Modelling employee retention in small and medium-sized enterprises and large enterprises in a dynamically changing business environment

Purpose – This study aims to explore the holistic context of organisational staff retention in small, medium and large organisations. It also aims to identify the factors affecting the retention of organisations of different sizes


Introduction
In the modern business environment, employee retention has become a key factor in competitiveness and achieving sustainable growth (Allen et al., 2010).Research has shown that companies that are able to flexibly adapt and change their structures and processes are the ones that can succeed in the market.A motivated, high-performance company is built on the foundation of organisational innovation (Esso †sy and Vink oczi, 2018).When key employees leave an organisation, both direct and indirect costs must be taken into account.The direct costs associated with workers' exit, such as recruitment, training and administrative costs, result in additional expenditures, while the indirect costs of reduced productivity have a negative impact on the growth of organisations (Gialuisi and Coetzer, 2013).
The relationship between staff retention and organisational size is a complex and multifaceted phenomenon.Researchers have analysed the factors influencing employee retention (Pat o, 2017), but less attention has been paid to the phenomena arising from organisational size differences.While it is generally accepted that different sizes of organisations can have different levels of employee turnover, identifying the specific factors behind this relationship requires a deeper understanding.Some studies have found that organisations face different challenges in employee retention based on their size (Ikram et al., 2021;Labatmedien_ e et al., 2007;Li et al., 2019;Sjöberg and Sverke, 2000;K oczy et al., 2022).Some researchers point to the limiting effect of a hierarchical organisational structure, which encourages workers to seek new opportunities elsewhere, as a possible explanation for the higher turnover rate in larger organisations (Li et al., 2019).Other studies have identified less frequent individual recognition and bureaucracy in larger organisations as factors limiting employee retention (Ikram et al., 2021;Buics and Eisingern e, 2020).
According to Cardon and Stevens (2004), socialisation processes in the workplace also differ between organisations of different sizes, with smaller organisations having stronger employee relations and a more cohesive workplace community.New employees are involved in meetings and community events sooner in these organisations, and subordinates are not isolated from senior managers, which allows for a more personal, less formal style of leadership and communication (Down, 2010; IJOA Labatmedien_ e et al., 2007;Storey and Greene, 2010).However, an important challenge for smaller organisations is that they have limited financial and human resources and therefore cannot compete with larger companies in terms of compensation and benefits, and their HR departments are smaller and have an operational rather than a strategic function (Cardon and Stevens, 2004).Moreover, employees in smaller organisations are less likely to have access to training and development opportunities (Cassell et al., 2002;Storey and Greene, 2010) and their career development opportunities are limited due to the flatter organisational structure (Arnold et al., 2002;Storey and Greene, 2010).Small and medium-sized employers are also characterised by the fact that employees are often not tied to specific job tasks or roles, but are required to rotate between roles and tasks according to organisational requirements (May, 1997).Smaller organisations are therefore characterised by multitasking rather than by specific job functions (Cardon and Stevens, 2004).This also happens in the innovation-driven, the start-up companies (K ezai and Konczos Szombathelyi, 2021).
Larger organisations use a wider range of career development opportunities to compensate for their hierarchical structure, which can contribute to more effective retention as employees can see the potential for growth and development within the organisation.Based on Smerek and Vetr akov a's (2020) study, multinational managers adopt a more versatile approach to the career development of their employees, allowing for more flexibility in job roles within the company and emphasising the promotion of a collective corporate culture, which is highly influential in retaining employees.Their resource surplus can also be observed in terms of human resources, as they tend to have more advanced human resource strategies that focus on efficiency and cost reduction.Unlike small organisations, they are more likely to have supportive HR organisations and defined retention strategies that can help address employee problems and improve retention rates (Cegarra-Leiva et al., 2012;Newman et al., 2011).This situation persisted even during the global coronavirus (Dajnoki et al., 2023).
A peculiarity of the international literature on this topic is that it examines the sub-areas of organisational employee retention, and, thus, does not provide an answer to the question of the optimal retention strategy.The literature suggests the importance of the moderating effect of organisational size, but the main shortcoming of the research is that the specificities of employee retention arising from size differences have not been explored.To our knowledge, no study to date has previously examined a holistic approach to employee retention in the context of the moderating effect of organisational size.This study aimed to understand the whole phenomenon of staff retention, using a holistic approach that incorporates organisational psychology and management approaches.Unlike previous studies, this paper does not aim to analyse the effect of changes in a single factor but to model real organisational circumstances and examine the combined effect of multiple factors.According to Rai et al. (2019) and Barrick and Zimmerman (2005), the literature on employee retention has mainly focused on the factors associated with employee turnover and much less is known about the factors that strengthen employees' intention to stay.There have been few studies at the organisational level of analysis to investigate whether and how employers can reduce turnover.This study aims to contribute to the international literature on the subject by taking a holistic approach to the factors that influence employee retention, such as intention to leave, organisational commitment and normative commitment.Rynes et al. (2002) stated that there is a significant knowledge gap in the outcomes of positive or negative job satisfaction.To fill this gap, this study was based on Spector's (1985) Job Satisfaction Survey.To our knowledge, no previous study has taken a holistic approach to understanding the factors influencing employee retention, including factors such as intention to leave, organisational commitment and normative commitment, especially with the participation of Hungarian employees.

Literature review
The relationship between well-being at work, employee retention and engagement is a complex and interdependent relationship.Research has shown that well-being at work is significantly related to employee retention and engagement.Brunetto et al. (2012) found that a sense of satisfaction and well-being positively influences employee organisational commitment.

Dimensions of well-being at work
Well-being at work is an increasingly important issue in today's organisations, referring to the physical, emotional and social aspects of workers' lives.Well-being actions contribute to the overall health and happiness of employees at work (Hennicks et al., 2022;Keller and Ercsey, 2023).Early research on human behaviour at work also examined how individuals' work-related behaviour and attitudes are influenced by the satisfaction of people's needs (Mobley, 1977;Spector, 1985).Employee well-being is one of the most important factors influencing long-term employee engagement and retention.Mutual respect among employees also plays a role in creating well-being at work.The absence of this has negative consequences for all parties (Juh asz et al., 2022).
Employees who are satisfied with their job are more likely to remain loyal to their employer and less likely to switch jobs (Madigan and Kim, 2021;Bombiak, 2023;Moln ar and Csehn e Papp, 2023).So organisational satisfaction is a measure of how much we like or dislike our job (Spector, 1997).Research has shown that there is a causal relationship between well-being and job satisfaction (Nguyen et al., 2020) and that the two concepts form a complementary cluster.Research shows that dealing intensively with employees' psychological well-being and attachment to the organisation can stimulate a more positive atmosphere (Jaškevi ciūt_ e et al., 2023).This helps to foster an innovative, caring and creative culture (Jalil et al., 2021;Nguyen et al., 2020) that improves workforce retention.According to Pint er et al. (2022), digital financial inclusion contributes significantly to sustainable employment in high-income economies, especially in upper-middle-income and less significant economies.In addition to the above factors, Pint er et al. (2022) showed that digital financial inclusion contributes to sustainable employment in high-income economies, especially in upper-middle-income economies.
This study explains the factors influencing employee well-being along the lines of Spector's (1985Spector's ( , 1997) ) satisfaction factors (Figure 1).Spector defined nine satisfaction dimensions: pay, promotion, fringe benefits, contingent rewards, supervision, co-workers, operating procedures, nature of work and communication.
Pay is one of the most controversial areas of well-being research (Alieva and Powell, 2022;Ferry et al., 2023;Massingham and Tam, 2015;Valaei and Rezaei, 2016), but many researchers agree that fair and competitive pay systems contribute to organisational commitment (Dargahi and Veysi, 2021;Valaei and Rezaei, 2016).Research on executive pay strategies has found that rewarding managers appropriately improves organisational performance in the short term and reduces the likelihood of adverse corporate performance in the long term (Ferry et al., 2023).This is particularly noticeable in crises, when it may be necessary to modify expectations related to the standard of living (Z eman et al., 2021).The next group of studies focused on the relationship between pay and employee skills.It was found that there is a direct and positive relationship between employee skills and wages, as the development of employee skills leads to wage improvements.Employees with higher skills are more likely to get important jobs, and in return, employers reward them with higher pay to make them feel valued (Massingham and Tam, 2015;Valaei and Rezaei, 2016): H1. Pay satisfaction (a) has a positive effect on organisational commitment and (b) has a negative effect on turnover intention.

IJOA
Fringe benefits play a key role in both attracting and retaining workers (Baran and Sypniewska, 2020).Salary is only one of the employee retention factors (T oth et al., 2023).Fair and personalised benefit systems increase employees' sense of fairness and commitment.Studies have shown that when employees perceive their employer's benefits as unfair, it negatively impacts engagement, performance and retention (Laundon et al., 2019).Therefore, managers should provide benefits to employees that encourage a higher level of engagement and quality of work (Baran and Sypniewska, 2020): H2. Satisfaction with benefits (a) positively affects organisational commitment and (b) negatively affects turnover intention.
Contingent rewards are central to the debate on incentive schemes (Laundon et al., 2019).As an external motivating factor, rewards are already important in childhood, and their role remains in adulthood, although it changes somewhat (T oth et al., 2022).Research has found a strong correlation between contingent rewards and employee engagement, with stronger effects than financial incentives (Fröber and Dreisbach, 2016).Employees who have a negative perception of performance-related rewards from their employer will not be motivated to perform at a high level (Jilani and Juma, 2015).An increasing number of companies in this Central Eastern Europe (CEE) have recognised that total cash compensation and benefits should be addressed as one (Po or et al., 2019): H3. Satisfaction with contingent rewards has (a) a positive effect on organisational commitment and (b) a negative effect on turnover intention.
Opportunities for promotion, professional development and career development within the organisation can increase engagement (Dargahi and Veysi, 2021).Research has confirmed the moderating effect of promotion in changes in engagement and increasing motivation (Masika and Juma, 2015).Promotion is, therefore, positively linked to all aspects of organisational Modelling employee retention commitment.Valaei and Rezaei (2016) point out that high-performing employees should have a fair chance of promotion.Equity theory also notes that employees think in terms of comparisons, so they compare their promotion opportunities with those offered by other employers.Therefore, in case of injustice, the intention to leave the organisation is amplified in the employees.Busari et al. (2017) conducted research among organisations using planned promotion programs, which supports the negative relationship between promotion and fluctuation.This requires the involvement of employees in the evaluation process for promotions, which should be based on objective criteria (Cahyadi et al., 2022): H4. Satisfaction with promotion (a) positively affects organisational commitment and (b) negatively affects turnover intention.
Effective supervision and a supportive leadership style can help to increase employee commitment, (Dargahi and Veysi, 2021) as people-centred leadership plays a very important role in employment (Baran and Sypniewska, 2020).When people-centred management is present in an organisation, a healthier organisational culture is created (Paltu and Brouwers, 2020), and employees are more willing to participate in tasks and can overcome difficulties more easily (Boontantrapiwat and Kitcharoen, 2022).Consequently, the lack of peoplecentred management is not only associated with low levels of commitment, but also with lower levels of direct involvement in work in such an organisation (Baran and Sypniewska, 2020) and change processes can also face ongoing obstacles (Lee and Shin, 2023).Research has shown that the performance and behaviour of managers is a key determinants of the quality of the workplace climate, which predicts the need for continuous monitoring and evaluation.If employees experience toxic leadership, it can deteriorate the culture of the organisation, leading to increased turnover (Paltu and Brouwers, 2020): H5. Satisfaction with supervision (a) positively affects organisational commitment and (b) negatively affects turnover intention.
Not only supervisors, but co-workers as well, contribute significantly to the development of loyalty and attachment to the organisation (Kmieciak, 2022).Empirical studies show a negative relationship between support from co-workers and voluntary turnover intention (Karatepe, 2012;Limpanitgul et al., 2014).Employees who experience help and support from their co-workers will feel more connected to the workplace community, which will encourage them to maintain organisational commitment.Organisational commitment acts as a mediator between support from co-workers and turnover intentions.When employees have the appropriate support from their colleagues, they develop a greater sense of belonging, loyalty, commitment and emotional attachment to the organisation, which reduces turnover (Alieva and Powell, 2022).It can be assumed that the influence of co-workers on the attachment to the organisation may be stronger than that of managers.This effect is particularly noticeable in large organisations, where the relationship between supervisors and subordinates is weaker than in small organisations (Limpanitgul et al., 2014): H6. Satisfaction with co-workers (a) positively affects organisational commitment and (b) negatively affects turnover intention.
Operating procedures are the documented processes that an organisation uses to define how different tasks and processes are carried out.The predictability of operational processes enhances employees' commitment to the organisation.Research shows that employees' responses can vary depending on their previous experience of change processes (Kar acsony et al., 2023).

IJOA
Well-designed and effective operational procedures, processes and systems enable employees to work more efficiently and effectively, while reducing stress and prolonged working hours (Hanaysha, 2016).Results from Valaei and Rezaei (2016) indicate a positive relationship between operational procedures and organisational commitment.However, lower satisfaction levels with operational procedures prompt a review of the effectiveness of current procedures and regulations (K ezai and Konczos Szombathelyi, 2021;Valaei and Rezaei, 2016).When redesigning organisational processes, managers need to consider the impact of technological innovation and digitalisation on employees.Indeed, while such changes can undoubtedly lead to significant efficiency savings for the organisations, they can also have serious implications for employee well-being (Moore et al., 2022): H7. Satisfaction with operating procedures has (a) a positive effect on organisational commitment and (b) a negative effect on turnover intention.
The nature of work, and motivational factorthe nature and content of the work, the sense of achievement, recognition and responsibilityhelps employees to find their value-making role in the organisation (Raziq and Maulabakhsh, 2015).According to the study by Lee and Shin (2023), organisational and managerial support has an impact on employees' sense of responsibility, regardless of their job tasks.With the increase in working from home, organisations need to pay particular attention to the content of work, which goes hand in hand with the protection of workers' mental health.Empirical research shows that isolation and the social environment in the workplace can affect workers' well-being and productivity (Catan a et al., 2022).However, to enhance the employee experience, employers must do everything they can to ensure that their employees have a fun and enjoyable time at work (Valaei and Rezaei, 2016).A solution could be the development of gamified HR systems, whose motivational aspects also contribute to productive work.Gamified systems lead to a more pleasant and interesting work environment, further increasing satisfaction and engagement (Silic et al., 2020).However, employee autonomy and responsibility hold significant value.Giving employees the freedom to carry out tasks that they find useful fosters loyalty.The organisation should empower employees by giving them decisionmaking autonomy, thereby fostering a sense of belonging and encouraging retention (Ghosh et al., 2013): H8. Satisfaction with the nature of work (a) has a positive effect on organisational commitment and (b) has a negative effect on turnover intention.
Internal communication is effective when employees know what is happening in the organisation and are aware of its goals and processes.To achieve organisational goals, clear and unambiguous communication is needed, with a full explanation of the related work assignments to employees (Valaei and Rezaei, 2016).Regular communication and networking allow employees to keep each other informed about work processes and their daily activities.Internal communication development includes the implementation of employee relations activities and annual team-building exercises (Boontantrapiwat and Kitcharoen, 2022).Employee engagement can be enhanced by involving immediate superiors and subordinates in the dialogue as a team (Seymour and Geldenhuys, 2018), for which modern management methods such as agile management provide an excellent opportunity (Helozía and Luiz, 2020): H9. Satisfaction with communication has (a) a positive effect on organisational commitment and (b) a negative effect on turnover intention.

Organisational commitment
Workplace well-being factors have a significant impact on employee engagement and turnover intention, and ultimately on organisational staff retention.As with other aspects of organisational behaviour, the study of employee retention covers a broad spectrum.
Engagement is most often defined as an employee's commitment, loyalty to the organisation and willingness to make an effort on behalf of it (Al Dalahmeh et al., 2020).
Organisational commitment is a concept that expresses the commitment of employees to the organisation, which leads them to prioritise the interests of the organisation over their individual interests (Kmieciak, 2022).Research on the subject has evolved over time, but the Allen and Meyer three-component model (affective, continuance and normative commitment) has become accepted in both academia and business (Allen and Meyer, 1990;Meyer and Allen, 1991).According to the model's approach, affective commitment is based on the individual's personal desire to be part of the organisation, which is mainly based on the emotional attachment to the organisation.In contrast, continuous commitment is primarily based on the common interests between the individual and the organisation.In this case, the individual, weighing up the advantages and disadvantages of being part of the organisation, decides to stay because the disadvantages of leaving are greater than the disadvantages of remaining part of the organisation.
Normative commitment is based on the belief that employees are accountable to the organisation; that is, employees must be committed to the organisation (Limpanitgul et al., 2014), which is related to the perception of norms.While some research has confirmed that the three factors together determine the employee's relationship with the organisation (Allen and Meyer, 1990;Meyer and Allen, 1991;Meyer et al., 2002), there have been several research findings that refute some of the relationships in the Allen-Meyer three-component model (Dargahi and Veysi, 2021;Helozía and Luiz, 2020;Labatmedien_ e et al., 2007).In an alternative approach, there has been research investigating the degree to which organisational measures impact the three dimensions of engagement.The findings indicate a positive correlation between organisational measures and normative commitment (Helozía and Luiz, 2020).Ghosh et al. (2013) concluded that the underlying focus behind any retention strategy of an organisation must, thus, be to enhance the normative commitment of its workforce.To mitigate the practical issues associated with the model, Kim et al. (2016) introduced the notion of organisational commitment by using a measurement scale that evaluates commitment to the organisation in a multifaceted manner.Thus, in constructing the model for this study, we integrated the approaches of Kim et al. (2016), and Allen and Meyer (1990)  negatively affects turnover intention.

Turnover intention.
Turnover intention can be defined as the intention or desire of an individual to voluntarily leave his or her current job (Nguyen Ngoc et al., 2022;Paltu and Brouwers, 2020).Research has shown that employee turnover intentions can be influenced by both individual and organisational factors.These include work motivation, work engagement, quality of life at work and organisational commitment (Massingham and Tam, 2015).An organisational approach suggests that factors such as leadership practices, organisational climate, support from colleagues and superiors and development opportunities have a significant impact on turnover intention (Boontantrapiwat and Kitcharoen, 2022).In this context, Kundu and Lata (2017) found that a supportive work environment, including perceived climate, supervisory relationships, peer group interactions and perceived organisational support, has a positive impact on employee retention.

IJOA
Turnover intentions are a complex phenomenon, which is underpinned by a lack of, or inadequate levels of, several workplace well-being factors.Hence (Cegarra-Leiva et al., 2012), turnover is one of the variables most closely related to job dissatisfaction and, thus, plays a crucial role in the construction of the research model (Figure 1).Many studies and theories have been developed to understand employee turnover intention, its causes and consequences.Early literature focused mainly on attitudes towards work, and narrowly defined satisfaction (Mobley, 1977), and later research focused on commitment and engagement related to work (Brown, 1996).A large literature on turnover intention (Sjöberg and Sverke, 2000) found that job involvement and organisational commitment were negatively related to turnover intention.Therefore, the related hypothesis is as follows: H11. Organisational commitment negatively affects turnover intention.
If workers have a positive perception of the measures in place in their organisation, they are less likely to leave their jobs (Li et al., 2019).Therefore, turnover intention acts as an intermediate link, which supports the view that the moderating effect should be examined as a combined effect of several organisational factors.This literature has validated the moderating effect of turnover intention but has the limitation of only examining the effect of job involvement and commitment with respect to turnover.In comparison (Newman et al., 2011), it took a more complex approach to understanding turnover intentions, analysing the combined impact of several HR initiatives.The holistic approach of the present study is confirmed by the results of Sjöberg and Sverke ( 2000) and Newman et al. (2011), which suggest that organisational commitment and employee and managerial support, among other factors, have a moderating role on turnover intention.The conceptual framework of the study is summarised in Figure 1.

Sample and data collection
In this study, we further developed a previously published model (Gelencs er et al., 2023) by creating two new models adapted to small and medium-sized enterprises (SMEs) and large organisations.The empirical research is based on an online questionnaire survey in which a total of 549 Hungarian employees participated between April and June 2023.During the sampling, the main emphasis was on the participation of employees from heterogeneous organisational backgrounds, so no sectoral narrowing was applied, for example.The questionnaire was designed to be interpreted by people working in organisations of different sizes, in different sectors, industries, jobs and positions.Participation was subject to two conditions: respondents had to have at least one year of work experience and active employment at the time of completion.To this end, the survey started with the following screening questions: Q1. Do you have at least one year of work experience?Are you currently in active employment?
The questionnaire was based on statements that had previously been used effectively in international research and published in scientific journals.The translation of the statements in international research was carried out with the help of a professional translator.After the translation, the claims were tested with a small sample (n ¼ 9) based on the recommendation of Reynolds et al. (1993).Based on the results of the pre-test, there was no reason to modify the questionnaire.Based on the screening questions and due to the missing data, the results

Modelling employee retention
of 38 questionnaires had to be deleted, so a total of 511 complete responses were received (the response rate was 93%).
Based on the demographic data of the respondents, it can be concluded that 64% of the participants are female and 36% are male.Most respondents have tertiary education: 41.9% of the participants have a college degree and 14.5% have a university degree.Among respondents with secondary education, the majority have a high school diploma (28.4% of the total sample).The proportion of respondents with a primary, vocational or technical school education and an academic degree is negligible.About 42.9% of the participants belong to Generation Y (born between 1980 and 1994), 29.7% were born in 1995 or later (Generation Z) and 25% belong to Generation X (born between 1965 and 1979).Boomer generation respondents born between 1946 and 1964 make up only 2.4% of the total sample.
Regarding the respondents' current jobs, 68.1% of the respondents work in the private sector, 24.1% in the public sector and 7.8% in the non-profit sector.Most respondents (54.6%) work in services, 41.7% in industry and only 3.7% in agriculture.The size of the organisation was defined according to the European Commission definition as follows: The category of micro, SMEs includes enterprises with fewer than 250 employees, and organisations with more than 250 employees are classified as large enterprises (European Commission, 2003).Based on this categorisation, 53.8% of the respondents work for large organisations and 46.2% work for micro, small and medium-sized organisations.In terms of the type of job, there is a predominance of white-collar jobs (72.4%) compared to blue-collar jobs (27.6%).Regarding the current management position of the respondents, the majority (74%) are in a subordinate position, 12.1% in middle management, 10.4% in team-leading positions and only 3.5% in senior management.

Measures
In addition to questions about demographic data, respondents completed a structured questionnaire containing 58 statements.They were asked to rate their responses to the statements on a five-point Likert scale (1strongly disagree, 5strongly agree).All statements in the questionnaire were taken from previously used surveys (Allen and Meyer, 1990;Kim et al., 2016;Newman et al., 2011;Sjöberg and Sverke, 2000;Spector, 1985).The survey, which includnuted 58 statements, consisted of two main chapters (employee wellbeing, and employee retention) and additional sub-chapters.The dimensions and factors examined during the data collection are shown in Table 1.
The first part of the questionnaire was based on Spector's (1985) Job Satisfaction Survey, which examined the subject through nine factors with a total of 36 statements (4 positive or negative statements per factor).

IJOA
The second part of the questionnaire aims to examine the issue of employee retention.The related sub-chapters were normative commitment, organisational commitment and intention to quit.The examination of the issue of normative commitment was based on the survey of Allen and Meyer (1990), which included a total of eight (also positive and negative) statements.Organisational commitment was measured using seven statements based on Kim et al. (2016), and the seven statements used to measure intention to leave were based on Newman et al. (2011), Sjöberg andSverke (2000) and Wayne et al. (1997).

Partial least squares structural equation modelling
Partial least squares structural equation modelling (PLS-SEM) is a path analysis that has been increasing in popularity since the 2000s (Memon et al., 2021).The method is most prevalent in the social sciences (e.g.HR and marketing research), as it involves building a model that illustrates the relationships between latent and observed variables.In most cases, research using the PLS-SEM method uses the SmartPLS program.
According to Hair et al. (2011), PLS-SEM distinguishes between inner and outer models.In the inner model, there are no causal relations, but it separates endogenous and erogenous constructs.In the erogenous model, there are no structural relationships between the latent variables, whereas in the endogenous model, the structural relationships between the latent variables are explained by structural relationships between the other constructs.In the outer model, there are one-way relationships between the latent variable and the assigned indicators.Multiple connections are not allowed, so all indicators are assigned to a latent variable.The strength of the relationship between the indicator and the latent variable is shown by the outer loading.
According to Dash and Paul (2021), the application of the PLS-SEM methodology consists of the following steps: Individual constructs: definition of latent variables.
Preparing for CFA (confirmatory factor analysis): confirmatory factor analysis is a practical test of theoretical models suitable for detecting relationships between indicators and latent variables.
Running CFA: includes verification of reliability, validity, convergent and discriminant validity.
Structural modelling: running SEM, as a result of which we get the values of the regression coefficients as well as the graphical model.Drawing conclusions and checking hypotheses based on the model.

Measurement model
Examining the measurement model includes internal consistency reliability, convergent and discriminant validity.Internal consistency reliability assesses the extent to which the items measure a specific latent construct.Following the recommendation of Hair et al. (2017), we assessed internal consistency reliability by ensuring that both Cronbach's a and the composite reliability (CR) are higher than 0.70 and below 0.95.The operating procedures construct was deleted due to the low Cronbach's a value.The results presented in Table 1 demonstrate that the remaining constructs exhibited Cronbach's a and CR values surpassing the threshold of 0.7, indicating a strong internal consistency of the measures.

Modelling employee retention
Convergent validity is the second measure used to evaluate the measurement model, which examines the degree to which a measure positively correlates with alternative measures of the same constructs (Hair et al., 2017).Convergent validity was assessed through the generation of 5,000 bootstrapping samples in PLS.The assessment of convergent validity requires checking the outer loading values of the items and the average variance extracted (AVE).As recommended by Hair et al. (2017), indicators with weaker outer loadings can be retained if other indicators with high loadings explain at least 50% of the variance.In total, six items were deleted in the case of SMEs, two items from the co-workers construct (Cow2 and Cow4), three items from the normative commitment construct (Nc1, Nc7 and Nc8) and one from the turnover intention construct (Ti7).As for large companies, a total of five items were deleted, one from the benefits construct (Ben4), one from the co-workers construct (Cow2), two from the normative commitment construct (Nc1, Nc7) and one from the turnover intention construct (Ti7).The deleted items are shown in the Table 12.
As shown in Table 2, the AVEs of the latent variables were between 0.505 and 0.945 in the case of SMEs, and between 0.517 and 0.750 for large companies.All AVEs are greater than the 0.5 standard value, indicating that the reflective measurement variables had favourable convergent validity.
Discriminant validity pertains to the extent to which the constructs used in the model are distinct from one another (Hair et al., 2017).Two methods were applied to evaluate discriminant validity.The first method used was the Fornell and Larcker (1981) criterion, which compared the correlation between constructs to the square root of AVE for each construct.To establish discriminant validity, the square root of the AVE for each latent variable must exceed the correlation value of the same construct (Fornell and Larcker, 1981).As evident from Tables 3 and 4, the square root values of AVE for each specific latent variable are higher than the correlation values provided in the corresponding rows and columns (Fornell and Larcker, 1981), thereby confirming satisfactory discriminant validity.
The second approach used to confirm discriminant validity was the Heterotrait-Monotrait (HTMT) ratio.While much research relies solely on the Fornell-Larcker criterion and crossloadings to assess discriminant validity (e.g.Hair et al., 2012), Henseler et al. (2015) have demonstrated that these criteria perform inadequately in revealing discriminant validity issues, therefore instead, researchers should use the HTMT criterion.Elevated HTMT values indicate a concern with discriminant validity.Based on simulations and prior research, Henseler et al. (2015) recommend that HTMT values should not exceed 0.90 when the path model incorporates conceptually similar constructs.For more distinct constructs, a more conservative threshold value of 0.85 is suggested.As observed from Tables 5 and 6, the HTMT values confirm that the present study fulfils all the assumptions of discriminant validity.
4.1.1Multicollinearity.Before evaluating the structural model, in addition to validity and reliability, it is essential to examine multicollinearity.Multicollinearity can be assessed using the variance inflation factor (VIF).Burns and Burns (2008) suggest that a VIF value greater than 10.0 indicates the presence of multicollinearity.Hair et al. (2014) recommend a cut-off value of 5.0 for multicollinearity, while other researchers propose even lower values such as 3.33 (Diamantopoulos and Siguaw, 2006).The VIF values for all items are presented in Tables 7 and 8.In our case, all of the VIFs for the indicators were below 3.33, indicating no evidence of multicollinearity between the latent constructs.

Structural model
This study assessed the structural model using the method of 5,000 bootstraps in Smart-PLS software.Tables 9 and 10 represent the results of the bootstrapping procedure, which includes the magnitude of the effects between latent variables.Bootstrapping is any test or IJOA

Modelling employee retention
metric that uses random sampling with replacement (e.g.mimics the sampling process) and falls into the broader class of resampling methods (Efron, 2003).In the tables, identical relationships between latent variables are highlighted.The tables show the values of the correlation coefficients between the latent variables for both the original (n ¼ 511)

IJOA
non-random sample and the bootstrapped sample (n ¼ 5,000).The results show that the difference between the correlation coefficients for the two samples is measured in thousands of units, confirming that the original sample is sufficiently reliable.Correlation coefficients with a negative sign indicate negative relationships between latent variables.As for large companies, the SRMR value for the saturated model was 0.065, and 0.069 for the estimated model.

Modelling employee retention
In the next step, we evaluated the R 2 values, according to Chin (1998), the values of R 2 must be > 0.1 or zero to be acceptable.Based on the results, the R 2 values are higher than zero.According to Chin (1998), the following correlations can be distinguished based on the R 2 values of the endogenous latent variables: R 2 < 0.19very weak; 0.19 <¼ R 2 < 0.33weak; 0.33 <¼ R 2 < 0.67moderate; R 2 >¼ 0.67substantial.
Based on this, the constructs in the case of SMEs can be categorised as follows: very weak correlation (promotion and supervision); weak correlation (benefits, nature of work and coworkers); moderate correlation (communication, contingent rewards, pay and turnover intention); and substantial correlation (organisational commitment).

IJOA
For large companies, the categories are the following: very weak correlation (promotion); weak correlation (benefits, supervision and communication); and moderate correlation (nature of work, co-workers, turnover intention, pay, contingent rewards and organisational commitment).
Figure 2 represents the results of the path analysis for small and medium-sized organisations, while Figure 3 illustrates the results for large organisations.

Discussion
To the best of the authors' knowledge, this paper is one of the few studies that seeks to address one of the main HR challenges of our time, the issue of organisational workforce maintenance.The study aimed to identify the factors affecting the retention of employees in SMEs and large organisations, and the relationships between them.The research contributes

Modelling employee retention
to the development of theoretical and practical knowledge on the subject.The scientific and practical value of the research lies in the fact that it examined the combined effect of several factors using a holistic approach.To explore the differences in the size of the organisation, two models were constructed to illustrate the impact of eight satisfaction factors on employee turnover through three dimensions (normative commitment, organisational commitment and turnover intention).The holistic approach of the study fills a significant research gap, as the research carried out so far on the subject has only studied a few factors.The main practical contribution of the research is that it has identified the critical factorswith the involvement of actively employed workersthat can have a decisive impact on the effectiveness of organisations' retention measures.The study calls for a rethinking and reform of the retention measures applied by some employers, which could be beneficial for organisations in terms of cost-effectiveness and competitiveness, and for employees in terms of progress towards wellbeing at work.The theoretical models were created based on Allen and Meyer (1990), and Kim et al. (1997), and tested using the PLS-SEM method.It was found that the discriminant and convergent validity of both models is adequate, so the theoretical model fits properly with the empirical data, and there is no multicollinearity between the variables.The model adapted for small and medium-sized organisations can identify a total of 30 significant relationships, while the model for large organisations can identify 32 significant relationships.
In both cases, organisational commitment has the highest explained variance of the latent variables tested, which is 0.712 for small and medium-sized organisations (R 2 ¼ 0.712) and 0.669 for large organisations (R 2 ¼ 0.669), meaning that the models are able to explain about 70% of the factors that influence employees' organisational commitment.In both models, employees' commitment to the organisation is influenced by normative commitment and satisfaction with the nature of the job, benefits and co-workers.However, a difference between the two models is that, unlike in small and medium-sized organisations, in large organisations supervision does not have a significant effect on employee engagement, while communication does.This result may be explained by the fact that in smaller organisations, the relationship between managers and employees may be closer, which may influence the employees' attachment to the organisation.In large organisations, however, the effectiveness of internal communication to convey organisational goals adequately plays a more important role in employee retention due to size.This result supports the finding of Limpanitgul et al. (2014) that the relationship between superiors and subordinates is weaker in large organisations.Table 12.

IJOA
Based on the results for smaller organisations, it can be concluded that co-workers contribute more to organisational commitment than managers.Therefore, based on the models, the variables nature of work, normative commitment, benefits and co-workers have a critical function in terms of organisational commitment, regardless of organisational size.If employees perceive these factors negatively, it will also entail a decrease in commitment.
There is also a significant variance in turnover intention, which is (R 2 ¼ 0.607) for SMEs, and (R 2 ¼ 0.582) for large companies, that is, the models can explain approximately 60% of the factors that influence workers' turnover intention.Of the latent variables, all variables that are negatively associated with turnover intention reduce the likelihood of leaving the organisation.Based on the results, the most significant negative association can be observed between organisational commitment and turnover intention for both models.This result confirms the finding of Sjöberg and Sverke (2000) that organisational commitment reduces

Modelling employee retention
the likelihood of quitting.The results show that in small and medium-sized organisations, employee organisational commitment has a stronger negative correlation (r ¼ À0.499) with turnover intention than in large organisations (r ¼ À0.388), i.e. a decrease in organisational commitment contributes more to turnover intention in smaller organisations than in larger ones.In small and medium-sized organisations, employees' intention to quit is reduced by normative commitment and satisfaction with contingent rewards and internal communication.Satisfaction with supervision, normative commitment, nature of work and pay have a significant negative impact on the intention to quit of employees working in large organisations.Based on the results, it can be concluded that organisational commitment and normative commitment, regardless of organisational size, have a significant, negative impact on employees' intention to quit.This can be explained by the fact that a normatively committed employee feels a moral obligation to maintain his/her organisational affiliation IJOA (McDonald and Makin, 2000), and therefore a decrease in this leads to a loosening of the employee's moral commitment to the organisation, which increases turnover intention.In addition, in small and medium-sized organisations, the positive perception of contingent rewards and communication reduces the intention of employees to quit, and in large organisations, the increase in satisfaction with supervision, nature of work and pay has a negative impact on the intention of employees to quit.Therefore, in smaller organisations, improving reward and communication systems can be effective in retaining employees who intend to leave, while in large organisations, strengthening the relationship between managers and employees, providing more meaningful and challenging tasks and improving pay can be effective.Among the latent variables of the model for small and medium-sized organisations, the explained variance of contingent rewards (R 2 ¼ 0.501) and the explained variance of pay (R 2 ¼ 0.592) show a medium strength (r ¼ 0.343) correlation.The explained variance of contingent rewards is significantly higher for large organisations (R 2 ¼ 0.652) and smaller, but also positive, for pay (R 2 ¼ 0.607).The coefficient of correlation between the latent variables is also medium but higher (r ¼ 0.433).The results suggest that the explained variance of the contingent rewards and pay variables is significant for both models and that employee satisfaction with contingent rewards has a stronger impact on satisfaction with pay in large organisations than in small and medium-sized organisations.This correlation supports the finding of Cardon and Stevens (2004) that large organisations operate more diverse incentive and compensation schemes, which, in addition to base pay, make a significant contribution to the financial well-being of employees.This result is supported by the fact that benefits also contribute more to satisfaction with pay in large organisations (r ¼ 0.338) than in smaller organisations (r ¼ 0.297).
A total of 22 equal effects can be identified between the latent variables of the models (highlighted in Tables 8 and 9), which when compared show that the strongest effect is observed between organisational commitment and turnover intention for small and medium-sized organisations (r ¼ À0.499), while for large organisations it is observed between contingent rewards and pay (r ¼ 0.433).In small and medium-sized organisations, the weakest effect is between normative commitment and turnover intention (r ¼ À0.137), while in large organisations it is between co-workers and organisational commitment (r ¼ 0.134).For small and medium-sized organisations, the second strongest relationship is between the nature of work and organisational commitment (r ¼ 0.479), while for large organisations the strength of this effect (r ¼ 0.336) is only eighth in the ranking of significant effects.Presumably this difference, based on Labatmedien_ e et al. (2007), is due to the fact that employees in small and medium-sized organisations are able to develop more personal relationships in the course of carrying out work processes, so a negative change in these relationships has a stronger negative impact on engagement than in large organisations.This finding is also supported by the fact that in small and medium-sized organisations, employee relations contribute more to organisational commitment (r ¼ 0.195) than in large organisations (r ¼ 0.134).
In small and medium-sized organisations, the effect of promotion on satisfaction with benefits is (r ¼ 0.448), while in large organisations this value is lower (r ¼ 0.357).A similar effect can be identified in relation to pay: compared to large organisations (r ¼ 0.135), in smaller organisations, employees' satisfaction with promotion opportunities contributes more to satisfaction with pay (r ¼ 0.220).From this, we can conclude that in the case of large employers, promotion does not affect benefit and salary satisfaction as much as in smaller organisations.This is also likely to be due to the more complex and flexible benefits offered by large companies, which are available in some jobs regardless of the job level.In this Modelling employee retention context, it is important to note that the impact of satisfaction with supervision on contingent rewards is higher for small and medium-sized organisations (r ¼ 0.202) than for large organisations (r ¼ 0.138).This may stem from the closer manager-subordinate relationships in smaller organisations, which can lead to more often use personalised motivational tools by immediate superiors.
This may stem from the fact that immediate superiors more often use personalised motivational tools in smaller organisations, which is more prominent due to the closer manager-subordinate relationship in SMEs.
Satisfaction with promotion opportunities had a positive effect on satisfaction with supervision in both modelsfor small and medium-sized organisations (r ¼ 0.335) and for large organisations (r ¼ 0.390)which suggests that promotion is a motivational tool that should be used regardless of the size of the organisation, as it has a positive effect on manager-employee relations.Finally, the level of pay is not related to organisational commitment in any of the models and only has a small negative effect (r ¼ À0.102) on the intention to leave large organisations.

Implications of the study
The study has several practical implications for HR professionals and managers.Based on the findings, it can be concluded that managers should view employee retention as a longterm investment in the future growth and competitiveness of the organisation and strive to implement it at a strategic level.New trends require a change in mindset directly from HR professionals, emphasising the need for a personalised management approach that focuses on individual needs rather than the traditional HRM approach (Agarwala, 2003).It is, therefore, recommended that HR leaders align their practices to enhance organisational commitment and competitiveness with the needs of individual employee groups (Nutov and Hazzan, 2014).
The findings of the study support that the variables of the nature of work, normative commitment, benefits and co-workers have a significant effect on employees' commitment to the organisation.Organisational commitment and normative commitment also influence turnover intention.Based on the findings, the most effective way to increase commitment to the organisation is to increase satisfaction with the nature of work.Therefore, HR professionals and managers are advised to pay special attention to providing employees with meaningful, interesting and challenging tasks, as this can increase not only their commitment but also their performance.To improve performance, managers need to give their subordinates a greater degree of autonomy.This can facilitate more objective performance appraisals and make employees more satisfied, innovative and independent in their task performance (Samuel and Chipunza, 2009).In addition to the nature of work, employees' normative commitment contributes significantly to their commitment to the organisation.Employers should therefore focus more on measuring normative commitment when selecting potential employees, as it can have a significant impact on employees' organisational commitment and their intention to leave in the future.To retain employees in the long term, it is essential to pay special attention to regularly reviewing the competitiveness of benefit packages within the compensation system and positioning them against competitor offerings (Sishuwa and Phiri, 2020).According to the findings, satisfaction with co-worker relationships contributes to employee commitment to the organisation in both small medium and large organisations.Therefore, managers should strive to improve employee relations.Possible methods to achieve this could include organising team-building programmes or, in the event of problems, using conflict resolution, personality awareness or communication training.

IJOA
Since the effectiveness of organisational retention policies is fundamentally determined by the characteristics of the organisation and the needs of the workforce, it is imperative to conduct regular employee surveys to assess satisfaction and organisational commitment to identify areas for improvement (Sishuwa and Phiri, 2020).For this purpose, organisations need HR professionals who know how to conduct effective employee attitude surveys, identify correlations from the survey and formulate relevant recommendations (Saari and Judge, 2004).Given that comprehensive surveys covering the entire organisation require a significant investment of time and resources, management is advised to consider employing research-oriented HR professionals.

Conclusions
Using a holistic approach, this study aimed to explore the specificities of organisational staff retention arising from differences in organisational size.By adapting approaches from organisational psychology and management, the research, in contrast to previous studies, did not seek to analyse the effect of a single factor change, but rather to model real organisational circumstances and investigate the combined effect of several factors.As a result of this research, two models have been developed that make a significant contribution to understanding the interrelationships that affect the staff retention of SMEs and large organisations.
Based on statements in the existing literature, a total of 11 hypotheses were formulated and empirically tested with a sample of 511 employees.The models generated using the PLS-SEM method showed a total of 62 significant correlations between the latent variables studied, with 22 identical correlations present in both models.The results of the path analysis have allowed the hypotheses formulated based on previous research to be tested, which, in addition to extending the theoretical framework of the subject, also allows important practical conclusions to be drawn.
H1, regarding pay satisfaction H1a positively affecting organisational commitment and H1b negatively affecting turnover intention, is only partially supported by the results.Satisfaction with pay did not show a significant relationship with organisational commitment in any of the models, and only in large organisations did it have a small negative effect (r ¼ À0.102) on intention to leave.This leads to the conclusion that the base pay does not motivate workers to commit to the organisation in the longer term and has only a small influence on their decision to leave.
According to H2, satisfaction with benefits H2a positively affects organisational commitment and H2b negatively affects turnover intention.Among the sub-hypotheses, H2a is supported by the results, while the negative effect of satisfaction with benefits on turnover intention is not statistically confirmed.The results suggest that benefits have a significant impact on employees' organisational commitment regardless of the size of the organisation, so it is worth paying special attention to the competitiveness of benefits components when designing a remuneration strategy for retention.
H3 is also only partially acceptable based on the results.Satisfaction with contingent rewards did not show a significant relationship with organisational commitment, only in the case of small and medium-sized organisations did it showcase a statistically significant negative effect (r ¼ À0.177) on turnover intention.Thus, the results suggest that neither pay nor contingent rewards contribute to employees' organisational commitment, but there are differences in the factors that influence the intention to quit based on organisational size differences.While in smaller organisations, a pay raise can reduce the incentive for employees to leave, employees in large organisations are more motivated to stay by the enhancement of contingent rewards.

Modelling employee retention
H4, which states that satisfaction with promotion H4a has a positive effect on organisational commitment and H4b has a negative effect on turnover intention, is rejected due to lack of a significant relationship.Therefore, based on our results, promotion does not have a significant effect on labour retention, which contradicts the findings of Busari et al. (2017).
H5, which states that satisfaction with supervision H5a has a positive effect on organisational commitment and H5b has a negative effect on turnover intention, is also only partially supported by the results, again highlighting differences by organisational size.Supervision has a significant positive effect (r ¼ 0.085) on the commitment of employees in small and medium-sized organisations, but this is not true for employees in large organisations.The difference by organisation size also indicates that the quality of the relationship with the manager does not contribute to reducing voluntary turnover in smaller organisations, but has a significant negative effect (r ¼ À0.219) on employees' intention to leave in large organisations.
According to H6, satisfaction with co-workers has H6a a positive effect on organisational commitment and H6b a negative effect on turnover intention.The results only confirm the positive impact on organisational commitment, which can be found in both smaller and larger organisations.It can, therefore, be concluded that the quality of employee relations and the workplace community are shown to contribute to employees' commitment to the organisation, which confirms the findings of Kmieciak (2022).However, the results do not support the findings of Karatepe (2012) and Limpanitgul et al. (2014) suggesting that positive employee relations reduce employees' turnover intention.
H7 was discarded as the operating procedures construct had to be removed from the model due to the low Cronbach's a value.
According to H8, satisfaction with the nature of work has H8a a positive effect on organisational commitment and H8b a negative effect on turnover intention.This hypothesis is supported for large organisations, as the nature of work has a negative effect on turnover intention (r ¼ À0.119) while having a positive effect on organisational commitment (r ¼ 0.336).For small and medium-sized organisations, on the other hand, only the effect on organisational commitment (r ¼ 0.479) can be confirmed.
H9, according to which satisfaction with communication has H9a a positive effect on organisational commitment and H9b a negative effect on turnover intention, is also partially supported.For small and medium-sized organisations, the effectiveness of internal communication significantly reduces the intention of employees to leave (r ¼ À0.148), but this finding cannot be confirmed for large organisations.For larger employers, the effectiveness of organisational communication has a significant effect on employees' organisational commitment (r ¼ 0.303).
Results from both models confirm H10 according to which normative commitment H10a positively affects organisational commitment and H10b negatively affects turnover intention.The results, thus, support the statement that employees' moral beliefs about maintaining organisational affiliation fundamentally influence their organisational commitment (Allen and Meyer, 1990;Meyer and Allen, 1991;Meyer et al., 2002).
H11 is also accepted as the results for both models support the assumption that organisational commitment has a negative effect on turnover intention.Thus, the results confirm the negative relationship between the constructs, in consensus with Sjöberg and Sverke (2000).
In conclusion, the study presents the results of a survey conducted among Hungarian employees.The hypotheses related to employee satisfaction factors (H1-H9) are only partially accepted, which draws attention to the fact that the general literature findings are not always valid when approached from a specific perspective, which in this case was the IJOA organisational size difference.At the organisational level, therefore, differentiated retention measures are recommended to ensure effective staff retention, taking into account both organisational and employee characteristics.Nevertheless, regardless of organisational size, the results show that the nature of work, normative commitment, benefits and co-workers variables have a significant effect on employees' commitment to the organisation, and that organisational commitment and Normative commitment have an effect on intention to leave.This result draws attention to the fact that these factors should form the basis for measures to retain staff.

Limitations and future research suggestions
Naturally, like all studies, this one has its limitations as well.The models created as a result of this research are suitable for identifying differences due to organisational size differences, but there is no empirical evidence yet on the use of the scales for homogeneous groups of employees.The research does not address employee characteristics such as age, gender, job title and position, which may also influence the results.Thus, in the future, a research area can be identified in which research is carried out among different groups of employees.Among groups of employees, particular attention should be given to those over 50 years of age.The increase in average life expectancy has extended the active work-life span, which may result in the emergence of new career paths with specific job requirements, presenting employers with new practical challenges during this period.In addition, research on this age group is relevant from a mental health perspective since burnout syndrome is more common at this stage of life.The study has also revealed several new discoveries regarding the effects of organisational strategies.Of significant interest is the nature of work which plays a key role, regardless of the size of the organisation.This theme underscores a crucial area of research that is especially affected by the revolution in the present technology-driven work atmosphere.Among other things, empirical data are scarce regarding the effect of AI on mental health, which prompts queries regarding work content.
Figure 1.Conceptual framework (Figure 1): H10.Normative commitment (a) positively affects organisational commitment and (b) fighting at work Figure 2. Results of the path analysissmall and medium-sized organisations (n ¼ 236) Figure 3. Results of the path analysislarge organisations (n ¼ 275)

Table 2 .
Sources: Created by the authors; PLS-SEM-generated results

Table 11
Henseler et al. (2016)ed model results.Standardised root means square (SRMR) values were used to examine the model's fit.As perHenseler et al. (2016), the SRMR should be < 0.08.This study exhibits an adequate level of model fit because in the case of SMEs, the SRMR value for the saturated model was 0.068 and for the estimated model it was 0.075.