employees.

Organisational structure versus job stress
The present study showed that organisational structure has a significant positive effect on job stress. The alternative hypothesis (H9) was proved true. These results support the conclusions by Kanten, Kanten and Gurlek (2015) that employees in high performance organisations (those that follow the divisional or decentralised structures) possess greater levels of organisational trust than employees in traditional hierarchical organisations, thereby leading to reduced stress levels.
Conversely, Cummings and Worley (2015) acknowledged that goals that traditional organisations (those following formalised and functional structures) tend to focus on are primarily how well the company is doing (business goals), whereas high performance organisations' goals tend to be more related to customer satisfaction, improving employee skills, as well as adapting to change within the workplace. The former structure is likely to result in increased stress, whilst the latter is likely to result in reduced stress and intention to quit amongst employees. Considering that B&B workspaces are more divisional and decentralised in structure, the current results are not surprising.

Conclusions and recommendations
This study investigated the impact of selected individual and organisational factors on intention to quit amongst employees in one of the districts in South Africa. Specifically, the study measured job satisfaction, organisational commitment and job stress (individual factors), as well as HR practices, QWE and organisational structure (organisational factors) and their impact on intention to quit. The results supported that organisational factors have a positive significant effect on intention to quit amongst B&B employees.
In terms of individual factors, the results showed that there were no significant relationships between the selected factors, namely, job satisfaction, organisational commitment and job stress and intention to quit amongst B&B employees. These results were contradictory to the research hypotheses and literature reviewed which found that the selected individual factors and intention to quit were significantly correlated.
It is, therefore, recommended that B&B owners/managers put more focus on improving organisational factors, so as to allow effective implementation of retention strategies. For example, continuously creating and maintaining a conducive work environment within B&B establishments might result not only in minimising intentions to quit but also improved business reputation and performance.
Also, effective HR practices, such as satisfactory compensation, could also play an important role in assisting managers retain their skilled employees as it is evident that satisfactory compensation packages and benefits influence employee motivation, loyalty and result in low levels of intentions to quit amongst employees (Khaleefa & Al-Abdalaat 2017). It is also recommended that owners or managers of B&Bs conduct needs-analysis within their organisations to identify areas for improvement. This could prevent high rates of intention to quit within these establishments.
Future studies might consider conducting similar research in other districts or provinces in South Africa with larger sample sizes for more robust findings. Further studies could also consider demographic factors, such as age, which are likely to influence intention to quit amongst employees in B&B establishments. It could also be interesting to contrast how intentions to quit differ amongst employees in B&B establishments against those in large companies.

Limitations of the study
The study was based on B&B establishments in a selected district in the Free State province of South Africa. This means that the results might have limited applicability to other similar institutions. The other limitation lies in the fact that there was a discriminant validity concern in the final measurement model for all constructs (organisational factors, intention to quit and individual factors). It is, however, encouraging to note that despite the discriminant validity concern, convergent validity was met, therefore, the results of the present study might be a basis upon which similar studies can be established.