Does learning empowerment matter? Development and initial validation of the work-related learning empowerment scale

As continuous and lifelong learning through and for work have become increasingly important to meet the challenges of the changing world of work (Kraiger and Ford 2021), traditional approaches—often characterized by formal learning—are being replaced or complemented by more learner-centered forms of development. More emphasis has been put on flexible learning forms, such as informal (Decius et al. 2019) and self-regulated learning (Sitzmann and Ely 2011). The concept of New Learning has emerged to reflect this shift, drawing conceptual inspiration from New Work (Bergmann 2019; Schermuly 2019). Psychological empowerment has been used as the theoretical foundation of New Learning (Decius et al. 2022; Foelsing and Schmitz 2021). From this psychological perspective, New Learning places strong emphasis on the autonomy and motivation of the individual, highlighting the agency of the learner. Learners are viewed as psychologically empowered actors who perceive learning opportunities in their environment, form learning intentions, and engage in formal, informal, or self-regulated learning activities (Decius et al. 2022).

Taken together, New Learning understands learning as “a socially-embedded process in which the protean and empowered learner seeks and utilizes learning opportunities to engage in meaningful formal, informal and self-regulated learning to achieve subjectively valuable learning outcomes and personal growth. New Learning takes place in a new societal and organizational context brought forth by digitalization and characterized by dynamic change, uncertainty, and complexity” (Decius et al. 2022, p. 5232). Despite its growing practical relevance, New Learning remains a relatively new and still under-theorized concept. Therefore, we want to investigate how psychological empowerment can be specifically operationalized and measured in the domain of work-related learning.

Psychological empowerment with its dimensions meaning, competence, self-determination (sometimes also referred to as autonomy), and impact (Spreitzer 1995) has not only been scholarly studied across various organizational contexts as demonstrated by several meta-analyses (e.g., Seibert et al. 2011; Llorente-Alonso et al. 2024), but has also been increasingly recognized in organizational practice (Schermuly 2019; Foelsing and Schmitz 2021). Several studies have confirmed the originally proposed four-factor structure of Spreitzer’s empowerment scale (e.g., Spreitzer 1995; Albar et al. 2012; Boudrias et al. 2004). On the other hand, some studies also question this structure as other factor structures have also been found (e.g., Hancer et al. 2005; Fulford and Enz 1995). Additionally, Spreitzer herself formulated the question, if “the potency in empowerment [is] in the gestalt or the individual dimensions” (Spreitzer 2008, p. 64) and conceptualized empowerment as a second-order factor to account for the shared variance of the individual dimensions and connected the dimensions additively (Spreitzer 1995). Seibert et al.’s meta-analysis (2011) yielded evidence for the identification of psychological empowerment as a unitary second-order latent construct, leading the authors to conclude that psychological empowerment can be conceptualized as “a single second-order construct made up of the four cognitions of meaning, competence, self-determination, and impact” (p. 995).

Building on the studies of psychological empowerment and the theoretical foundations of Thomas and Velthouse (1990) as well as Spreitzer (1995), we define work-related learning empowerment (WRLE) as learners’ positive motivational beliefs regarding their work-related learning (such as, meaning, competence, self-determination, and impact). To date, very few studies have examined psychological empowerment in the context of learning (Frymier et al. 1996; Creon and Schermuly 2022). Only Frymier et al. (1996) investigated the factorial structure by applying the concept of psychological empowerment to the educational classroom setting. They based their conceptualization of student empowerment on Thomas and Velthouse’s (1990) model–which is similar to that by Spreitzer (1995)–with the four factors of meaningfulness, competence, choice (self-determination), and impact, and developed an instrument to assess psychological empowerment in the classroom. However, contrary to the assumptions, in their results choice did not emerge as a separate factor, yielding a three-factor solution, with choice items loading on the impact factor. They did not find a second-order factor. In this respect, the dimensional structure in learning contexts is not yet clearly understood.

Based on the above, we formulate our first research question:

To theorize the outcomes of WRLE, we draw on the sustainable career framework by De Vos et al. (2020), which conceptualizes career sustainability through a dynamic fit between person and context, reflected in the three dimensions of happiness, health, and productivity. This model offers a useful lens to structure the expected effects of WRLE, as it emphasizes the long-term alignment of individual needs, well-being, and performance—that are not only related to learning and career.

WRLE may contribute to sustainable career outcomes by enabling individuals to take ownership of their development. When employees feel psychologically empowered to learn, they experience their learning as meaningful and aligned with personal values, which supports job satisfaction—a core indicator of happiness (Fisher 2010). In terms of health, WRLE is expected to strengthen work engagement, a key psychological resource. Work engagement reflects a positive work-related state characterized by vigor, dedication, and absorption (Schaufeli et al. 2002). Engaged employees feel energized and connected to their work (Bakker et al. 2014), which signals resilience and mental well-being—core aspects of the health dimension. Furthermore, by promoting proactive learning behaviors—including formal, informal, and self-regulated learning—WRLE may enhance productivity and employability in dynamic work contexts (Decius et al. 2024b).

Therefore, we examine whether WRLE can be linked to positive outcomes across all three sustainability domains.

The data collection for this study was part of a larger multi-wave online study on the topic of work-related learning. The data was collected from June 2024 to February 2025 on the Prolific platform. Online panel data have been shown to yield comparable results to conventional data collection procedures and are economical for scale development (Lutz 2016; Porter et al. 2019; Walter et al. 2019).

For the present paper, we used data from two times of measurement (T1 = December 2024 and T2 = February 2025) which were 8 weeks apart. The sample used in the present study consisted of N = 258 working adults. Participants were compensated £ 3.33 (T1) resp. £ 2.96 (T2) (approx. $ 3) for their time. To ensure data quality, we excluded data from participants who did not pass the two attention checks (e.g., “Please select ‘agree’ to ensure data quality.”). In this final sample, 44.96% of participants identified as female, the mean age was 44.1 years (SD = 11.1 years). The sample was highly educated with 15.5% having a high school diploma or equivalent degree and 82.2% having at least a bachelor’s degree. The participants worked in services (30.23%), industry (21.71%), the public sector (18.61%) trade and commerce (10.10%) or indicated that they worked in another sector (18.21%).

We used JASP (version 0.19.1; JASP Team 2024) for most statistical analyses. Only structural equation modeling (SEM) was conducted with R (version 4.3.1) and lavaan (version 0.6-19). For the scale development we used a two-step approach. In the first step, we randomly divided the sample into two equal groups (each N = 129). With the first half of the sample, we performed an exploratory factor analysis with parallel analysis to check the factorial structure. We then used the second half of the sample to confirm the factorial structure. SEM with the WRLES (T1) and the outcomes examined (T2) was then calculated using the complete sample.

The descriptive statistics of the twelve items measuring WRLE in the T1 sample can be seen in the supplementary material at https://osf.io/am9rz. First, exploratory factor analysis (EFA) was calculated using the first half of the sample (N = 129). The Bartlett’s test (X² = 1641.058, df = 66, p < 0.001) as well as the Kaiser-Meyer-Olkin (KMO) index (overall KMO = 0.85, smallest individual KMO = 0.75) indicated that the sample is suitable for conducting an EFA (Hutcheson and Sofroniou 1999). Principal axis analysis with maximum likelihood estimation was used as a factor extraction method and promax rotation as an oblique rotation technique. A parallel analysis revealed three factors, as shown in Table 1. It can be seen that all three competence items composed a single factor (factor 3). Factor 2 is composed from all three meaning items (EP04_03, EP04_02, EP04_01) and one impact item (item EP04_10), with the meaning items loading significantly higher on the factor. Factor 1 is composed of all three self-determination items (EP04_08, EP04_09, EP04_07), which also have the highest factor loadings, and two meaning items (EP04_11 and EP04_12).

In a second step, confirmatory factor analysis (CFA) was conducted with the second half of the sample (N = 129) to test the factorial structure we found in the exploratory analysis (i.e., EFA, model 1 in Table 2). This method provides various fit indices to assess how well the proposed model represents the observed data. The fit indices were within the acceptable range with a CFI = 0.93 and an SRMR = 0.078 according to accepted cut-off values specified (Hu and Bentler 1999; Kline 2023), although the RMSEA = 0.138 was considerably above the limit. The reliability (McDonald’s ω) values were very good (Factor 1: ω = 0.947; Factor 2: ω = 0.956; Factor 3: ω = 0.925).

Additionally, we computed a four-factor model according to the factor structure assumed by Spreitzer (model 2 in Table 3). This resulted in a similar model fit with slightly better fit indices (CFI = 0.95, SRMR = 0.047, RMSEA = 0.118). Because Model 1 and Model 2 are nested, we compared these models statistically. Even though, Model 2 (AIC = 3886.09, BIC = 4006.20) had a significantly better model fit (∆χ² = 43.04, ∆df = 3, p < 0.001) than Model 1 (AIC = 3923.14, BIC = 4034.67) we preferred the three-factor model (Model 1) for two reasons: First, Model 1 is less complex und thus more parsimonious in terms of number of factors, and second, because further analysis revealed that some of the factors showed latent correlations above 0.80 (i.e., correlations between factor 4 with factors 1 and 3). This may indicate problems with discriminant validity (Rönkkö and Cho 2020).

Therefore, in order to obtain the final scale (i.e., the WRLES) with a clear factor structure that is also short and therefore application-friendly, we additionally calculated a three-factor model based on Model 1, in which we only included the three items with the highest loadings per scale (see factor loadings in Table 1). Thus, we excluded items EP04_11, EP04_12, and EP04_10 (i.e., the items of the impact factor according to Spreitzer) before calculating the model. These items contained phrases that expressed not only impact but also self-determination (e.g., EP04_11: “great deal of control,” EP04_12: “enabled me to have significant influence over what happened at work”) and were therefore not particularly discriminating. This model (model 3) showed a very good model fit, with all fit indices considered falling within the cutoff limits (CFI = 0.99, SRMR = 0.030, RMSEA = 0.067). Since the items of the three scales of the final factor structure correspond exactly to the items of Spreitzer’s concept, we name the factors accordingly: Factor 1 = Self-Determination, Factor 2 = Meaning, Factor 3 = Competence. The reliability values were also very good (Factor 1/WRLES Self-determination: ω = 0.941; Factor 2/WRLES Meaning: ω = 0.973; Factor 3/WRLES Competence: ω = 0.925).

For examining convergent and discriminant validity, we calculated the values for average variance extracted (AVE; Fornell and Larcker 1981) and shared variance (SV) (i.e., the inter-construct squared correlation) across the factors. If the AVE values are greater than the SV value between two factors of interest, this provides evidence for discriminant validity (Farrell 2010). As can be seen in Table 3, the AVE values of all three factors are at least 0.80, thus exceeding the 50% rule of thumb (Hair et al. 2010). We therefore can assume convergent validity of the WRLES. Furthermore, the SV values range between 0.30 and 0.37 (see Table 3). This means that all AVE values are greater than the corresponding SV estimates, which indicates discriminant validity (Hair et al. 2010).

In order to provide researchers with an additional measure that reflects general work-related learning empowerment and can be used in larger studies that do not allow for the nine-item version, we also offer a short scale consisting of three items. This makes sense for questions that focus less on the influence of learning empowerment with its sub-facets and more on the whole, as the correlation between the three factors is between 0.5 and 0.6. This measure consists of one item per factor, with the item with the highest loading in Model 1 (items with two asterisks, see Table 1) being selected. The three-item short Work-Related Learning Empowerment Scale (WRLES-3) has a reliability of ω = 0.767.

To investigate the predictive validity of the WRLES we used SEM. Prior investigation revealed that there were many missing values in the formal learning scale (between N = 89 for item 1 and N = 139 for item 5) because the participants indicated that they did not participate in such an activity. Therefore, we excluded this scale before calculating the model. A model with WRLES Self-determination, WRLES Meaning, and WRLES Competence at t1 and the t2 outcomes for happiness (i.e., job satisfaction), health (i.e., work engagement), and productivity (i.e., informal and self-regulated learning) had a good model fit (Χ² = 676.540, df = 255, p < 0.001; CFI = 0.93, RMSEA = 0.080; SRMR = 0.058). The loadings are displayed in Fig. 1. WRLES Self-determination had only significant relations with job satisfaction (β = 0.146, p = 0.028) and work engagement (β = 0.154, p = 0.015), WRLES Meaning had significant (p < 0.001) relations with all outcomes, while WRLES Competence had no significant relations with the outcomes.

Full size image

An additional SEM with WRLES‑3 as a predictor instead of the three WRLES factors revealed a poor model fit (Χ² = 542.856, df = 143, p < 0.001; CFI = 0.89, RMSEA = 0.104; SRMR = 0.065). Only after model modifications (allowing correlations of the informal learning items IL3 and IL4 as well as IL4 and IL5), the model fit was acceptable (Χ² = 399.110, df = 141, p < 0.001; CFI = 0.93, RMSEA = 0.084; SRMR = 0.055). In this model the WRLES‑3 had significant relations to job satisfaction (β = 0.626, p < 0.001), work engagement (β = 0.757, p < 0.001), informal learning (β = 0.655, p < 0.001), and self-regulated learning (β = 0.605, p < 0.001). Correlations of all variables can be seen in the online supplement (https://osf.io/am9rz).

This study makes several theoretical contributions. First, our results contribute to the theoretical foundation of New Learning by applying the concept of psychological empowerment to the field of work-related learning. Our findings specifically show that learning-related beliefs are related to relevant learning and career outcomes. Our findings substantiate the role of psychological empowerment as a foundational mechanism of New Learning (Decius et al. 2022). The findings also support the idea that psychological empowerment is not just relevant in general work settings but—driven by the WRLES meaning dimension—also plays a key role in work-related learning behavior, especially in dynamic and self-directed environments (Kauffeld et al. 2025). Second, we contribute to the ongoing debate on the dimensionality of psychological empowerment. While many studies treat it as a unitary construct (e.g., Seibert et al. 2011), our study shows first indications that, within the learning domain, a parsimonious three-factor structure—comprising self-determination, meaning, and competence—offers the best empirical and conceptual fit. The impact dimension, often central in work contexts (Spreitzer et al. 1997), did not emerge clearly. The effects of the WRLES subscale meaning are particularly noteworthy, as this was the only subscale found to be associated with all outcomes considered. This can be explained by the central role of meaning in maintaining motivation and commitment (e.g., Deci and Ryan 2000; Hackman and Oldham 1976). In the context of work-related learning, meaningfulness can help individuals connect learning activities to their personal identity and long-term career goals, thereby promoting sustainable employability and well-being. These findings align with Spreitzer’s (2008) view that empowerment may function both as a gestalt—we would add that in a learning context, this may be driven by the meaning dimension—and through distinct dimensions with different antecedents and outcomes. Third, the WRLES‑3 short scale allows work-related learning empowerment to be assessed as an overall construct, reflecting the theoretical idea of a higher-order empowerment factor. This dual approach—distinguishing subdimensions while enabling holistic assessment—can inform future research on motivation, learning, and self-development at work.

This study also has implications for human resource development practitioners in organizations. First, we offer starting points for practitioners what psychological empowerment can mean in the context of work-related learning and why it is important. In a dynamic world of work, the responsibility for learning will increasingly shift towards the learner, requiring a sense of psychological agency and motivation on the part of the learner (Beier et al. 2025; Decius et al. 2022). Second, the WRLES offers a practical tool to assess and foster psychological empowerment in learning processes. By identifying which dimensions of empowerment are lacking—be it self-determination, competence, or meaning—targeted interventions can be designed to strengthen learning motivation and initiative. The aspect of meaning can, for example, be addressed through a learner’s direct supervisor as well as on the organizational level through respective communication measures, such as including learning in value or strategy statements. Building learning competence could for example already be included into onboarding programs for new employees (Mitschelen and Kauffeld 2025), or supported through guided learning circles in which learners can explore and reflect on different learning strategies. These individual-level interventions can complement organizational-level interventions, such as building communities of practice (see Kortsch et al. 2023a) or fostering a strong learning culture (Kortsch and Kauffeld 2019). A learning culture can also compensate for a lack of psychological empowerment in New Learning (Kortsch et al. 2023b). This underscores the importance of cultivating a supportive learning culture as a key lever for work-related learning.

This study has several limitations. First, the sample consisted mainly of highly educated participants from an online panel, which may limit the generalizability of the findings. Future research should aim to replicate the results across different occupational groups and cultural contexts. Second, the impact dimension, originally proposed by Spreitzer (1995), could not be clearly distinguished in the learning context and was excluded from the final model. However, this may be due to conceptual overlaps with self-determination items, which future studies should investigate more systematically. The question also arises here as to whether learning behavior and learning transfer are confounded in these items. Third, due to a high number of missing responses, formal learning had to be excluded from the outcome analysis. Although bivariate correlations suggest meaningful associations, structural modeling was not possible. One plausible reason is the short interval of eight weeks between measurement points, during which most participants did not engage in formal learning. Previous research has also reflected on this well-known measurement problem of formal learning, and we support the call for the development of a more accurate operationalization of formal learning (Decius et al. 2024b). Until such advancements are made, future studies should consider extending the time frames to better capture such learning activities. Finally, it should be noted that the study focused exclusively on work-related (psychological) learning empowerment. Future studies should also incorporate the original scale for psychological empowerment in order to demonstrate the incremental validity of the WRLES compared to general psychological empowerment.

In terms of future research directions, psychological empowerment could be further examined in structured learning settings and across different learning formats. Moreover, the role of psychological empowerment in cognitive and metacognitive learning processes deserves closer attention. The Self-regulated Informal Learning Cycle (SILC; Decius et al. 2024a) offers a promising framework for this, as psychological empowerment may facilitate the intentional and reflective phases of informal learning. An open question remains regarding which levers organizations can utilize to enhance WRLE. One potential approach, aligned with previous research (Decius et al. 2024c), is to consider the suitability of different learning method clusters—emerging from self-directed e‑learning as well as classroom-based learning in both physical and virtual settings—from a self-determination perspective as a starting point. Finally, the topic of environmental sustainability in learning contexts—so far largely neglected—could be explored more deeply. As sustainable behavior becomes more relevant in the workplace (Zacher et al. 2023), future studies might examine how empowered learners engage in “green learning” and contribute to long-term organizational and ecological goals (Kauffeld et al. 2025; Lorber et al. 2025).