The theoretical framework of this study encompasses several constructs that, although related, have distinct conceptual natures. Retention refers to an organisation’s ability to maintain its workforce and prevent voluntary turnover. Talent retention, in contrast, refers specifically to the retention of employees who possess critical skills, high performance or strategic value for the organisation. Employee retention intention, the outcome variable in this study, refers to employees’ self-reported intention to remain with their organisation. Well-being is a broad construct that includes emotional, psychological and social dimensions (Ryan et al., 2019), while workplace well-being reflects these dimensions in professional settings. Work engagement is a motivational state characterised by vigour, dedication and absorption. Flourishing denotes an optimal level of psychological functioning associated with personal growth, purpose and positive relationships. These constructs are interconnected within the proposed theoretical model, as effective socialisation experiences can foster positive psychological states that, in turn, influence attitudes and behaviours related to their intention to remain in the organisation.

For the purpose of this study, employee retention is operationalised as employees’ intention to remain with their organisation, a widely used behavioural indicator in organisational research. While the literature distinguishes between general employee retention and talent retention, the present study focuses on employees’ intention to remain, thereby ensuring conceptual alignment between the theoretical framing and the measurement instrument.

The study employed a cross-sectional design and a quantitative methodological approach. Data were gathered through an online questionnaire distributed via professional networks and email, following a snowball sampling strategy. Participants were fully informed about the purpose of the study, the confidentiality procedures and the anonymous handling of their responses, and they provided informed consent prior to taking part. Although this approach allowed us to access a diverse group of employees, the reliance on convenience and snowball sampling limits the representativeness of the findings and restricts their generalisability across industries and cultural contexts (Podsakoff et al., 2012).

A total of 401 employees participated in the study, 57.9% of whom were women. Participants’ ages ranged from 21 years to 65 years (M = 38.18; standard deviation [SD] = 11.32). The sample comprised individuals occupying different organisational roles, generational cohorts and tenure levels (Table 1). Despite this diversity, the sample was geographically concentrated and did not account for contextual factors such as organisational culture, remuneration or leadership style, all of which may influence retention-related outcomes. In this study, retention outcomes refer to employees’ general intention to remain, not specifically to talent retention outcomes. These constraints should be taken into consideration when interpreting the findings.

Data analysis was conducted using SPSS and AMOS. Prior to testing the hypotheses, multivariate normality was assessed according to Mardia’s (1980) criterion, and acceptable levels of skewness and kurtosis were observed. To evaluate the potential presence of common method bias, Harman’s single-factor test was performed, indicating that a single factor accounted for 49% of the variance, which is below the 50% threshold (Fuller et al., 2016). Nevertheless, given the well-documented limitations of this procedure (Podsakoff et al., 2003), common method bias cannot be fully ruled out. Future research should adopt more rigorous techniques, such as confirmatory factor analysis-based common latent factor models.

Reliability and validity were examined using Cronbach’s alpha, composite reliability (CR) and average variance extracted (AVE). Discriminant validity was assessed by comparing AVE with the maximum shared variance (MSV). While correlations between flourishing, workplace well-being and work engagement were high, AVE values exceeded MSV, supporting their empirical distinctiveness. Nevertheless, their conceptual overlap should be recognised as a limitation.

An exploratory analysis was also conducted to examine whether employee retention intention differed across sociodemographic and professional groups using independent-samples t-tests and one-way ANOVAs.

Hypotheses were tested using Hayes’s PROCESS macro (Model 6; Hayes, 2018) with 5.000 bootstrap samples to estimate indirect effects and their 95% confidence intervals. This method is considered a best practice for mediation testing, providing more robust evidence than reliance on Sobel tests alone. Given the cross-sectional design, however, causal inferences should be drawn with caution, and longitudinal studies are recommended to confirm the temporal sequence of variables. Consistent with Cohen’s (2013) guidelines, the magnitude of explained variance was interpreted using the conventional thresholds for R² (0.02 small, 0.13 medium, 0.26 large).

Ethical clearance to conduct this study was obtained from the Instituto Superior de Gestão Ethics Committee (Ref. No. CIG_0010.14/2025).

Prior to testing the hypotheses, the data were examined to assess normality and potential common method bias (Ventura-León et al., 2023). Mardia’s coefficient indicated acceptable levels of skewness and kurtosis (Mardia, 1980; Mardia et al., 2024). To assess common method bias, Harman’s single-factor test was performed, indicating that a single factor explained 49% of the variance, below the 50% threshold. This result suggests that common method bias is unlikely to pose a serious concern; however, the limitations of this diagnostic method mean that bias cannot be entirely ruled out.

Reliability and validity analyses confirmed adequate psychometric properties. Cronbach’s alpha values ranged from 0.72 to 0.93. Composite reliability and AVE exceeded the recommended cut-off values (CR > 0.70; AVE > 0.50). Maximum shared variance was lower than AVE, supporting discriminant validity, even in the presence of high correlations among constructs. Table 2 presents descriptive statistics, reliability indices and bivariate correlations among the study variables. The hypothesis-testing results presented below are based on regression-based analyses conducted with PROCESS, rather than on these correlations. This distinction clarifies that the correlational matrix serves only descriptive purposes, while the inferential findings stem from the regression and mediation models.

A confirmatory factor analysis was conducted to determine whether the observed variables adequately represented the underlying latent constructs (Widaman & Helm, 2023). The fit indices indicated that the model demonstrated an acceptable fit to the sample data, based on standard cut-off criteria (comparitive fit index [CFI] and Tucker-Lewis Index [TLI] ≥ 0.90; root mean square error of approximation [RMSEA] ≤ 0.08). The RMSEA confidence interval (LO90–HI90) reflects the lower and upper bounds of the 90% confidence interval for the RMSEA estimate, indicating the precision of the fit. Fit indices were as follows: χ² (212) = 2.37, p < 0.001, Goodness-of-Fit Index (GFI) = 0.91, CFI = 0.96, TLI = 0.95, RMSEA = 0.05 (LO90 = 0.05, HI90 = 0.06). The hypotheses derived from the literature review were subsequently tested.

Regression analyses (Table 3) showed that flourishing positively predicted employee retention intention (β = 0.600, p < 0.001), supporting H1. Flourishing also significantly predicted workplace well-being (β = 0.822, p < 0.001) and work engagement (β = 0.789, p < 0.001), supporting H2a and H2b. Both workplace well-being (β = 0.592, p < 0.001) and work engagement (β = 0.641, p < 0.001) positively predicted retention, supporting H3a and H3b.

Sequential mediation analysis using Hayes’s PROCESS macro (Model 6; Hayes, 2018) with 5000 bootstrap samples and 95% confidence intervals revealed significant indirect effects through workplace well-being (B = 0.242, 95% CI [0.109, 0.375]) and work engagement (B = 0.280, 95% CI [0.171, 0.390]). The sequential indirect pathway from flourishing to employee retention intention via workplace well-being and work engagement was also significant (B = 0.085, 95% CI [0.038, 0.145]), supporting H4.

When comparing the indirect pathways, the effect of flourishing on employee retention intention through workplace well-being alone (B = 0.254) emerged as stronger than the sequential mediation effect (B = 0.085). The indirect effect through work engagement alone (B = 0.222) was also stronger than the sequential pathway (Table 4). These findings indicate that while the sequential model is statistically valid, simple mediation mechanisms account for a larger proportion of the indirect effect. This nuance refines our understanding of how flourishing translates into employees’ intention to remain and suggests that workplace well-being, in particular, plays a dominant mediating role.

The model accounted for 36% of the explained variance in employee retention intention. While this represents a meaningful proportion of the explained variance, it also indicates that a substantial part of retention behaviour is driven by other factors not included in the model, such as leadership, compensation or organisational culture. This should be considered when interpreting the results. According to Cohen’s (2013) effect size guidelines (R² = 0.02 small, R² = 0.13 medium, R² = 0.26 large), the explained variance observed in this study (R² = 0.36) falls within the range typically interpreted as a large effect.

In addition to testing the hypothesised model, we conducted an exploratory analysis to examine whether employee retention intention differed across sociodemographic and professional groups. Prior research has shown that retention outcomes are influenced not only by personal psychological resources but also by demographic and organisational factors such as age, tenure and job role (Ganguli & Padhy, 2023; Riekhoff & Järnefelt, 2022; Snyman & Ferreira, 2023). To address this, mean-comparison tests were performed across generational cohorts, educational qualifications, job roles and organisational tenure. Gender and age were excluded from this analysis, as they had already been examined as control variables in the main model.

All constructs were assessed using a 5-point Likert scale. Higher values correspond to greater flourishing, workplace well-being, work engagement or intention to retain. The results of the mean-comparison procedures are reported in Table 5.

The findings revealed that employee retention intention was significantly higher among Baby Boomer employees (M = 4.20, SD = 0.45) compared with Generation X (M = 3.95, SD = 0.48) and Millennials (M = 3.78, SD = 0.56), F_{(3, 397)} = 12.24, p < 0.001. Employees in managerial positions (M = 4.12, SD = 0.52) also reported higher retention than non-managerial employees (M = 3.89, SD = 0.50), t₍₃₉₉₎ = 5.03, p < 0.001. Likewise, employees with 10 or more years of organisational tenure (M = 4.18, SD = 0.49) reported stronger retention intentions than those with fewer than 10 years (M = 3.87, SD = 0.54), t₍₃₉₉₎ = 6.11, p < 0.001. No significant differences were found regarding educational level, F_{(3, 397)} = 1.42, p = 0.236.

While these analyses are exploratory and do not permit causal inference (Hayes, 2018), they highlight the importance of considering both personal and organisational factors when designing retention strategies. Specifically, younger cohorts and early-career employees may require targeted interventions to strengthen their intention to stay.

Given the results, it can be concluded that participants from the Baby Boomer Generation, those in managerial positions and those with 10 or more years of tenure exhibit the strongest willingness to stay within the organisation.

The present study offers several implications for HRM. Firstly, the central role of workplace well-being suggests that organisations should prioritise initiatives that foster a healthy and supportive environment. Examples include policies to promote work–life balance, programmes to reduce stress and burnout and practices to enhance psychological safety (Gallup, 2019; Murthy, 2022; Sirgy & Lee, 2018). Because workplace well-being emerged as the strongest mediating mechanism, these initiatives may represent an especially effective pathway for strengthening employees’ intention to remain in the organisation.

Secondly, the role of work engagement as both an independent mediator and part of the sequential pathway highlights the importance of job design and motivational strategies. Human Resource practitioners should create opportunities for employees to experience vigour, dedication and absorption, for example, by granting autonomy, recognising achievements and providing meaningful tasks (Gomes & Marques, 2025; Saks, 2021).

The identified demographic differences indicate that retention policies must be tailored to employee profiles. Younger generations may benefit more from career development programmes and mentoring, while senior employees may value stability and recognition of their experience. Managers themselves, who displayed particularly strong retention intentions, should be supported as key actors in cascading workplace well-being and work engagement throughout their teams.

Overall, these findings emphasise that investing in flourishing, workplace well-being and work engagement is not only an ethical responsibility but also a strategic necessity. This is particularly relevant in competitive and resource-constrained labour markets, including emerging economies, where retaining skilled employees is a pressing organisational challenge.

The present study provides relevant contributions to research on positive psychology and HRM. From a theoretical perspective, it extends previous research by integrating flourishing, workplace well-being and work engagement into a single sequential mediation framework. While flourishing has been studied in relation to each of these constructs individually (Imran et al., 2020; Redelinghuys & Rothmann, 2020), fewer studies have considered their combined and sequential effects on retention. By testing this process model, the study refines current knowledge of how personal psychological resources translate into organisational outcomes.

In addition, the findings highlight the relative strength of alternative mediating mechanisms. The indirect effect through workplace well-being alone was stronger than the sequential pathway, suggesting that workplace well-being is a more proximal predictor of retention, whereas work engagement provides incremental explanatory value by capturing behavioural manifestations of flourishing. This distinction advances theoretical clarity by explaining the respective roles of affective and behavioural processes within employee retention. Moreover, the present research contributes to the ongoing discussion of discriminant validity, showing that flourishing, workplace well-being and work engagement, despite being correlated, remain conceptually and empirically distinct.

From an applied perspective, the findings highlight the need to establish organisational environments that promote employees’ workplace well-being. Given that workplace well-being emerged as the strongest mediator of the flourishing–retention link, organisations should prioritise initiatives such as fostering healthy working conditions, providing psychological support and promoting work–life balance. These practices can strengthen employees’ satisfaction and attachment, reducing turnover intentions (Murthy, 2022; Sirgy & Lee, 2018). Work engagement has also been shown to be an important mechanism, suggesting that managers should invest in practices that enhance vigour, dedication and absorption, for example, through recognition, autonomy and meaningful job design (Bakker & Demerouti, 2017; Saks, 2021).

Finally, the complementary analysis indicated that retention intentions vary by generation, job role and organisational tenure. This implies that HR strategies should not be uniform but instead be adapted to different employee profiles. For instance, targeted initiatives may include structured mentoring for early-career employees, personalised development opportunities or flexible work arrangements tailored to employees’ life stages.

Overall, the findings provide both theoretical refinement and practical insights, contributing to a deeper understanding of the psychological and demographic determinants shaping employee retention intention.

Although the present investigation offers relevant contributions, some constraints must be acknowledged. Data were obtained through self-report questionnaires, which may introduce common method bias. Harman’s single-factor test suggested that this was not a major issue; however, this diagnostic has recognised limitations. More rigorous techniques, including confirmatory factor analysis with a common latent factor or the use of multi-source data, would strengthen future research (Podsakoff et al., 2012)

The sampling strategy also imposes restrictions. The use of a convenience and snowball sample of Portuguese employees’ limits representativeness and narrows the applicability of the findings to other sectors and cultural settings (Redelinghuys & Rothmann, 2020). Broadening data collection to probability samples and incorporating more diverse international contexts would enhance external validity.

Another methodological issue concerns the measurement of retention. The employee retention questionnaire (Salman et al., 2014) had one item removed to improve reliability, thereby increasing internal consistency but reducing content validity. Future research should adopt alternative validated scales to capture the construct more comprehensively.

Furthermore, the constructs of flourishing, workplace well-being and work engagement were strongly correlated. Although AVE and MSV analyses supported their distinctiveness, the overlap between them raises concerns about discriminant validity. Exploring alternative modelling strategies, such as bifactor or second-order models, could provide stronger evidence of empirical differentiation.

Moreover, the study examines general retention outcomes rather than talent-specific retention processes. Future research should differentiate between these constructs, as the drivers of general employee retention may differ from those influencing the retention of high-performing or critical talent.

The cross-sectional nature of the research precludes firm conclusions about causality (Hayes, 2018). Although the mediation patterns align with theoretical assumptions, the temporal sequencing of the variables cannot be definitively determined. Longitudinal and experimental designs would allow a more rigorous test of causality.

Finally, contextual variables such as organisational culture, leadership style and compensation practices were not considered, even though they are likely to influence retention. Including such factors in future models would contribute to a broader and more integrated understanding of the conditions under which flourishing fosters employees’ intention to stay. By overcoming these limitations, future investigations may further extend the current work’s contribution and more precisely clarify the interplay among personal resources, workplace experiences and retention outcomes.