Introduction

Employee turnover imposes significant costs on organizations. Upon an employee’s resignation, the organization must navigate a protracted process involving recruiting, hiring, and training a replacement—an undertaking accompanied by substantial tangible expenses (Moynihan & Landuyt, 2008; Tziner & Birati, 1996). In addition to the quantifiable costs linked to this replacement cycle, voluntary turnover brings about intangible costs including loss of knowledge, decreased customer satisfaction, disruptions in organizational performance, and diminished morale among remaining employees (Dess & Shaw, 2001; Felps et al., 2009; Hausknecht et al., 2009). Considering these multifaceted costs, it is understandable that scholars and practitioners have given particular attention to comprehending the reasons behind employee departures since the early 20th century (Hom et al., 2017).

Since the pioneering work of March and Simon in 1958, researchers have strived to predict voluntary turnover using various models. A recent meta-analysis of studies on voluntary turnover studies (Holtom et al., 2008) show that more than 50 broad antecedents have been recognized as scientifically valuable predictors of voluntary turnover. While factors such as personality, job satisfaction, job characteristics, and job stress have traditionally been extensively studied in turnover research, the organizational context, including aspects such as organizational culture and employees’ working environment, has only recently garnered recognition from researchers (Holtom et al., 2008; Hom et al., 2017). The term “context” can be defined as “stimuli and phenomena that surround and exist in the environment external to the individual” (Mowday & Sutton, 1993).

Given the pivotal role of context in studies of organizational behavior, it is noteworthy that researchers have not adequately delved into the impact of contextual factors on voluntary turnover. Unlike existing turnover studies, which mainly focus on specific contextual factors like organizational justice (Hemdi & Nasurdin, 2007), trust (Jabeen & Isakovic, 2018), and organizational culture (Campbell & Im, 2016), this study undertakes a more comprehensive examination and endeavors to address a significant gap in the literature by examining the combined influence of various workplace characteristics on turnover intention. Through a comprehensive review of workplace contextual factors and their effects on turnover intention, this study aims to reduce undesired voluntary turnover by gaining a nuanced understanding of the broader organizational context.

Theoretical Backgrounds

Turnover is defined as the extent of individual movement across the membership boundary of a social system (Price, 1977, P. 3). When an employee decides to cross the membership boundary of an organization, it is termed voluntary turnover, whereas involuntary turnover occurs when an employer makes this decision, such as through firing or layoffs. This study specifically centers on voluntary turnover, with turnover intention serving as a proxy for this phenomenon. Turnover intention refers to an individual’s thoughts about leaving their current organization. Although turnover intention may not always lead to actual turnover, research suggests a substantial relationship between the two (Park & Shaw, 2013), with turnover intention frequently regarded as an alternative measure for actual turnover (Price, 2001). Indeed, turnover intention has emerged as a widely acknowledged proxy for actual voluntary turnover in studies conducted in both the public and private sectors (Bertelli, 2007; Caillier, 2011; Cho & Lewis, 2012; Cohen et al., 2016; Moynihan & Landuyt, 2008; Pitts et al., 2011).

Determinants of Voluntary Turnover

The cost of replacing a departing employee is estimated to be approximately twice their annual salary (Allen et al., 2010). This significant expense has motivated researchers’ efforts over the past century to understand the reasons behind employee departures (Hom et al., 2017). Resulting in over 1,500 academic studies, these investigations have identified 50 broad antecedents of voluntary turnover (Holtom et al., 2008). Determinants of voluntary turnover are typically categorized into external environmental, work-related organizational, and individual characteristic factors (Mobley et al., 1979). External environmental factors encompass perceived alternatives (such as job availability) and the unemployment rate (e.g., Anderson & Milkovich, 1980; Carsten & Spector, 1987; Fields, 1976; Lee et al., 2017). Research indicates a positive correlation between employees’ turnover and turnover intention with job availability and a negative correlation with the unemployment rate (Peters et al., 1981). Work-related organizational factors include various aspects like job satisfaction (Medina, 2012), development and growth opportunities (Weng et al., 2010), promotion (Yücel, 2012), pay (Irvine & Evans, 1995), perceived autonomy (Gillet et al., 2013), goal clarity (Davis & Stazyk, 2015), and job stress (Arshadi & Hayavi, 2013). In general, turnover or turnover intention tends to decrease when employees experience high job satisfaction, ample development and growth opportunities, increased promotion prospects, better compensation, greater autonomy, clearly defined goals, or minimal job stress (Griffeth et al., 2000). Individual characteristic factors encompass demographic variables such as gender (Sabharwal, 2015), education level (Lambert et al., 2001), tenure (Um & Harrison, 1998), race (Jones & Harter, 2004), age (Emiroğlu et al., 2015), and marital status (Kim et al., 2012). Typically, turnover or turnover intention tends to decrease when employees are female, less educated, long-tenured, white, older, or married (Griffeth et al., 2000).

Remarkably, the exploration of job turnover in public sector settings did not receive significant attention until the new millennium (Lee & Jimenez, 2011). Determinants of turnover for public employees have been identified across various areas, including job characteristics (Kim, 2012), human resource management practices (Tinti et al., 2017), person-organizational fit (Ballinger et al., 2016), and public service motivation (Bright, 2007).

Contextual Factor Studies

The organizational context has only recently garnered attention from turnover researchers, despite its significant impact on the occurrence and interpretation of organizational behavior (Rubenstein et al., 2018, p. 38). Traditionally, turnover researchers have focused on examining the effects of selected contextual factors on voluntary turnover, considering both the organizational context level and the person-context interface (Holtom et al., 2008; Rubenstein et al., 2018). Factors at the organizational context level that have captured turnover researchers’ attention include organizational support (Amarneh et al., 2021), engagement aggregated (Harter et al., 2003), organizational citizenship behavior (Mallick et al., 2014), organizational size (Guan et al., 2014), and diversity level (Choi, 2013). Generally, researchers have found that turnover or turnover intention decreases when employees experience high organizational support, exhibit high engagement, engage in organizational citizenship behaviors, work in larger organizations, and experience lower levels of diversity.

Person-context interface factors encompass elements such as offered rewards (Miao et al., 2013), organizational justice (George & Wallio, 2017), trust (Ellickson & Logsdon, 2001), and organizational culture (Egan, 2008). Researchers generally found that employees’ turnover or turnover intention decreases when employees are satisfied with rewards, experience justice and trust in their organization, or when a positive organizational culture (e.g., high-performing or learning culture) prevails.

Workplace and Demographic Factors

The study employed the Merit Principles Survey 2016 Data, providing a comprehensive perspective on various workplace aspects. Utilizing exploratory factor analysis, we grouped 20 workplace variables into four distinct factors. Further elaboration on the methodology of the factor analysis will be provided in the subsequent methods section. In this section, our focus shifts to proposing hypotheses centered around each of the identified factors and demographic variables, which will serve as focal points in our regression analyses.

Workplace Factors

Happy and Innovative Working Climate (Factor 1)

Studies suggest that when employees feel valued, their inclination to leave decreases (Mancuso et al., 2010). For example, expatriate educators are more inclined to stay in their current positions when they experience a positive work environment, characterized by recognition from both peers and management (Odland & Ruzicka, 2009). An innovative work environment encourages employees to explore fresh ideas, nurturing a heightened sense of psychological empowerment and job satisfaction (Hsu & Chen, 2017). As a result, an innovative work environment has been linked to a reduction in turnover intention (Yeun, 2014).

  • H1: Employees working in happy and innovative working climate tend to have a low level of turnover intention.

Feeling Valued and Trusted (Factor 2)

Employees who perceive low support are less likely to feel valued, and this has been linked to higher turnover, particularly in the retail sector (Shanock & Eisenberger, 2006). The sense of not feeling valued is identified as a significant factor contributing to voluntary turnover among U.S. child welfare employees (Nittoli, 2003). Conversely, when employees feel trusted, they experience higher levels of autonomy, fostering a desire to remain in their current organizations (Dirks & Skarlicki, 2004).

  • H2: When employees feel valued and trusted, they tend to have a low level of turnover intention.

Coworker Support and the Spirit of Camaraderie (Factor 3)

The relationship with coworkers and their support play a crucial role in determining organizational departure (Feeley, 2000). Employees who maintain positive relationships with coworkers and receive support from them exhibit lower turnover intentions (Young, 2015). Furthermore, a strong sense of camaraderie also diminishes turnover intentions (Lopes Morrison, 2005).

  • H3: When the spirit of camaraderie exists in an organization and employees receive support from coworker, they tend to have a low level of turnover intention.

Opportunities for Growth and Development (Factor 4)

Employees often depart organizations in pursuit of better growth and development opportunities (Quarles, 1994). For instance, the availability of career advancement prospects within current organizations fosters organizational commitment, potentially reducing turnover intentions among employees of public accounting firms (Nouri & Parker, 2013).

  • H4: When employees are satisfied with opportunities for growth and development, they tend to have a low level of turnover intention.

Demographic Factors

In this study, various demographic factors were included in regression analyses, revealing that turnover intention generally decreases due to certain demographic factors. Employees with longer tenure (Griffeth et al., 2000), a managerial status (Farris, 1969), higher salary (Sturman & Trevor, 2001), an older age (van Hooft et al., 2021), union membership or a teleworker status (Peters et al., 2004) typically exhibit lower levels of turnover intention. Conversely, turnover intention tends to increase among female employees (Karatepe et al., 2008), racial minorities (Xue, 2015), or those with higher levels of education (Choi, 2006).

  • H5-1: Employees with longer tenure show a lower level of turnover intention than employees with shorter tenure.

  • H5-2: Employees with higher managerial status show a lower level of turnover intention than employees with lower managerial status or non-managerial status.

  • H5-3: Employees with higher salary show a lower level of turnover intention than employees with lower salary.

  • H5-4: Employees with older age show a lower level of turnover intention than employees with younger age.

  • H5-5: Employees with a labor union membership show a lower level of turnover intention than employees without a labor union membership.

  • H5-6: Teleworkers show a lower level of turnover intention than non-teleworker.

  • H5-7: Female employees show a higher level of turnover intention than male counterparts.

  • H5-8: Employees with racial minority status show a higher level of turnover intention than employees with racial majority status.

  • H5-9: Employees with more education show a higher level of turnover intention than employees with less education.

Methods

This study employed ordinary least squares (OLS) regression analyses to investigate the influence of various workplace characteristics on turnover intention. Before conducting the OLS regression, exploratory factor analysis was utilized to uncover underlying factors among the 20 workplace variables. The data for this study were obtained from the Merit System Protection Board (MSPB)’s 2016 Merit Principles Survey (MPS) dataset ‘path 2.’ The sample consisted of 14,473 full-time civilian federal employees from 24 federal agencies, with a response rate of 38.7% for the dataset ‘path 2’ (Merit System Protection Board, 2016).

Major Variables

  1. 1)

    Dependent variable

In this study, turnover intention was used as the dependent variable. While turnover intention may not perfectly align with actual turnover, a strong correlation between the two exists (Belkhir, 2009), and turnover intention has remained a key focus in turnover research (Cohen et al., 2016). Consequently, turnover intention frequently acts as a proxy measure for actual turnover in public administration literature (Moynihan & Landuyt, 2008), as well as in general turnover studies (Griffeth et al., 2000). In the 2016 MPS dataset, survey participants were asked to indicate their agreement level regarding the possibility of transitioning to a different occupation or line of work, with responses rated on a scale from 1 (strongly disagree) to 5 (strongly agree).

  1. 2)

    Independent variables

Various aspects of the workplace for federal employees served as independent variables in this study. Survey participants were asked to express their level of agreement on diverse workplace variables, rated on a scale from 1 (strongly disagree) to 5 (strongly agree). To identify underlying dimensions among the 20 workplace variables, an exploratory factor analysis was conducted. In determining the number of factors, both eigenvalues and an eigenvalue scree plot were considered, as suggested by methodologists (Ferguson & Cox, 1993; Hayton et al., 2004). While one factor exhibited an eigenvalue greater than one (see Appendix Table 4), the eigenvalue scree plot indicated that the slope of the graph did not change significantly after the fourth factor (see Appendix Fig. 1). Following the scree test guideline (DeCoster, 1998; Yong & Pearce, 2013), it was recommended to retain all factors until the slope displayed minimal change. Consequently, this study identified four factors. Further details regarding the factor analysis can be found in Appendix. After the factor analysis, appropriate names were assigned to the extracted factors, acknowledging that the given names may not fully encapsulate the meaning of all component variables in each factor. The four attained factors, along with their component variables, are outlined below. The meanings of the component variables in each factor are detailed in Appendix Table 5.

  • Factor 1 (happy and innovative working climate): w9, w10, w11, w12, w14, w16, w17

  • Factor 2 (feeling valued and trusted): w2, w4, w5, w6, w7, w8

  • Factor 3 (coworker support and the spirit of camaraderie): w1, w3, w8, w13, w15

  • Factor 4 (opportunities for growth and development): w19, w20

The internal consistency or reliability of these four factors was evaluated by computing Cronbach’s alpha values. The Cronbach’s alpha values for all four factors fell within the range of 0.85 to 0.95 (refer to Appendix Table 5). According to the standards for internal consistency (Cronbach, 1951; Nunnally, 1978), a set of variables is considered to have satisfactory internal consistency or reliability when the Cronbach’s alpha value exceeds 0.7.

  1. 3)

    Demographic variables

Nine demographic variables were utilized in this study, with numbers assigned to each variable as follows.

  1. (1)

    Years with current agency (d2)

    1: 3 years or less, 2: 4 years or more,

  2. (2)

    Supervisory status (d4)

    1: non-supervisor, 2: team leader, 3: supervisor, 4: manager, 5: executive

  3. (3)

    Union membership (d8)

    0: non-union membership, 1: dues-paying union membership

  4. (4)

    Salary level (d10)

    1: $74,999 or less, 2: $75,000-$99,999, 3: $100,000-$149,999, 4: $150,000 or more

  5. (5)

    Racial minority (d11)

    0: non-minority, 1: minority

  6. (6)

    Gender (d12)

    0: male, 1: female

  7. (7)

    Age group (d15)

    1: 39 and under, 2: 40 and over

  8. (8)

    Education level (d16)

    1: less than AA degree, 2: AA or BA degree, 3: graduate degree

  9. (9)

    Teleworker status (d18)

    0: non-teleworker, 1: teleworker

Findings

Descriptive Statistics and Correlations

Mean values and standard deviations for both workplace and demographic variables are presented in Tables 1 and 2. Table 1 demonstrates that all correlations between workplace variables were significant and positive (p <.01), and likewise, all correlations between workplace variables and turnover intention were significant and negative (p <.01). In essence, turnover intention tends to decrease when employees express agreement or strong agreement with various workplace variables.

Table 1 Descriptive statistics and correlations for workplace variables
Full size table
Table 2 Descriptive statistics for demographic variables
Full size table

On average, federal employees did not demonstrate a high level of turnover intention (mean = 2.33 out of 5). Among the 20 workplace variables, the top five variables with which employees reported a high level of agreement were as follows: “I understand how I contribute to my agency’s mission” (w18, mean = 4.17), “My judgment is trusted and relied on at work” (w6, mean = 3.93), “I feel needed and depended on at work” (w4, mean = 3.88), “I like the quality of relationships I have with my coworkers” (w13, mean = 3.82), and “I feel comfortable being myself at work” (w17, mean = 3.78). It is worth noting that three of these variables (w18, w6, and w4) belong to Factor 2 (Feeling valued and trusted), while none are from Factor 4 (Opportunities for growth and development).

Conversely, the bottom five variables with which employees reported a low level of agreement were: “I am able to share my true thoughts and feelings at work” (w11, mean = 3.38), “I feel encouraged to try new things in my work” (w10, mean = 3.43), “I feel fully appreciated at work” (w9, mean = 3.44), “There is a culture of openness and support for new or different perspectives in my work unit” (w3, mean = 3.45), and “I feel cared about personally at work” (w12, mean = 3.46). Remarkably, four of these variables (w11, w10, w9, and w12) belong to Factor 1 (Happy and innovative working climate).

In Table 2, it is noted that, on average, survey participants remained in their current agencies for more than 4 years and held positions as team leaders or higher-level managers (mean = 2.25). Moreover, 15% of survey participants were union members, 33% were racial minorities, 42% were female, and 56% had the option to telework. Additionally, the average salary range fell between $75,000 and $150,000, the average age exceeded 40 years, and the typical education level was an associate’s (AA) or bachelor’s (BA) degree.

Workplace Variables, Demographics, and Turnover Intention

OLS regression analyses were conducted to determine which aspects of workplace characteristics contribute to decreased turnover intention and how demographic variables influence turnover intention. Variance inflation factor (VIF) was assessed during regression runs to detect potential multicollinearity issues arising from high correlations among workplace variables (see Table 1). Severe multicollinearity in regression analysis can compromise the statistical significance of each independent variable and render the results unreliable (Mansfield & Helms, 1982). However, the average VIF was found to be 2.79. According to suggested guidelines (Mansfield & Helms, 1982; Miles, 2005), multicollinearity is not a concern in regression analysis if the average VIF value is below 10.

The regression analyses began with only demographic variables. As depicted in Model 1 (Table 3), turnover intention decreased as salary increased (coefficient: − 0.170, p <.001), as employees aged (coefficient: − 0.094, p <.05), or when employees had the option to telework (coefficient: − 0.057, p <.05). Conversely, turnover intention increased if employees were union members (coefficient: 0.102, p <.01) or belonged to racial minorities (coefficient: 0.343, p <.001).

Table 3 Regression analysis results
Full size table

In Model 2 (Table 3), the four workplace variable factors (consisting of 20 variables) were incorporated with the demographic variables from Model 1. Notably, more demographic variables demonstrated significant effects on turnover intention in this model. In addition to the demographic variables that showed significant effects in Model 1, turnover intention decreased with longer tenure at current agencies (coefficient: − 0.096, p <.05), or among female employees (coefficient: − 0.059, p <.05). Conversely, turnover intention increased when employees held higher managerial status (coefficient: 0.051, p <.001). However, telework status and education level did not exhibit a significant effect on turnover intention.

In summary, out of the nine hypotheses regarding the effects of demographic variables, four were supported. Longer tenure (H5-1), higher salary (H5-3), and older age (H5-4) were associated with lower turnover intention, while racial minority status (H5-8) was linked to higher turnover intention. The effects of higher managerial status (H5-2), union membership (H5-5), and being female (H5-7) contradicted the predictions in the hypotheses. Unexpectedly, turnover intention increased when employees held higher managerial status or were union members, but decreased when employees were female. These unexpected findings will be discussed in the last section.

Regarding the effects of workplace variables, Model 2 revealed that three out of seven variables in Factor 1, three out of six variables in Factor 2, one out of five variables in Factor 3, and two out of two variables in Factor 4 had significant and negative effects on turnover intention. In other words, turnover intention significantly decreased when employees agreed or strongly agreed with those workplace variables. Overall, H1 and H2 were partially supported, H3 was not supported, and H4 was supported. Specifically, turnover intention significantly decreased when employees agreed or strongly agreed with Factor 4 (Opportunities for growth and development), whereas turnover intention was not affected by most variables in Factor 3 (Coworker support and the spirit of camaraderie). Among Factor 1 variables, turnover intention significantly decreased when employees agreed or strongly agreed with “I feel fully appreciated at work” (w9), “I feel comfortable talking to my supervisor about the things that matter to me at work” (w14), and “I feel comfortable being myself at work” (w17). For Factor 2 variables, turnover intention significantly decreased when employees agreed or strongly agreed with “My judgment is trusted and relied on at work” (w6), “I feel valued at work” (w7), and “I understand how I contribute to my agency’s mission” (w18).

Managerial Implications and Contributions to the Literature

This study identifies specific workplace factors that significantly influence turnover intention. Factor 4, which pertains to opportunities for growth and development, emerged as the most influential, indicating that investing in employee development can significantly reduce turnover intention. Conversely, Factor 3, related to coworker support and camaraderie, showed the least impact. This suggests that while social support is important, it may not be as influential in retaining employees as opportunities for professional growth. HR managers can use these findings to tailor their strategies and initiatives to address turnover. Emphasizing and enhancing opportunities for growth and development, such as training programs, career advancement paths, and skill-building opportunities, can be particularly effective in reducing turnover intention. Additionally, fostering a positive and innovative working climate (Factor 1) and ensuring employees feel valued and trusted (Factor 2) are also crucial for mitigating turnover intention.

This study also highlights the influence of demographic variables on turnover intention, including tenure, salary, age, race, managerial position, union membership, and gender. Understanding these demographic trends can help managers identify at-risk groups and implement targeted retention strategies. For example, recognizing that employees promoted to higher managerial positions may experience increased turnover intention due to heightened stress can guide the design of leadership development and support programs. In summary, this study underscores the importance of considering diverse workplace factors and demographic variables in understanding and addressing turnover intention. By leveraging these insights, organizations can develop targeted HRM practices that foster employee engagement, satisfaction, and retention.

The findings of this study make a significant contribution to the existing literature on turnover intention and workplace factors. The study fills a gap in the turnover literature by examining a wide array of organizational contextual factors that have not received significant attention previously. Unlike prior studies that often focused on specific factors such as organizational justice or culture, this research considers all contextual factors simultaneously. This holistic approach provides a comprehensive understanding of how various organizational elements interact to influence employee turnover intentions. By identifying which contextual factors have a more pronounced effect on turnover intention, the study offers valuable insights for organizations seeking to implement targeted retention strategies. For example, it highlights the significant impact of opportunities for growth and development (Factor 4) in reducing turnover intention. Understanding the relative importance of different factors can help organizations allocate resources more effectively to address turnover. While acknowledging the need for caution in interpreting findings, the study’s comprehensive approach to examining workplace factors and turnover intention enhances the generalizability of its results. By considering a diverse range of demographic variables and organizational contexts, this study provides insights that can be applicable across various industries and settings.

Discussion and Conclusion

This study aimed to identify workplace factors that could potentially alleviate turnover intention. Initial correlation analyses suggested a general inverse relationship between turnover intention and various workplace factors. However, further examination through Ordinary Least Squares (OLS) regression analyses revealed that turnover intention was primarily influenced by Factor 4 (opportunities for growth and development), while Factor 3 (coworker support and camaraderie) had the least impact. Additional findings from the OLS regression analyses indicated that Factor 1 (a positive and innovative working climate) and Factor 2 (feeling valued and trusted) exerted a partial influence on turnover intention. These results emphasize the nuanced interplay of workplace elements in shaping employee attitudes toward turnover.

This study also delved into the impact of major demographic variables on turnover intention. Consistent with existing literature, employees with longer tenure, higher salaries, or older age exhibited a lower level of turnover intention, while those from racial minority backgrounds displayed a higher level of turnover intention. However, certain findings contradicted initial hypotheses. Contrary to the prediction in hypothesis H5-2, turnover intention increased when employees were promoted to higher managerial positions. Despite the general tendency for managers to stay in their current organizations due to higher investment (Farris, 1969), the study suggests that the heightened stress associated with elevated managerial roles, as indicated by Cavanaugh et al. (2000), may outweigh the perceived value of their investment, potentially leading to resignation. Additionally, turnover intention showed an unexpected increase among union members, contrary to the prediction in hypothesis H5-5. This finding supports an argument that union members typically experience lower job satisfaction than non-members (Hammer & Avgar, 2017). Despite the protective measures offered by unions (Freeman & Medoff, 1984), this study suggests that the reported lower job satisfaction among union members might contribute to their decision to resign. Contrary to the prediction in hypothesis H5-7, female employees exhibited a lower level of turnover intention than their male counterparts. This unexpected finding may be attributed to the observed higher job satisfaction among female employees, despite typically facing working conditions and rewards inferior to those of their male counterparts, as noted in some studies (e.g., Clark, 1997). The implication is that the presence of high job satisfaction among female employees serves as a mitigating factor against turnover intention, influencing their decision to remain in their current roles.

Despite the valuable contributions made by this study, it is essential to approach the interpretation of its findings with caution. Several factors warrant consideration. Firstly, while the focus on a limited number of workplace contextual factors and their impact on turnover intention is helpful, it is important to note that there is no universally agreed-upon set of workplace factors. The definition of workplace contextual factors may vary based on the survey questions posed. Conducting studies across diverse settings is crucial to establish more commonly agreed-upon workplace factors, enhancing the generalizability of findings. Secondly, the applicability of the findings in this study to public employees at different levels of government, such as state and local governments, or in various countries, remains uncertain. Given that the MPS data included only federal employees in the U.S., further research in different settings is necessary before extrapolating the findings to other contexts. Expanding the scope of investigation to include a broader range of participants will contribute to a more comprehensive understanding of the relationships between workplace contextual factors and turnover intention. Overall, this study advances our understanding of turnover intention by shedding light on the nuanced interplay of workplace elements and their impact on employee attitudes. Its findings offer actionable insights for HR practitioners and provide a foundation for future research in this field.