Introduction
Work meaningfulness, understood as the degree to which employees feel work is important, personally significant, worthwhile (Caillier, Reference Caillier2021; Lysova, Allan, Dik, Du, & Steger, Reference Lysova, Allan, Dik, Du and Steger2019; Woods & Sofat, Reference Woods and Sofat2013), and contributing positively to society (Bailey, Yeoman, Madden, Thompson, & Kerridge, Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019), has been viewed as a critical psychological state. This state, driven by core job dimensions, has been shown to lead to numerous positive outcomes, including job satisfaction (Steger, Dik, & Duffy, Reference Steger, Dik and Duffy2012), work motivation (Caillier, Reference Caillier2021), performance (Nikolova & Cnossen, Reference Nikolova and Cnossen2020), and employee’s intention to stay (Sánchez-Cardona, Vera, & Marrero-Centeno, Reference Sánchez-Cardona, Vera and Marrero-Centeno2021). It has also been found to enhance beneficial work outcomes (Dechawatanapaisal, Reference Dechawatanapaisal2022) while exhibiting a negative relationship with turnover (Charles-Leija, Castro, Toledo, & Ballesteros-Valdés, Reference Charles-Leija, Castro, Toledo and Ballesteros-Valdés2023; Nikolova & Cnossen, Reference Nikolova and Cnossen2020; Scroggins, Reference Scroggins2008).
In 2008, a study of employees across various industries found meaningful work to be more strongly related to intentions to leave than job satisfaction (Scroggins, Reference Scroggins2008). However, in a review of 100 years of turnover literature, meaningfulness is not mentioned at all; it is the workplace antecedent of job satisfaction that is deemed to play a major role in the turnover intention process (Hom, Lee, Shaw, & Hausknecht, Reference Hom, Lee, Shaw and Hausknecht2017), with a significant and negative correlation with turnover intentions for knowledge workers (Hofaidhllaoui & Chhinzer, Reference Hofaidhllaoui and Chhinzer2014; Huang, Reference Huang2011).
Thus, while companies attempt to provide work that is meaningful and employees search for meaningful jobs (Rigby & Ryan, Reference Rigby and Ryan2018), researchers emphasize the importance of advancing and refining the conceptual and theoretical frameworks underpinning the understanding of work meaningfulness (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019; Blustein, Lysova, & Duffy, Reference Blustein, Lysova and Duffy2023). In reviewing 71 empirical studies on work meaningfulness, the authors concluded that further research is needed to explore the significance of various job design features to enhance our understanding of the relationship between job design and work meaningfulness (Bailey et al., Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019).
Our aim is thus to analyze the relationship between work meaningfulness and both turnover intentions and job satisfaction in a less-studied group, that of knowledge workers, by examining the role of job design and contextual factors. Our setting is an electric commercial vehicle manufacturer, specifically a division that had been experiencing higher turnover. By focusing on a single division at one point in time, we assess the current attitudes and behaviors of the employees in their specific context to offer a predictive model that could be tested in other contexts. We also contribute to the turnover literature, which lacks comprehensive studies investigating the role of meaningful work (Scroggins, Reference Scroggins2008), although the recent study of meaningful work and turnover in the context of the COVID-19 pandemic is one of the few exceptions (Heath, Williams, & Luse, Reference Heath, Williams and Luse2024).
We also enhance the body of research on job satisfaction, a critical aspect of well-being, by addressing the need for an increased focus on employee well-being amidst changes in the nature and context of work (Guest, Reference Guest2017). While some research shows that meaningful work has important implications for employee well-being (Allan, Batz-Barbarich, Sterling, & Tay, Reference Allan, Batz-Barbarich, Sterling and Tay2019), scholars lament the scarcity of research regarding this relationship (Bailey et al., Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019). In our current context of knowledge workers, our results suggest that work meaningfulness serves as a more powerful predictor of job satisfaction. Nevertheless, it has a significantly strong negative relationship with turnover intentions. A surprising result concerns both higher levels of turnover intentions and lower job satisfaction among supervisory-level employees. We address this in a later section.
To expand the limited body of research on work meaningfulness that applies self-determination theory (SDT) (Allan, Autin, & Duffy, Reference Allan, Autin and Duffy2016; Frank Martela et al., Reference Martela, Gómez, Unanue, Araya, Bravo and Espejo2021; Nikolova & Cnossen, Reference Nikolova and Cnossen2020), we utilize SDT as our theoretical framework. The results of a study of working adults in 30 European countries showed that the nonmonetary aspects of work, such as relatedness, autonomy, and competence (which correspond to SDT’s basic psychological needs), have a 4.6 times stronger association with the meaningfulness of work than income, job insecurity, benefits, and working hours (Nikolova & Cnossen, Reference Nikolova and Cnossen2020). Researchers have recently encouraged the application of SDT to guide further research on work meaningfulness (Martela et al., Reference Martela, Gómez, Unanue, Araya, Bravo and Espejo2021; Nikolova & Cnossen, Reference Nikolova and Cnossen2020).
The rest of this paper is structured as follows: After this introduction, we briefly describe our theoretical framework, look at the variables of work meaningfulness, job autonomy, job satisfaction, and turnover intentions, and state our hypotheses. We then describe the study population sample and the measures used, followed by the results of the partial least squares structural equation modeling (PLS-SEM) analyses. The study closes with a discussion of its main theoretical and practical implications, limitations, and future lines of research.
Theoretical lens: SDT
SDT is an empirically based theory of human behavior that is concerned with how social-contextual factors support or thwart individual thriving through the satisfaction of the three basic psychological needs for competence, autonomy, and relatedness (Ryan & Deci, Reference Ryan and Deci2017). SDT scholars maintain that satisfying these needs is the basis for a meaningful life, that life purposes or goals may not be experienced as meaningful if they do not satisfy basic psychological needs and furthermore, meaning is viewed as an outcome of basic need satisfaction (Ryan & Deci, Reference Ryan and Deci2017). We propose that the same principles apply to work meaningfulness.
Scholars maintain that individuals must satisfy these basic needs to develop and function optimally (Deci & Ryan, Reference Deci and Ryan2011). The primary focus of SDT is the individual and how factors such as work context and management style support or hamper the individual’s motivation and well-being (Rigby & Ryan, Reference Rigby and Ryan2018). SDT proposes that a positive work environment (Trépanier, Forest, Fernet, & Stress, Reference Trépanier, Forest, Fernet and Stress2015) and the social environment (Grouzet, Vallerand, Thill, & Provencher, Reference Grouzet, Vallerand, Thill and Provencher2004) advance well-being by satisfying these psychological needs, thus ensuring autonomous work motivation. Within SDT, meaning is considered an outcome of the natural inherent growth processes of intrinsic motivation.
Although Person-Environment Fit theory is similar in that an individual’s thriving is contingent upon the degree of congruence between personal attributes and contextual factors, SDT provides explanatory mechanisms for this type of fit on well-being and other positive individual-level outcomes (Greguras, Diefendorff, Carpenter, & Troster, Reference Greguras, Diefendorff, Carpenter, Troster and Gagné2015). Experts argue that trait-based fit is not a sufficient condition for meaningful work (Lysova et al., Reference Lysova, Allan, Dik, Du and Steger2019). Furthermore, researchers emphasize that, according to SDT, not only is a minimum level of motivation necessary to derive any utility from meaning, but the satisfaction of the three basic psychological needs also nurtures and maintains meaningfulness (Nikolova & Cnossen, Reference Nikolova and Cnossen2020; Ryan & Deci, Reference Ryan and Deci2017). We believe that SDT is better suited as our focus is on human motivation and how well environments can support human psychological needs.
The first of the three needs is competence, which refers to the need to be proficient and effective in one’s work, as feelings of effectiveness nourish peoples’ selves (Ryan & Deci, Reference Ryan and Deci2017). The need for autonomy concerns the extent to which individuals experience their behavior as fully self-endorsed and compatible with their values. Autonomous actions can more fully engage employees’ talents, abilities, and energies (Ryan & Deci, Reference Ryan and Deci2017). Finally, relatedness concerns feeling connected to others, a sense of belonging, and mutual respect (Deci et al., Reference Deci, Ryan, Gagné, Leone, Usunov and Kornazheva2001).
According to SDT, employees have been shown to be healthier, happier, and work harder when efforts to satisfy their basic psychological needs are supported in their place of work (Deci, Olafsen, & Ryan, Reference Deci, Olafsen and Ryan2017). The results are higher levels of autonomous motivation, which translates to employee well-being and talent retention, among other positive outcomes (Rigby & Ryan, Reference Rigby and Ryan2018). SDT suggests that intrinsic motivation can be seen as a process through which individuals create meaning (Ryan & Deci, Reference Ryan and Deci2017). The results of a study of employees from a wide range of occupations and industries concluded that being internally motivated at work (resulting from need satisfaction) may be essential to experiencing meaningful work (Allan et al., Reference Allan, Autin and Duffy2016).
As a number of researchers base their understanding of meaningful work on SDT (Nikolova & Cnossen, Reference Nikolova and Cnossen2020) we now turn to the primary variable of interest in our study, work meaningfulness.
Work meaningfulness
Almost half a century ago, researchers proposed the now famous theory of job design (the Job Characteristics Model), which positioned work meaningfulness as a psychological state resulting from the core job design characteristics of skill variety, task identity, task significance, autonomy, and feedback (Hackman & Oldham, Reference Hackman and Oldham1975).
Decades later, Kahn stressed the importance of meaningful connections at work and identified five dimensions of meaningful connections among employees: task accomplishment, career development, sense-making, provision of meaning, and personal support. He discerned that these dimensions enabled employees to build relationships that met cognitive, relatedness, growth, and other needs (Kahn, Reference Kahn, Dutton and Ragins2007). Other researchers have identified a sense of belonging and unity with others as a fundamental element of meaningfulness (Lips-Wiersma & Morris, Reference Lips-Wiersma and Morris2009; Rosso, Dekas, & Wrzesniewski, Reference Rosso, Dekas and Wrzesniewski2010), which is equitable to SDT’s need for relatedness. Making the theoretical connection to SDT, scholars claim that employees cannot benefit from meaning without satisfying the psychological needs of competence, autonomy, and relatedness (Nikolova & Cnossen, Reference Nikolova and Cnossen2020).
More recently, meaningful work has attracted much attention among scholars (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019, Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019; Lysova, Fletcher, & El Baroudi, Reference Lysova, Fletcher and El Baroudi2023) and is considered an important pillar of the work lives of people around the globe (Blustein et al., Reference Blustein, Lysova and Duffy2023). However, its conceptualization is trickier; scholars have identified 36 definitions of meaningfulness (Martela & Pessi, Reference Martela and Pessi2018). Nevertheless, scholarship has advanced considerably by bringing to light its multi-dimensionality and presenting a framework that links the individual, job, organizational, and societal dimensions as factors that influence meaningful work (Lysova et al., Reference Lysova, Allan, Dik, Du and Steger2019).
Building upon extant research that identifies job autonomy as a central tenet of work design (Dysvik, Kuvaas, & Gagné, Reference Dysvik, Kuvaas and Gagné2013) and its potential to foster motivation and well-being (Väänänen, Toivanen, & Lallukka, Reference Väänänen, Toivanen and Lallukka2020) alongside acknowledging its essential role in the effective functioning of knowledge workers (Drucker, Reference Drucker1999), we have selected job autonomy as the independent variable in our model. Notably, in a recent study involving mental health workers, autonomy emerged as a significant predictor of job satisfaction, stress, and turnover intention (Hood & Patton, Reference Hood and Patton2022).
Job autonomy
Job autonomy is the extent to which a job allows freedom, discretion, and independence to schedule work, make decisions, and choose the methods to perform tasks (Humphrey, Nahrgang, & Morgeson, Reference Humphrey, Nahrgang and Morgeson2007; Van den Broeck, Ferris, Chang, & Rosen, Reference Van den Broeck, Ferris, Chang and Rosen2016). SDT’s equivalent of autonomy, it primarily influences work outcomes via autonomy need satisfaction (Laguerre & Barnes-Farrell, Reference Laguerre and Barnes-Farrell2024), and is considered an important job resource that may contribute to achieving and sustaining positive individual-level outcomes (Clausen, Pedersen, Andersen, Theorell, & Madsen, Reference Clausen, Pedersen, Andersen, Theorell and Madsen2022). As a resource, job autonomy is also known to protect employees from psychological strain as it allows individuals to control basic aspects of their working conditions, such as organizing their work tasks and scheduling their breaks (Karasek, Reference Karasek1979). It is essential in complex jobs as employees are required to make decisions and exercise judgment (Chung-Yan, Reference Chung-Yan2010).
Redesigning jobs to provide greater autonomy has been demonstrated to benefit employee well-being (Clausen et al., Reference Clausen, Pedersen, Andersen, Theorell and Madsen2022; Guest, Reference Guest2017), and scholars have found that greater relative autonomy makes companies great places to work (Ryan & Deci, Reference Ryan and Deci2017). Providing autonomy support, in turn, facilitates the satisfaction of the needs for competence and relatedness (Ryan & Deci, Reference Ryan and Deci2017). SDT researchers emphasize that supporting autonomy in the workplace yields a high level of motivation, which allows employees to employ problem-solving skills and cognitive flexibility more deeply (Gagné & Deci, Reference Gagné and Deci2005). Scholars also point to the multidimensionality of autonomy and stress its inherently social and interdependent nature, especially in relation to knowledge work (Väänänen et al., Reference Väänänen, Toivanen and Lallukka2020).
As the need for autonomy has also been found to be connected to work meaningfulness (Humphrey et al., Reference Humphrey, Nahrgang and Morgeson2007; Kubiak, Reference Kubiak2020; Martela & Riekki, Reference Martela and Riekki2018; Zheng, Wu, & Graham, Reference Zheng, Wu and Graham2020) and is a key need in SDT (Karkkola, Kuittinen, & Hintsa, Reference Karkkola, Kuittinen and Hintsa2019), we propose our first hypothesis:
Hypothesis 1: Job autonomy positively influences work meaningfulness.
Meta-analytic findings have shown that work meaningfulness serves as a mediator in the relationship between job characteristics and numerous outcomes (Parker, Reference Parker2014), among them job satisfaction (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019). Recognizing that employee well-being is key to employee retention, we turn to the concept of job satisfaction.
Job satisfaction
Employee well-being is a broad term and encompasses both physical and psychological health. Here, we focus on the psychological aspect and context-specific well-being, in particular, job satisfaction (Arnold, Turner, Barling, Kelloway, & McKee, Reference Arnold, Turner, Barling, Kelloway and McKee2007). Job satisfaction has been defined as an expression of approval of a work environment (Locke, Reference Locke and Dunn1976) and positive feelings about the employee’s particular role (Harris, Cooper–Thomas, Smith, & Cheung, Reference Harris, Cooper–Thomas, Smith and Cheung2022). It is considered a positive emotion and serves to buffer the harmful effects of stress (Fredrickson, Reference Fredrickson2001). Several studies have also revealed a strong negative relationship between job satisfaction and turnover intentions (Chung, Jung, & Sohn, Reference Chung, Jung and Sohn2017; Jung, Reference Jung2012; Richer, Blanchard, & Vallerand, Reference Richer, Blanchard and Vallerand2002).
Recent research points to the continued importance of job satisfaction, and scholars call for an emphasis on its determinants (Judge, Zhang, & Glerum, Reference Judge, Zhang, Glerum, Sessa and Bowling2020). Empirical findings support the notion that work meaningfulness enhances job satisfaction (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019; Choi et al., Reference Choi, Au, Hui, Leung, Chiu and Yang2021).
As job autonomy is an important determinant of well-being (Clausen et al., Reference Clausen, Pedersen, Andersen, Theorell and Madsen2022) and work meaningfulness has been found to predict job satisfaction (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019), we present our next hypothesis.
Hypothesis 2: Work meaningfulness positively influences job satisfaction.
In a study of nurses that examined the relationship between workload and job satisfaction, the mediating variable of meaningfulness served as a motivator for the nurses and enhanced their perceived job satisfaction (Al-Hakim, Zhang, Jin, & Sevdalis, Reference Al-Hakim, Zhang, Jin and Sevdalis2022). In a time-lagged study of Generation Y knowledge workers who employed job crafting, the variable of work meaningfulness also served as a mediator, improving the likelihood of them staying with the organization (Malik & Malik, Reference Malik and Malik2024).
We predict that job autonomy, as an element of job design and a valuable job resource shown to enhance intrinsic motivation as per SDT (Liu, Jameel Ahmed, Anjum, & Mina, Reference Liu, Jameel Ahmed, Anjum and Mina2024), will foster work meaningfulness and that work meaningfulness, as the examples above illustrate, is crucial for achieving positive individual outcomes such as job satisfaction or the intention to remain with an organization.
Building on our first two hypotheses and following a segmentation approach (Rungtusanatham, Miller, & Boyer, Reference Rungtusanatham, Miller and Boyer2014), we propose a hypothesis for the mediation effect of work meaningfulness:
Hypothesis 3: Work meaningfulness positively mediates the relationship between job autonomy and job satisfaction.
While a greater focus on well-being is crucial in its own right, an important benefit of job satisfaction is that it has also been found to reduce turnover, a key concern for companies.
Turnover intentions
An organization is adversely affected when employees intend to leave, whether they finally end up departing or not, as it can decrease their fulfillment of obligations and thus negatively affect their individual and organizational performance (Jung, Reference Jung2012). When employees do leave, the resulting work disruptions, losses of organizational knowledge, and emotional effects on co-workers can have a strong negative impact on an organization (Ju & Li, Reference Ju and Li2019).
Both scholars and practitioners have been speculating for decades about the reasons that lead employees to leave organizations and how to prevent production disruption and loss of key employees and organizational memory (Christensen & Knardahl, Reference Christensen and Knardahl2022). The cost of replacing employees has been estimated to be between 90% and 200% of an employee’s annual salary (Allen, Bryant, & Vardaman, Reference Allen, Bryant and Vardaman2010). Knowledge workers are among the most expensive employees to replace and are vital to a firm’s success due to the value of their intellectual capital (Hofaidhllaoui & Chhinzer, Reference Hofaidhllaoui and Chhinzer2014).
Turnover intention is important as it predicts actual turnover or the likelihood that someone will leave an organization in the near future (Caillier, Reference Caillier2021). Despite its predictive value, there has been a tendency among scholars to overlook the influence of working conditions on turnover intentions (Christensen & Knardahl, Reference Christensen and Knardahl2022). In a study comparing working adults who graduated in 2008 with those who graduated in 2022, researchers found that the COVID-19 pandemic has prompted individuals to seek meaningful work opportunities in firms that emphasize their health and well-being (Donald, Reference Donald2023). Interviews with Generation Z professionals point to a sense of purpose at work as the most important driver for their happiness while also placing great importance on a collaborative and inclusive work environment, job autonomy, access to training and development, a competitive salary package, and career progression opportunities (Donald, Reference Donald2024).
Furthermore, the authors of a 2-year turnover study in Norway encourage researchers to focus less on personal factors and attitudes and more on environmental characteristics that may reduce turnover. The results point to factors such as experiencing clarity and harmony between goals, empowering leadership, and a social climate characterized by high levels of trust, support, and fairness (Christensen & Knardahl, Reference Christensen and Knardahl2022). According to scholars, job characteristics do not directly impact turnover intentions; instead, emotional reactions precede intentions to quit (Agarwal & Gupta, Reference Agarwal and Gupta2018). Given the motivating potential of meaningfulness in employee contexts (Albrecht & Su, Reference Albrecht and Su2012), its experience as an affective state that promotes positive emotions (Carton, Reference Carton2018), and a recent study pointing to meaningful work and its negative influence on turnover intention (Heath et al., Reference Heath, Williams and Luse2024), we propose our fourth hypothesis:
Hypothesis 4: Work meaningfulness is negatively related to turnover intentions.
Scholars have demonstrated that work meaningfulness is not a static but rather a continuous state, which involves a deeper level of intrinsic motivation (Malik & Malik, Reference Malik and Malik2024). Other researchers claim that meaningful work can attract and retain talent and sustain employees’ intrinsic motivation (Bailey et al., Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019). We predict that as long as job autonomy enhances work meaningfulness, work meaningfulness will (negatively) impact employee turnover intentions. Thus, we see work meaningfulness as the key to understanding the relationship between job autonomy and turnover intentions.
Building on our first hypothesis regarding the relationship between job autonomy and work meaningfulness, we propose our last hypothesis for the mediating effect of work meaningfulness:
Hypothesis 5: Work meaningfulness negatively mediates the relationship between job autonomy and turnover intentions.
Before viewing our results, let us look at our conceptual model (Figure 1), our setting, and our measurements.
Methodology
Setting and participants
The setting for this study was a division of a commercial electric vehicle manufacturing company in the United States. We were informed that a survey on culture and values had been administered 6 months earlier. The survey results pinpointed several areas that needed attention, among them, insufficient onboarding, a lack of accountability, long working hours, and a sense of confusion regarding individual roles. The company’s leadership advised us that a number of people had recently left the company and that a substantial number of new employees had recently been hired.
This particular division of knowledge workers consisted mainly of engineers, although there were also CAD designers, test technicians, test engineers, software developers, technical publishers, and project managers. As work meaningfulness also refers to the context within which the work is performed (Wang & Xu, Reference Wang and Xu2019) and a company that strives to help humanity by providing clean energy solutions would, in our view, be regarded as significant and beneficial, we deemed our choice all the more interesting for this study. Furthermore, this division was not only comprised of engineers but also included employees occupying various roles across three hierarchical tiers (non-supervisory, team leader, and department manager), rendering it not only representative of the organization but also reflective of the broader electric transportation sector.
A brief explanation for the survey (‘… to follow up and delve deeper into the results of the previous culture survey’) along with a link to the survey on LimeSurvey was sent by email to division managers. The managers then forwarded it to their groups of employees. Although the surveys were voluntary, employees were encouraged to participate to have their voices heard. Several open-ended questions were included for this purpose, to allow the company to implement changes based on employee feedback. Participants were assured of complete anonymity, and only aggregate data was reported back to the company (information regarding gender, age, length of service, and supervisory status was excluded).
Out of a total of 140 employees, 129 answered the surveys. Nine incomplete surveys were excluded, resulting in 120 valid surveys (86% return rate). This high response rate reduced the possibility of selection bias. One executive (the only one holding that position) and two employees who did not specify their supervisor status were excluded from our survey for a total of 117 surveys.
Several recent cross-sectional studies published in top journals (Q1) have successfully used small sample sizes (114 or less) to examine relationships involving psychological need satisfaction or work meaningfulness (Haffer, Haffer, & Morrow, Reference Haffer, Haffer and Morrow2021; Hood & Patton, Reference Hood and Patton2022; Karkkola et al., Reference Karkkola, Kuittinen and Hintsa2019).
Considering the most complex regression in our model is generated by five predictors, we used the G*Power (Faul, Erdfelder, Buchner, & Lang, Reference Faul, Erdfelder, Buchner and Lang2009) tool. Assuming a medium effect size (f 2 = 0.15), we conducted a priori analysis to compute the required sample size. As a result, we would initially need a minimum sample of 92 cases to obtain a power of 0.80 and an alpha level of 0.05. Therefore, our sample (n = 117) is technically sufficient to test our research model. Furthermore, we performed a post hoc analysis, taking into account our final sample of 117 cases. As a result, our sample allows us to achieve a power level of 0.904 (alpha = 0.05, f 2 = 0.15) and is enough to detect path coefficients with effect sizes (f 2) greater than 0.115 (alpha = 0.05, power = 0.80).
Table 1 presents the descriptive statistics for the participants.
N = 117.
Measures
The items for each measure (job autonomy, work meaningfulness, job satisfaction, and turnover intentions) were assessed on a 5-point Likert scale, where 1 = strongly disagree and 5 = strongly agree. As employee perceptions of what is happening to them at work are more accurate predictors of their attitudes, behavior, and well-being than managers’ reports (Boxall & Macky, Reference Boxall and Macky2014), our level of measurement was the individual employee.
Job autonomy
JA was measured using a nine-item scale (Morgeson & Humphrey, Reference Morgeson and Humphrey2006). A sample item was ‘This job provides me with significant autonomy in making decisions.’
Work meaningfulness
WM was measured using four items from the 2016 Merit Principles Survey. Caillier used the scale in his study on workplace aggression and its impact on work meaningfulness (Caillier, Reference Caillier2021). A sample item was ‘The work I do is meaningful to me.’
Job satisfaction
JS was measured using a three-item job satisfaction scalevalidated by Bowling and Hammond (Reference Bowling and Hammond2008), and strongly recommended by these authors as a tool for assessing global job satisfaction. The only modification was to convert the third item into a positive question. A sample item included, ‘All in all, I am satisfied with my job.’
Turnover intentions
TRN was measured using a two-item scale adapted from (Colarelli, Reference Colarelli1984). A sample item included: ‘I frequently think about quitting my job.’
Control variables
We used age, tenure, and supervisor status to control for potential confounding effects. Age was measured in five 10-year increments, while tenure was divided into four increments. Supervisor status was measured in three categories: non-supervisor, team leader/supervisor, and department/division manager.
Data analysis
We applied PLS-SEM (Ciavolino, Aria, Cheah, & Roldán, Reference Ciavolino, Aria, Cheah and Roldán2022) using Smart PLS version 4 to analyze our data (Ringle, Wende, & Becker, Reference Ringle, Wende and Becker2022) and the SEMinR package in R (Ray, Danks, & Calero Valdez, Reference Ray, Danks and Calero Valdez2022) to compute the predictive contribution of the mediator (PCM). Following the guidelines proposed by researchers on the use of PLS-SEM, our choice was based on testing a theoretical framework from a causal-predictive perspective and a relatively small sample of 117 employees (Hair, Risher, & Ringle, Reference Hair, Risher and Ringle2019). Scholars have also praised the greater statistical power of PLS-SEM (compared to covariance based SEM or CB-SEM) as it increases the likelihood of identifying a statistically significant relationship when one is present in the population (Hair, Hult, Ringle, & Sarstedt, Reference Hair, Hult, Ringle and Sarstedt2022). As our model is based on variables that respond to a common factor model, we used the consistent PLS tool to test our model and the consistent bootstrapping tool. In terms of hypothesis testing, scholars recommend that whenever the common factor model holds, researchers should use consistent PLS over regression on sum scores or traditional PLS (Dijkstra & Henseler, Reference Dijkstra and Henseler2015).
Common method bias
Common method bias (CMB) refers to the phenomenon where respondents’ answers are influenced by the manner in which questions are asked, meaning the measurement instrument itself affects their responses rather than their actual beliefs or attitudes (Magalhães-Teixeira, Roldán, & Leal Millán, Reference Magalhães-Teixeira, Roldán and Leal Millán2024). To prevent the occurrence of CMB, we followed Podsakoff’s recommendations during the design of the surveys (Podsakoff, MacKenzie, & Podsakoff, Reference Podsakoff, MacKenzie and Podsakoff2012). Additionally, we conducted a collinearity test based on variance inflation factors following the guidelines provided by Kock (Kock & Lynn, Reference Kock and Lynn2012). According to the authors, a variance inflation factor value greater than 3.3 in the inner model would indicate pathological collinearity and most likely be affected by common method bias. As the variance inflation factor values in this test ranged between 1.349 and 2.715, the model was deemed to be free of common method bias.
Results
Assessment of the measurement model
Researchers recommend a two-step procedure for PLS-SEM assessment, which begins with evaluating the measurement model, followed by assessing the structural model (Hair et al., Reference Hair, Hult, Ringle and Sarstedt2022). To assess the measurement model, we examined indicator reliability, internal consistency reliability, convergent validity, and discriminant validity. All loadings were generally above 0.70, as recommended by Hair et al. (Reference Hair, Hult, Ringle and Sarstedt2022). Cronbach’s α, composite reliability, and Rhoa estimates were all above 0.7 and below 0.95, indicating internal consistency reliability (Guenther, Guenther, Ringle, Zaefarian, & Cartwright, Reference Guenther, Guenther, Ringle, Zaefarian and Cartwright2023). The measures for the average variance extracted supported convergent validity as they were all above 0.5 (Cepeda, Roldán, Sabol, Hair, & Chong, Reference Cepeda, Roldán, Sabol, Hair and Chong2024). The results are displayed in Table 2.
CR: composite reliability; AVE: average variance extracted.
To assess discriminant validity, we evaluated the heterotrait–monotrait values as recommended by Franke and Sarstedt (Franke & Sarstedt, Reference Franke and Sarstedt2019). Our results confirmed that all the heterotrait–monotrait values were below the threshold of 0.85 (Table 3), thus establishing discriminant validity (Henseler, Ringle, & Sarstedt, Reference Henseler, Ringle and Sarstedt2015). To further test discriminant validity, we used the Fornell and Larcker criterion (Fornell & Larcker, Reference Fornell and Larcker1981). Discriminant validity was also supported since the square root of the average variance extracted values was greater than their correlations with the other constructs.
JA: job autonomy; MNG: work meaningfulness; JS: job satisfaction; TRNV: turnover intention. The diagonal elements (bold) are the square roots of the average variance extracted (AVE) values. The Fornell–Larcker criterion is in the lower left corner, and the heterotrait–monotrait ratios (HTMT; italics) are in the upper right corner. Off-diagonal lower elements are the correlations between constructs.
We also applied bootstrap-based saturated model fit tests (standardized root mean square residual [SRMR] and dULS) for confirmatory factor analysis (Henseler, Reference Henseler2017). The two indicators, which measure the difference between the observed correlation matrix and the model-predicted values, were lower than the respective HI95 or HI99 values from the saturated model. This suggests support for our proposed measurement model, as the observed disparity was not significant (Table 4).
SRMR: standardized root mean squared residual; dULS: unweighted least squares discrepancy; HI95: bootstrap-based 95% percentile; HI99: bootstrap-based 99% percentile.
Assessment of the structural model
From a causal perspective, we analyzed the overall fit of the hypothesized model to obtain empirical evidence for the research model (Henseler, Reference Henseler2021) (Table 4). The SRMR achieved a value of 0.075, below the standard cut-off of 0.08. Next, we conducted two bootstrap-based tests of the overall model fit (SRMR and dULS). Our results showed that all discrepancy values were below the 95th or 99th percentile. Consequently, the postulated model cannot be rejected, as it is likely true.
Next, to ensure that our PLS-SEM results were robust, we applied the Gaussian copula approach to test for endogeneity (Hult et al., Reference Hult, M., Hair, Proksch, Sarstedt, Pinkwart and Ringle2018) to identify potential omitted variables. We focused on our two dependent variables, the most complex regressions of the hypothesized model: MNG and JA regressed to JS, and MNG and JA regressed to TRNV. We tested JA and MNG, which may demonstrate endogeneity, and found that they had a nonnormal distribution. The Cramér–von Mises test indicated that both predecessors were not normally distributed (Table 5). Next, we performed a Gaussian copula analysis in SmartPLS and executed a two-tailed bootstrap analysis of 10,000 samples. None of the four copula terms was statistically significant at the 5% probability of error level, leading us to conclude that our results are unaffected by possible endogeneity concerns (Table 5), mainly due to omitted constructs that could correlate with one or more predictor constructs and both dependent constructs.
JA: job autonomy; MNG: work meaningfulness; JS: job satisfaction; TRNV: turnover intention; GC: Gaussian copula. p-value based on bootstrapping (n = 10,000 subsamples) using a two-tailed test.
Having found evidence to support the research model from a causal point of view, we now assess the basic parameters of the structural model. The first step is to ensure that there are no multicollinearity problems between the antecedent variables of the dependent constructs. To do this, we evaluate the predecessor variables’ variance inflation factor values. Since the highest value we found is 1.561, we can confirm that there are no multicollinearity problems in our structural model.
Table 6 and Figure 2 include the key structural parameters of our research model, including the relationships postulated and two nonhypothesized relationships, which are necessary to calculate the mediation effect (Nitzl, Roldan, & Cepeda, Reference Nitzl, Roldan and Cepeda2016; Rasoolimanesh, Wang, Roldán, & Kunasekaran, Reference Rasoolimanesh, Wang, Roldán and Kunasekaran2021). We used consistent bootstrapping (10,000 samples) to generate path values and confidence intervals. We first tested the model without the control variables and excluded the nonhypothesized paths. Our path results were similar to those outlined below; only the MNG path was higher before adding the control variables and the nonhypothesized paths (−0.470 vs. −0.389). The R 2 values were the same for MNG (0.338), lower for JS (0.459 vs. 0.558), and lower for TRNV (0.221 vs. 0.425) while remaining statistically significant. Therefore, including our control variables did not significantly alter the results. Furthermore, all the signs of the path coefficients coincided with our hypotheses.
JA: job autonomy; MNG: work meaningfulness; JS: job satisfaction; TRNV: turnover intentions; VIF: variance inflation factor; PBCI: percentile bootstrap confidence interval. Bootstrapping based on n = 10,000 subsamples. The hypothesized effects are evaluated using a one-tailed test for a Student’s t distribution (CI 90%). The nonhypothesized effects and those of the control variables are assessed by using a two-tailed test (CI 95%).
The significance of the paths, as well as the f 2 values and the R 2 values for the dependent variables after testing the model with the control variables and nonhypothesized paths, are shown in Table 6 and Figure 2. The three direct effects were statistically significant. This means that H1(+), H2(+), and H4(−) were supported. Consequently, JA positively influences MNG, while the latter has a positive significant effect on JS and a negative impact on TRNV. Regarding the control variables, as seen in Table 6, supervisor status was statistically significant in predicting both lower job satisfaction and higher turnover intentions. Length of service also had a statistically significant positive effect on turnover intentions. The R 2 values for our endogenous latent variables were moderate for MNG (0.508), JS (0.558), and TRNV (0.425) (Hair et al., Reference Hair, Hult, Ringle and Sarstedt2022). Thus, the R 2 values showed that 42.5% of turnover intention and 55.8% of job satisfaction were explained by our model, indicating a substantial in-sample predictive power. We also assessed the f 2 effect sizes and compared them with the cut-off points of 0.02, 0.15, and 0.35, as suggested by Cohen (Reference Cohen1988). We found that the effect size of work meaningfulness on job satisfaction was large (f 2 = 0.583), and the effect size of work meaningfulness on turnover intentions was medium (f 2 = 0.173).
To assess our model’s predictive performance, we followed the procedure outlined by Shmueli and colleagues and evaluated the out-of-sample predictive power of the model using PLSpredict (Shmueli et al., Reference Shmueli, Sarstedt, Hair, Cheah, Ting, Vaithilingam and Ringle2019). We performed k-fold cross-validation with k = 3 subgroups to ensure at least 30 cases in each holdout sample. In the first stage, we observed that all Q 2predict values for all indicators of dependent variables were positive (Table 7). In the second step, because the prediction errors were symmetrically distributed, we used the root mean squared error to assess the degree of prediction error. In this case, the root mean squared error PLS values produced smaller prediction errors than the root mean squared error linear model values (Table 7), thus indicating a high level of out-of-sample predictive performance of the model. These results indicate that the research model could yield generalizable findings for other data sets and potentially equivalent contexts.
LM: linear regression model; RMSE: root mean square error.
Finally, we performed the cross-validated predictive ability test (CVPAT) (Sharma, Liengaard, Hair, Sarstedt, & Ringle, Reference Sharma, Liengaard, Hair, Sarstedt and Ringle2023) using the overall indicator average and the linear model prediction benchmarks to evaluate the predictive accuracy of the model. We began by comparing the average loss difference when faced with PLS-SEM predictions versus the naïve indicator average for the overall model. As shown in Table 8, our model had a significant and negative loss difference. Next, we compared the PLS-SEM predictions with the linear model prediction benchmark. The difference was also negative and significant for the overall model. Consequently, from the point of view of the CVPAT analysis, our model has strong predictive validity. This is in line with the conclusions raised by the previous PLSpredict analysis. Therefore, our model would produce generalizable results that could be applied to other similar populations.
We analyzed the mediation effects of our research model (Figure 1), adopting a double causal-predictive approach (Cepeda et al., Reference Cepeda, Roldán, Sabol, Hair and Chong2024). From an explanatory point of view, we sought to understand how and why JA affects both JS and TRNV through an intervening variable, MNG. We included both direct effects (c’ and d’) as nonhypothesized relationships to evaluate the mediating effects (Nitzl et al., Reference Nitzl, Roldan and Cepeda2016). We estimated the total, direct, and indirect effects, controlling for the three control variables. Researchers advise using bootstrap results for testing indirect effects directly from the PLS software to avoid fixed measurement issues (Nitzl et al., Reference Nitzl, Roldan and Cepeda2016). Statistical mediation is considered significant when confidence intervals exclude zero. The results of the mediation analyses (Table 9) revealed significant indirect effects for ab 1 and ab 2. Furthermore, neither direct effect (c’ and d’) was significant. Therefore, H3(+) and H5(−) were supported, and there was only indirect mediation, which implies that MNG fully mediated the relationships between, on the one hand, JA and JS and, on the other, JA and TRNV. In addition, we estimated effect size measures to assess the magnitude of both indirect effects. We first calculated the variance accounted for (VAF) index (Nitzl et al., Reference Nitzl, Roldan and Cepeda2016), which compares the indirect and direct effects to the total impact. We found that 78.4 % of the influence of JA on JS was due to the indirect effect of MNG. Only 21.6 % was direct. At the same time, 66.9% of the effect of JA on TRNV was indirect. Both results reinforce the mediating role of MNG. Next, we turned to the effect size ν, a metric proposed by Lachowicz, Preacher, & Kelley (Reference Lachowicz, Preacher and Kelley2018). The ν indicator can be interpreted following the threshold proposed by Gaskin, Ogbeibu, & Lowry (Reference Gaskin, Ogbeibu, Lowry, Latan, Hair and Noonan2023): 0.01 (small), 0.04 (medium), and 0.09 (large). Table 9 shows that the indirect effect ab 1 achieved a ν value of 0.132, which means a large mediation effect. For its part, ab 2 reached a medium effect size (ν = 0.051). Both results aligned with the results achieved with the VAF assessment and underlined the outstanding mediating role of MNG in both indirect effects.
JA: job autonomy; MNG: work meaningfulness; JS: job satisfaction; TRNV: turnover intention. Total, direct, and indirect effects are calculated, considering supervisor status, age, and length of service as control variables on JS and TRNV, respectively. PBCI: percentile bootstrap confidence interval. Bootstrapping was based on 10,000 subsamples. The indirect effect is assessed by applying a one-tailed test (PBCI 90%). Total and direct effects are assessed using a two-tailed test. VAF: variance accounted for.
In addition, we assessed the two mediating relationships in our research model using a predictive approach. We designed two models that allow us to determine the predictive contribution of MNG as a mediator variable for each dependent construct (i.e., JS [Hypothesis 3] and TRNV [Hypothesis 5]). In this vein, we used the PCM metric (Danks, Reference Danks2021) (Table 10). The SEMinR package (Ray et al., Reference Ray, Danks and Calero Valdez2022) was used to estimate the PCM metrics for each indicator of JS and TRNV, respectively, considering 3 folds and 5,000 repetitions. Regarding the indirect effect of JA on JS via MNG, the three indicators of JS achieved PCM metrics above zero, with two of them showing a solid PCM (over 0.17). This means that MNG contributed substantially and provided strong confirmatory predictive evidence for the mediation effect on JS. Regarding the role of MNG as a mediator in the relationship between JA and TRNV, PCM estimates were also above zero but around 0.05 for the two manifest variables. This result provided almost moderate but confirmatory predictive evidence for the mediation effect of MNG. Both results provided additional post hoc evidence to support the generalizability of MNG as a mediator, as postulated by Hypotheses 3 and 5. Furthermore, our findings justified the added complexity of including MNG as a mediating variable, as it produced an improved predictive accuracy of both JS and TRNV (Danks, Reference Danks2021).
MNG: work meaningfulness; JA: job autonomy; JS: job satisfaction; TRNV: turnover intention; PCM: predictive contribution of the mediator.
Discussion
Theoretical contributions
This study aimed to investigate the role of work meaningfulness among knowledge workers and its direct and indirect effects on turnover intentions and job satisfaction, framed by SDT and its key driving variable, job autonomy.
According to Hypothesis 1, our study supports previous literature identifying job autonomy as a key predictor of work meaningfulness, explaining 33.6% of the variance (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019). Consistent with prior research (Bailey et al., Reference Bailey, Yeoman, Madden, Thompson and Kerridge2019), our findings also show that work meaningfulness on its own plays an important role in predicting job satisfaction, thus supporting Hypothesis 2. Researchers agree that work meaningfulness is a ‘positive, subjective, individual experience’ in relation to work, and emerging literature ties it to positive outcomes (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019).
Regarding the mediated hypotheses, our results indicate that work meaningfulness fully mediates the relationship between job autonomy and job satisfaction (Hypothesis 3) and between job autonomy and turnover intentions (Hypothesis 5). These findings deepen our understanding of how contextual and job design factors can facilitate psychological need satisfaction and, together with work meaningfulness, significantly predict job satisfaction and turnover intentions. As scholars note, specific psychological experiences are necessary for work to be perceived as meaningful (Martela et al., Reference Martela, Gómez, Unanue, Araya, Bravo and Espejo2021), and employees must attain a minimum level of motivation to experience any utility from meaning (Nikolova & Cnossen, Reference Nikolova and Cnossen2020).
Regarding Hypothesis 4, which predicts a negative relationship between work meaningfulness and turnover intentions, our results are in contrast to those of a meta-analysis that pointed to work meaningfulness performing more effectively in promoting well-being than in reducing negative emotions where the effects may be indirect rather than direct (Allan et al., Reference Allan, Batz-Barbarich, Sterling and Tay2019). Our results support the strength of work meaningfulness as a direct predictor of lower turnover intentions among knowledge workers, explaining 18.2% of the variance while showing that it is an even stronger predictor of job satisfaction.
In terms of our control variables, a notable and unexpected finding, although not hypothesized, was the increase in turnover intentions among supervisory-level employees (explaining 11.5% of the variance) and the decrease in job satisfaction (5.1% variance). This trend was similarly observed, although to a lesser extent, in employees with longer tenure. Thus, our study revealed that the impact of work meaningfulness on these work outcomes may vary for employees depending on their level of responsibility.
In our group of study participants, issues such as insufficient onboarding, lack of accountability, role ambiguity, and excessive working hours were raised. These challenges may frustrate the basic psychological needs of both supervisory and longer-tenured employees, who face excessive workloads due to inadequate training of new hires, a lack of well-defined roles, and the resulting absence of accountability. This frustration may explain the higher turnover intentions observed among these employees in our study. It may also clarify why our results differ from those of a study using a representative national sample from the Korean Labor and Income Panel Study. After following more than 1,500 people over an 11-year period, researchers found that longer job tenure was associated with lower turnover intention (Ju & Li, Reference Ju and Li2019).
Employees with low control of outcomes at work due to dynamics such as those mentioned above may experience competence, and autonomy need frustration. Combined with low reciprocity and social support at work, their third need, relatedness, may also be frustrated (Gonzalez, Niemiec, & Williams, Reference Gonzalez, Niemiec, Williams and Gagné2015). An important factor to consider is that the tasks of knowledge workers, especially engineers, are based on mutual interdependencies; their work requires specialized input from various individuals, including co-workers and clients (Väänänen et al., Reference Väänänen, Toivanen and Lallukka2020). Therefore, it is crucial to support their need for relatedness, as it is essential for them to do their job effectively.
Looking at the darker side of work meaningfulness, employees who find their work meaningful may end up working long, erratic hours that may affect their work-life balance and undermine their close relationships outside of work (Bailey et al., Reference Bailey, Lips-Wiersma, Madden, Yeoman, Thompson and Chalofsky2019; Oelberger, Reference Oelberger2019). Although research results point to belonging and unity as core elements of meaningfulness (Lips-Wiersma & Morris, Reference Lips-Wiersma and Morris2009; Rosso et al., Reference Rosso, Dekas and Wrzesniewski2010), employees who experience work meaningfulness may feel a sense of belonging to the company. Nevertheless, other aspects of relatedness, such as collaboration, will not flourish without efficient and well-implemented processes, clearly defined roles, and a strategic road map. Good job design can support a firm’s employees and can pave the way for goals such as effective teamwork, reducing work overload, and supporting resource allocation (Cooman, Stynen, Broeck, Sels, & Witte, Reference Cooman, Stynen, Broeck, Sels and Witte2013; Grant, Fried, Parker, & Frese, Reference Grant, Fried, Parker and Frese2010). In order for this to occur, however, the firm must prioritize meeting employees’ needs for autonomy, competence, and relatedness and, based on our results, ensure the inclusion of supervisory-level employees.
Researchers warn against explaining the meaning of work on single antecedents as employees draw meaning from many sources (Rosso et al., Reference Rosso, Dekas and Wrzesniewski2010). The dynamic multilevel framework of factors fostering meaningful work recently put forth by scholars (Lysova et al., Reference Lysova, Allan, Dik, Du and Steger2019) proves that the nature of work is complex and dynamic and that changes in the job and organizational context impact how employees interact with their work environments. Our model, framed within SDT, offers a more integrative approach that supports both the antecedents and strengthens the outcomes of work meaningfulness. Scholars argue that a foundational grounding in basic psychological needs and motivational quality sets the scene for a highly motivated workforce and its ensuing positive outcomes for both employees and the firm.(Rigby & Ryan, Reference Rigby and Ryan2018).
At the same time, autonomous motivation within SDT is strongly related to outcomes such as job satisfaction, well-being, attendance, and lower turnover (Gagné & Deci, Reference Gagné and Deci2005). This motivation is influenced by environmental factors via their effect on perceptions of competence, autonomy, and relatedness (Dysvik et al., Reference Dysvik, Kuvaas and Gagné2013; Grouzet et al., Reference Grouzet, Vallerand, Thill and Provencher2004). In a meta-analysis of 99 organizational workforce studies across several cultures, the satisfaction of the needs for autonomy, competence, and relatedness was associated with lower turnover intentions (Van den Broeck et al., Reference Van den Broeck, Ferris, Chang and Rosen2016). Satisfying these needs paves the way to fully experiencing work meaningfulness, its positive outcomes, and those resulting from autonomous motivation.
Practical implications
The causal-predictive nature of PLS-SEM allows researchers to obtain recommendations for practice (Hair et al., Reference Hair, Hult, Ringle and Sarstedt2022). The results of our study have several implications for reducing turnover intentions in workplaces. First and foremost, work meaningfulness should be an essential element of retention models. Even its greater strength in predicting job satisfaction is necessary as it has been shown to significantly influence knowledge workers’ choices to leave or remain with a firm (Chung et al., Reference Chung, Jung and Sohn2017; Jung, Reference Jung2012; Richer et al., Reference Richer, Blanchard and Vallerand2002).
Our study also demonstrated the importance of job autonomy as a driver of work meaningfulness. However, autonomy has been shown to have an illusory character for knowledge workers, as their work processes are embedded in a number of social and organizational relationships (Väänänen et al., Reference Väänänen, Toivanen and Lallukka2020). Manufacturing companies with knowledge workers should adopt self-managing teams not only to increase the overlap of employees’ knowledge (Foss, Minbaeva, Pedersen, & Reinholt, Reference Foss, Minbaeva, Pedersen and Reinholt2009) but also to encourage teamwork and increase collaboration to satisfy the employees’ relatedness needs. A study in the electric light industry and the telecommunications industry in China and Japan concluded that due to their effects on job satisfaction and turnover intention among knowledge workers, the motivating work characteristics of learning, autonomy, and significance should be considered a vital dimension of job characteristics. The author encouraged further research in different industries (Huang, Reference Huang2011). These characteristics mirror SDT’s needs of competence and autonomy with significance included as a dimension of work meaningfulness.
Engineers have also been found to have a higher turnover rate than other professionals predominantly due to management ineffectiveness and a lack of satisfaction with supervisors (Hofaidhllaoui & Chhinzer, Reference Hofaidhllaoui and Chhinzer2014). Thus, training managers to support their team’s basic psychological needs is also imperative since employees who receive support from their superiors report, on average, higher levels of satisfaction (Blechman, Tóth-Király, Nadon, Fernet, & Morin, Reference Blechman, Tóth-Király, Nadon, Fernet and Morin2022). Managers who work to support their employees’ needs also invest in the organization and its culture by fostering talent retention (Rigby & Ryan, Reference Rigby and Ryan2018). Giving constructive feedback, recognizing employees, acknowledging different points of view, and supervising them in a way that is perceived as interest in the worker rather than reducing autonomy contributes to motivation and positive outcomes (Ariely, Kamenica, & Prelec, Reference Ariely, Kamenica and Prelec2008), as does the frequency and quality of communication and fomenting group cohesion (Blechman et al., Reference Blechman, Tóth-Király, Nadon, Fernet and Morin2022).
Facilitating employee access to knowledge within and outside the organization is especially crucial for knowledge workers who constantly need to improve their skills. In a study of engineers across six firms in the United States, researchers found that the aspects of a job most valued by engineers were solving a problem, learning, and working in a team (Anderson, Courter, Mcglamery, Nathans-Kelly, & Nicometo, Reference Anderson, Courter, Mcglamery, Nathans-Kelly and Nicometo2010).
Framing work meaningfulness within SDT could be an important step forward by ensuring high-quality motivation, which translates to employee retention and improves job satisfaction. In a three-wave sample of working adults, the results of an SDT-based conceptual model in which chronological age served as the moderator suggested that competence and autonomy-supportive HR practices were the most beneficial for employees regardless of their career stage (Laguerre & Barnes-Farrell, Reference Laguerre and Barnes-Farrell2024).
Limitations and directions for future research
Despite its implications, this study has several limitations that should be addressed and that call for future investigations. First, although the research model explains more than half of the variance of work meaningfulness on job satisfaction and more than 40% of work meaningfulness on turnover intention, the influence of individual-level endogenous factors such as comparable alternative employment opportunities should be considered. Second, due to the nature of the industry, our sample was male-dominant. It would be interesting to replicate the study in a more gender equal setting. Third, while the sample involved only one division of one firm, it included a number of occupations and thus was representative of the company. Considering the increasingly warranted focus on work meaningfulness and its impact on turnover intention and job satisfaction, these results may be salient for knowledge workers in other firms and industries and we thus encourage scholars to test this model in different environments.
In terms of future research directions, there are some exciting possibilities. First, examining the three higher-order categories of autonomous work recently put forth by Parker (Parker & Knight, Reference Parker and Knight2024), namely, timing autonomy, method autonomy, and decision-making autonomy, it would be useful to study the impact of each on work meaningfulness. Second, expanding the model to include measures for the satisfaction of competence and relatedness needs and comparing the results in terms of age, length of service, and supervisor status would be helpful, as the strength of particular needs may change over time.
Third, recent research points to the dynamic nature of turnover intentions (Jeong & Lee, Reference Jeong and Lee2023). Although our study was conducted at one point in time to gain current insight specific to the high turnover context of this particular company, conducting longitudinal studies that include additional organizational initiatives (apart from job autonomy) could shed more light on how to best support employee needs and sustain work meaningfulness to ensure employee well-being and positive work outcomes. Fourth, as leadership support is vital to ensuring the satisfaction of employees’ autonomy, competence, and relatedness needs, further research could consider different types of leadership as moderators to compare their effects on the relationship between work meaningfulness and turnover intentions. Fifth, further exploring the role of work meaningfulness within SDT for supervisory-level employees would be a valuable line of research, particularly given their critical and hard-to-replace roles and our results, which pointed to adverse outcomes for this particular group. Sixth, replicating this study in other countries would be beneficial in adding to our understanding of work meaningfulness across diverse cultural contexts. Finally, future research could adopt a qualitative approach to ensure that employees’ voices are heard. This would enable researchers to explore in greater depth the underlying reasons for negative outcomes, such as increased turnover intention among supervisors and longer-tenured employees, and to develop strategies to prevent the loss of such valuable employees.
Conclusions
While previous studies have identified autonomy as a strong predictor of turnover intentions, our findings suggest that work meaningfulness, driven by job autonomy, predicts both lower turnover intentions and higher job satisfaction. Our results indicate that work meaningfulness, while a stronger predictor of job satisfaction than turnover intentions, should still be considered a crucial element in understanding and addressing turnover dynamics.
From an organizational perspective, the results of the present study highlight the importance of the organization’s effort in catering to the needs of its employees. The strong relationships between job autonomy and work meaningfulness, between work meaningfulness and both turnover intentions and job satisfaction, underscore the crucial role of work meaningfulness. In work that is inherently meaningful both due to its complexity and its contribution to society (i.e., clean energy), fomenting and sustaining work meaningfulness remains crucial to retaining employees and enhancing their well-being. To achieve this, a firm’s leadership and culture must prioritize supporting the three psychological needs of employees at all levels. The results of our study should apply not only to the manufacturing sector but also to other industries where knowledge workers are employed.
Acknowledgement
Funding for open access publishing: Universidad Pablo de Olavide.
Data availability statement
The data that support the findings of this study are available from the corresponding author, J.W. upon reasonable request.
Conflict of interest
The authors report that there are no competing interests to declare.
Joanna Wandycz-Mejias is the founder and director of the International Center at Pablo de Olavide University. In possession of two master’s degrees and with 20+ years of experience managing people both in Spain and in the US, she is currently in the final stage of a doctoral degree program in Business Administration at Pablo de Olavide University. Her main interest is organizational behavior and self-determination theory. She is a passionate advocate for catering to the needs of employees to ensure employee motivation and well-being.
José L. Roldán is a full professor of Management in the Department of Business Administration and Marketing at the Universidad de Sevilla (Spain). He received his doctorate from the Universidad de Sevilla in Business Administration, with a specialization in information systems. He has been a visiting professor at different European and Latin American Universities. His current research interests include technology acceptance models, business intelligence, knowledge management, social network sites, and partial least squares (PLS). His recent contributions have been published in the British Journal of Management, Journal of Business Economics and Management, International Business Review, European Journal of Information Systems, International Small Business Journal, Computers in Human Behavior, Computer Standards & Interfaces, Industrial Marketing Management, and Internet Research, among others.
Alvaro Lopez-Cabrales is a full professor of Human Resource Management in the Business Administration Department, Pablo de Olavide University (Seville), where he obtained his doctorate degree in 2003. He has been a Visiting Scholar at Texas A&M University in the United States. His current work focuses on research in human capital, employment relationships, and sustainability in HRM. He has published his research in international journals such as Human Resource Management, R&D Management, Personnel Review, the British Journal of Management, The International Journal of Human Resource Management, Employee Relations, Human Resource Management Review, and Corporate Social Responsibility & Environmental Management, among others.