This work has three objectives: the first is to analyse the relationship between learning at an individual level and innovative behaviour; the second is to analyse the mediation effect of learning at the group level between individual-level learning and innovative behaviour and the third is to analyse the moderation effect of job autonomy between individual-level learning and innovative behaviour.
A questionnaire survey was given to 303 employees (144 Colombian and 159 Spanish), and the hypotheses were tested using structural equation modelling.
The results are significant and indicate that individual-level learning is significantly related to innovative behaviour. Group-level learning mediates the relationship between individual-level learning and innovative behaviour, whilst job autonomy positively moderates the relationship between individual-level learning and innovative behaviour. The control variable “country culture” is not statistically significant. This provides valuable insights into the role of culture in innovation.
This study offers a novel contribution by integrating individual and group learning dynamics within a multilevel framework to explain innovative behaviour while empirically validating the moderating role of job autonomy. It extends the organisational learning and innovation literature by indicating that job autonomy significantly amplifies the positive effect of individual learning on innovation. Its theoretical value lies in the combined application of transactive memory systems and job demands–resources (JD-R) models, while its empirical contribution is reinforced by a cross-cultural analysis and the use of partial least squares structural equation modelling (PLS-SEM) methodology.
1. Introduction
Facilitating innovation at work is key for organisational success and survival, as innovation enables organisations to adapt and respond proactively to market challenges (Anderson et al., 2014). Organisations typically seek and encourage the active contribution of all employees in the innovation process (Cangialosi et al., 2020), and this active participation of employees has been identified as innovative behaviour (Ekmekcioglu and Öner, 2024), by which we mean behaviour aimed at the intentional generation, promotion and implementation of new and useful ideas, processes, products or procedures within a role, group or organisation (Janssen, 2000).
The literature indicates that organisational learning influences a company’s innovation activity (Jiménez-Jiménez and Sanz-Valle, 2011). This is because organisational learning allows the company to identify opportunities, adapt processes, and generate creative solutions to the challenges of the competitive environment (Argote, 2011). Organizational learning has been analysed at different levels, including the individual, the group and the organisational levels (Ellström, 2010). Individual learning is particularly related to innovative behaviour since this behaviour is identified with an individual’s willingness to generate ideas and apply them in the workplace. Furthermore, it is expected that employees who exhibit innovative job behaviour will engage in creative thinking and propose original and valuable ideas to improve current organisational processes (Ahmed et al., 2024). The first research objective in this paper is to analyse the relationship between individual learning and innovative behaviour.
Continuing with the previous proposal, and based on the Transactive Memory Systems (TMS) theory (Wegner, 1986) which suggests that learning processes at the individual and collective levels (e.g. group learning) are interrelated, we propose our second research objective: to analyse the mediating effect of group-level learning between individual-level learning and innovative behaviour. In other words, we analyse how individual learning affects innovative behaviour through group learning (Anderson and Lewis, 2014).
If individual-level learning relies on the group to influence innovation, one question to be resolved is how individual learning can have a stronger influence on innovative behaviour. This proposal leads us to the third research objective: to analyse the moderating effect of job autonomy between individual-level learning and innovative behaviour. Job autonomy refers to an employee’s degree of freedom and discretion to make work-related decisions, including task planning, choice of methods, and time management (Hackman and Lawler, 1971). When employees have a high level of autonomy, they are more likely to engage in experimentation, reflection, and problem-solving, which fosters individual learning (Fagerlind et al., 2013). This proposal is based on the Job Demands–Resources (JD-R) model (Bakker and Demerouti, 2007). In the JD-R model, job autonomy is a resource in the psychosocial job environment. Resources in the psychosocial job environment are expected to reduce the strain associated with job demands, improve workers’ ability to achieve their work goals and stimulate workers’ growth and learning (Clausen et al., 2022).
Prior research has not carried out a thorough multilevel empirical analysis that integrates individual and group learning with the moderating role of job autonomy in fostering innovation, particularly across different cultural contexts. This research provides an original perspective by integrating individual learning, group learning, and job autonomy into a multilevel model to explain innovative behaviour. In addition, this work advances the literature on learning and innovation (Widmann and Mulder, 2018) in two important ways. First, it analyses the relationship between learning at an individual level and learning at a group level and the ways in which the two types of learning can complement each other. Secondly, it examines how the relationship between individual learning and the development of innovative behaviour can be strengthened through job autonomy. Organisations facing the demands of the current climate of change (Bakker and Demerouti, 2014) should provide a working environment in which employees feel committed, well-resourced and healthy (Stansfeld and Candy, 2006). This externally created environment needs to be revised and developed through employee-initiated changes (so-called “job crafting”) (Vogt et al., 2015). Based on the JD-R model, job crafting comprises proactive changes initiated by an employee to shift their job demands and resources to match their personal needs, values and capabilities (Tims et al., 2012).
The hypotheses were tested on a sample of employees from Colombia and Spain, which strengthens the research results by including “country culture” as a control variable in the analysis model. When considering different cultures, one tries to avoid ethnocentrism. In this context, this would be a matter of not treating behaviour towards innovation as universal, which we achieve because our study involves employees from Colombia, which has a more collectivist culture, and employees from Spain, which has a more individualistic culture, allowing us to extend our results to different cultures. This aspect provides more originality to the work by enabling the exploration of possible cultural differences in job autonomy (van Hoorn, 2016).
2. Theoretical framework and development of hypotheses
2.1 Theoretical framework
We use TMS theory to argue for the mediating role of group learning in the relationship between individual learning and innovative behaviour. TMS is a collective system for storing and retrieving information that gives individuals access to more information than they possess alone (Argote and Guo, 2016). In terms of learning, individuals benefit from the group and thus reinforce innovation. Drawing on JD-R (Bakker and Demerouti, 2007), which posits that resources can moderate the negative impact of demands on well-being in the demands–resources interaction, we propose the moderating effect of job autonomy. In our case, job autonomy is a resource that facilitates innovation, which is considered a job demand.
2.2 Learning at individual level, and the development of innovative behaviour
Learning at the individual level refers to the process by which individuals generate new ideas and knowledge from existing ideas and knowledge (Barba Aragón et al., 2014). Anderson and Lewis (2014) highlighted the relationship between individual-level learning and innovation. Learning at the individual level provides people with the skills and mindsets that they need to generate and apply new ideas. Through learning, individuals also improve their ability to analyse and solve problems, which allows them to advance and innovate (Li et al., 2024), to the extent that new solutions to challenges are found through learning. In turn, Yang et al. (2016) showed that employees’ orientation towards learning goals was positively associated with innovative behaviour. The following hypothesis is therefore proposed:
Individual-level learning is positively related to innovative behaviour.
2.3 The mediation effect of group-level learning
Group learning is critical to the development of innovation because when ideas are shared and feedback is received in the group, this facilitates the generation of novel ideas. The different interactions between people in a group allow the group members to combine their perspectives and knowledge, encouraging them to face problems and find solutions. The relationship between individual learning and group learning has received more attention than that between organisational learning and either of the other two levels of learning. Barker and Neailey (1999), for example, indicated that individual learning follows group learning, and group learning leads to innovation. Group learning is about encouraging team members to reflect on their learning; later, leaders bring together the group’s learning, which is an opportunity for innovation. It could be argued that individual learning requires group interactions if innovative behaviour is to develop, which leads us to propose the following hypothesis:
Group-level learning mediates the relationship between individual-level learning and the development of innovative behaviour.
2.4 The moderating role of job autonomy
Individual innovation behaviour, as opposed to team- or organisational-level innovation, is based on the individual’s commitment to generating and applying new ideas and approaches in the workplace (Wu et al., 2014). At the individual level, innovation can arise from experimentation and a willingness to improve existing processes. There should, therefore, be a relationship between innovative behaviour and job autonomy, because employees with autonomy can make decisions to address their work problems. Autonomy also increases employees’ sense of ownership of and responsibility for their work, which motivates them to be innovative (Nazir and Islam, 2019).
Job autonomy is a key facilitator in influencing the strength of social exchanges, and it is expected that the positive effects of better work engagement on employees’ innovative behaviour are more significant when those individuals experience a high level of job autonomy (Garg and Dhar, 2017). Previous studies have also examined job autonomy as a moderating variable that positively affects some favourable working attitudes, such as organisational citizenship behaviour (Runhaar et al., 2013). Giebels et al. (2016) viewed job autonomy as a moderating variable and emphasised the relationship between a proactive personality and task conflict control. Based on these arguments, the following hypothesis is proposed:
Job autonomy moderates the relationship between individual learning and innovative behaviour.
2.5 Country culture as a control variable
Although our work is not intended to debate culture in the broad sense, it is commonly accepted that culture is a set of customs, traditions, values and beliefs held by ethnic groups and nations (Hofstede, 2003), and that cultural dimensions may explain the behaviours of individuals and companies (Chen et al., 2015). Although Colombia and Spain share a language and some customs, there may be cultural differences between the two countries which reflect their history and diversity and give each its own identity. Colombian culture may have a greater sense of community and mutual support than Spanish culture, and this may, in turn, influence the relationship between learning and innovation. In a culture with a strong sense of community (like Colombia), collaboration between people can be essential for learning development; however, in a more individualistic context (like Spain), personal incentives can drive innovation.
Figure 1 illustrates the theoretical model.
The flow begins with a circle on the left labeled “Individual Learning (I L).” From this circle two arrows arise: an upward arrow labeled “a” arises and points to a circle on the top labeled “Group Learning (G L),” and another is a right‑pointing arrow labeled “H 1 open parenthesis plus close parenthesis equals C dash, and H 2 open parenthesis plus close parenthesis equals I L, right pointing arrow, G L, right pointing arrow, I B equals a times b subscript 1,” which points to another circle on the right labeled “Innovative Behavior (I B).” Below this right‑pointing arrow, a circle labeled “Job Anatomy (J A)” is present with an upward arrow labeled “H 3 open parenthesis plus close parenthesis equals b subscript 3” pointing to this right‑pointing arrow. A circle labeled “Country Culture (C C)” is presented above the I B circle and points to it. From the “Group Learning (G L)” circle, a downward arrow labeled “b subscript 1” points to the “Innovative Behavior (I B)” circle.Research model and hypotheses Source: Authors’ own work
The flow begins with a circle on the left labeled “Individual Learning (I L).” From this circle two arrows arise: an upward arrow labeled “a” arises and points to a circle on the top labeled “Group Learning (G L),” and another is a right‑pointing arrow labeled “H 1 open parenthesis plus close parenthesis equals C dash, and H 2 open parenthesis plus close parenthesis equals I L, right pointing arrow, G L, right pointing arrow, I B equals a times b subscript 1,” which points to another circle on the right labeled “Innovative Behavior (I B).” Below this right‑pointing arrow, a circle labeled “Job Anatomy (J A)” is present with an upward arrow labeled “H 3 open parenthesis plus close parenthesis equals b subscript 3” pointing to this right‑pointing arrow. A circle labeled “Country Culture (C C)” is presented above the I B circle and points to it. From the “Group Learning (G L)” circle, a downward arrow labeled “b subscript 1” points to the “Innovative Behavior (I B)” circle.Research model and hypotheses Source: Authors’ own work
3. Methods
3.1 Participants
Ethical considerations were borne in mind throughout this research. The main ethical objectives were to protect the research participants’ right to privacy and ensure the honesty of our data collection and analysis. The participants in the study were 303 employees from Colombia and Spain. This is an incidental sample, to which the research team had good access. The characteristics of this sample are presented in Table 1.
Participants in the study
| N | % | |
|---|---|---|
| Country | ||
| Colombia | 144 | 47.5 |
| Spain | 159 | 52.5 |
| Firm Size | ||
| 1 to 9 employees | 56 | 18.5 |
| 10 to 49 employees | 56 | 18.5 |
| 50 to 99 employees | 46 | 15.2 |
| +99 employees | 145 | 47.8 |
| Sector | ||
| Primary | 22 | 6.6 |
| Industry | 148 | 48.8 |
| Service | 135 | 44.6 |
| Gender | ||
| Man | 138 | 45.5 |
| Woman | 165 | 54.5 |
| Educational level | ||
| No college degree | 173 | 57.1 |
| College degree | 130 | 42.9 |
| Job category | ||
| Senior manager | 27 | 8.9 |
| Manager | 109 | 36.0 |
| Middle manager | 82 | 27.1 |
| Employee | 85 | 37.0 |
| N | 303 | 100 |
The sample comprised employees from Colombia (47.5%) and Spain (52.5%). Participants were employed across firms of varying sizes, with 47.8% working in companies with over 99 employees. Sectoral representation included the primary (6.6%), secondary (48.8%), and tertiary (44.6%) sectors. Most companies were multinationals or affiliated with multinational groups, operating in technology and auditing (Colombia) or the chemical and construction industries (Spain). Regarding personal characteristics, 45.5% of participants were male and 54.5% female; 42.9% held a university degree. Job roles ranged from senior management (8.9%) to employees (37.0%), with an average age of 39.4 years (SD = 11.1; range = 18–72).
3.2 Measurement of variables
Five variables were analysed. Four variables (individual learning, group learning, job autonomy and innovative behaviour) were measured using a 7-point Likert scale, where 1 represented “strongly disagree” and 7 represented “strongly agree”. The country culture variable was dichotomous (Colombia or Spain).
Learning at the individual level was measured, following the work of Barba Aragón et al. (2014), through seven items. These items identified the individual’s behaviour and commitment to learning in the organisation. Example items included: “I try to open my mentality to see things differently and open new paths” and “I have a clear sense of the direction of my work”.
The work of Barba Aragón et al. (2014) was also used to measure learning at the group level, with seven items referring to the mutual commitment to learning and development of groups in the company. Example items included: “Our work groups have the right people involved to address problems” and “Our company groups are prepared to rethink decisions when presented with new information”.
The work of Wang et al. (2019) was followed to measure innovative behaviour, using six items that referred to individual problem-solving skills for developing and implementing new ideas, strategies, products and services. Example items included: “I usually look for the necessary means to implement my new ideas” and “Whenever possible, I look for new techniques, processes or technologies to improve my work”.
The work of Fuller et al. (2006) was followed to measure job autonomy. Three items were considered to measure the degree to which the employees had an important role in scheduling their work. Example items for measuring this construct included: “In my company, I have sufficient autonomy to do my job” and “In general, I have considerable opportunity for independence and freedom to do my job”.
3.3 Data analysis
Partial least squares structural equation modelling (PLS-SEM) was run with partial regressions to obtain composite scores that minimise the residual variances in the relationships between composites and indicators (Tenenhaus et al., 2005). This approach was chosen for two main reasons (Ghasemy et al., 2022): the existence of compounds within the model estimated in Mode A and the fact that we are using mediation analysis.
Mode A corresponds to the reflective model, in which the indicators are considered manifestations of the latent construct. PLS-SEM is a recommended approach for mediation analysis, as it makes no assumptions about the distribution of variables and can be applied to small sample sizes with great confidence (Hair et al., 2017).
3.4 Common method bias
In questionnaire-based studies, the possibility of common method bias (CMB) should be considered when the same individual responds to questions on the dependent and the independent variables (Podsakoff et al., 2012). Initially, we attempted to avoid psychological CMB by separating the predictor and criterion variables and ensuring anonymity in the responses (Podsakoff et al., 2012). Furthermore, at the beginning of the questionnaire we included detailed instructions about the purpose of the research (Jordan and Troth, 2020). A multicollinearity test was also carried out. All resulting variance inflation factors (VIF) were less than 3.3. Therefore, CMB did not contaminate our model (Kock, 2015) (Table 2). In addition, we performed Harman’s single factor test (Jordan and Troth, 2020), in which all items in the study were subjected to principal component analysis (PCA) with an unrotated factor solution. The aim was to determine whether a single factor explained more than 50% of the shared variance between the items. In our case, only 39.93% of the variance was explained by a single factor, indicating that there was no single factor problem (Tehseen et al., 2017).
Measurement model results
| VIF | Loading | α Cronbach | CR | AVE | |
|---|---|---|---|---|---|
| Group learning | 0.870 | 0.901 | 0.605 | ||
| Appreciate the perspectives of group members | 3.194 | 0.824 | |||
| Share work group successes | 2.223 | 0.799 | |||
| Adaptability of work groups | 2.005 | 0.778 | |||
| Adaptation to new circumstances of work groups | 2.010 | 0.815 | |||
| *Promote empathy in the work group | 1.648 | 0.599 | |||
| Effective conflict resolution in the group | 2.139 | 0.746 | |||
| Adaptation of people to the work group | 1.586 | 0.703 | |||
| Individual learning | 0.781 | 0.850 | 0.533 | ||
| Pride in the work | 2.179 | 0.771 | |||
| Perception of achievement at work | 2.058 | 0.713 | |||
| Sense of direction at work | 1.965 | 0.753 | |||
| *Awareness of critical issues affecting work | 1.217 | 0.413 | |||
| New idea generation | 2.255 | 0.764 | |||
| Perceived psychological security at work | 1.348 | 0.704 | |||
| *Open mind to new ideas | 1.263 | 0.532 | |||
| Innovative behaviour | 0.788 | 0.856 | 0.546 | ||
| Defending of one’s own ideas in front of others | 2.059 | 0.701 | |||
| Perception of an innovative person | 2.341 | 0.822 | |||
| Proactivity in innovation | 1.431 | 0.722 | |||
| Proactivity in the generation of ideas | 3.297 | 0.794 | |||
| Proactivity in searching for resources | 2.203 | 0.707 | |||
| *Development of plans to implement new ideas | 1.555 | 0.577 |
Note(s): VIF: variance inflation factor, CR: Composite Reliability, AVE: Average Variance Extracted. * Items have low factor loadings, and were deleted accordingly
4. Results
4.1 Measurement model results
Our constructs were estimated in Mode A. Traditional internal consistency, reliability and validity measures were applied (Henseler et al., 2016). The indicators and dimensions met the reliability requirements because their loadings exceeded 0.7 (see Table 2), except for the item “Promote empathy in the workgroup” in the group learning construct and the items “Awareness of critical issues affecting work” and “Open mind to new ideas” in the individual learning construct; these items were dropped in the final analyses to guarantee the convergent validity of the scales. The composite reliability measure indicated that all the constructs were reliable because the measures were above 0.7. All constructs also achieved convergent validity, as the values of their average variance extracted (AVE) were higher than 0.5 (Hair et al., 2019).
Table 3 shows the discriminant validity criteria. Following the Fornell–Larcker criterion (Fornell and Larcker, 1981), for which the square root of the AVE must be greater than the correlation between the construct and any other construct, and the heterotrait–monotrait correlation criterion (HTMT) (Henseler et al., 2015), which allows the evaluation of discriminant validity in SEM based on variance, all the constructs achieved discriminant validity (Hair et al., 2017) or, in other words, all constructs were empirically different.
Measurement model. Discriminant validity
| Fornell-Larcker criterion | Heterotrait-monotrait ratio (HTMT) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GL | IL | IB | Cc | GL | IL | IB | Cc | ||
| GL | 0.778 | GL | |||||||
| IL | 0.631 | 0.730 | IL | 0.755 | |||||
| IB | 0.533 | 0.741 | 0.787 | IB | 0.615 | 0.809 | |||
| Cc | 0.012 | 0.035 | 0.047 | 1.000 | Cc | 0.037 | 0.046 | 0.056 | |
Note(s): GL: Group learning, IL: Individual learning, IB: Innovative behaviour, Cc: Country-culture
4.2 Structural model
Table 4 shows the main parameters of the two models analysed; this allows the three hypotheses to be tested. Model 1 shows the total effect of individual learning on innovative behaviour, which is statistically significant (c = 0.74***). Model 2 shows the effect of individual learning on innovative behaviour when group learning intervenes in the model (c′ = 0.67**). This effect remains statistically significant, which allows us to accept H1, thus supporting the relationship between individual learning and innovative behaviour.
Structural model results
| Relationships | Model 1 | Model 2 | Model 3 | Model 4 | f2 | Support |
|---|---|---|---|---|---|---|
| R2GL = 0.40 | R2GL = 0.40 | R2GL = 0.40 | ||||
| R2IB = 0.55 | R2IB = 0.56 | R2IB = 0.56 | R2IB = 0.62 | |||
| H1: IL → IB | (c) 0.74*** (20.36) (0.68; 0.80) | (c′) 0.67*** (10.62) (0.67; 0.80) | (c′) 0.61*** (7.89) (0.55; 0.77) | (c′) 0.70*** (10.91) (0.66; 0.83) | Yes | |
| IL → GL = a | 0.63*** (14.52) (0.56; 0.70) | 0.63*** (14.53) (0.56; 0.70) | 0.63*** (14.61) (0.56; 0.70) | |||
| GL → IB = b1 | 0.11* (1.87) (0.02; 0.21) | 0.08 (1.50) (0.01; 0.18) | 0.09 (1.57) (0.06; 0.19) | |||
| JA → IB = b2 | 0.11* (1.93) (0.01; 0.21) | 0.20*** (3.42) (0.11; 0.30) | ||||
| H3: IL × JA → IB = b3 | 0.18*** (4.22) (0.12; 0.26) | 0.08 | Yes | |||
| Control variables | ||||||
| Cc | 0.02 (0.25) (−0.13; 0.16) | 0.04 (0.57) (−0.08; 0.17) | 0.01 (0.17) (−0.11; 0.18) | 0.01 (0.16) (−0.08; 0.13) |
Note(s): GL: Group learning, IL: Individual learning, IB: Innovative behaviour, JA: Job autonomy, Cc: Country-culture. Bootstrapping based on n = 5,000 subsamples. (based on t (4,999), one-tailed test) t (0.05; 4,999) = 1.645; t (0.01, 4,999) = 2.327; t (0.001, 4,999) = 3.092; (based on t (4,999), two-tailed test); t (0.05, 4,999) = 1.960, t (0.01, 4,999) = 2.577; t (0.001, 4,999) = 3.292. *p < 0.05; **p < 0.01; ***p < 0.001
The critical condition for determining this mediating effect is the test of the significance of a × b1 (Hayes, 2009). We used SmartPLS to obtain the value of this indirect effect (a × b1 = 0.07**), which was significant (see Table 5). This result supports H2. Consequently, a partial mediation of group learning in the relationship between individual learning and innovative behaviour was assumed, because the direct (H1 = c′) and indirect effects (H2 = a × b1) were significant (Baron and Kenny, 1986). Therefore, group learning influences the relationship between individual learning and innovative behaviour.
Summary of mediating effect tests
| Total effect on IB (model 1) | Direct effect on IB (model 2) | Indirect effect on IB (model 2) | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| BCCI | BCCI | BCCI | ||||||||||||||
| Path | t | Lower | Upper | Path | t | Lower | Upper | Point estimate | t | Lower | Upper | Sig | VAF | |||
| IL (c) | 0.74*** | 20.36 | 0.67 | 0.79 | H1: IL (c′) | 0.67*** | 10.62 | 0.68 | 0.80 | H2: ab1 (via GL) | 0.07* | 1.72 | 0.01 | 0.14 | Yes | 9.37% |
| Control variables | ||||||||||||||||
| Cc | 0.02 | 0.25 | −0.13 | 0.17 | 0.02 | 0.25 | −0.08 | 0.17 | ||||||||
Note(s): GL: Group learning, IL: Individual learning, IB: Innovative behaviour, JA: Job autonomy, Cc: Country culture. BCCI: Bias corrected confidence interval. VAF: Variance accounted for. Bootstrapping based on n = 5,000 subsamples. (based on t (4,999), one-tailed test) t (0.05; 4,999) = 1.645; t (0.01, 4,999) = 2.327; t (0.001, 4,999) = 3.092; (based on t (4,999), two-tailed test); t (0.05, 4,999) = 1.960, t (0.01, 4,999) = 2.577; t (0.001, 4,999) = 3.292. *p < 0.05; **p < 0.01; ***p < 0.001
The variance accounted for (VAF) (Hair et al., 2017) was also calculated to establish the size of the indirect effect (a × b1) relative to the total effect (c). When the VAF is between 20 and 80%, it reveals an expectation of partial mediation. In our case, according to this criterion, mediation was not indicated because, at 9.37%, the VAF was less than 20%.
The product indicator technique was then used to test the moderation hypothesis (H3: b3) that job autonomy moderates the path between individual learning and innovative behaviour (Chin et al., 2003). Model 3 includes job autonomy and Model 4 adds the interaction term (individual learning × innovative behaviour = b3) (Table 4). The result supports hypothesis H3 (b3 = 0.18***) (Table 4, Model 4). Finally, Table 4 summarises the significance and intensity of the moderating effect. Cohen’s (1988) f2 shows the significance of this change; it is between 0.02 and 0.15, which represents a small effect size. Therefore, this moderating variable affects the strength or direction of the relationship between the individual-level learning variable and innovative behaviour.
Although the cultural differences between Colombia and Spain could have affected the proposed model (Hernández Izquierdo and Marchesi, 2021), the control variable “country culture” was not statistically significant. However, other factors, such as the nature of the variables analysed, could have influenced the results obtained. In this sense, innovation could be as crucial in Colombia as it is in Spain because, in both countries, the development of innovative employee behaviour would increase the competitiveness of critical sectors.
5. Discussion
Authors such as Pasamar et al. (2019) have analysed the relationship between learning and innovation development; delving into these analyses, this work proposed three hypotheses. The first hypothesis was that individual learning and the innovative behaviour of the individual were related, and this hypothesis was supported. As the literature indicates, learning, particularly at the individual level, affects the development of innovation (Anderson and Lewis, 2014). Individual learning refers to the process by which a person acquires knowledge and skills, whilst innovative behaviour is the ability of a person to generate and apply ideas, which may be the result of the acquisition of knowledge and skills.
The second hypothesis was that group-level learning mediates the relationship between individual learning and innovative behaviour; this hypothesis was also supported. Although this effect should be interpreted with some caution, because the significance of the effect was low and the VAF was not significant, our findings represent an important contribution to the literature on learning in organisations because they relate learning at different levels to the development of innovation. This means that, in order to develop innovative behaviour, people could improve their individual learning using the experiences of others that they then internalise as their own. This would be in line with the approach of Edmondson et al. (2007), who suggested that, by viewing teams as problem-solving and information-processing systems, the concept of group learning indicates how individuals generate, share and combine knowledge within the group and explains the factors that influence the outcome of their learning behaviour. Wilson et al. (2007) also pointed out that groups have access to more information storage than do individuals, or at least that it is possible for groups to have more information, depending on who dominates the storage power at the group versus the individual level.
The third hypothesis was that job autonomy moderates the relationship between individual learning and innovative behaviour. This hypothesis was also supported, as the inclusion of job autonomy in our model positively enhances the relationship between individual learning and innovative behaviour. In the academic literature, researchers have recognised job autonomy as an important variable that significantly influences employees’ levels of motivation (Peng et al., 2010); motivation helps to improve individual innovative behaviour.
Confirming these three hypotheses has helped to fill the gap in the research that had been noted. Our data show that the group learning supports learning at an individual level to affect innovation, and that this learning is strengthened when the employee has control over his or her job.
Regarding the “country culture” control variable, it has been pointed out that different cultures can have different impacts on the development of innovation (Tian et al., 2024), but in our case this variable was not statistically significant. There could be several causes for this result. For example, cultural differences may not be significant enough to have an observable effect. This result would be in accordance with van Hoorn’s proposal (2016) that the disparities in job autonomy associated with different countries are not very marked since employers and employees face greater disparities in job autonomy when they change sectors. Another reason that could explain this result is that internal factors, such as company culture, leadership or management practices, may have a greater impact on innovative behaviour than the national culture. In relation to this, works such as that by Naqshbandi and Tabche (2018) have shown that there is a relationship between leadership, learning culture, and innovation development. Further research should determine whether company culture influences employees’ innovative behaviour more than country culture. For example, the participants in our study were employees of multinational companies in which there may be some homogeneity in employees’ innovative behaviour regardless of their country of residence.
6. Theoretical implications
This work contributes to the development of the Transactive Memory Systems theory (TMS; Wegner, 1986) by demonstrating that individual learning requires a group to generate innovation through the specialisation and communication that facilitate access to shared cognitive resources. In turn, it is based on the Job Demands–Resources model (JD-R; Bakker and Demerouti, 2007) and shows how job autonomy acts as a resource that enhances the relationship between learning and innovation, highlighting the relevance of job crafting (Petrou et al., 2012) in organisational contexts that favour learning and adaptation. Overall, the study proposes a dynamic vision of innovative behaviour as a process that begins with learning and culminates in the creation of ideas and solutions (Janssen, 2004; Manzi-Puertas et al., 2024).
7. Practical implications
The findings of this study offer relevant practical implications for managers seeking to foster innovation among their employees. In line with the results of Naqshbandi et al. (2024), it is recommended that practices based on the AMO model (ability, motivation, and opportunity) are implemented to cultivate a learning culture, which is essential for innovation (Naqshbandi et al., 2023). Such initiatives would include training programmes, internal learning networks, incentives for employees who participate in training activities, and tasks that require the application and expansion of prior knowledge. Likewise, establishing flexible organisational structures and resource management practices tailored to teams’ needs is recommended (Choi, 2024), as this fosters a dynamic environment in which knowledge is shared and applied creatively.
Furthermore, job autonomy is a crucial factor in facilitating organisational learning. Consistent with the work of Dam (1995), employees should play a central role in information acquisition, both individually and collaboratively, thereby strengthening innovative capacity. According to Galbraith (1974), decentralising decision-making improves information processing and reduces costs. Therefore, human resources practices such as recognition, training, and empowerment positively influence organisational commitment (Bhatnagar, 2007) and innovation (Chowhan, 2016). Communication and feedback are also critical in this regard (Canet-Giner et al., 2020).
8. Limitations and future research
Despite some promising findings, the present study has limitations that could be addressed in future studies. First, it was conducted using a cross-sectional design, so any inference of causality is necessarily limited. To infer causality, more research is needed, with other research designs, such as experimental and longitudinal research (Cangialosi et al., 2020). Second, the observed relationships between the variables may be biased because all the data were collected from the same sources using self-reporting measures. However, the measures were designed to allow the participants to respond without interference (Podsakoff et al., 2012). Finally, country culture was considered as a control variable. In other works, organisational culture could be further analysed to test whether company culture is statistically significant.
