Which way of learning benefits your career? The role of different forms of work-related learning for different types of perceived employability

ABSTRACT

Building and maintaining employability is important for employees and organizations in a changing world of work. Research considered work-related learning as an antecedent of employability but provided ambiguous findings. We attribute this inconsistency to the multidimensionality of the constructs involved. Drawing on Conservation of Resources Theory, we investigated the links of three different forms of work-related learning (i.e., informal, formal, and self-regulated learning) with both internal (i.e., within the organization) versus external (i.e., in the general labour market) and qualitative (i.e., finding a better job) versus quantitative (i.e., finding another job) perceived employability. Our two-wave study (N = 307 employees) showed that informal learning positively relates to qualitative and quantitative internal employability, but not to any type of external employability. Conversely, self-regulated learning is positively linked to qualitative external employability, but not to quantitative external or any type of internal employability. Formal learning (i.e., training and workshops) surprisingly did not contribute to employability at all. Furthermore, we did not find that learning forms conceptually associated with more resource investment regarding effort would show stronger relations with employability. We discuss learning-related strategies to foster employability and address the missing association between formal learning and employability, calling for more sophisticated measures of formal learning.

KEYWORDS:

• Employability
• work-related learning
• informal learning
• training
• self-regulated learning

Introduction

Just a few decades ago, it was common to spend one’s entire professional career with one employer. Changes in the world of work, especially driven by digitalization and globalization, have resulted in the need for organizations and employees to be more flexible today (Regan & Delaney, 2011). Employees are increasingly responsible for their careers, which is reflected in psychological processes of identity awareness, adaptability, and agency (Hall et al., 2018). The modern world of work is knowledge-driven, so employees never stop learning and must continue to develop (Cascio & Montealegre, 2016; Noe et al., 2014). This is an important way they can maintain or increase their own employability in the long term, resulting in positive personal and organizational outcomes (e.g., Hogan et al., 2013; Philippaers et al., 2017).

Employability refers to “an individual’s chance of getting a job in the internal and/or external labor market” (Forrier & Sels, 2003, p. 106). We focus in our study on perceived employability, which is the “individual’s perception of available employment opportunities” (Forrier et al., 2015, p. 57). The advantage of perceived employability over objective employability measures is that employees consider both individual capabilities and contextual factors such as labour market conditions when assessing their own perceived employability (Berntson et al., 2006; De Cuyper et al., 2011; Harari et al., 2021). Furthermore, research argues that employees act as “free agents” based on their perceptions, and thus perceived employability affects their labour market behaviour (Forrier et al., 2015; Houben et al., 2021; Vanhercke et al., 2014).

Meta-analytically, prior work-relevant knowledge, skills, and abilities (KSAs) have been identified as the strongest predictor of employability¹—however, KSA development shows only a moderate relationship with employability (Harari et al., 2021). Researchers nevertheless assumed that employability can be developed through work-related learning (Houben et al., 2021; Nelissen et al., 2017; Van der Heijden et al., 2009). One rationale is based on Conservation of Resources (COR) Theory (Hobfoll, 1989). From this perspective, employability is considered an important occupational resource, whereas work-related learning can be viewed as the process of resource investment to maintain or improve one’s employability (Forrier et al., 2018; Van Hootegem et al., 2019; Vanhercke et al., 2015).

However, neither employability nor work-related learning are homogeneous constructs, and effects may differ depending on which facets are investigated: Work-related learning can be classified into three distinct learning forms: informal, formal, and self-regulated learning (Cerasoli et al., 2018; Decius, 2020; Kyndt & Baert, 2013; Sitzmann & Ely, 2011). Employability can be divided into internal (i.e., perceived employment opportunities with the current employer) and external employability (i.e., perceived employment opportunities with another employer) (e.g., Houben et al., 2021; see Rothwell & Arnold, 2007, for the conceptual basis of this categorization). A second distinction can be made between quantitative (i.e., finding some job) and qualitative employability (i.e., finding a better job than the current one; De Cuyper & De Witte, 2010) which has received less attention in the literature. Research on the role of learning for employability has largely neglected these distinctions (for exceptions see Van der Heijden et al., 2009, who distinguished between formal and informal learning, and Houben et al., 2021, who distinguished between formal and informal learning, as well as internal and external employability).

This study aims to investigate differential relationships between informal, formal, and self-regulated learning with the four types of employability according to the distinctions between quantitative versus qualitative and interval versus external employability, respectively. While all forms of learning may generally promote employability, they may do so to different degrees, because (a) the different types of learning may reflect different resource mechanisms and differ in the required investment effort (see Cerasoli et al., 2018; S. Chung et al., 2021; Sitzmann & Ely, 2011), and (b) though both represent resources, the different types of employability may differ in their comparative value to the individual – whereas quantitative employability reflects maintaining the status quo (i.e., resource conservation), qualitative employability signals a potential improvement from one’s current job and thus additional resource gain. We thus generally expect different forms of work-related learning to relate to all four types of employability (i.e., quantitative, qualitative, internal, and external). But we also expect the more resource-intensive forms of learning (e.g., self-regulated learning) to relate more strongly to qualitative employability.

Our study offers three main contributions to the literature: First, by differentiating several types of employability, we offer a more nuanced understanding of the extent to which certain dimensions of employability are influenced by work-related learning. Especially the distinction between qualitative and quantitative employability will provide new insights into which type of learning contributes to expectations of additional resource gain (i.e., qualitative employability) in comparison to resource conservation (see Halbesleben et al., 2014). Differential relationships with internal versus external employability may be less straightforward to predict, but may also point to different resource mechanisms, as general occupational expertise tends to relate more strongly to external employability, whereas job-specific skills tend to cater to internal employability (see Van Harten et al., 2022).

Second, by differentiating several forms of work-related learning, we extend the view of which form of learning contributes distinctively to employability. Based on COR Theory (Hobfoll, 1989), we argue that the three work-related learning forms differ in effort and goal-directedness. We hypothesize that forms of learning associated with higher effort will lead to greater employability gains because people are willing to invest more resources if they expect to gain more valuable resources in return (Halbesleben et al., 2014).

Third, our study considers both work-related learning and employability holistically with methodological rigour. We aim to further unravel the link between work-related learning and employability, using a structural equation model to simultaneously account for the different facets and thus their conceptual overlap and interaction (see Bednall & Sanders, 2017; De Cuyper & De Witte, 2010; S. Richter et al., 2020). Moreover, we take into account that quantitative employability is conceptually included in qualitative employability (De Cuyper & De Witte, 2010) by separating the respective effects of work-related learning on qualitative employability versus quantitative employability, and further distinguish between internal and external employability. Using a two-wave design, we contribute to a small number of longitudinal studies on work-related learning and employability (e.g., Houben et al., 2021; Nelissen et al., 2017).

In terms of practical implications, differential mechanisms associated with different forms of learning would suggest varying levers for employability development, as for example meta-analyses show different predictors for the work-related learning forms (see Blume et al., 2010; Cerasoli et al., 2018; Sitzmann & Ely, 2011).

Work-related learning and employability: a conservation of resources perspective

Our study is based on the COR Theory (Hobfoll, 1989). COR Theory views individual well-being as a function of resources, i.e., objects, conditions, energies, or personal characteristics that people value or that serve to acquire valued resources (Hobfoll, 1989). COR Theory states that people strive to obtain, protect, and accumulate resources, but to do so, they must invest resources (Hobfoll et al., 2018). Halbesleben et al. (2014) further specified the definition of resources by defining their value by the extent to which individuals perceive them as helpful in achieving their goals. From this perspective, employability can be seen as an important occupational resource, which helps individuals to secure and/or improve access to employment and develop their careers, whereas work-related learning can be viewed as the goal-oriented process of resource investment to maintain or improve one’s employability (De Vos et al., 2020; Forrier et al., 2018; Van Hootegem et al., 2019; Vanhercke et al., 2015).

Work-related learning as an umbrella term includes informal, formal, and self-regulated learning (Decius, 2020). Informal learning occurs problem-induced during the work process (Marsick & Volpe, 1999). The learning content results from particular challenges of the work task and is therefore job-specific. Informal learning is a multidimensional construct that includes facets such as trying and applying own ideas, model learning, feedback, and reflection (Decius, Knappstein, et al., 2021). For instance, employees learn about procedures handled within the organization or special product knowledge. Informal learning fosters resources conducive to employability such as job-specific skills, job involvement, or self-efficacy (Cerasoli et al., 2018; Decius, Schaper, et al., 2021). Formal learning is mostly standardized and often provided by external instructors or institutions (Cerasoli et al., 2018). Structure, learning objectives, time and location are usually predetermined (Kyndt & Baert, 2013). In addition to courses on methodological content, for example on how to use standard software, training can also teach generic competencies such as communication or presentation skills. Self-regulated learning is characterized by learners setting their own learning goals, monitoring their achievement, and counteracting deviations (Sitzmann & Ely, 2011; Zimmerman, 1990). Thus, the learning process is more autonomous than in the case of pre-structured formal learning and problem-induced informal learning. However, the learner has to pay more attention to suppressing learning distractions and sticking to their own learning goals, because neither a teacher nor an immediate situational demand ensures this (Decius & Decius, 2022).

The nature of employability

Employability, typically defined as the individual ability to find and maintain employment, is a complex multidimensional construct (Forrier & Sels, 2003; Fugate et al., 2004; Hillage & Pollard, 1998). Employability research can be broadly categorized into input-based and outcome-based approaches (Forrier et al., 2018): Input-based research conceptualizes employability in terms of characteristics of the individual, focusing on personal assets and resources thought to increase people’s job chances (e.g., competencies or attitudes). Outcome-based approaches aim to capture these job chances directly. Furthermore, objective indicators (e.g., job transitions) can be distinguished from subjective approaches measuring employees’ individual beliefs about their employment chances (Forrier et al., 2018 see also Berntson et al., 2008). In this study, we follow the outcome-based approach and investigate employees’ perceived employability. These perceptions of employability reflect people’s assessment of their long-term employment prospects on both internal and external labour markets beyond the security of their current job (H. Chung & Van Oorschot, 2011). As such, the concept of employability is inherently tied to sustainable careers: Identifying antecedents of employability perceptions advances our understanding of an essential personal resource to ensure happy, healthy, and productive careers in line with employees’ values and needs (see De Vos et al., 2020; Kirves et al., 2014).

We consider four different types of employability along two dimensions (see Table 1; De Cuyper & De Witte, 2010). First, employability can be differentiated into a quantitative and qualitative component: Quantitative employability refers to the perceived chances of finding a similar job or another job in general. Given volatile labour markets and the primacy of loss principle in COR theory (Hobfoll et al., 2018), feeling secure about one’s access to employment, that is, to conserve the resources provided through work by avoiding unemployment, is an important protective resource for many employees. Accordingly, employability plays an important role in the context of job insecurity (e.g., De Cuyper et al., 2012; Van Hootegem et al., 2019). Additionally, qualitative employability may be an even more valuable resource in the context of career development, because it refers to one’s chance of finding a better job in the future. What constitutes a “better” job is in the eye of the beholder and does not necessarily imply upward career advancement in the traditional sense (see De Cuyper & De Witte, 2010): Qualitative employability can also refer to potential improvements in resources other than status or salary, such as working conditions (e.g., more flexible hours or a shorter commute compared to the current job). Thus, qualitative employability should not be equated with upward job mobility. Rather, perceptions of qualitative employability entail employees’ expectations of gaining additional resources compared to their current job and may relate to both upward and lateral job changes. The two types of employability are logically connected, as qualitative employability is incremental to quantitative employability: In order to see better employment opportunities, workers need to perceive general employment opportunities in the first place. It should be noted that actual transitions to a “better” job can be more ambivalent and associated with stress and simultaneous resource drain as employees adapt to new environments and demands (see Fletcher & French, 2021; Lipshits-Braziler & Gati, 2019). However, perceived employability as defined in this paper does not refer to actual job transitions, but to employees’ expectations thereof. We therefore assume that when reflecting about their hypothetical chances of a better job, employees will largely focus on the potential for resource gain, especially when these reflections are contrasted with quantitative employability expectations.

Table 1. Overview of four employability types.

	Quantitative	Qualitative
Internal	I could easily find another job in my company.	I could easily find a better job in my company.
External	I could easily find another job outside my company.	I could easily find a better job outside my company.

Note. Typology based on De Cuyper and De Witte (2011), p. 155.

Additionally, the literature distinguishes between employability in the internal versus external labour market (Forrier & Sels, 2003; Rothwell & Arnold, 2007). The focus of internal employability lies on the individual’s perceived job opportunities inside the current company and is thus more relevant for employees’ in-house career (e.g., perceived opportunities to find a similar job in a different branch or department, or to advance in the internal hierarchy). On the other hand, external employability refers to the individual’s perceived job opportunities outside the current company and is thus more relevant for employees’ occupational mobility. Unlike the quantitative-qualitative dimension, internal and external employability are not conditional upon one another, and one is not by definition associated with better employment than the other. Nevertheless, it is worthwhile to distinguish between internal and external employability, because the two have been shown to be empirically distinct and relate differently to antecedents and consequences (Forrier et al., 2015; Houben et al., 2021; Van den Broeck et al., 2014; Van Harten et al., 2022). For example, in a study with Belgian employees, internal employability predicted subsequent internal job transitions while decreasing the likelihood of external job transitions, whereas external employability predicted external job transitions only (Forrier et al., 2015). Given our theoretical reasoning and previous research, we therefore differentiate between four types of employability along the dimensions quantitative-qualitative and internal-external, respectively, to uncover potentially differential associations with different types of work-related learning.

Empirical findings on the work-related learning – employability link

Regarding previous studies on the relationship of work-related learning and employability, some have investigated the links between learning and input-based operationalizations of employability, that is, personal assets and characteristics beneficial for one’s career (e.g., corporate sense or occupational expertise; Crans et al., 2021; Van der Heijden et al., 2009; Van der Klink et al., 2014). To the best of our knowledge, only two studies have investigated the relationships between work-related learning and employability in terms of employees’ perceived job chances, which is the focus of this paper (i.e., outcome-based; see Forrier et al., 2018): Considering informal learning, Wiśniewska et al. (2021) reported an association of some learning facets (e.g., feedback) on employability, while for others (e.g., learning from one’s mistakes and observing others) no link was found. Houben et al. (2021) differentiated between informal and formal learning, as well as between internal and external employability in a three-wave study. They found reciprocal effects between formal learning and internal employability, but not with regard to external employability. Furthermore, they found no effect of informal learning on either internal or external employability.

To the best of our knowledge, the impact of self-regulated learning on employability has not yet been studied. The only study approaching this direction is Raemdonck et al. (2012), reporting that higher levels of self-directedness in learning (which overlaps with self-regulated and informal learning) correspond with a higher likelihood of advancing to a superior position (vertical occupational mobility, which overlaps with employability). Other studies on employability and career research do not focus on single forms of work-related learning. Juhdi et al. (2010), for instance, examined the link of career management practices (such as mentoring and coaching) and training on employability. However, no clear distinction can be made here between formal, informal, and self-regulated learning. In sum, the few existing studies suggest that the specific links between different forms of work-related learning and employability are underexplored, and that the ambiguous evidence requires further research.

Differential relationships regarding the different types of employability

We expect that successful work-related learning results in increased employability: When employees invest time and effort in learning, they expand their professional knowledge and skills and thus gain individual capabilities. However, differentiating between different types of learning and employability, respectively, we argue that a COR perspective can help develop a more nuanced understanding of how and when work-related learning contributes to employability perceptions.

On the one hand, different types of learning cater to different types of resources and differ in their extent of goal-directedness and resource investment for the employee (see Cerasoli et al., 2018; S. Chung et al., 2021; Sitzmann & Ely, 2011). On the other hand, different forms of employability likely differ in their value as resources: Although resources are generally thought to have positive effects, people do differentiate between more or less valuable resources, depending on what they perceive as helpful in attaining a given goal (Halbesleben et al., 2014). Following this logic, both quantitative and qualitative employability are important resources for employees, but qualitative employability may be even more valuable, because it relates to expectations of resource gain by finding better jobs in contrast to a job in general (see De Cuyper & De Witte, 2010). Quantitative employability can thus be understood as resource conservation as it reflects people’s ability to protect the status quo, especially when their job may be insecure (De Cuyper et al., 2012; Silla et al., 2009). Given the increasing requirements for continuous learning in today’s working world (Cascio & Montealegre, 2016; Noe et al., 2014), one can argue work-related learning is necessary for employees to keep their skills up to date and thus maintain their access to employment in general (i.e., quantitative employability). Distinguishing and controlling for differences between quantitative and qualitative employability thus provides insights when and how work-related learning leads to expectations of resource gain, beyond conservation: As employees expand their KSAs or learn new skills, this would be reflected in an effect on qualitative employability, because qualitative employability signals a potential for improvement (De Cuyper & De Witte, 2010). Empirical results on this are not yet available, as research on work-related learning and employability has not distinguished between the quantitative and qualitative dimension.

In contrast to quantitative versus qualitative employability, it is not straightforward to argue why either internal or external employability should be more or less valuable in general. Both internal and external employability may conceivably entail different kinds of resource gain: When internal employability is high, employees feel secure about maintaining or even improving their job with their current employer, without losing the benefits of organizational tenure and embeddedness. On the other hand, externally employable employees are less dependent on a single employer, thus fostering security in the broader labour market, and in the long run of their career (H. Chung & Van Oorschot, 2011). At the same time, external employability can improve employees’ internal bargaining power when the organization wants to retain them (see Ng & Feldman, 2007). Empirically distinguishing between internal versus external employability is thus warranted, because the two types reflect distinct constructs with differential relationships to antecedents and outcomes (Forrier et al., 2015; Houben et al., 2021; Van den Broeck et al., 2014; Van Harten et al., 2022). For instance, general occupational expertise tends to relate more strongly to external employability, whereas job-specific skills tend to relate more strongly to internal employability (see Van Harten et al., 2022). We therefore distinguish between the four types of employability and hypothesize the following for the different types of work-related learning:

H1:

Informal learning positively relates to a) quantitative internal employability, b) quantitative external employability, c) qualitative internal employability, and d) qualitative external employability.

H2:

Formal learning positively relates to a) quantitative internal employability, b) quantitative external employability, c) qualitative internal employability, and d) qualitative external employability.

H3:

Self-regulated learning positively relates to a) quantitative internal employability, b) quantitative external employability, c) qualitative internal employability, and d) qualitative external employability.

Differential relationships regarding the different forms of work-related learning

As we argued above, qualitative employability seems to be more valuable to the employee than quantitative employability, as it refers to the expectations of a resource gain through the perception of being able to find a better job in contrast to a job in general. Because of this high relevance of qualitative employability, we focus on this employability type when we consider differential effects of different forms of work-related learning on employability in the following. In terms of work-related learning, the different learning forms differ in their extent of goal-directedness and resource investment for the employee (see Cerasoli et al., 2018; S. Chung et al., 2021; Sitzmann & Ely, 2011). According to COR Theory, people are willing to invest more effort in resource acquisition the more valuable resources they expect to gain (Halbesleben et al., 2014). To understand when and how learning benefits employees’ expectations of resource gain beyond resource conservation, we also test which type of learning contributes most strongly to qualitative employability in particular. We therefore argue that the more effortful and goal-directed types of learning (i.e., formal and self-regulated) will contribute more strongly to qualitative employability compared to the less effortful informal learning: Though controlled by the learner, informal learning happens spontaneously, directed towards immediate problems and short-term goals as they occur on the job, and thus requires less deliberate effort, time or goal-setting from the employee than other forms of learning (Decius, 2020; Marsick & Volpe, 1999).

Regarding formal learning, we expect employees to participate only if the content goes beyond their previous knowledge because training interventions are time-consuming and costly (Blume et al., 2010). In other words, the goal of training is a more strategically-oriented qualitative improvement of competencies which requires employees to invest more time and deliberate effort than informal learning (see Kyndt & Baert, 2013). Thus, formal learning should foster resources perceived as more conducive to employees’ career goals than informal learning, and consequently have a stronger effect on qualitative employability.

Regarding self-regulated learning, we assume that employees who have freedom over their goal setting will primarily choose goals that lead to a qualitative improvement in their competencies. Since self-regulated learning is cognitively demanding and requires more deliberate effort than pre-structured formal trainings, COR Theory would predict that an employee will engage only if there is a reasonable expectation of a positive return (Halbesleben et al., 2014). This would be the case if the employee expects higher chances of getting a better job after learning, but perhaps not just a job in general. In comparison to formal learning, the outcomes of self-regulated learning should be perceived as even more conducive to goal attainment, because it serves self-selected goals, which should thus be of particular value for the employee (Locke & Latham, 2002).

To sum up, we expect that the relationships specified in H1–H3 will be most strongly pronounced for self-regulated learning, because self-regulated learning requires the most deliberate effort in investing resources such as time and energy, followed by formal learning, then informal learning. As the invested effort in resource acquisition can signal these resources’ value (Halbesleben et al., 2014), we hypothesize the following:

H4:

Self-regulated learning will have a stronger positive association with qualitative a) internal and b) external employability than formal learning.

H5:

Formal learning will have a stronger positive association with qualitative a) internal and b) external employability than informal learning.

When investigating these hypotheses, we also include quantitative employability in the model to assess the incremental value of learning for qualitative employability, because qualitative employability implies quantitative employability. Regarding internal versus external employability, existing theory or research does not suggest clear predictions as to which form should be more valuable to employees. Nevertheless, we investigate the internal-external dimension separately, because differential effects may still occur and indicate differential resource mechanisms (Houben et al., 2021; Van den Broeck et al., 2014; Van Harten et al., 2022). Figure 1 provides a graphical overview of all assumed hypotheses.

Figure 1. Conceptual Model of the Relationships Between Work-related Learning and Employability.

Note. H = Hypothesis. H1, H2, and H3 refer to the impact of work-related learning on employability. H4 and H5 refer to the expected differences in the strength of these associations, meaning that the larger the font size of the respective work-related learning construct, the larger the effect.

Methods

Sample

We surveyed a final sample of 307 employees in Germany (31.9% male, 68.1% female; age: M = 40.6 years, SD = 12.9 years, Min = 20 years, Max = 64 years, 2 individuals refused to state their age) on a voluntary basis using an online questionnaire, between July and December 2021. Participants were acquired through the second author’s professional network. A broad range of industries where the participants came from was ensured; the biggest share came from people in the care industry (41.7%). The remaining people came from mixed office settings (58.3%). The reason behind this wide approach was to come up with a diverse sample and to be able to generalize the results as much as possible. Consistent with the applicable ethical guidelines of the authors’ academic institutions, participants were informed that their participation was voluntary and anonymous.

Considering the educational background, 30.3% finished a vocational education, and 69.7% got a graduate academic degree (including two individuals who are still studying). For the sake of anonymity, we established the following seven categories to measure work experience: less than 3 years (23.1%), 3–5 years (11.1%), 6–10 years (9.1%), 11–15 years (11.1%), 16–20 years (9.8%), 21–25 years (9.4%), and more than 25 years (26.4%). 59.9% worked full-time. On average, the weekly working time in a home office setting was 11.3 hours (SD = 15.1 hours, Min = 0, Max = 60 hours).

Each participant took part twice in the survey, with an interval of approximately four weeks between both identical surveys. To match the answers from time 1 and time 2, we used a pseudonymization code asking for the first two letters of the first name of the participant’s mother, the first two letters of the participant’s father and the first two letters of the own place of birth. 67% of the participants at time 2 filled in the code completely and correctly relatable to the code at time 1, resulting in the final sample of N = 307. The pseudonymization code was requested at the end of both questionnaires. Incomplete questionnaires could therefore not be matched and had to be excluded. We thus had no issues with missing values, also because the online questionnaire was programmed in a way that participants needed to respond to all items.

Measures

Informal learning was measured with the 24-item scale from Decius et al. (2019), using a 6-point Likert scale: 1 = Not agree at all, 2 = Largely not agree, 3 = Rather not agree, 4 = Rather agree, 5 = Largely agree, 6 = Fully agree. A sample item is “I ask my colleagues about the methods and tricks they use at work” (subscale vicarious feedback).

Formal learning was measured with the 4-item scale from Grosemans et al. (2020), using the original 7-point Likert scale: 1 = Never, 2 = Rarely – Once or twice in the previous six months, 3 = Occasionally – Monthly, 4 = Often – A few times each month, 5 = Very often – weekly, 6 = Very often – A few times each week, 7 = On a daily basis. Due to pandemic-related widespread virtual work and cutting back on formal face-to-face formats (Hurley, 2021), the items were provided with an extension, e.g., “I take part (virtually or in presence) in professional workshops”.

Self-regulated learning was measured with the 6-item subscale “task strategies” from the Self-Regulated Learning at Work Questionnaire by Fontana et al. (2015), using the original 5-point Likert scale: 1 = not at all true for me, 2 = sometimes true for me, 3 = quite true for me, 4 = true for me, 5 = very true for me. A sample item is “I write down a plan to describe how I hope to achieve my learning goals”.

Employability was measured with the 16-item scale (four subscales) by De Cuyper and De Witte (2011), using a 5-point Likert scale: 1 = Do not agree at all, 2 = Disagree rather, 3 = Partly/partly, 4 = Agree rather, 5 = Agree completely. A sample item for internal quantitative employability is “I am optimistic that I would find another job with this employer, if I looked for one”. External employability items used “elsewhere” instead of “with this employer”; qualitative employability items used “a better job” instead of “another job.”

English scales were translated into German using a translation-backtranslation process; this applied to all scales except informal learning, where a bilingual version is available.

Controls

We controlled for gender (1 = male, 2 = female), age as a continuous variable, and education (1 = completed vocational training, 2 = academic degree) as typical structural determinants of employability perceptions, which may also affect employees’ opportunities for workplace learning (see Berntson et al., 2006; Harari et al., 2021; Kyndt & Baert, 2013; Kyndt et al., 2009). Since we conducted our survey across different occupations during the COVID-19 pandemic and the extent of home office opportunities can vary greatly among these, we also included home office hours per week (from the second survey) and employment status (1 = part-time, 2 = full-time) as controls.

Analysis strategy

For hypothesis testing, we used structural equation modelling (SEM) to holistically examine the associations between the three work-related learning forms (time 1) and the four employability types (time 2) in one model. As a robustness check, we computed three latent cross-lagged panel (CLP) models² for informal, formal, and self-regulated learning with the four employability types respectively to control for reversed causality (i.e., the possible effect of employability on learning). In each of these models, we modelled the effects of the respective learning form on the four employability types while controlling for reversed causality paths, i.e., the effects of the four employability types on the learning form.

Using the package lavaan (version 0.6–9; Rosseel, 2012) in R (version 4.1.0), we performed latent SEM calculations, with effects-coding as scaling method. The manifest items represented the structural model indicators, with exception of informal learning with 24 items, where we averaged six items each according to the Dynamic Model of Informal Learning (Tannenbaum et al., 2010) and thus used four item means as indicators. The ratio of participants (N = 307) to indicators (36, without controls) is close to the 10:1 value recommended by Hair et al. (2010), i.e., 8.5:1.

We considered skewness and kurtosis of indicators as proxies for normal distribution. As the values at both measurement time points for skewness (−0.98 to 1.28) and kurtosis (−1.08 to 1.94) are in a small range (only the fourth item of formal learning deviates, skewness = 1.45, kurtosis = 2.97), we can assume approximately normal distribution (Kline, 2016), except for formal learning.³ We therefore used the maximum likelihood method for SEM estimation.

We tested hypotheses 1, 2, and 3 in the full model including five control variables (see controls section). For hypotheses 4 and 5, which involve comparisons of two path coefficients each, we computed additional models based on the full model, adding the constraint of equating these two paths. If the model fit of the additional model is significantly worse than the initial model with freely estimated paths, a significant difference between the path coefficients can be assumed. We would the reject the assumption that the path coefficients are equal.

In evaluating the models, we relied on the following parameters recommended by Kline (2016): Comparative Fit Index (CFI), Standardized Root Mean Square (SRMR), and Root Mean Square Error of Approximation (RMSEA). According to Weston and Gore (2006), a decent model should have at least the following values: CFI ≥.90, RMSEA ≤.10, and SRMR ≤.10; a good model fit can be assumed for these values: CFI ≥.95, RMSEA ≤.06, and SRMR ≤.08. We compared the model fit of the initial model and the additional models using Chi² difference test (Kline, 2016).

Results

The descriptive statistics of both surveys are presented in Table 2. Below the boldface diagonal are the correlations of the first measurement point, above the diagonal the correlations of the second measurement point. The reliability values (McDonald’s Omega, see diagonal in Table 2) are in a satisfactory range between .82 and .95, with exception of self-regulated learning (.72/.75).

Table 2. Descriptive statistics and correlations among time 1 variables (beneath the diagonal) and time 2 variables (above the diagonal).

	Variable	M t1	M t2	SD t1	SD t2	Skew t1	Skew t2	Kurtosis t1	Kurtosis t2	1	2	3	4	5	6	7	8	9	10	11	12
1	Informal learning	4.35	4.35	0.55	0.59	−0.25	−0.05	0.07	0.00	.88/.90	.19***	.50***	.19***	.22***	.10	.27***	.04	−.04	−.25***	.17**	.01
2	Formal learning	2.39	2.40	0.99	0.98	1.37	1.46	2.84	3.31	.35***	.87/.92	.14	.07	.09	.01	.07	.18**	−.08	−.14*	.24***	.13*
3	Self-regulated learning	3.19	3.21	0.66	0.70	−0.04	0.08	0.07	−0.12	.42***	.33***	.72/.75	.16**	.13*	.13*	.14*	−.01	.01	−.04	.14*	.02
4	Internal qualitative emp.	2.48	2.50	0.90	0.95	0.42	0.72	0.12	0.34	.22***	.11	.18**	.92/.95	.60***	.42***	.28***	−.07	−.01	−.21***	.00	.06
5	Internal quantitative emp.	3.33	3.35	0.95	0.98	−0.28	−0.36	−0.23	−0.34	.28***	.14*	.22***	.60***	.84/.87	.24***	.41***	.04	−.10	−.22***	.07	.03
6	External qualitative emp.	2.96	2.96	0.95	0.94	0.24	0.12	−0.38	−0.31	.08	.07	.13*	.49***	.38***	.94/.95	.58***	−.05	−.05	−.23***	.04	.14*
7	External quantitative emp.	3.88	3.90	0.92	0.88	−0.78	−0.69	0.13	0.00	.25***	.14*	.18**	.37***	.50***	.62***	.89/.89	−.02	−.02	−.28***	.13*	.05
8	Home office hours per week	10.36	11.32	14.37	15.09	1.15	1.01	0.08	−0.44	.14*	.16**	.07	−.03	.10	−.07	−.00	-	−.14**	−.22***	.25***	.18**
9	Gender (1 = male, 2 = female)	-	-	-	-	-	-	-	-	−.01	−.12*	.03	−.01	−.12*	−.06	−.07	−.14*	-	−.02	−.06	−.29***
10	Age (in years)	40.57	-	12.59	-	0.09	-	−1.39	-	−.29***	−.06	−.10	−.17**	−.22***	−.21***	−.32***	−.17**	−.02	-	−.12*	−.11
11	Vocational Education (1 = completed vocational training, 2 = academic degree)	-	-	-	-	-	-	-	-	.18**	.30***	.21***	.09	.13*	.03	.16**	.25***	−.06	−.12*	-	.07
12	Employment status (1 = part-time, 2 = full-time)	-	-	-	-	-	-	-	-	.02	.17**	−.00	.04	.04	.06	.09	.16**	−.29***	−.11	.07	-

Note. N = 307. emp. = Employability; t1 = time point 1; t2 = time point 2 (approximately four weeks after t1); internal consistency reliabilities (McDonald’s Omega) are reported on the diagonal (in bold) for t1/t2. Omega for informal learning refers to the 24-item version; the four mean indicators show ω = .71/.75. SD Skew t1 = 0.67, t2 = 0.66; SD Kurtosis = 0.88, t2 = 1.05. For demographics, correlations at t1 are shown; home office hours from t2 are used. Informal learning was surveyed on a 6-point Likert scale, formal learning on a 7-point Likert scale, and all other constructs on a 5-point Likert scale.

*p < .05, **p < .01, ***p < .001.

We ran the structural equation model first without controls, then in a second step with controls (see Table 3 for values separated by model; see Figure 2 for standardized path coefficients of the controlled model). The uncontrolled model showed the better model fit, χ² (384) = 860.535, p < .001; CFI =.918; RMSEA =.064, 90% CI = [.058, .069]; SRMR =.052, compared with the controlled model, χ² (514) = 1173.278, p < .001; CFI =.889; RMSEA =.065, 90% CI = [.060, .070]; SRMR =.065. This is probably due to the higher model complexity (514 df vs 384 df) of the controlled model. For hypothesis testing, we draw on the results of the controlled model.

Figure 2. Effects of Work-related Learning on Employability.

Note. N = 307. Values shown are standardized parameter estimates. Gender, age, vocational education, employment status (part-time vs full-time), and home office hours per week were included as control variables. Manifest indicators of the latent constructs, error terms, correlations between work-related learning constructs at time 1, and correlations between employability constructs at time 2 (approximately four-week interval) are omitted for the sake of clarity. *p < .05.

Table 3. Uncontrolled and Controlled Models of the Effects of Work-related Learning on Employability.

Predictor Variable	Unstandardized Estimate
	Qualitative Internal E.	Quantitative Internal E.	Qualitative External E.	Quantitative External E.
Model 1
Informal learning	.48*	.56**	−.03	.45*
Formal learning	−.03	.00	−.02	.04
Self-regulated learning	.17	.14	.29✝	.03
Model 2 with Controls
Informal learning	.38*	.44*	−.19	.28
Formal learning	.01	.01	−.02	.04
Self-regulated learning	.18	.17	.31*	.03
Gender (1 = male, 2 = female)	−.01	−.22	−.10	−.06
Age (in years)	−.01***	−.01***	−.02***	−.02***
Vocational Education (1 = completed vocational training, 2 = academic degree)	−.12	−.04	−.05	.17
Employment status (1 = part-time, 2 = full-time)	.09	−.04	.25*	.03
Home Office hours per week	−.01*	−.00	−.01*	−.01*

Note. N = 307 (Model 1)/305 (Model 2). E. = Employability. The table shows unstandardized coefficients. For the standardized coefficients of the controlled model, see Figure 2. For work-related learning and demographic controls, time 1 (t1) data are used; for employability and home office hours, time 2 (t2) data are used (approximately four weeks after t1). Informal learning was surveyed on a 6-point Likert scale, formal learning on a 7-point Likert scale, and all other constructs on a 5-point Likert scale.

*p < .05, **p < .01, ***p < .001, ✝p = .064.

H1, H2 and H3 assumes that the three work-related learning forms lead to both internal and external qualitative as well as quantitative employability. For informal learning, the SEM results in Figure 2 indeed show significant positive relationship from informal learning to quantitative (β = .23) and qualitative (β = .19) internal employability, but no significant link to quantitative (β = .16) and qualitative external employability (β = −.10). These findings support H1a and H1c, but not H1b and H1d. For formal learning, the results show no significant links, neither to quantitative (β = .01) and qualitative (β = .01) internal employability, nor to quantitative (β = .05) and qualitative external employability (β = −.02). Thus, H2a – d are not supported by the data.

In terms of self-regulated learning, we did not find significant associations with quantitative (β = .11) and qualitative (β = .11) internal employability nor with quantitative external employability (β = .02). However, we found a significant positive relation between self-regulated learning and qualitative external employability (β = .19). This suggests a reverse pattern compared to informal learning: informal learning fosters internal employability, self-regulated learning fosters (qualitative) external employability. Thus, support emerges for H3d but not for H3a – c.

H4 and H5 are comparative hypotheses that follow the rationale that among the three work-related learning forms, self-regulated learning should have the strongest association with qualitative employability, followed by formal learning, followed by informal learning. According to H4, the link of self-regulated learning to qualitative internal and external employability should be significantly higher than the link of formal learning.

A descriptive glance at both path coefficients reveals that the association with internal qualitative employability is indeed a little stronger for self-regulated learning (β = .11) than for formal learning (β = .01); however, both paths are not significant. A comparison of the additional model – in which both path coefficients were set equal – to the initial model shows no significant difference between the two models, ΔChi² = 0.938, Δdf = 1, p = .333. Thus, there is no empirical support for H4a. The coefficient of the association with external qualitative employability is noticeably larger for self-regulated learning (β = .19, significant) than for formal learning (β = −.02, not significant). However, the model comparison between the additional and initial model shows no significant difference, ΔChi² = 3.541, Δdf = 1, p = .060, so H4b is not supported. The coefficient of the association with internal qualitative employability is lower for formal learning (β = .01, not significant) than for informal learning (β = .19, significant), contrary to the hypothesis. The two models also do not differ significantly, ΔChi² = 2.483, Δdf = 1, p = .115, arguing against H5a. The coefficient of the association with external qualitative employability is not significant for either formal learning (β = −.02) or informal learning (β = −.10); moreover, contrary to the hypothesis, both path coefficients are negative. There is no significant difference between the two models, ΔChi² = 0.526, Δdf = 1, p = .468; therefore, H5b is not supported.

Supplemental analysis

Calculation of the three separate CLP models including the five control variables – as a robustness check to control for reversed causality – revealed predominantly nonsignificant paths, as in the full model without reversed causality. The acceptably fitting CLP models for formal learning—χ² (855) = 1.573.470, p < .001; CFI =.934; RMSEA =.052, 90% CI = [.048, .057]; SRMR =.067—and for self-regulated learning—χ² (1.035) = 1.872. 597, p < .001; CFI =.918; RMSEA =.052, 90% CI = [.048, .055]; SRMR =.064—neither showed significant links of learning to the four employability types nor of employability to learning. Only the CLP model for informal learning, also having an acceptable model fit—χ² (855) = 1.655.354, p < .001; CFI =.922; RMSEA =.055, 90% CI = [.051, .059]; SRMR =.068—showed a positive significant link of informal learning to qualitative internal employability (β = .12, p = .046). This represents additional support for H1c. Furthermore, there was a negative significant relationship for the reverse direction of qualitative internal employability with informal learning (β = −.14, p = .045).

Discussion and implications

In our two-wave study, we examined the differential effects of formal, informal, and self-regulated learning on quantitative internal and external employability, as well as qualitative internal and external employability, while controlling for sociodemographic characteristics, education, and employment characteristics. Three hypotheses were supported by the data: Informal learning showed positive relations with both quantitative and qualitative internal employability (H1a and H1c, respectively), but not with either type of external employability (H1b and H1d were not supported). Conversely, self-regulated learning showed a positive link to only qualitative external employability (H3d), but not to any other type of employability (H3a – c were not supported). The relationship between informal learning and qualitative internal employability proved to be stable in a separate CLP model under control of reversed causality effects; the links of informal learning with quantitative internal employability and of self-regulated learning with qualitative external employability, however, did not stand up to this robustness check. Formal learning showed no significant relationship with any type of employability (H2a – d were not supported). Contrary to our expectations, no differences in effects strength emerged between the different learning forms and the two types of qualitative employability (H4a – b and H5a – b were not supported). In the following, we discuss the results for the three work-related forms of learning in more detail.

The effect of informal learning on employability

Informal learning was consistently related to both quantitative and qualitative internal employability, but not related to either type of external employability. This finding can be explained by the nature of informal learning. Since informal learning is driven by problems and challenges arising from the work process (Marsick & Volpe, 1999; Segers et al., 2018), the employee acquires knowledge specifically tailored to the organization, e.g., work process knowledge in dealing with machines or software programs. In addition, informal learning outcomes are directly tangible at work and could strengthen job-specific self-efficacy as a personal resource. This could lead to an increase in internal employability. The specific competencies are only marginally transferable to other organizations; for instance, knowledge about the informal hierarchy and social interaction within a specific corporate culture is not useful when moving to another company. Moreover, the increased self-efficacy could be domain- and role-specific (i.e., role breadth self-efficacy; Parker, 1998). Therefore, if employees have learned something informally, they may assess their internal employability highly, but therefore do not automatically believe themselves to be externally employable. Future research should consequently consider self-efficacy as a mediator between informal learning and employability.

Moreover, the supplemental CLP analysis showed a significant negative inverse effect of qualitative internal employability on informal learning. This suggests that employees who perceive high employability in terms of getting even better jobs with their current employer engage in lower levels of informal learning. One explanation could be that these individuals have a higher level of competence, which both positively influences their assessment of employability and gives them less reason to learn more, for example, through typical informal learning activities such as trying things out on their own, seeking feedback, and reflection (Decius, Knappstein, et al., 2021). This is supported by evidence that KSAs meta-analytically show a considerable relationship with employability (Harari et al., 2021). Future studies might include current skill level as a possible mediator between employability and informal learning, or at least control for skill level and hierarchical position in the organization.

The effect of formal learning on employability

The nonexistent effects of formal learning on internal and external employability, on the other hand, are puzzling. We would not go so far in our conclusion as to deny that training and workshops have any effect on employability, even if the results of our study might suggest this. Instead, we see a fundamental problem in the methods currently used to operationalize formal learning. Difficulties arise especially when the aim is not to evaluate a single training course – e.g., in a quasi-experimental design – but, as in our study, to investigate differential effects of different work-related forms of learning.

Previous research particularly employed self-report scales that query the frequency of participation in formal courses (e.g., Grosemans et al., 2020; Pajo et al., 2010). Problematically, memory bias can occur, especially with longer time periods, i.e., the employee may only recall certain trainings and repress others. In addition, employees have different perceptions of which events can be considered formal training and which cannot. In the grey area are, for instance, mentoring and network meetings, which are formally organized, but which are legitimately counted by some authors as part of informal learning opportunities (e.g., D. Richter et al., 2014). Moreover, asking for frequency leads to a test-theoretical floor effect, as employees usually participate in trainings only rarely, so that it is difficult to differentiate between individuals with a low and high level of formal learning. Shorter periods, such as the four-week interval in our study, could amplify this effect.

Frequency of attendance at a formal event also says little about learning behaviour and learning success. Training design features have a notable effect on training effectiveness (Arthur et al., 2003), as does motivation to learn (Colquitt et al., 2000). An employee who is intrinsically motivated to attend a single event may acquire more knowledge than a colleague who attends five events with little motivation or only out of obligation. We therefore call for the development of new methods – whether scales or other assessment approaches – that consider not only frequency but also criteria such as length and intensity of training, degree of formalization, and fit of content to the job, e.g., by using experience sampling methods on a weekly basis.

Furthermore, the COVID-19 pandemic has affected many dynamics, especially with regard to formal learning. Many courses had to be cancelled due to public meeting restrictions, only some got replaced by virtual formats. People have obviously taken part in formal learning activities to a lesser extent than before. The skewness and low mean scores of formal learning (see Table 2) may also signal this. In other words, when there is a comparatively only low participation in formal learning activities, the effects stemming from formal learning cannot be that strong compared to more frequent other forms of learning (possibilities for informal and especially self-regulated learning have not been restricted so much by the pandemic).

Another explanation for the lack of effect of formal learning on employability could be the temporal perspective: Possibly, the four-week interval between the two measurement points we chose was too short to unfold the effect of learning on employability – or, on the contrary, it was too long. Previous research has suggested, for example, that formal training can result in a short-term boost (i.e., six-week follow-up) in perceived employability, but that this boost may not be stable in the long term (i.e., one-year follow-up) (Akkermans et al., 2019). Although our time lag of four weeks was even shorter than the six-week period in this study, we cannot be sure of the exact timing of training participation because of the potential measurement inaccuracy mentioned above. In any case, further research is needed on this issue and the dynamics of employability development (see also our suggestions for calculating the “optimal time interval” in the limitations section).

The effect of self-regulated learning on employability

The differential effects of self-regulated learning on internal and external employability can be understood by looking at processes of self-regulated learning. In self-regulated learning, learners decide for themselves which learning goals to set and monitor their goal attainment (Sitzmann & Ely, 2011; Zimmerman, 1990). Because of the high investment of cognitive resources for this monitoring, on the one hand, and because of the high autonomy regarding goal selection, on the other hand, learners should especially select learning goals that will benefit them in the long run and independent of employers in the labour market – especially in times when employer changes are commonplace (Hall et al., 2018). Thus, we can explain that self-regulated learning leads to qualitative external employability. The absence of the effect of self-regulated learning on quantitative external employability contrasts with our theoretical reasoning that if someone can obtain a better job (i.e., qualitative employability), they can, by default also obtain a different job (i.e., quantitative employability). One explanation could be that employees engage in self-regulated learning to a greater extent only when they aim for specific better positions or job roles elsewhere and their employer does not offer the learning opportunities to acquire the necessary competencies. Such specific learning activities might then not contribute to a general sense of quantitative employability – perhaps employees would even feel overqualified for comparable jobs in such cases. Overall, it seems as if qualitative employability does not always build on quantitative employability in the way we expected. Future research could delve deeper to people’s understanding of and cognitions about different employability facets to resolve this issue.

The absence of the effect of self-regulated learning on qualitative internal employability could result because learning situations aimed at internal employability might be already covered by informal learning. It would be consistent with COR Theory that individuals invest only as few resources as necessary if they can still achieve resource gains by doing so (see also Halbesleben et al., 2014). Future research could investigate – for example in controlled (online) experiments – whether employees consciously use self-regulated learning, especially when opportunities for informal learning (which is integrated into the work process and thus less demanding) are unavailable or difficult to achieve.

Theoretical implications

Our results show that not every form of learning may benefit every type of employability. In line with previous research, this finding underlines the importance to differentiate between different forms of learning and employability, respectively, to uncover differential relationships with antecedents and outcomes (Forrier et al., 2015; Houben et al., 2021; Van den Broeck et al., 2014; Van Harten et al., 2022). We were particularly interested in the distinction between quantitative and qualitative employability: The difference between chances to find a better job versus chances to find another job in general may point to different mechanisms in terms of resource gain versus resource conservation. This distinction has so far received less attention in employability research than the internal – external dimension (De Cuyper & De Witte, 2010; Van den Broeck et al., 2014).

However, we did not find empirical support for our expectation that work-related forms of learning that are conceptually associated with more effort and goal-directedness would have a stronger effect on employability. Based on COR Theory, we had assumed that individuals invest more resources in learning behaviours when they gain qualitative employability as an outcome. The higher the investment, the higher the return should be (see also Halbesleben et al., 2014). Yet we did not find stronger effects on qualitative employability for self-regulated learning compared to formal learning, or for formal learning compared to informal learning, despite the differences in cognitive effort (i.e., resource investment) between these different learning forms (Decius, 2020; Sitzmann & Ely, 2011).

One reason for the absent effects of formal learning could be that training participation is not always voluntary and thus training content is less goal-oriented from an employee perspective than we have assumed.⁴ When managers send employees to training, the course may not reflect their interests and perceived needs. This is usually not done out of malice or ignorance, but because HR management has the strategic needs of the organization in mind, not exclusively the personal benefit of the employees, for instance in the case of mandatory corporate design training to strengthen the external image of the organization. For learning outcomes, however, this can make a difference: An evaluation of voluntary versus nonvoluntary participation in life skill and employability training programs in three further education contexts using interviews and observations suggested the importance of voluntary and motivated training participation for learning outcomes (O’Grady & Atkin, 2005). Future studies on formal training and employability should therefore consider participation voluntariness.

The overall picture which emerges from our findings is rather that informal learning fosters employees’ expectations of both resource conservation and resource gain, but only for their in-house career. Consequently, employees may not only protect their status quo, but even improve their job chances with their employer. Nevertheless, because the skills acquired informally are less transferrable, they do not benefit in the external labour market. It could be that informal learning is more cognitively resource-intensive than research has previously suggested (Cerasoli et al., 2018; Decius et al., 2019; Marsick & Volpe, 1999; Tannenbaum et al., 2010), and it may be worthwhile to test our assumed mechanisms by explicitly measuring resource investment or the degree of goal-directedness in learning.

However, to understand the resources-based mechanisms between work-related learning and employability, the key question may not be how much effort employees invest and how much resource value they expect in turn, but rather what kind of resources matter for which type of employability, for instance, job-specific versus transferrable skills (see Van Harten et al., 2022). This may also explain why self-regulated learning seems to help employees to improve their chances of resource gain outside their current organization, but not their internal standing. The resources employees gain from self-regulated learning are in line with their own needs and career goals, but this may not always coincide with the employers’ needs or social norms by significant others (Baert et al., 2006). It is also conceivable that employees engage more in self-regulated learning when they perceive a misfit between their own learning needs or career goals and the opportunities their employer can offer. Consequently, they would focus on gaining resources which help them make an external career transition (De Vos et al., 2021). Future research could delve deeper into these mechanisms by testing specific investments and resource gains as mediating mechanisms between different forms of learning and employability, respectively.

Furthermore, more research on the construct validity of different employability facets seems warranted, especially since the distinction between quantitative and qualitative employability has been less frequently investigated than the internal-external distinction (De Cuyper & De Witte, 2010; Van den Broeck et al., 2014). Qualitative studies could also dig deeper into employees’ construal of quantitative and qualitative employability with respect to expectations of resource conservation, gain, or even loss, while taking into account previous experiences with actual job transitions (see Fletcher & French, 2021; Lipshits-Braziler & Gati, 2019).

Additionally, the mechanisms between learning and employability may be context-specific and subject to boundary conditions: For example, COR Theory (Hobfoll et al., 2018) postulates that individuals can better focus on further resource gain when their existing resources are secure, whereas they enter a defensive mode under conditions of threat. Consequently, the level of job security or the current economic climate may play a role, such that work-related learning may become more salient for quantitative employability when one’s job is threatened, and more salient for qualitative employability when one’s job is secure (see Van Hootegem et al., 2019).

Overall, the effects of work-related learning on perceived employability turned out to be much smaller and more inconsistent than we had assumed. This is in line with meta-analytic findings that while KSAs are a strong predictor of employability, KSA development (i.e., learning) has only a moderate association with employability (Harari et al., 2021). We thus question the general proposition that work-related learning always leads to employability. Further research seems necessary on the situations and boundary conditions under which work-related learning can increase employability.

Limitations

Our study provides valuable insights for research but does not come without limitations. Despite using a longitudinal design, we can only approximate causality because the sample size is too small to provide sufficient power for a full CLP model – though causality testing is only possible in controlled experimental settings anyway (Stone-Romero & Rosopa, 2008). However, we calculated three separate CLP models whose results corroborated at least the effect of informal learning on qualitative internal employability.

We also relied on self-reported data from a convenience sample, which may be accompanied by biases in measurement accuracy. However, we mitigated the risk of common method bias by using the procedural strategy of applying two measurement time points (see Podsakoff et al., 2003). Because we were interested in capturing perceived employability, and because employees consider individual abilities and contextual factors such as labour market conditions when assessing their own employability (De Cuyper et al., 2011; Forrier et al., 2015), self-report was the method of choice here anyway. On the one hand, this can make the concept of employability, especially qualitative employability, quite ambiguous, because what one person considers better may be very different from what another person considers better. On the other hand, we still believe in a common subset of employability: we argue that employees may have a similar understanding of some basic factors constituting a good job (e.g., safe working conditions or adequate pay), although at the individual level, everyone has their own priorities. This is like the value of money in society: some people prefer to have as much money as possible, while others are fine with less. Still, everyone needs a certain amount of money to cover some basic expenses that cannot be avoided. Research could look more deeply into the distinction between general and individual factors of employability.

To reduce ambiguity in the concept of (qualitative) employability, future studies could also examine whether our results can also be replicated if objective employability measures were used, e.g., in long-term studies focusing job applications and job change behaviour. Then, however, labour market conditions would also need to be considered, which employees already implicitly take into account when stating their perceived employability. In addition to the employee perspective, future research should also more thoroughly incorporate the employer perspective, which our study, like most employability studies too, has largely sidestepped (Fugate et al., 2021).

Furthermore, scholars could attempt to operationalize informal and self-regulated learning more objectively than we did, although the use of observations, for example, is limited because of the high proportions of reflection especially in self-regulated learning (Decius & Decius, 2022; Sitzmann & Ely, 2011). In work-related learning research, there is a general lack of studies with high internal validity, e.g., experimental (laboratory) studies. We also detailed our critique of operationalizing formal learning via frequency scales in the first section of the discussion.

Another limitation is that we only used one subscale (i.e., task strategies) from the Self-Regulated Learning at Work Questionnaire by Fontana et al. (2015) to measure self-regulated learning. We considered the use of task strategies – a central component of self-regulated learning that, among other criteria, allows for differentiation from informal learning (Decius & Decius, 2022; Zimmerman, 1990)—to be particularly relevant for employability, as it reflects the employee’s handling of job-related task demands, which are also important for the employee’s search and application for internal positions and jobs on the external labour market. Nevertheless, future research could also consider other facets of self-regulated learning (e.g., elaboration strategies and critical thinking; Fontana et al., 2015 see also Zimmerman, 1990).

The latent controlled model slightly failed to meet the criteria for an acceptable model fit – in this case, the collected data apparently did not fit the hypothesized model structure well. However, this is also due to the high model complexity, as we modelled paths from all five demographic control variables on all four employability types. We used a four-week interval between the two measurements. The rationale for this choice was that the period is sufficiently long to expect changes in work-related learning processes (Decius, 2020). Other studies on employability also used four-week intervals (e.g., Ayala Calvo & Manzano García, 2021; Yizhong et al., 2017), but some studies also applied longer intervals of several months (e.g., Houben et al., 2021). The choice of time interval might also have influenced effect sizes positively for informal learning and negatively for formal learning: Learning transfer is sometimes problematic and lengthy in formal training, whereas informally learned knowledge can be immediately integrated into the work process (Blume et al., 2010; Marsick & Volpe, 1999). Since there has been little research on the time course of the learning’s effects on employability, future research could conduct “shortitudinal” pilot studies to determine the “optimal time interval” for this relationship (Dormann & Griffin, 2015, p. 489).

Practical implications

In terms of practical implications, our findings suggest different levers for improving different forms of employability. In contrast, broad approaches that invest in any type of learning in the hope of improving employability per se may not work for either employees or organizations. The differences that emerged between internal and external employability are in line with the notion in previous research that promoting internal employability is to a greater extent the responsibility of the employer, whereas employees themselves are responsible for their external employability (Van den Broeck et al., 2014). For organizations, the distinction between internal and external employability is crucial: If learning leads to internal employability, the employer benefits from the individual’s job-tailored competencies. If learning leads to external employability, there is a risk that the individual will leave the employer – a situation discussed in research under the term “management paradox” (De Cuyper & De Witte, 2011). Following this logic, a key implication of our study is that organizations should support their employees in informal learning to strengthen their internal employability and consequently, employee retention (see Forrier et al., 2015). For strategic human resource management, this might entail a shift in focus from traditional training approaches to creating opportunities for learning on the job through work design. At an intermediate level, supervisors can promote informal learning, for instance, by designing conducive conditions such as a positive learning culture (Kortsch et al., 2019), by providing positive role models (Cerasoli et al., 2018), and by offering organizational and managerial support including supervisor feedback (Decius, Schaper, et al., 2021; Hilkenmeier, Goller, et al., 2021; Zia et al., 2022).

Such facilitative and confidence-building interventions have additional positive effects besides promoting learning: Arnold and Staffelbach (2012) showed that employability is negatively related to perceived job insecurity, which can be interpreted as a protective buffer effect. They also found that the higher the employee’s trust in the employer, the stronger this protective effect. This is consistent with recent research that trust within organizational collaborations is a key construct for achieving work-related goals and building organizational human capital (Hilkenmeier, Fechtelpeter, et al., 2021).

According to our findings, self-regulated learning only increases external employability. This holds with the caveat that the effect of self-regulated learning on external employability, unlike the effect of informal learning on internal employability, did not withstand the robustness check when controlling for reversed causality in the CLP model. However, we caution organizations to discourage their employees from self-regulated learning-also because it is associated with other positive outcomes for competence development (Keith & Frese, 2005; Schulz & Stamov Roßnagel, 2010). Moreover, there is no empirical support for companies’ concerns that their best employees will leave if they invest in their employability (Forrier et al., 2018).

We sing a similar tune with respect to formal learning. Even though our study found no effects of formal learning on employability, we still recommend that organizations invest in formal training and workshops, as training is associated with positive effects for skill acquisition (Arthur et al., 2003) and increases organizational performance (Combs et al., 2006). Moreover, it is debatable whether the absence of the effect of formal learning on employability can also be partly explained by measurement artefacts (see in detail in the first section of the discussion).

As an implication for employees, we would like to encourage learners to consider which type of career resource they aim to improve: If the goal is to protect or improve their job chances with their employer, they should engage in informal learning. In order to identify and take advantage of opportunities for informal learning, participation in training to increase self-learning skills and self-directed learning orientation could be a starting point (Cerasoli et al., 2018; Decius, Schaper, et al., 2021). In contrast, employees who wish to improve their job chances in the external labour market and perhaps reduce dependency on their current employer, learning informally on the job may not be enough. Rather, they should take time to reflect what they want and need to learn, set goals for themselves, and thus also engage in self-regulated learning. For promoting self-regulated learning, employees can exercise setting appropriate goals for themselves and enhancing their self-efficacy, as these factors have been meta-analytically shown to be the most influential predictors of self-regulated learning in the workplace (Sitzmann & Ely, 2011).

Conclusion

Both organizations and employees can benefit from positive effects of work-related learning on employability – but it depends on the particular work-related learning form, as differential effects of learning exist on different employability types. In terms of proactive human resource development and employee self-development, our study shows that informal learning (conducive to internal employability) and self-regulated learning (conducive to external employability) can not only increase one’s confidence to find any other job (i.e., quantitative employability), but even improve the perceived opportunities to find a better job (i.e., qualitative employability).

Acknowledgements

We would like to thank Theresa Knappstein for her support in data processing.

Disclosure statement

We have no known conflict of interest to disclose. Since only questionnaires without sensitive questions and no interventions were used, no permission from the local ethics committee had to be obtained. Data are available from the authors upon reasonable request.

Notes

1. When we refer to employability in the following, we mean perceived employability, unless otherwise stated.

2. Due to model complexity in the structural equation model, the sample size was not sufficient to compute an overall latent cross-lagged panel model with all three work-related learning types and all four employability types.

3. An additional Jarque Bera test confirmed the violation of the normal distribution assumption for the mean of the four items of formal learning, χ² (2) = 193.63, p < .001.

4. We thank an anonymous reviewer for this thought.