Capturing change: a quasi-longitudinal qualitative approach to motivation and engagement in workplace e-learning

The rapid expansion of digital learning environments has generated increasing research on adult engagement with e-learning, particularly in workplace contexts (Ally et al., 2022; Beier et al., 2025; Theelen et al., 2022). While studies that address the concepts of motivation, engagement and knowledge retention exist, methodological approaches often rely on single-time-point surveys or interviews. These cross-sectional designs provide useful snapshots but fail to capture how learning evolves in real time, a limitation noted across both educational and workplace learning research (Gašević et al., 2015; Heikkinen et al., 2025).

In the context of Industry 5.0, where human–machine collaboration, human-centricity and continuous upskilling are emphasised as central to workforce resilience (Breque et al., 2021; Poláková et al., 2023), the temporal dimension of learning becomes particularly salient. Learners' experience with e-learning is not static; motivation may rise or fall during the completion of a module. Engagement can fluctuate with design features, and knowledge retention can only be meaningfully assessed after skills are consolidated in practice (Xie et al., 2025). Hilpert et al. (2023) call for the use of complexity frameworks to better account for the dynamic and interdependent nature of engagement in technology-mediated contexts, while Fan et al. (2024) demonstrate how generative artificial intelligence can alter motivational processes through ‘metacognitive laziness’.

Avci and Kula (2022) posit that online engagement develops through interlinked cognitive, emotional and contextual dimensions, reinforcing the need for approaches that capture change across time. Hilpert et al. (2023) extend this perspective by conceptualising engagement as an emergent property of interconnected systems. Building on these foundations, Saari et al. (2024) highlight how workplace practices mediated by mobile technologies are inherently temporal, with agency shaped by evolving rhythms of use. Additionally, Fan et al. (2024) demonstrated how motivation is shaped and reshaped in real time by technological affordances. These studies demonstrate that learning trajectories are flexible and evolving, reinforcing the methodological case for quasi-longitudinal approaches. In response, this study introduces and applies the Temporal Evaluation Framework for Workplace E-Learning (TEF-WEL) to structure evaluation across three phases: baseline, midpoint and post-module. Grounded in the principles of qualitative longitudinal research (QLR), TEF-WEL provides a structured means of capturing how learner experience unfolds over time in workplace e-learning contexts.

Accordingly, this study was guided by the following research question: How do adult learners' motivation, engagement and knowledge retention evolve over time as they engage with an e-learning module?

This literature review synthesises research on workplace e-learning, experiential learning and qualitative longitudinal inquiry to establish the theoretical and methodological foundations for examining how adult learner motivation, engagement and knowledge retention evolve over time.

Research into e-learning for adults has expanded in recent decades, reflecting the demand for scalable, flexible and technology-mediated workplace training (Beier et al., 2025). The extant literature has prioritised evaluative outcomes of workplace e-learning, most commonly captured through post-course measures, rather than examining how learner experiences develop and change during learning itself (Bergdahl et al., 2024; Xie et al., 2025). Cross-sectional surveys and one-off post-course interviews continue to dominate (Huang et al., 2024), providing useful but ultimately static accounts of learning. Such designs capture perceptions at a single point in time, obscuring the iterative processes through which adults construct meaning, develop confidence and integrate learning within organisational and career contexts (Hamori, 2023).

The preponderance of cross-sectional designs can be traced to the evaluation culture of workplace training. Organisations have often prioritised efficiency metrics such as completion rates, knowledge quizzes or satisfaction scores (Kirkpatrick and Kirkpatrick, 2006). While these indicators provide useful signals, they risk conflating training outputs with outcomes, assuming that knowledge retained immediately post-training equates to long-term capability. This critique echoes broader concerns in qualitative research that static designs oversimplify complex processes (Guest et al., 2020; Hermanowicz, 2013).

Notably, earlier workplace e-learning scholarship has already highlighted the risks of placing too much interpretive responsibility on learners. For example, Hase (2009) argues that digital workplace learning increasingly requires learners to adopt self-determined, autonomous approaches, placing responsibility for pacing and meaning-making on the learner rather than the instructor. Similarly, Alby et al. (2025) highlight the complex routines, role-related demands and organisational constraints that define professional environments, warning that the associated overly reductionist analyses risk obscuring insights that are essential for understanding practice. While motivation, engagement and knowledge retention are widely studied, the vast majority of workplace e-learning research remains cross-sectional, with few studies tracing change over time (Corden and Millar, 2007; Huang et al., 2024). This gap highlights the need for temporally sensitive designs that can capture the development of workplace learning across sequences of reflection, experimentation and application.

Experiential learning perspectives further highlight the need for temporally sensitive approaches. Kolb's (1984) experiential learning cycle: concrete experience, reflective observation, abstract conceptualisation and active experimentation, frames learning as cyclical and iterative. Similarly, Schön's (1983) concepts of “reflection-in-action” and “reflection-on-action” emphasise the evolving sense-making that underpins professional practice. This resonates with Tomlinson and Souto-Otero's (2025) study, which shows that graduates conceptualise meaningful work not only in terms of career outcomes but also in relation to personal growth and contribution, underscoring the role of meaning-making in professional learning trajectories.

Best practice workplace e-learning aligns with these models, with learners shifting from initial exposure, through experimentation, to application and reflection over time (Pimblett et al., 2025). This dynamic orientation positions workplace learning as a process of continual adjustment rather than a discrete event. Recent research indicates that scenario-based digital training fosters both skill development and professional identity formation by supporting iterative cycles of reflection and re-application (Illeris and Ryan, 2020; Wang and Lajoie, 2023).

A quasi-longitudinal qualitative approach operationalises these theoretical insights by tracing learner reflections not just at isolated points but across a sequence of interactions. In doing so, it captures the evolving processes of meaning-making that experiential theories foreground, but which cross-sectional methods frequently miss.

Theory and recent workplace studies highlight the temporal dimension of learning; qualitative longitudinal research (QLR) remains underrepresented in digital learning research (Neale, 2019; Treanor et al., 2021). Neale (2019) and QLR Saldañ;a (2003) provide an established tradition for examining how experiences, practices and perspectives develop over time. Originating in the fields of sociology and education, it has been used to trace change over extended periods, illuminating not only what changes but also how change unfolds (Hermanowicz, 2013). Applications span health research, where QLR has examined adaptation to chronic illness across care trajectories (Corden and Millar, 2007) and youth studies, where it has captured transitions from education into employment and the shifting meanings associated with these stages (Thomson and McLeod, 2015).

Despite its widespread use elsewhere, QLR has been applied more cautiously in digital learning contexts. Reviews of longitudinal learning research suggest that many studies remain short-term or prioritise behavioural metrics over interpretive depth, limiting insight into learners' meaning-making processes (Harju et al., 2019). In a more recent work, Heikkinen et al. (2025) demonstrated the potential of longitudinal qualitative approaches for understanding digital learning as a dynamic process. Building on this foundation, scholars are increasingly calling for temporally sensitive qualitative approaches that can capture shifts in engagement and motivation over time, particularly in digitally mediated learning environments (Avci and Kula, 2022; Treanor et al., 2021). In workplace e-learning contexts, where training occurs in compressed timeframes and recall can fade rapidly, a quasi-longitudinal design offers a pragmatic adaptation of QLR principles, enabling exploration of reflection, application and evolving learner experience within organisational constraints.

Bergdahl et al. (2024) posit that Learning Analytics (LA) are increasingly used to capture the temporal development of engagement by tracking behaviours such as clicks, time-on-task or module progression. Supporting this, Honson et al. (2024) demonstrated how LA can provide continuous, data-informed insights to improve engagement and support in large digital learning contexts. Such approaches illustrate the potential for iterative evaluation methods that align with the temporal design adopted in this study. Multimodal learning analytics approaches (MMLA) seek to integrate behavioural, cognitive and emotional indicators, offering valuable insights into shifting patterns of learner engagement (Seufert et al., 2023). However, while LA can highlight what learners do, it cannot capture why motivation fluctuates or how learners interpret design features. For this reason, LA are best understood as a complement to, rather than a substitute for, interpretivist qualitative approaches that foreground meaning-making and change over time (Neale, 2019; Saldañ;a, 2003). By privileging digital traces of behaviour over learners' interpretations, analytics risks flattening complex processes of engagement into surface-level indicators.

This study positions itself at the intersection of LA and interpretive, qualitative approaches. By adopting a quasi-longitudinal qualitative design, it provides the interpretive depth missing from LA while retaining sensitivity to temporal dynamics. In doing so, it bridges the strengths of analytics with the narrative richness of QLR, offering a more holistic account of how engagement and motivation evolve in workplace e-learning.

The urgency of adopting temporally sensitive research is reinforced by the paradigm of Industry 5.0, which reframes digital transformation around human-centricity, resilience and sustainability (Breque et al., 2021; Shahidi Hamedani et al., 2024). A parallel shift is evident in Education 5.0, which similarly foregrounds learner-centred, adaptive and holistic approaches to knowledge development (Shahidi Hamedani et al., 2024). Recent studies reinforce this reorientation. Engström et al. (2024) show how AI-driven systems reshape workplace learning by intertwining technological sensemaking with human interpretive practices. Complementing this, Valtonen and Holopainen, 2025 demonstrates that digital transitions can either foster or undermine employee well-being and resilience.

Together, these perspectives reinforce the need for temporally sensitive research designs that capture learning as it unfolds. In this context, workplace e-learning must go beyond compliance training or narrow knowledge transfer to cultivate adaptability, critical reflection and resilience as enduring capabilities (Breque et al., 2021; Poláková et al., 2023). Quasi-longitudinal methods, in contrast, foreground learning as an unfolding process, aligning evaluation with the human-centric and developmental values of Industry 5.0 (Breque et al., 2021). In situating quasi-longitudinal qualitative research within this policy paradigm, this research demonstrates how methodological innovation can directly support strategic priorities for future skills and workplace transformation.

The theoretical framing of this study draws on two key perspectives. Primarily, Cognitive Load Theory (CLT) (Sweller et al., 2011) highlights the importance of sequencing and structuring content over time, recognising that extraneous, intrinsic and germane load fluctuate as learners progress through material. Concurrently, experiential and reflective learning theories (Illeris and Ryan, 2020; Kolb, 1984; Schön, 1983) emphasise cyclical processes of sense-making, experimentation and reflection, which align closely with the staged experiences afforded by branching scenarios and multimodal digital design (Pimblett et al., 2025). Alongside these theories, QLR (Neale, 2019; Saldaña, 2003) provides a methodological precedent for treating temporality and change as central to inquiry, offering tools to investigate how learner motivation, engagement and knowledge retention evolve over time. Together, these perspectives position quasi-longitudinal qualitative research as a methodologically and theoretically appropriate approach to investigating workplace e-learning, situating the study within the wider Industry 5.0 agenda of adaptability, human-centricity and resilience (Breque et al., 2021; Poláková et al., 2023).

The study is conducted within an interpretivist paradigm recognising that learners' experiences are socially constructed and contextually situated (Creswell and Poth, 2018; Lincoln and Guba, 1985). A qualitative case study approach (Stake, 1995) and a quasi-longitudinal qualitative design are employed, involving repeated interviews over several weeks to examine how motivation, engagement and learning experiences developed during a workplace e-learning intervention (Neale, 2019; Saldaña, 2013). While conventional QLR can span years (Hermanowicz, 2013), workplace learning and development typically occur within compressed timeframes. Rather than positioning quasi-longitudinal inquiry as a compromise, this study advances it as a methodological innovation suited to workplace e-learning contexts, where training cycles are short and timely insight is essential for organisational adaptation (Deng, 2024).

Aligning with Stake's (1995) conception of case study research, this study examines a bounded qualitative case situated within a specific workplace e-learning module. The case comprises an online e-learning module delivered in two sequential parts, each incorporating multimodal elements including text-based content, short videos, quizzes and branching scenarios, which are used to test decisions in a “safe failure” environment (Pimblett et al., 2025). The e-learning module is designed to enhance problem-solving skills, which are increasingly recognised as a fundamental competency within Industry 5.0 (Poláková et al., 2023).

The case is bounded by the e-learning module, as well as by the participant cohort engaging with it during the study period. Although participants are recruited from multiple industries and occupational roles, all engage with the same standardised learning module, designed for cross-industry applicability.

Participants (n = 10) were recruited through purposeful sampling from the researcher's professional and extended networks, targeting individuals with recent experience of workplace e-learning across multiple industries. This approach enabled the selection of information-rich cases (Patton, 2015) with variation in role, age and prior e-learning experience, while maintaining a sample size appropriate for in-depth, longitudinal qualitative analysis (Creswell and Poth, 2018).

Participants range in age from their mid-20s to mid-60s, with a balanced gender distribution (five male, five female) and an age balance (five age 20–40 and five 41+). The participants occupied a range of administrative, technical and managerial roles, reflecting the heterogeneous nature of contemporary workplace learning. Inclusion criteria required participants to be in continuous employment for a minimum of two years and actively engaged in digital workplace training during the study period. All participants were in their current role for at least six months, supporting contextual stability across the three interview stages.

Overall, the sample was designed to maximise information power (Malterud et al., 2015) by balancing heterogeneity with the feasibility of sustained engagement. Each participant completed three interviews, generating thirty interviews in total. All participants provided informed consent and were assigned pseudonyms for data analysis. Table 1 summarises participant characteristics.

This study was approved by the institutional research ethics committee. All participants provided informed consent prior to participation. The study adhered to the Chartered Association of Business Schools (2015) ethical principles for research involving professionals in contextual workplace settings.

Semi-structured interviews were conducted in three stages between November 2024 and January 2025, reflecting a quasi-longitudinal qualitative design (Hermanowicz, 2013). All interviews were conducted online via Microsoft Teams, recorded with consent and transcribed verbatim for analysis and thematic coding (Braun and Clarke, 2006, 2021). Details of each interview phase, timing and areas of exploration can be found below.

• Phase 1 (baseline): Prior to engagement with the module. Questions explored prior experiences of e-learning, expectations and initial confidence levels.

• Phase 2, three weeks after phase 1 (mid-point): After learners completed half the module, questions focused on motivation, engagement, knowledge retention and their experiences of the module so far.

• Phase 3, three weeks after phase 2 (post-module): Post completion. Questions explored perceived relevance, transferability and reflections on changes in motivation, engagement and knowledge retention.

Intervals of three weeks between the modules (totalling six) balanced feasibility within workplace contexts while allowing participants to reflect on both immediate experiences and their subsequent application over time. This temporal structuring reflected the principles of experiential and reflective learning theories, which emphasise cycles of exposure, experimentation and application (Kolb, 1984; Schön, 1983).

Reflexive thematic analysis was employed to explore patterns of motivation, engagement and knowledge retention across the three phases of data collection (Braun and Clarke, 2006, 2021). Analysis was supported by NVivo software, which facilitated systematic organisation and comparison of codes across participants and timepoints. A secondary analytic pass was conducted in Excel as recommended by Saldaña (2013) to enhance reflexivity and reduce the risk of surface-level themes.

Analysis followed Braun and Clarke's six-phase process (Braun and Clarke, 2006, 2021), beginning with familiarisation through repeated transcript reading and memo writing, followed by line-by-line coding to generate descriptive and interpretive codes. Codes were iteratively refined and clustered into candidate themes, which were reviewed, defined and written up with illustrative extracts to capture both within-case trajectories and cross-case patterns. Consistent with reflexive thematic analysis, analytic sufficiency was understood in terms of the stabilisation of codes and themes across phases (Braun and Clarke, 2021), resulting in three overarching themes and nine sub-themes.

Temporality was treated as an analytic lens rather than a contextual backdrop (Neale, 2019). Following Saldaña's guidance on coding over time (2013), codes were traced within cases across baseline, mid-module and post-module phases, before being compared across cases to identify shared and divergent learning trajectories. This quasi-longitudinal approach enabled thematic development to accommodate the shifting ways in which learners construct meaning rather than isolated responses.

Rigour was supported through reflexive journaling (Berger, 2013), which was particularly important given the primary researcher's dual role as module designer and interviewer, aligning with practitioner–researcher reflexivity in interpretive case study research (Doyle, 2019). A second researcher reviewed coding structures to enhance dependability, while rich contextual description supported transferability (Lincoln and Guba, 1985).

Findings are presented in relation to the research question, tracing how learner motivation, engagement and knowledge retention evolved across the e-learning module. The quasi-longitudinal design enabled identification of three interlinked temporal themes: shifting motivation, evolving engagement and retention and transfer of knowledge into workplace practice. Each theme is explored across the three temporal phases, illustrating how participant perspectives unfolded over time.

Figure 1 provides an overview of the three overarching themes and associated subthemes, highlighting their temporal progression and points of interaction. The findings indicate that these themes are not discrete categories but dynamic and interdependent processes. Motivation shape engagement, as initial perceptions of relevance influence how actively learners participate in interactive tasks. In turn, engagement can reinforce motivation when learners experience success, relevance or enjoyment. Knowledge retention and transfer do not emerge automatically from engagement but are conditional and often fragile, requiring opportunities for application, reflection and scaffolding. When learning is successfully applied in practice, retention and transfer act as catalysts that strengthen intrinsic motivation and deepen subsequent engagement. Together, these relationships illustrate the cyclical and iterative nature of workplace learning, in which motivation, engagement and retention evolve and inform one another over time.

The sections that follow examine each theme in turn, supported by illustrative extracts that show how participants' experiences developed across the baseline, mid-module and post-module interviews.

At baseline, many participants express disengagement rooted in compliance-driven expectations. P1 reflects: “I have to do e-learning every year … I make a big sigh every December when I see it”, which highlights what others describe as compliance dread. Similarly, P5 characterises mandatory health and safety training as “just so tedious … next page, next page, read 100 pages, answer two questions”.

By the midpoint of the interviews, however, motivation often shifts when participants encounter interactivity. P3 explained: “I like engaging things, maybe a couple of quiz questions, then some information and videos or a variety of ways of teaching”. P2 reinforces this sense of relevance when describing a workplace first aid course: “I felt quite motivated … because some of the stuff I knew was going to be relevant to the workplace more so than others”.

By the final phase, motivation is framed less as obligation and more as application. P9 noted: “I probably retained more because it's my job, or at least because I had knowledge on it”, signalling a reframing from externally imposed to intrinsically meaningful.

Early expectations of interactivity are mixed, with some expressing scepticism. P4 comments: “Some bits feel like extra, optional reading. Is that an extra section I don't need to do? It's hard to know”. P1 similarly expresses frustration at “click-to-reveal” formats, calling them “like something you did in primary school”.

As learners progress, however, interactivity becomes valued when it provides immediate feedback or safe exploration. P6 reflects: “The most memorable bit for me has always been the branching scenarios, the what if. Even if I think I know the right answer, I go back and check what the wrong answer is”. P7 considers the impact of quizzes as engagement: “The quizzes are good because they test your knowledge and if you get one wrong you think, oh yeah … layered learning isn't it.”

By the post-module stage, interactivity had become central to knowledge construction. P2 recalls: “The active questions were the most engaging … You were having to constantly apply it, and it was encouraging you to make notes.” Likewise, P10 praises the flexibility of quizzes that allow reattempts, calling them both “reassuring and motivating”.

Initial expectations of retention are low. P8 admitted: “Normally I forget most of it by the time I'm back at my desk”. P3 echoed this challenge, stating: “I made notes, it wasn't something I instantly recalled, but they jogged my memory”. Several other participants share similar difficulties, with P5 confessing, “Have I remembered the information? No, not particularly well”. P4 apologises for their lack of recall, stating, “You'll have to excuse me, I don't remember it that well,” and P2 adds, “I can't lie, I can't remember much of the content”. With seven of the ten participants voicing comparable challenges, sustaining recall beyond a few weeks appears unrealistic.

By the mid-point, features such as summaries and diagrams begin to support knowledge retention. P6 highlights: “Definitely at the end I find it really useful for it to do like an executive summary … if you don't remember anything else, these are the key 5–6 points you need to remember”. Similarly, P5 recalls: “I remember the 5 Whys (process flow) … something insightful, something you can use”.

In the final interviews, participants describe the transfer into practice. P9 explains their adaptive strategy: “I always write at the top of the page what I'm learning … I can flip back to it if there's something that I need in the future”. P2 similarly emphasises application: “I always try to put knowledge to use to retain it … that way you digest it and put it into action. That means I'll hold it for longer”. These difficulties with sustaining recall highlight the need for evaluation approaches that capture learning processes as they occur, a point returned to in the discussion when considering the value of quasi-longitudinal design in workplace contexts.

These findings resonate with experiential and reflective learning theories, which show that knowledge is consolidated through iterative cycles of reflection, experimentation and use (Kolb, 1984; Schön, 1983). Participants' strategies, such as revisiting notes, drawing on summaries and applying concepts in workplace tasks, reflect these processes, reinforcing why retention and transfer emerge over time rather than at a single point.

Although individual trajectories varied, three cross-case patterns emerged. First, motivation consistently shifts from compliance-driven to relevance-driven. Second, engagement deepens as learners move from scepticism to experimentation to reflection. Third, retention becomes evident not through abstract recall but through embodied transfer into professional practice. Together, these findings illustrate the value of a temporal interpretivist approach in capturing the process of learning, rather than merely its endpoints. This, in turn, demonstrates that motivation, engagement and retention evolve in staged and interdependent ways, underscoring the value of quasi-longitudinal methods for tracing not only what changes but also exploring how such change occurs.

Collectively, these findings demonstrate that motivation, engagement and knowledge retention operate as interdependent and cyclical processes rather than linear stages. Learners' motivation shaped how they engaged with interactive elements, and meaningful engagement, in turn, reinforced both motivation and retention. Retention was sustained when learning was applied in practice, completing a feedback loop that renewed motivation. This temporal and iterative relationship forms the foundation of the interpretive model presented in the discussion that follows.

This study demonstrates that learner motivation, engagement and knowledge retention in workplace e-learning are not fixed states but dynamic processes that shift as learners progress through digital modules. By adopting a quasi-longitudinal design, the research captured not only what learners experienced at isolated points but also how their attitudes and strategies shifted across time. This demonstrates that quasi-longitudinal approaches are not merely pragmatic but can serve as a methodological bridge between cross-sectional surveys and fully longitudinal studies, offering a scalable template for temporal research in digital education. By structuring time as an analytical dimension, this study advances methodological debates beyond isolated case applications and towards a transferable approach that can be replicated across diverse e-learning contexts. These insights align with the broader principles of QLR, which foreground temporality, change and process as central objects of inquiry (Neale, 2019; Saldaña, 2013).

Much of the existing e-learning literature relies on cross-sectional evaluations, typically post-training surveys, one-off interviews or quantitative pre–post-test designs (Bergdahl et al., 2024; Huang et al., 2024). While these approaches provide efficient measurement, they tend to flatten the complexity of learner experience into static snapshots. The findings from this study highlight the risks of such a reduction. Learners described motivation rising and falling at different points of the module, engagement peaking in interactive scenarios but waning during dense text segments and knowledge retention becoming meaningful only once new skills were applied in practice. These nuances would have been invisible in a single time-point design. In this respect, the study not only corroborates but extends the calls by Neale (2019), Thomson and McLeod (2015), Harju et al. (2019) and Treanor et al. (2021) for digital education researchers to move beyond one-off evaluation metrics and embrace temporally sensitive designs that capture the evolving nature of learner experience.

These insights are embedded in the TEF-WEL framework, which operationalises temporal evaluation across phases and provides a structured model for capturing these dynamics in practice. The findings suggest that retention in workplace e-learning is fragile, with most participants reporting that content was quickly forgotten. This reinforces the value of adopting a quasi-longitudinal qualitative approach, which allows fading recall, reinforcement and transfer processes to be observed while they are still active.

The temporal focus of this study also offers new ways of linking empirical research to established learning theories. CLT (Sweller et al., 2011) emphasises that instructional design must carefully manage intrinsic, extraneous and germane load; however, these dimensions are not experienced in a uniform way. The quasi-longitudinal findings revealed that extraneous load was felt most strongly in the early stages of module engagement, while germane load increased as learners became more comfortable navigating scenarios. This view aligns with Kim et al. (2025), who advocate for a cognitive philosophy that recognises learning as a dynamic, context-sensitive process rather than a static cognitive transaction. Kolb's Experiential Learning Cycle (1984) provides a useful interpretive lens for understanding how reflective observation and abstract conceptualisation occur at different points in a learning journey. Supporting previous research, several participants described initial uncertainty giving way to confidence once they had tested their knowledge in branching scenarios, exemplifying the iterative, staged nature of experiential learning (Pimblett et al., 2025). By situating these processes within a temporal framework, the study bridges theoretical insights with empirical evidence that would have been obscured in static designs.

In practice, evaluations of workplace e-learning should move beyond end-of-course feedback and knowledge checks to include staged reflections or follow-up interviews that track learner development across weeks or months. This echoes findings from Honson et al. (2024), who similarly emphasise the importance of continuous, data-informed feedback in shaping digital learning experiences. Such approaches are better aligned with the Industry 5.0 emphasis on lifelong learning and human–machine collaboration, where learning is less about discrete modules and more about continuous adaptation in fluid environments.

The findings from this study highlight that workplace e-learning should not be treated as a single, static event, but rather as a process unfolding over time. Motivation, engagement and knowledge retention shifted at different points in the module, underscoring the need for evaluation and design practices that are equally dynamic. For practitioners, this generates three key implications.

First, organisations should move beyond reliance on end-of-course surveys or knowledge checks, which provide only a snapshot of learner experience. McCarthy et al. (2021) highlight the design–reality gap in e-learning platforms, showing how intended learning designs often diverge from actual learner practices; this underscores the limitation of static, end-point evaluations and strengthens the case for temporal approaches such as TEF-WEL, which can capture evolving learner experiences across time and during learning interventions. Embedding evaluation at multiple points enables the capture of rising and falling engagement, moments of struggle and the delayed relevance of knowledge. This temporal lens generates richer insight into where training is working and where targeted intervention is required.

Second, instructional design should explicitly align with the learner's temporal journey. Early stages may need stronger scaffolding to reduce extraneous cognitive load, mid-points may require interactive boosts to sustain motivation and post-training activities should emphasise application and reflection. Designing with these trajectories in mind allows for stronger transfer of learning into workplace practice. Third, this approach resonates with the broader Industry 5.0 agenda, which prioritises human-centric adaptability over mechanistic compliance. Engström et al. (2024) demonstrate how AI-driven systems are reshaping workplace learning by intertwining technological sensemaking with human interpretive practice, reinforcing the importance of human-centric and adaptive evaluation approaches such as TEF-WEL in Industry 5.0 contexts. Valtonen and Holopainen, 2025 similarly highlights that digital transitions can either foster or hinder employee well-being and resilience, depending on how they are managed, underscoring the value of evaluation frameworks like TEF-WEL that emphasise adaptability, confidence and human-centricity.

In practice, this means treating evaluation not as an administrative afterthought but as a developmental process that nurtures resilience, confidence and adaptability. Quasi-longitudinal methods, whether in the form of staged reflections, pulse surveys or follow-up interviews, offer a pragmatic way for organisations to embed these principles without the resource intensity of fully longitudinal studies. To operationalise these insights, this study develops and applies a practical model, the TEF-WEL (Figure 2). This framework operationalises these insights by mapping shifting motivation, evolving engagement and retention/transfer across distinct phases, providing organisations with a structured way to anticipate and support these dynamics. To the author's knowledge, TEF-WEL is the first framework to explicitly structure quasi-longitudinal evaluation within workplace e-learning contexts.

By applying TEF-WEL, practitioners can shift from static, end-point evaluations to dynamic, process-oriented approaches. Rather than solely assessing whether training was effective upon completion, TEF-WEL enables organisations to monitor how learning unfolds and to intervene mid-course when necessary. This facilitates more responsive instructional design, enhances inclusivity, particularly for neurodivergent or less digitally confident learners, and provides more robust evidence of impact for stakeholders.

TEF-WEL advances the earlier methodological approach proposed by Neale (2019) and Saldaña (2013) by adapting the principles of QLR to digital workplace learning. Unlike traditional models such as Kirkpatrick and Kirkpatrick's (2006) Four-Level Model, which privilege static post-training metrics, TEF-WEL foregrounds temporality, learner adaptability and process-oriented evaluation. By structuring evaluation across baseline, midpoint and post-module phases, it positions time as an analytic dimension rather than a contextual backdrop. As such, TEF-WEL functions not only as a practical tool but also as a methodological contribution, reframing digital learning evaluation as a temporal, interpretive process with potential application across sectors, training types and mixed-method designs.

This study advances methodological debates by offering a “right-sized” quasi-longitudinal design that addresses limitations in both cross-sectional and fully longitudinal approaches to workplace e-learning. Cross-sectional evaluations, as noted by Bergdahl et al. (2024) and Huang et al. (2024), flatten learner experience into static outcomes, while fully longitudinal studies remain impractical in organisational settings.

Building on Neale's (2019) argument that time should be treated as a core analytic dimension, this study demonstrates how temporality can be operationalised within the short cycles typical of workplace learning. It also extends Saldaña's (2013) longitudinal coding principles by applying them to compressed, practice-based learning trajectories. In doing so, the study shows how quasi-longitudinal qualitative methods can capture the fluid, iterative nature of learner experience and positions temporality as essential rather than peripheral in digital education research.

As a bounded case study, the findings are contextually situated and should be interpreted accordingly (Stake, 1995). Focussing on a single e-learning module with one learner group enabled depth but limited transferability across industries, training types and cultural contexts. The small sample (n = 10), while appropriate for QLR, further constrains the diversity of perspectives represented.

Data were generated through interviews, meaning behavioural data were not captured directly. Future research could integrate LA, digital trace data or observational methods to triangulate interpretive insights. Although the quasi-longitudinal design enabled temporal comparison, the compressed six-week timeframe prioritised short-term reflections.

Future research could extend the TEF-WEL framework through larger, multi-method and longer-span studies across varied organisational and cultural contexts. Baseline motivation and engagement may differ across national settings due to cultural norms and workplace values, as reflected in frameworks such as Hofstede's cultural dimensions (Hofstede, 2001), making cross-cultural applications particularly valuable. Delayed follow-ups could examine how learning trajectories evolve over months rather than weeks, while integration with LA or adaptive AI-driven systems offers a promising avenue for strengthening temporal evaluation.

This paper makes a threefold contribution. First, it advances the work of Neale (2019) and Treanor et al. (2021) by demonstrating how learner experiences in workplace e-learning unfold dynamically across short, practice-based timeframes, evidencing the value of temporality in a context that has received little qualitative longitudinal attention. Second, it extends Saldaña's (2013) principles of longitudinal qualitative analysis by showing how quasi-longitudinal methods can surface shifts in motivation, engagement and knowledge retention that remain invisible in single-time-point evaluations, while adapting his guidance to the compressed cycles typical of workplace training. Third, it offers a methodological contribution by reframing workplace e-learning evaluation as a dynamic, process-oriented activity rather than a static endpoint, operationalised through the TEF-WEL framework developed in this study.

By developing and applying the TEF-WEL framework, the study provides organisations with a practical model for integrating temporal insight into e-learning evaluation. This aligns with the human-centred commitments of Industry 5.0 and supports more adaptive and resilient approaches to workplace learning. More broadly, the paper positions time-sensitive, process-oriented evaluation as a necessary shift for research and organisational policy in digital education.