Uncovering Gender and Temporal Dynamics: Career Resources Impacting Career Success

Abstract

(1) Background/Purpose: Talent management research has typically focused on early-career entrants or high-potential employees, leaving mid-career professionals underexplored despite their pivotal role in organisational continuity and leadership pipelines. This study examines whether the principles of Conservation of Resources (COR) theory apply to careers, testing whether career resources predict objective and subjective career success, and whether gender differences emerge. (2) Study Design/Methodology/Approach: A three-wave survey of 543 individuals employed in the United Kingdom (UK) (mean age 39) was analysed using Latent Growth Modelling and hierarchical regression to capture the temporal dynamics of career resources and their links to success. (3) Findings: Subjective career success declined overall, but increased among participants with high human capital, environmental resources, career self-management behaviours, and baseline motivation. Gender differences were found: human capital and self-management were stronger predictors for men, while environmental resources were more important for women. Objective success was predicted by human capital only for women, while private-sector employment and subjective success were the strongest predictors for men. (4) Originality/Value: Our unique contribution advances understanding of mid-career dynamics among women and men, highlighting critical implications for talent management. Some, but not all, predictions of COR theory are supported. Women and men experience the benefits of resources differently. Whilst career resources were critical for career success, caring responsibilities were not, irrespective of gender. Organisations must recognise that subjective career success needs resources to sustain it and move beyond one-size-fits-all approaches by tailoring development, mobility, and support systems to gendered and career-stage-specific needs.

1. Introduction

A resource-based view sits at the core of contemporary talent management (TM) and career research. Conservation of Resources (COR) theory posits that people strive to obtain, retain, and protect valued resources—objects, conditions, personal characteristics, and energies (Hobfoll, 1989, 2001, 2011; Hobfoll et al., 2018).

At the organisational level, TM stakeholders benefit from leveraging such key resources to motivate and develop their staff. The Career Resources Framework (CRF) specifies four actionable resource domains—human capital, environmental supports, motivational resources, and career self-management (CSM)—that map cleanly onto HR levers for development, mobility, and retention (Hirschi, 2012; Hirschi et al., 2018).

Taken together, COR and the CRF provide an empirically grounded, practice-relevant lens for TM: if HR knows which resources matter, when they change, and for whom they are most impactful, organisations can design interventions that prevent loss spirals, catalyse gain spirals, and build sustainable internal pipelines.

This resource-centric lens is urgently needed. TM is increasingly intermediated by vendors: talent acquisition has become the largest outsourced HR function by revenue (Cappelli & Schwartz, 2024), and 77% of vacancies were filled externally in 2021 (Society for Human Resource Management, 2022). At the same time, employer investment in training and development has waned since COVID-19, straining internal talent pipelines, particularly impacting women who were more likely to be laid off and furloughed for longer than men during the pandemic (Dang & Nguyen, 2021; Jones & Cook, 2021). Conceptually, “talent” remains contested, varies by stakeholder, and is too often restricted to elite early-career pools (McDonnell et al., 2023; Vardi & Collings, 2023). These trends pose a risk of disadvantaging some groups of employees in achieving personal and organisational success.

Career success itself is inherently temporal and dual-faceted: it represents desirable work-related outcomes achieved across an evolving sequence of experiences (Arthur et al., 1989, 2005). Achieving externally tangible and usually organisation-centric metrics (salary, promotions, hierarchical status) is considered objective career success (OCS). Additionally, individuals’ personal evaluation of their accomplishment of work-related outcomes that they personally desire is termed subjective career success (SCS) (Stebbins, 1970; Van Maanen & Schein, 1977). Contrary to a common narrative that “traditional” OCS has faded in relevance, both OCS and SCS continue to matter and to co-evolve—often with SCS leading OCS over time (Abele & Spurk, 2009a; Spurk et al., 2019). Therefore, integrating both facets of success is essential in both empirical research and practical interventions.

Crucially, gendered processes shape both the distribution of resources and the extent to which those resources translate into success. The Unequal Attribute/Unequal Effect framework clarifies that gender can (a) precede resources (women and men differ in resource levels; Unequal Attribute) and/or (b) moderate resource-to-outcome relationships (identical resources convert into different levels of success by gender; Unequal Effect) (Frear et al., 2019). Empirical examples abound: women frequently report needing to invest more to achieve comparable outcomes (Santos et al., 2024), networking pays off differentially by gender (Wanigasekara et al., 2023), and equivalent work experience yields smaller salary gains for women (Evers & Sieverding, 2014). Social norms (e.g., the ideal-worker model), caregiving expectations, and informal gatekeeping (e.g., “old boys’ networks”) further constrain women’s access to resources, amplifying COR-consistent loss spirals and dampening gain spirals (Crompton & Sanderson, 2024; Williams, 2001; Yates & Skinner, 2021). Against this backdrop, persistent gender pay gaps across sectors (Borrett & Strauss, 2025; Office for National Statistics, 2024) underscore why OCS remains an equity-critical outcome even as SCS expands in prominence.

Despite this accumulated knowledge, TM research remains largely static: cross-sectional designs dominate, gender is too often omitted from analysis, and mid-career talent—pivotal for continuity and leadership pipelines—receives far less attention than early-career “high potentials” (Al Ariss et al., 2014; Festing & Schäfer, 2014; Marlapudi & Lenka, 2024). This neglect is consequential. COR principles are time-sensitive: primacy of loss predicts short-horizon declines when resource gains are absent. However, resource investment and gain paradox imply that well-timed HR interventions (e.g., targeted learning, structured internal mobility, sponsorship, job challenge, organisational career support) can prevent loss spirals and initiate gains, especially for under-resourced groups (Blokker et al., 2019). But identifying the right time windows and which levers to pull requires time-lagged evidence on mid-career workers, not just early-career entrants.

The present study addresses prior research limitations directly. Using a three-wave, full-panel design with 30-day lags, it tests whether CRF resources predict SCS and OCS over short horizons consistent with COR mechanisms, and whether effects differ by gender among mid-career adults. Specifically, it examines (H1) whether SCS declines in the absence of resource gains (primacy of loss), (H2) whether loss spirals emerge when initial SCS is low and resources are scarce, (H3) whether resource investment predicts higher SCS over time, (H4) whether gain spirals manifest among high-resource individuals, and (H5) whether human capital most strongly predicts contemporaneous OCS. In line with the Unequal Attribute and Effect (RQ1 and RQ2) propositions, the study also compares resource levels and resource-success linkages across women and men.

This study makes a distinctive contribution to career and talent management research by integrating career-stage, temporal, and gendered perspectives within a single empirical design. Whereas prior research on career resources and success has been dominated by cross-sectional designs, early-career samples, and static comparisons between women and men, the present study focuses explicitly on mid-career professionals and examines how career resources unfold and convert into success over short-time horizons consistent with Conservation of Resources theory. Using a three-wave full-panel design and latent growth modelling, the study provides one of the first direct tests of primacy of loss, loss spirals, and gain spirals in careers, rather than assuming these processes implicitly. In addition, by applying the Unequal Attribute and Unequal Effect framework, the study moves beyond documenting gender differences in resource levels to demonstrate how identical resources yield different objective and subjective career outcomes for women and men. In performing so, it advances a dynamic, resource-based understanding of sustainable talent management that is sensitive to both career stage and gendered conversion mechanisms.

By embedding COR and the CRF at the front of the contribution, the study offers clear pathways for sustainable TM. For HR leaders, the findings translate into actionable levers. For example, practical applications include diagnosing human capital gaps associated with pay and promotion inequities; strengthening environmental resources via sponsorship, transparent opportunity structures, and job challenges; building motivational resources (confidence, clarity, involvement) through coaching; and equipping employees with CSM behaviours (networking, exploration, learning). Strategically, the study shows when these levers are most effective (short-cycle resource windows) and for whom (gender-sensitive, mid-career focus). Conceptually, it responds to the Administrative Sciences Special Issue—Rethinking Talent Management for Sustainable Organisations—by shifting TM from static, elite identification to dynamic resource architectures that support equitable and sustainable career progressions (Afiouni, 2019; De Vos et al., 2020; McDonnell et al., 2023; Vardi & Collings, 2023).

2. Theoretical Framework

This paper adopts a resource lens to theorise how career success develops, fluctuates, and becomes unequally distributed across gender, particularly among mid-career employees. At its core is the COR theory, which posits that individuals strive to obtain, retain, foster, and protect resources (Hobfoll, 1989). Resources are defined as objects or personal characteristics that are centrally valued by individuals, as they increase motivation and reduce stress (Hobfoll, 1989, 2001, 2011; Hobfoll et al., 2018). These resources cluster and travel in caravans within passageways (socio-organisational ecologies that foster or block resource creation) (Hobfoll, 1989; Hobfoll et al., 1998, 2018). This means that an organisation that places sustainable, resource-rich TM at its forefront for all workers maintains positive career success and wellbeing outcomes for all.

COR theory’s principles are inherently temporal: losses tend to be larger and faster than comparable gains (primacy of loss principle). Individuals invest resources to gain favourable outcomes, buffer the effect of demands on wellbeing, and recover from stress (resource investment principle). Therefore, individuals with a surplus of resources show better outcomes. However, gains are more likely to occur for individuals who already experience favourable conditions—a phenomenon termed the gain paradox principle—which creates resilience among some individuals and vulnerability among others.

Therefore, the COR theory explicitly suggests that individuals with more resources, due to a more privileged position in society, experience better outcomes than their less privileged counterparts (Holmgreen et al., 2017). This resilience may occur due to belonging to a historically advantaged group, such as men. Indeed, in a demanding environment, such as company restructuring, women self-reported higher levels of stress than their male counterparts (Karambayya, 1998) and surviving a layoff was significantly associated with subsequent major depression in women, but not in men (Andreeva et al., 2015). Furthermore, women are more likely to experience a demanding environment at work than men, such as the detrimental effects of a lack of organisational support for pregnancy (Hackney et al., 2025; Little et al., 2018), menstruation (Howe et al., 2023) and menopause (Atkinson et al., 2021; Alemu et al., 2025). Not only is resource loss resulting in exacerbated losses of psychological resources for women compared to men, but company mergers and restructuring are also more likely to result in higher layoffs of women than men (Collins, 2005), further depriving them of tangible resources, in this case, employment. This trend was also observed during the COVID-19 pandemic, when women were more likely to lose their jobs than men (Dang & Nguyen, 2021). Therefore, COR theory predicts that women are more vulnerable to resource loss due to their less privileged position in society compared to their male counterparts, a prediction supported by a plethora of cross-sectional research in the work domain.

These COR principles, with a gendered lens, ought to be examined in a time-lagged design to make them actionable for TM. We embed the COR theory within the CRF (Hirschi, 2012; Hirschi et al., 2018) to make it operationalisable for research and organisational purposes. The CRF specifies four domains that can be measured and developed: human capital (occupational expertise, job-market knowledge, soft skills), environmental resources (organisational career support, career opportunities, job challenge, social career support), motivational resources (career involvement, confidence, clarity), and CSM behaviours (networking, exploration, learning). The CRF thereby operationalises COR for careers, with a validated questionnaire that demonstrates measurement invariance across gender and predictive validity for career success over time (Haenggli et al., 2021; Haenggli & Hirschi, 2020; Hirschi et al., 2018).

Placing resources upfront also aligns with how organisations conduct TM. If HR can diagnose which CRF resources are lacking, when they change, and for whom they convert into outcomes, HR can target development (e.g., learning for human capital; coaching for clarity/confidence), mobility (transparent internal opportunities, job challenge), and sponsorship/social support (environmental passageways) to interrupt loss spirals and ignite gain spirals—the essence of sustainable TM.

2.1. Time as Mechanism

Time is integral to COR’s propositions, yet longitudinal career studies remain scarce (Akkermans et al., 2024; Spector & Meier, 2014). Evidence indicates that shorter lags (weeks) are sensitive to COR mechanisms: effects emerge at one month and six weeks, whereas longer intervals invite adaptation back to baseline (Haenggli et al., 2021; Diener et al., 2009; Matthews et al., 2014; Ritter et al., 2016). Moreover, testing spiral logic requires three or more waves; two-wave designs cannot model change trajectories or intercept-slope covariance (T. D. Allen et al., 2019; Curran et al., 2010). Latent Growth Modelling (LGM) is therefore appropriate as it estimates initial levels (intercepts), rates of change (slopes), and their covariance, permitting direct assessment of loss and gain spirals (Lance et al., 2000). Accordingly, this study employs three waves at four-week intervals—targeting the “sweet spot” for observing primacy of loss and early gain effects before adaptation dampens them.

2.2. COR Principles, CRF Resources, and Hypotheses

2.2.1. Primacy of Loss (H1)

Resource loss is more salient, immediate, and potent than resource gain (Hobfoll, 2001, 2011). In work contexts, depleted job resources prospectively predict burnout, and negative shocks disrupt employability linkages more than positive shocks enhance them (Blokker et al., 2019; Lesener et al., 2019). In careers, SCS can decline over multi-year horizons without compensating gains (Spurk et al., 2011). In a short-lag design, if career resources are not replenished, SCS should show a downward drift, which is the core implication of the primacy of loss principle. If human capital, environmental, motivational, and CSM resources are not increasing, SCS (a felt evaluation of one’s career trajectory) should decrease as minor setbacks accumulate and are overweighed. Consequently, individuals perceive and weigh losses more heavily than equivalent gains. SCS should decline when resource gains do not occur.

H1. 

Over time, SCS will show a decrease in the absence of career resources.

2.2.2. Loss Spirals (H2)

When resources are low, people have fewer means to counter losses. Each iteration reduces capacity further, and loss spirals (Hobfoll & Shirom, 2000). Two-wave evidence shows vicious cycles between job insecurity and exhaustion (De Cuyper et al., 2012) and three-wave evidence shows reciprocal amplification among work pressure, work–home interference, and strain (Demerouti et al., 2004). LGM allows a direct test: a lower initial SCS (intercept) should be associated with a more negative slope. Inadequate environmental supports (opaque opportunity structures, weak sponsorship), thin human capital, and low motivational resources (confidence/clarity) create risk passageways that foster loss cycles; low CSM behaviours limit access to countervailing opportunities.

H2. 

Over the two months of the study, SCS will show a loss spiral in the absence of career resources.

2.2.3. Resource Investment (H3)

Consistent with COR’s investment principle, individuals use existing resources to build new ones and recover from loss (Hobfoll et al., 1998). Job and personal resources predict engagement and wellbeing across time; resources buffer the effects of shocks and demands (Ali & Mehreen, 2022; Bakker & Demerouti, 2018; Barthauer et al., 2020; Hofer et al., 2021; Lesener et al., 2019). Together, the four resource domains in the CRF (Hirschi et al., 2018) reflect both the capacity and the act of investing in one’s career. Over weeks, such investments should lift SCS as individuals see progress, recognition, meaning, and control.

H3. 

Over the two months of the study, career resources will positively predict higher SCS.

2.2.4. Gain Spirals (H4)

Resource gains matter most during loss, yet those with fewer resources are least able to generate gains, thereby creating vulnerability (Hobfoll et al., 2018). In previous studies, gain spirals were supported for CSM together with SCS at one-month time lags (Haenggli et al., 2021). There is also evidence for all resource domains to lead to either OCS or SCS or both, albeit not in a full-panel design (Haenggli & Hirschi, 2020). High baselines in human capital, environmental support, motivational resources, and CSM behaviours should be associated with steeper positive SCS slopes, thereby creating a gain spiral (Hobfoll et al., 2018). Conversely, individuals starting with few resources may fail to initiate gain spirals, even when highly motivated to improve their situation.

H4. 

Over the two months of the study, SCS will show a gain spiral in the presence of high career resources.

2.2.5. Human Capital as a Key Resource for OCS (H5)

Resources cluster (i.e., correlate positively) and exist within passageways that foster or impede resource formation (Hobfoll, 2011). Of all CRF domains, human capital should be most tightly coupled to OCS (salary, promotion, level), conceptually and empirically (Haenggli & Hirschi, 2020; Ng et al., 2005; Ng & Feldman, 2014). Since occupational expertise, soft skills, and market knowledge are evaluated in pay/advancement decisions, human capital should show the strongest contemporaneous link with OCS, especially cross-sectionally in a short panel design.

H5. 

Individuals with high human capital career resources will show higher OCS cross-sectionally.

In sum, these five hypotheses collectively test how resource availability and change shape career success trajectories over short timeframes. However, the extent and strength of these effects are likely to differ by gender.

2.3. Gendered Mechanisms (RQ1, RQ2)

A central claim of this study is that gender shapes both resource levels and resource payoffs, especially in mid-career. Gender shapes both the levels of career resources and the degree to which those resources translate into success. The Unequal Attribute/Unequal Effect framework (Frear et al., 2019) captures two non-mutually exclusive pathways. Firstly, Unequal Attribute (RQ1) posits that women and men differ in resource levels (e.g., sponsorship access, confidence), reflecting historical socialisation, caregiving norms, and exclusion from informal networks (Abele & Spurk, 2009b; Pitan & Muller, 2020; Yates & Skinner, 2021). Secondly, Unequal Effect (RQ2) contends that gender moderates the relationship between resources and career success. Even at equal resource levels, achieving career success differs by gender. For example, male CEOs earn more if they are board members of other organisations, whereas their female counterparts are penalised for it (Malhotra et al., 2021; Shropshire, 2010). These pathways are COR theory-consistent. Gendered risk passageways (e.g., ideal-worker norms, caregiving burdens, biassed evaluation scripts) deplete environmental and motivational resources, making loss spirals more likely and gain spirals harder to ignite for women. Conversely, resource-rich passageways (sponsorship, challenge with support, equitable evaluation) enable conversion. The mid-career period is where multiple resource demands intersect (e.g., childcare, eldercare, career plateau risks); therefore, this career stage provides the ideal context for examining how gendered resource passageways shape success trajectories.

RQ1: How do career resource levels differ in men and women? (Unequal Attribute)

RQ2: How do career resources differentially predict the career success of men and women? (Unequal Effect)

2.4. Connecting the Theory to Talent Management Theory and Practice

TM scholarship emphasises attracting, developing, deploying, and retaining people whose contributions are pivotal to organisational performance (Al Ariss et al., 2014). Yet, heavy outsourcing of acquisition and the contraction of employer-funded development post-COVID risk an over-reliance on external markets, while internal resource caravans atrophy (Cappelli & Schwartz, 2024; Society for Human Resource Management, 2022). Moreover, “talent” remains contested and often narrowly operationalised (McDonnell et al., 2023; Vardi & Collings, 2023), with insufficient attention to mid-career contributors and gendered conversion mechanisms (Afiouni, 2019). Generational and psychological-contract perspectives stress that age/cohort shape expectations and retention (Festing & Schäfer, 2014), and that gender and experience can moderate TM-outcome relationships (Antony et al., 2023).

By specifying measurable resources (CRF), how they operate over short time windows (COR theory), and for whom their conversion differs (Unequal Attribute/Effect), this study provides a design template for sustainable TM.

2.5. Anticipated Contributions

The framework advances career and TM research in three ways. First, it specifies temporal parameters of COR processes in careers over weeks, not only years, using full-panel modelling suited to spiral logic. Second, it integrates CRF measurement with COR process theory to deliver actionable resource levers for HR linking human capital, environmental resources, motivation, and CSM to time-sensitive interventions. Third, it foregrounds mid-career and gendered conversion, addressing the dual gap of temporality and heterogeneity in TM scholarship. Collectively, the framework repositions TM as a dynamic, equity-attuned resource architecture that builds sustainable organisations.

3. Materials and Methods

3.1. Data Collection/Procedure

The study received ethical approval from Northumbria University (#3308) and was pre-registered on the Open Science Framework (OSF, https://osf.io/8q5mw/files/osfstorage, accessed 22 December 2025). A longitudinal, web-based survey with three waves spaced 30 days apart was employed, using the same time lag at each wave to minimise dropout and common-method bias and to enable Latent Growth Modelling (LGM) (Duncan & Duncan, 2009; Podsakoff et al., 2003). Questionnaires and ethics materials are available on OSF (Supplementary Materials Survey folder).

Eligibility: The study employed UK adults (≥18 years). Where participants held multiple jobs, they responded with reference to the job they performed most. Recruitment combined Prolific (paid £6/hour; usernames recorded for matching) and an opportunity sample (LinkedIn/Reddit; email addresses captured at T1 solely to send T2/T3 surveys and then deleted). Reminder messages/emails were sent at three and six days; non-response within seven days led to exclusion. At T2 and T3, participants were initially asked if they still held the same job as at T1, and if they did not, they were excluded from further study participation. Participants were not informed that the study examined gender differences; the full purpose was disclosed at T3 with debrief and a two-week withdrawal window. Data collection began in September 2023. Prior work indicates that Prolific data quality is comparable to opportunity samples and other panel vendors (Palan & Schitter, 2018; Peer et al., 2017, 2022).

3.2. Participants

Our working UK sample was chosen due to much research in this field being conducted in other parts of Europe, e.g., Germany, by Haenggli et al. (2021) and Blokker et al. (2019) in Germany, and on early careerists, focusing less on those individuals at the mid-career stage. At T1, 757 eligible participants responded (105 non-Prolific). Given the centrality of gender to the RQs and the very small non-binary subgroup (n = 6), these cases were excluded from inferential analyses due to the likelihood of error with such a small sample size compared to two larger sample sizes (Biemann & Kearney, 2010). Failure on two attention checks (“please tick ‘Strongly disagree’…”) led to removal, yielding N = 685 for analyses and invitations to T2 (n = 585; 85% response) and T3 (n = 543; 93% of T2).

Table 1. Demographic characteristics of the T1 sample, separated by men and women.

Variable	Women (n = 353)	Men (n = 332)	Total (N = 685)
Age (M, SD)	39.18 (11.4)	40.71 (11.7)	39.9 (11.5)
Caring responsibility (n, %)
None	180 (51.9%)	189 (58.2%)	369 (53.9%)
Parent and/or Carer	167 (48.1%)	136 (41.8%)	316 (46.1%)
Education (n, %)
Secondary school	31 (8.8%)	29 (8.7%)	60 (8.8%)
College/sixth form	76 (21.5%)	75 (22.6%)	151 (22.1%)
Undergraduate degree	132 (37.4%)	146 (44.0%)	278 (40.6%)
Master’s degree	90 (25.5%)	70 (21.1%)	160 (23.4%)
Doctorate	24 (6.8%)	12 (3.6%)	36 (5.3%)
Hours worked per week (M, SD)	31.37 (9.8)	35.87 (9.2) *	33.6 (9.8)
Tenure (n, %)
Up to 6 months	24 (6.8%)	13 (3.9%)	37 (5.4%)
Over 6 months–up to 1 year	26 (7.4%)	17 (5.1%)	43 (6.3%)
Over 1 year–up to 2 years	61 (17.3%)	46 (13.9%)	107 (15.6%)
Over 2 years–up to 5 years	73 (20.7%)	66 (19.9%)	139 (20.3%)
Over 5 years–up to 10 years	79 (22.4%)	82 (24.7%)	161 (23.5%)
Over 10 years–up to 20 years	74 (21.0%)	80 (24.1%)	154 (22.5%)
Over 20 years	16 (4.5%)	28 (8.4%)	44 (6.4%)
Sector (n, %)
Private	164 (46.5%)	221 (66.6%)	385 (56.2%)
Public	174 (49.3%)	101 (30.4%)	275 (40.2%)
Voluntary/Third 1	15 (4.3%)	10 (3.0%)	25 (3.7%)
Most common industry (n, %)	Education (86, 24.4%)	Education (43, 13.0%)	Education (129, 18.8%)
Most common occupation (n, %)	Administrative (69, 19.6%)	Operations/Production (57, 17.2%)	Administrative (92, 13.4%)

* Asterisks indicate a significant independent-samples t-test between men and women (* p < 0.001). ¹ Due to the low sample of participants within the voluntary sector, their sector was coded as blank, to permit the inclusion of this variable as a dichotomous control to the regression and LGM analyses.

summarises the demographic information. Women (n = 353) and men (n = 332) were similar in age, caring responsibilities, education, and tenure. Education was the most common industry overall, with women more often in administrative roles and men in operations and production. The only significant gender difference was hours worked per week (higher number for men), t(683) = 6.00, p < 0.001. Prolific vs. non-Prolific demographics were largely comparable; Prolific participants were older and worked more hours (age: t(109) = 7.00, p < 0.001; hours: t(683) = 3.00, p < 0.001). Residual invariance held across the Prolific/non-Prolific split for both the career resources and SCS measures (TLI/SRMR within acceptable bounds), supporting psychometric comparability.

3.3. Measures

We selected validated measures for each construct based on an extensive review of the literature, choosing tools with good reliability. Both multi-item scales used 5-point Likert responses (1 = Strongly Disagree to 5 = Strongly Agree). They also both showed configural invariance across time and scalar invariance by gender, supporting longitudinal and gender-group analyses. T2/T3 omitted demographics/job items and included a screening question about job change; those who had changed jobs were excluded (six at T2; two at T3). OCS was captured at T1; at T2/T3, respondents indicated any promotion/salary increase since the prior wave and then completed OCS items again, if applicable.

3.3.1. Career Resources Questionnaire

Career resources were measured with the 41-item Career Resources Questionnaire (CRQ) covering four domains (Human Capital, Environmental, Motivational, CSM), each measured with three or four items (Hirschi et al., 2018). Example item is as follows: “I have good knowledge of the job market (Job-Market Knowledge in the Human Capital domain). Reliability was high across waves (α = 0.87/0.87/0.89). A four-factor model fit better than a one-factor model at each wave (e.g., T1: Δχ²(6) = 537, p < 0.001), with acceptable CFI/TLI and RMSEA/SRMR ranges.

3.3.2. Subjective Career Success (SCS) Inventory

SCS was assessed via the 24-item Subjective Career Success Inventory (Shockley et al., 2016). It captures eight dimensions (Recognition, Quality Work, Meaningful Work, Influence, Authenticity, Personal Life, Growth and Development, and Satisfaction), each measured with three items. The stem for each item is as follows: “Considering my career as a whole…”, with an example item being “my career has been personally satisfying” (Satisfaction dimension). Reliability exceeded 0.70 at each wave. CFA supported both eight- and one-factor structures, with the one-factor solution fitting better (e.g., T1: Δχ²(138) = 371, p < 0.001), consistent with recent applications (Kundi et al., 2023).

3.3.3. Objective Career Success (OCS) Measures

OCS was captured via four metrics: seniority, salary, number of promotions in the current company, and in the career. Seniority was coded on a four-level ordinal scale (no managerial responsibility to top-level manager); “no hierarchy” cases (n = 25) were treated as missing for that variable. Annual salary was recorded as a continuous value (£). Respondents also indicated the London location (see adjustment below in Section 3.4.1). Promotions were defined as increases in hierarchical level and/or substantial job scope increases attributable to performance (not inflation or tenure). Counts of promotions within the current employer and across the career were collected on an ordinal scale ranging from 0 to 11+, with intervals of one. At T2/T3, promotion since the last wave triggered the re-collection of OCS information.

3.3.4. Demographic Variables

Gender (1 = woman, 2 = man) was the focal grouping variable. Age, caring responsibility, education (secondary school to doctorate), hours worked per week, tenure, sector (private/public; third sector excluded due to small sample size), industry, and occupation were recorded and used as controls, given known associations with OCS and SCS.

3.4. Data Analysis Strategy

All analyses were conducted in R 4.3.1; the script, anonymised data, and data treatment protocol are on OSF.

3.4.1. Control Variables

To minimise contextual heterogeneity, only UK-based workers were included. Since London salaries materially exceed those elsewhere, the salaries were decreased by 20% (based on London Living Wage Foundation (2025) guidance and Office for National Statistics (2019) data). Before adjustment, London salaries (M = £46,555, SD = £29,193) exceeded non-London (M = £31,734, SD = £18,294), t(676) = 7.00, p < 0.001. After adjustment (M = £37,244, SD = £23,354), the difference persisted but represented a small effect size, rather than large (Cohen’s d = 0.29 vs. 0.72 pre-adjustment). Promotion during follow-ups (yes/no) was also coded for use as a time-invariant control in LGM (70 cases).

3.4.2. Statistical Analysis

LGM tested time-lagged hypotheses for SCS (H1–H4). Following recommended practice, we first estimated univariate (unconditional) models for SCS and resources, comparing no-growth vs. linear-growth forms and unconstrained vs. residual-variance-constrained models (Burant, 2016). The best-fitting models were selected via standard fit indices and chi-square tests (Duncan & Duncan, 2009). The mean slope addressed H1 (SCS drift). The intercept–slope covariance was informed by the loss- and gain-spiral dynamics (H2, H4). A negative covariance would be expected, as losses should be faster than gains according to COR theory.

We then estimated conditional multivariate LGMs, adding time-invariant covariates (e.g., gender, age, and other controls) to predict intercepts/slopes and time-varying covariates (CRF domains) to predict SCS at each wave (H3). If gender significantly predicted parameters, we ran multiple-group models (women vs. men) to test RQ2 (Unequal Effects).

For H5, we used hierarchical regression with T1 OCS as the outcome. Because OCS indicators use different scales, salary, seniority, and promotions were log-transformed and combined into a composite (and also analysed separately). Assumptions (linearity, homoscedasticity, normality of residuals) were checked via standard diagnostic plots (see OSF, Method, Supplementary Materials folder). Gender-split regressions further informed RQ2. RQ1 (Unequal Attribute) was addressed by comparing baseline resource levels by gender.

3.4.3. Attrition Analysis

Attrition was modest (T1 N = 685, T2 N = 585 14.6% attrition, T3 N = 543 7.2% attrition). To avoid listwise deletion bias, models used Maximum Likelihood (ML), appropriate under MCAR/MAR, and recommended for LGM (Allison, 2012). Comparisons of T1 SCS and OCS (log composite) between completers and non-completers were non-significant. Little’s MCAR tests for SCS at T2/T3 were non-significant as well (T2: χ²(1) = 2.03, p = 0.154; T3: χ²(1) = 3.31, p = 0.07) supporting the MCAR assumption. ML therefore provides efficient, less biassed estimates for missingness in LGM than listwise deletion or imputation (Curran et al., 2010).

4. Results

4.1. Career Success Levels Among Women and Men

Men reported significantly higher OCS overall log metric than women (M = 13.72, SD = 1.64 vs. M = 12.88, SD = 1.69), t(652) = −6.00, p < 0.001. This pattern held across all OCS indicators (salary, seniority, promotions). On average, men earned £38,095 (SD = £21,607) compared to women’s £27,616 (SD = £15,384), t(676) = −7.00, p < 0.001. The difference remained significant for both full- and part-time employees (all t-tests can be found on the OSF, Results, Supplementary Materials folder).

By contrast, SCS did not differ significantly by gender in the overall sample (t(682) = 0.70, p = 0.50), nor within full-time or part-time groups.

This correlation between OCS and SCS did not differ significantly between men (r = 0.33) and women (r = 0.27), Fisher’s z = 0.86, p = 0.39, suggesting that both genders experience a similar association between objective and subjective success. Correlations between all variables, together with their descriptive statistics, can be found on OSF (Results, Supplementary Materials folder).

4.2. Unequal Attribute (RQ1)

Independent t-tests compared mean resource levels at T1 across the four CRF domains, shown in

Table 2. Descriptive statistics and t-tests to explore RQ1 across the career resource domains.

Resource	Mean (SD)			t-Test	Result
	Overall	Men	Women
Human capital resource domain	3.50 (0.56)	3.56 (0.55)	3.44 (0.55)	t(683) = −3.00, p = 0.01	Men significantly higher than women.
Occupational expertise	3.61 (0.77)	3.73 (0.74)	3.49 (0.79)	t(683) = −4.00, p < 0.001	Men significantly higher than women.
Job-market knowledge	2.95 (0.86)	3.03 (0.83)	2.86 (0.87)	t(683) = −3.00, p = 0.01	Men significantly higher than women.
Soft skills	3.93 (0.68)	3.91 (0.72)	3.96 (0.65)	t(683) = 1.00, p = 0.30	No significant difference.
Environmental resource domain	3.30 (0.77)	3.29 (0.77)	3.32 (0.77)	t(683) = 0.40, p = 0.70	No significant difference.
Career opportunities	3.02 (1.06)	3.06 (1.07)	2.99 (1.04)	t(683) = −0.09, p = 0.40	No significant difference.
Organisational career support	3.31 (0.98)	3.32 (0.96)	3.31 (1.00)	t(683) = −0.10, p = 0.90	No significant difference.
Job challenge	3.57 (0.88)	3.60 (0.88)	3.54 (0.88)	t(683) = −0.90, p = 0.40	No significant difference.
Social career support	3.31 (0.82)	3.18 (0.84)	3.43 (0.79)	t(683) = 4.00, p < 0.001	Women significantly higher than men.
Motivation resource domain	3.50 (0.69)	3.50 (0.68)	3.51 (0.69)	t(683) = 1.00, p = 0.90	No significant difference.
Career involvement	3.20 (1.00)	3.11 (1.05)	3.28 (0.95)	t(683) = 2.00, p = 0.03	Women significantly higher than men.
Career confidence	3.78 (0.66)	3.84 (0.64)	3.73 (0.69)	t(683) = −2.00, p = 0.03	Men significantly higher than women.
Career clarity	3.54 (0.92)	3.56 (0.90)	3.51 (0.94)	t(683) = −0.60, p = 0.50	No significant difference.
CSM resource domain	3.27 (0.73)	3.31 (0.76)	3.24 (0.69)	t(683) = −1.00, p = 0.20	No significant difference.
Networking	3.14 (0.95)	3.23 (0.99)	3.05 (0.90)	t(683) = −2.00, p = 0.01	Men significantly higher than women.
Career exploration	3.01 (0.96)	3.01 (0.98)	3.01 (0.95)	t(683) = 0.02, p = 1.00	No significant difference.
Learning	3.67 (0.74)	3.69 (0.77)	3.65 (0.71)	t(683) = −0.70, p = 0.50	No significant difference.

. Significant gender differences emerged primarily within human capital. Men reported higher overall human capital (M = 3.56 vs. 3.44), t(683) = −3.00, p = 0.01, specifically for occupational expertise and job-market knowledge, but not for soft skills.

Women scored higher on social career support (environmental domain) and career involvement (motivational domain), while men reported higher career confidence (motivational domain) and networking (CSM domain). Other resources did not differ significantly. These findings partially support the Unequal Attribute proposition (RQ1), particularly for the human capital resource domain.

4.3. Subjective Career Success (SCS)

4.3.1. Unconditional (Univariate) LGMs

LGMs were estimated for SCS and the four CRF resource domains. Details of model fit comparisons are available on OSF (Results, Supplementary Materials folder). Accepted models are summarised in

Table 3. Parameter estimates for the unconditional latent growth models.

Latent Growth Models	Intercept		Slope		Covariance
	Mean	Variance	Mean	Variance
SCS (constrained)	3.945 *** (0.022)	0.283 *** (0.018)	−0.044 *** (0.008)	0.009 ** (0.003)	0.008 (0.005)
Human Capital (unconstrained)	3.499 *** (0.021)	0.260 *** (0.019)	0.035 *** (0.008)	0.010 (0.007)	−0.008 (0.008)
Environmental (unconstrained)	3.301 *** (0.029)	0.475 *** (0.034)	−0.007 (0.010)	0.007 (0.012)	0.003 (0.014)
CSM (unconstrained)	3.271 *** (0.027)	0.418 *** (0.033)	−0.015 (0.033)	0.019 (0.013)	−0.006 (0.014)

Note: Standard errors are depicted in brackets. ** p ≤ 0.01, *** p ≤ 0.001.

. It omits the motivational resource domain, which was not found to fluctuate across the length of the study.

Results in Table 3 revealed significant mean intercepts for all constructs, indicating meaningful baseline variation across participants. Both SCS and human capital showed significant slopes: SCS declined over the study period (mean slope = −0.044, p < 0.001), supporting H1 (SCS will decline), while human capital increased (mean slope = 0.035, p < 0.001). Slope variance for SCS was significant, showing that participants differed in their rate of change.

Covariance between intercept and slope was non-significant for all constructs, indicating no evidence of loss spirals over time. Thus, H2 (SCS will show a loss spiral) was not supported.

4.3.2. Conditional (Multivariate) LGMs

Two conditional LGMs were estimated to test whether time-varying resources predicted SCS trajectories. Model 2, which included time-varying covariates (human capital, environmental, and CSM resources), fit better than Model 1 (ΔAIC = 420, ΔBIC = 389; RMSEA = 0.04, SRMR = 0.01, TLI = 0.97, CFI = 0.99).

All resource domains significantly and positively predicted SCS at each time point, except CSM at T1, providing strong support for H3 (career resources positively predict SCS). However, covariance between intercept and slope was non-significant (−0.005, p = 0.25), indicating no cumulative resource-driven gain spirals. Hence, H4 (SCS will show a gain spiral in the presence of high resources) was not supported.

Because gender was a significant predictor of SCS intercept and slope, the model was re-run separately for men and women to address RQ2 (Unequal Effect). Results are shown in

Table 4. Parameter estimates of the latent growth models, which predict SCS over time, for women and men separately.

	Women		Men
	Intercept	Linear Slope	Intercept	Linear Slope
Mean	1.169 *** (0.275)	−0.098 (0.148)	0.833 ** (0.298)	−0.070 (0.139)
Variance	0.093 *** (0.013)	0.010 ** (0.004)	0.096 *** (0.013)	0.003 (0.003)
Time-invariant predictors
Age	0.007 ** (0.003)	−0.000 (0.001)	0.003 (0.002)	0.000 (0.001)
Carer	0.037 (0.050)	−0.017 (0.026)	0.062 (0.050)	−0.005 (0.023)
Education	−0.033 (0.025)	0.011 (0.013)	0.052 (0.027)	0.001 (0.013)
Hours	0.001 (0.003)	−0.002 (0.002)	−0.000 (0.003)	−0.002 (0.001)
Tenure	0.023 (0.018)	−0.005 (0.010)	0.015 (0.020)	0.009 (0.009)
Sector	−0.125 * (0.050)	0.017 (0.026)	−0.005 (0.055)	0.023 (0.025)
Baseline motivation	0.350 *** (0.047)	−0.026 (0.027)	0.371 *** (0.048)	−0.021 (0.025)
Baseline OCS	0.008 (0.018)	−0.005 (0.009)	0.014 (0.019)	−0.007 (0.009)
Promotion	−0.000 (0.077)	−0.029 (0.040)	−0.050 (0.069)	−0.006 (0.032)
Time-variant predictors
Human capital T1	0.117 * (0.048)		0.205 *** (0.049)
Human capital T2	0.079 (0.041)		0.176 *** (0.043)
Human capital T3	0.177 *** (0.054)		0.167 *** (0.048)
Environmental T1	0.236 *** (0.036)		0.129 *** (0.035)
Environmental T2	0.236 *** (0.033)		0.121 *** (0.032)
Environmental T3	0.312 *** (0.036)		0.181 *** (0.035)
CSM T1	0.043 (0.042)		0.057 (0.037)
CSM T2	0.173 *** (0.037)		0.143 *** (0.036)
CSM T3	0.089 * (0.044)		0.169 *** (0.037)

Note: Unstandardised regression coefficients are reported, as recommended for LGM (Wang et al., 2019). Standard errors are depicted in the brackets. * p < 0.05, ** p ≤ 0.01, *** p ≤ 0.001. Covariance between the intercept and the slope was non-significant for both models and hence was not reported.

.

For both genders, model fit was acceptable (RMSEA = 0.07, SRMR = 0.02, TLI = 0.93, CFI = 0.96). Intercepts were significant (p < 0.01) for both groups, but slope variance was significant only for women (β = 0.010, p = 0.008), suggesting greater heterogeneity in women’s SCS trajectories.

Among time-invariant variables, age positively predicted SCS for women (β = 0.007, p = 0.0061) but not for men, while private sector employment enhanced SCS for women (β = −0.125, p = 0.013) but not for men. Baseline motivation was a strong predictor of SCS intercept for both genders (p < 0.001), consistent with COR’s resource investment principle.

Regarding time-varying predictors, the following was discovered:

• Environmental resources consistently predicted SCS for both groups at all waves, with stronger effects for women (e.g., T1: β = 0.236 for women vs. 0.129 for men, both p < 0.001).
• Human capital predicted SCS robustly across time for men (p < 0.001), but only at T1 (p = 0.014) and T3 (p = 0.001) for women, indicating uneven conversion of this resource to SCS.
• CSM effects increased for men across waves (T1 non-significant, T3 strongest p < 0.001), whereas women’s CSM effects fluctuated, being strongest at T2 (p < 0.001), but approaching non-significance at T3 (p = 0.042).

Overall, environmental resources were more salient for women’s SCS, while human capital and CSM were more stable predictors for men, partially supporting the Unequal Effect (RQ2).

4.4. Objective Career Success (OCS)

To test H5 (human capital predicts OCS), hierarchical regressions were conducted on T1 data using demographic and job variables, followed by resource domains. Gender remained a significant covariate even after controls (β = 0.115, p < 0.001). Adding career resources improved model fit (ΔR² = 0.03, p < 0.001) and human capital emerged as the strongest predictor of OCS (β = 0.206, p < 0.001). Environmental resources were marginal (β = 0.081, p = 0.053), while motivational and CSM domains were non-significant. These results overall support H5 (human capital will predict OCS cross-sectionally).

To explore Unequal Effects, regressions were run separately for men and women (see

Table 5. Hierarchical linear regression on men and women separately, with the log of the OCS metric as the outcome.

Predictor Variables	Women			Men
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
	β	β	β	β	β	β
Step 1—Demographic and Job Characteristics
Age	0.190 ***	0.185 ***	0.187 ***	0.179 ***	0.167 **	0.176 **
Carer	0.040	0.014	0.008	0.036	0.042	0.041
Education	0.221 ***	0.175 ***	0.170 ***	0.239 ***	0.233 ***	0.232 ***
Hours worked	0.334 ***	0.299 ***	0.280 ***	0.245 ***	0.220 ***	0.217 ***
Tenure	0.189 ***	0.195 ***	0.208 ***	0.292 ***	0.291 ***	0.287 ***
Sector	−0.101 *	−0.086	−0.090	−0.191 ***	−0.175 ***	−0.173 ***
Baseline SCS	0.140 **	0.035	0.007	0.202 ***	0.140 *	0.120
Step 2—Human Capital
Occupational Expertise		0.089	0.097		0.050	0.062
Job Market Knowledge		0.13 9 **	0.136 *		0.108 *	0.102 *
Soft Skills		0.136 **	0.147 **		0.008	0.007
Step 3—Environmental Resources
Career Opportunities			0.077			0.036
Organisational Career Support			0.048			0.022
Job Challenge			0.013			−0.018
Social Career Support			−0.064			0.002
R2	0.304 ***	0.355 ***	0.365 ***	0.398 ***	0.410 ***	0.412 ***
ΔR2		0.050 ***	0.010		0.012	0.002

* p < 0.05 ** p ≤ 0.01 *** p ≤ 0.001.

). For women, adding human capital (Model 2) significantly improved explanatory power (ΔR² = 0.05, p < 0.001), with job-market knowledge (β = 0.139, p < 0.01) and soft skills (β = 0.136, p < 0.01) predicting OCS. For men, human capital did not add explanatory power beyond demographics (ΔR² = 0.012, p > 0.05). Thus, H5 holds primarily for women, revealing an Unequal Effect: the same human capital yields greater OCS benefits for women than men.

Across demographic controls, age, education, hours worked, and tenure positively predicted OCS for both genders (p < 0.001). Private-sector employment benefited men (β = −0.191, p < 0.001) but not women, and baseline SCS predicted OCS for men (β = 0.202, p < 0.001) only. Caring responsibilities did not significantly influence OCS for either group.

Analyses of individual OCS metrics (salary, seniority, promotions) corroborated these patterns (full results can be found on OSF, Results, Supplementary Materials folder). Specifically, for women, job-market knowledge and soft skills predicted higher pay and promotions; for men, job-market knowledge related primarily to seniority. Overall, these findings support H5 for women and provide evidence for an Unequal Effect (RQ2).

5. Discussion

This study examined the temporal dynamics of career resources as predictors of career success and explored gender differences in these relationships, placing gender central to our understanding of the use of resources in achieving career success. A rigorous three-wave, time-lagged design followed the same working adults in the UK across two months, measuring their resources, SCS, and OCS. To our knowledge, ours is the first study to attempt to measure all variables over these time points (after Haenggli et al., 2021). Consistent with COR theory, SCS declined over time, but access to career resources enabled individuals to reverse this decline and maintain or enhance their career success. However, the hypothesised loss and gain spirals were not supported, as the LGM did not reveal significant covariance between the intercept and slope of SCS.

Significantly, and extending our understanding of how the Unequal Effects proposition applies to career resources (using the CRF) and their relationship to career success (RQ2; Frear et al., 2019), gender differences were identified. Men’s SCS benefitted more from human capital and CSM, while women’s SCS was more strongly influenced by environmental resources. CSM showed an inconsistent association with women’s SCS. In contrast, all career resources exhibited positive associations with SCS overall. This suggests that women’s SCS can be enhanced by organisational intervention, yet men’s appears to be improved through the collection of human capital and utilisation of career self-management, i.e., more individually driven.

In terms of OCS, only human capital had a significant effect (and only for women), again confirming Unequal Effects (RQ2). Interestingly, women reported lower levels of human capital overall than men (Unequal Attribute, RQ1). This advances our understanding of Frear et al.’s (2019) propositions, suggesting that both mechanisms (effect and attribute) may jointly contribute to persistent gender disparities. Our research also suggested that caring responsibilities had no effect on either SCS or OCS across genders. The following discussion integrates these findings within COR theory and the broader literature, emphasising their theoretical and practical implications.

5.1. Theoretical Contributions

The central and unique finding of this research is that career resources are critical for both women’s and men’s career success, while caring responsibilities are not. Previous research has often positioned motherhood as a key determinant of women’s career outcomes (Correll et al., 2007; Evers & Sieverding, 2014), yet changing social roles and shared childcare responsibilities (Offer & Kaplan, 2021) may be diminishing its explanatory power. It could also be that the internalisation of the ‘ideal worker’ norm (Williams, 2001) extends far beyond having caring responsibilities and more deeply relates to how women perceive themselves in the working world (i.e., their identity); however, equally, some studies do show a positive impact of motherhood on career outcomes such as time management, motivation and confidence (Torres et al., 2024). Nonetheless, gender disparities in OCS persist, with women self-reporting lower salaries, promotions, and seniority, even after controlling for working hours. It may be that there are intricacies and sensitivities in caring responsibilities that are worthy of further investigation (such as supportive mechanisms at work and home that can counter these, e.g., Frear et al., 2019, and whether experiences of caring responsibilities and their impact change over shorter time periods). Furthermore, our research suggests that we need to consider more than caring responsibilities in addressing the gender differences in career outcomes (Iqbal et al., 2025).

Although the correlation between SCS and OCS was similar across genders (small-to-medium, in line with Ng et al., 2005), men’s SCS translated more readily into OCS. Women thus face greater difficulty achieving both facets of success simultaneously. Specifically, human capital was the strongest predictor of women’s OCS, while environmental resources most strongly predicted their SCS. This is something that can be provided by workplaces. This duality reflects Unequal Effects: despite the motivational and environmental strengths women demonstrate, the unequal recognition and conversion of their human capital constrain OCS outcomes. Extending our understanding of Unequal Attributes, findings suggest lower self-reported human capital, meaning that women may undervalue or understate their expertise, reinforcing a perception gap that limits recognition and reward. Indeed, women self-reported significantly lower career confidence resources than men in this study, suggesting that their lack of self-efficacy in confidently reporting their human capital may pose a barrier to their OCS. Understating expertise is not a new phenomenon (e.g., Kunz & Prügl, 2019), but our research indicates that this could lead to lower OCS, which extends our understanding of gender role theory (Money, 1973) and the career outcomes that it impacts.

5.1.1. Temporal Dynamics (H1)

This study extends COR theory by elucidating the short-term temporal dynamics of career resources and SCS. The results demonstrate that both constructs fluctuate meaningfully over periods as short as one month, challenging the assumption that behavioural or attitudinal change in careers unfolds only over long durations or following career shocks (Akkermans et al., 2021). This shows that career satisfaction and resources evolve dynamically, even in stable employment contexts.

These findings indicate that career resources and satisfaction operate similarly to job resources, responsive to daily and monthly changes rather than static, long-term traits. For this research, we focused on monthly changes (in line with Haenggli et al., 2021), but recommend that future research examine daily changes for both constructs. Although it is currently hypothesized that career resources may fluctuate daily, empirical verification is required; the monthly fluctuations in resources measured here suggest that such high-frequency variability is likely. Consistent with prior longitudinal evidence of gradual satisfaction decline (Spurk et al., 2011), the present results show that such a decline can occur within two months. This underscores that, without renewed resource inputs, individuals’ SCS diminishes, supporting the COR principle that desirable outcomes cease when resources stagnate.

5.1.2. Loss and Gain Spirals (H2 and H4)

While the decline in SCS supports COR’s basic tenets, the expected loss and gain spirals were not observed. The absence of covariance between initial SCS and its trajectory suggests that individuals with low starting satisfaction did not deteriorate faster, nor did resource-rich individuals exhibit accelerated gains. This paper’s novel contribution is that we aimed to test loss and gain spirals in a robust way, using the covariance between the intercept and the slope; to our knowledge, this has not been performed before, yet we would encourage this approach. We expected a negative covariance between the intercept and the slop so individuals who start with a lower intercept of SCS at T1 were expected to show faster losses (loss spiral), whereas those with high SCS were expected to show gains but at a slower pace (i.e., a gain spiral). We did not observe this.

Two explanations are possible. First, previous studies reporting spirals (De Cuyper et al., 2012; Spurk et al., 2011) relied on unidimensional career satisfaction scales, whereas the multidimensional measure used here (Shockley et al., 2016) captures a broader construct encompassing multiple life domains, possibly dampening volatility, meaning that there is more fluctuation in the components of career satisfaction. Second, it is possible that spirals may manifest over shorter intervals (days or weeks) than the one-month lags used here, and this warrants further investigation. Initial evidence from diary studies (Volmer & Wolff, 2018; Zacher, 2015) suggests that this might be the case. Future research should therefore examine SCS and career resources using weekly or daily timeframes to capture loss dynamics and to identify at what time point changes are experienced.

5.1.3. Resource Accumulation (H3 and H5)

Despite the absence of spiral effects, this study provides robust support for resource accumulation as a driver of SCS. All four resource domains contributed positively to SCS, reinforcing prior meta-analytic evidence (Ng et al., 2005; Ng & Feldman, 2014). Notably, human capital accounted for a significant proportion of variance in women’s OCS, confirming that the CRF is a valid operationalisation for examining career success predictors over time.

5.1.4. Gendered Lens (RQ1, RQ2, H5)

Gendered analysis revealed asymmetrical patterns of resource effectiveness. Motivational resources benefited both genders, but environmental resources were more critical for women’s SCS, while human capital and CSM were stronger predictors for men. In OCS, human capital predicted success only for women, not for men, despite women reporting lower levels of it. This supports both Unequal Attributes and Unequal Effects.

The finding that CSM had inconsistent effects for women but growing benefits for men suggests gendered differences in how this resource is mobilised. Women may deploy CSM reactively (e.g., to engage in career exploration if they experience low SCS), while men apply it proactively (Akkermans & Hirschi, 2023). This points to a need for extending COR theory to consider gendered investment patterns: women’s resource utilization may aim to prevent loss, while men’s may focus on pursuing gain. Future qualitative studies should explore how women and men conceptualise and enact human capital and CSM, and how external perceptions shape their resource conversion into success.

5.2. Practical Implications

For UK HR leaders, this study underscores that career success is dynamic, not static, as SCS fluctuates over short timeframes. Therefore, organisations cannot assume stability between annual review cycles. To support equitable and sustainable career success, talent management systems should embed temporal sensitivity and gender awareness. It is possible that these implications relate to broader contexts than just the UK; however, this is the focus of the study, and further research would be warranted in other countries to understand whether these implications apply.

• Diagnose baseline resource profiles and risk passageways.
  Conduct periodic audits of career resources by gender at mid-career stages to identify loss-risk groups. Pay equity audits, skills inventories, and sponsorship mapping to reveal Unequal Attributes. For example, in the current age of AI, women have been found to adopt AI technologies less frequently than men (Tang et al., 2025), highlighting a potential gender gap in engagement with AI as a career resource. Within the career resource framework, this gap may place women at a disadvantage as organisations increasingly adopt AI-mediated talent management practices.
• Intervene through targeted, time-bounded resource building.
  Resource investment is programmable and should be linked to the UK Equality Act (2010). In particular, in the UK, discrimination on the basis of gender is unlawful; however, we argue that a one-size-fits-all approach to career development and talent management that does not account for gender could indirectly favour some genders over others (Iqbal et al., 2025). We therefore recommend that HR policies and practices strengthen Human Capital through modular learning and micro-credentials (over monthly or quarterly intervals, rather than yearly or longer development opportunities). This would benefit all, but particularly women’s OCS, in a UK context. In addition, consider how sponsorship works for women and men, but particularly to support women’s environmental resources, which are important for their achievement of SCS.
• Equitise resource conversion.
  Address Unequal Effects by ensuring that equivalent human capital and performance translate equitably into OCS. This requires transparent criteria, structured promotion calibration, and bias-resistant decision panels, and would conform with requirements of the UK Equality Act (2010), a one-size-fits-all approach, whilst appearing fair could actually be perpetuating inequities. Where conversion gaps persist, organisations should redesign career passageways to enhance visibility and fairness.
• Monitor temporal change.
  Implement frequent diagnostics (monthly or quarterly) to track SCS, job challenge, opportunity visibility, and sponsorship access. Detecting early drift in SCS enables rapid corrective support (e.g., coaching or stretch assignments). Our knowledge of UK contexts suggests that this rarely occurs, or when it does, it is sporadic and normally annual.
• Accelerate under-recognised mid-career talent.
  Identify resource-rich but under-visible employees and accelerate them through sponsorship and developmental roles, while identifying and supporting resource-poor groups with foundational access to networks and learning. It is possible that this also links to other protected groups within UK law, but also marginalises groups such as those with lower socio-economic status.

Collectively, these actions reframe UK TM from a static, annualised model toward a dynamic, equity-oriented resource architecture. They operationalise COR and CRF principles in practice, namely preventing resource loss spirals, fostering gain spirals, and building sustainable, gender-fair career pipelines.

5.3. Strengths, Limitations, and Future Research

This study addressed recent calls to investigate the dynamics of career success (Seibert et al., 2024) and to employ within-person methods (Akkermans et al., 2024). It also responded to demands for gender-focused analyses in career research (Balakhtar et al., 2024; Orser & Leck, 2010). Methodological strengths include the use of a multidimensional SCS measure, a full-panel design allowing analysis of change trajectories, and controls for job stability between time points, shorter than seen in previous work.

However, limitations exist. Working hours were only recorded at baseline, potentially overlooking workload variations that influence resources. Future research should measure working hours at each time point. Moreover, the study did not test reciprocal relationships between SCS and resources, which are expected in COR theory’s gain spirals. Future work should employ designs such as random intercept cross-lagged panel models (Hamaker et al., 2015; Usami, 2021) to capture both lagged and cross-lagged effects within individuals. We additionally made corrections based on London salaries, with the best available evidence; however, our estimate of 20% may have been too high or too low. For future research, it is important to ensure that the data at the time of completing the study is utilised to make such adjustments.

It is worth noting that there is very little evidence (especially from full-panel designs) and theorising as to how quickly these effects are unfolding. This study contributes additional evidence on the effect of time from a time-lagged full panel design with 30-day time lags, similar to Haenggli et al. (2021), who found significant autoregressive paths for SCS and CSM. Spiral effects were not found in the time lag utilised in our study (deliberately chosen to avoid adaptation effects that Ritter et al. (2016) suggest can occur with longer time lags). Therefore, our study points to future research perhaps needing to compare different time lags (like recently performed by S. J. Allen et al., 2025) and maybe even different analyses (e.g., RI-CLPM), but further interrogation of spiral effects is encouraged, so that we fully understand both how to measure them and when (in time) they occur.

Finally, in an area as complex as gender and career success, there are undoubtedly variables that impact career resources and outcomes that warrant further investigation. It is simply not possible to measure all potential variables; however, our contribution demonstrates a methodology and a unique perspective around gender, suggesting that we must do more than simply control for it in studies, because gender can differentially impact the availability of resources and the outcomes of utilising them.

6. Conclusions

This study provides empirical support for the foundational principles of COR theory while refining its temporal boundary conditions. Our study uniquely addressed gender differences in the utilization of career resources using a full panel design, not seen before in the literature in this field. Although loss and gain spirals were not observed, findings demonstrate that resource levels and SCS fluctuate over short periods (e.g., 30 days) and that, importantly for talent management, women’s and men’s objective and subjective career success may benefit differentially from specific career resources. These results advance theoretical understanding of resource dynamics and how the same resources can be utilised and experienced differently as a feature of gender. Additionally, our contribution suggests that we need further research to understand the impact of caring responsibilities on career success, because it appears that caring responsibilities alone are not the differentiating factor in our study. Finally, we highlight the need for gender-sensitive, time-responsive career interventions to foster sustainable TM for all.