1.1. Literature Review

In a highly globalized society like the current one, companies constantly generate new products and services to meet the consumption needs of their customers. This has caused significant changes in the markets for the consumption of goods and services, forcing business leaders to seek new ways of conducting business, and to encourage new leaders who have the social and personal skills necessary to face these challenges Arceo (2020). Even in countries like the U. S., where the study of leadership is a common practice in universities and research centers, 60% of its inhabitants consider that companies are run by people who are not suitable for the exercise of organizational leadership (Román-Rojas, 2020).

Currently, the academic environment affirms that leadership is a dyadic relationship between the leader and their followers and resides in the ability that the former possesses to articulate and manage the coordinated efforts of subordinates to reach specific goals. Therefore,

leaders need tofind ways toovercome thechanges in theorganizational

environment and develop competitive and sustainable companies where information is excessive and disorganized (Gonos & Gallo, 2013; Iqbal, Anwar, & Haider, 2015; Gez Romero, 2016; Benavides, 2017; Qian, Song, Jin, Wang, & Chen, 2018; Zhang, Song, Wang, & Guangjian, 2018; Sousa, 2019; Palazzeschi, Bucci, & Di Fabio, 2018; Cheung, Yeung, & Wu, 2018; Báez, 2019; Lez, De la Garza, & Zavala, 2020; Rojero-Jiménez, Gez-Romero, & Arrieta-Cabrales, 2022; Cahyani, Diana, & Jamiah., 2023).

Leadership has been widely studied by the administration, as well as by related disciplines, since the mid-seventeenth century when authors such as Carlyle (1840) and Galton (1869) laid the foundations of the theory of the Great Man, and therefore, of the studies of greatness, leadership, and the exercise of power in business organizations. Sin-ce then, numerous studies and investigations have been published delimiting the historical study of this variable over time. However, one way to categorize its evolution chronologically is the one presented by García-Solarte (2015), who divides the study of this variable in four main approaches:

1. Feature Approach: Based on the theories about the Great Man, and the contributions of classical management theory. This approach

defines the determining features of people who were supposedly

born with the gift of leadership.

2. Behavioral approach: It emphasizes behavioral theories of psychology, and its main authors are Lewin, Lippitt and White (1939), Likert (1961), Blake and Mouton (1964), who state that leaders

hip styles can declassified in two main aspects: An autocratic or

task aspect whose main goal is the achievement of institutional objectives, and its counterpart, a democratic aspect that fosters interpersonal relationships.

3. ContingentApproach:This approach arises from the difficulty ofthe behavioral approach defining a single leadership style that will solve

the problems of organizations in highly dynamic and competitive environments. This involved determining under which operational conditions each of the styles worked best, whether focused on tasks or people. There are two main models, the one developed by Fiedler (1971), who stated that the leadership style depended on the amount of power of the leader and the nature of the tasks.

The other model was developed by Hersey and Blanchard (1969) who established that, depending on the level of maturity of the worker, a certain style of leadership must be applied. Thus, autocratic leadership must be applied to less mature workers, while democratic models can be applied to mature workers.

4. Emerging approach: This approach is the most utilized during the last decade and its main theories are transformational leadership,

authentic leadership and learning leadership. The first of these

inherits the task and people approaches that were developed in the previously mentioned theories, in their transactional and transformational versions. Authentic leadership is the next evolutionary step of transformational leadership and tries to give a more ethical dimension to the exercise of power over subordinates by putting it not only at the service of organizational objectives, but also of the group interests of the workers. Finally, learning leadership promotes the need for companies willing to generate and transmit knowledge, both within the company and abroad.

It is precisely in this last approach where the present investigation focuses on, with the purpose of deepening into these theories and the possibility of generating a new leadership model that combines the main and predominant elements of transformational leadership, authentic leadership, and learning leadership.

1.2 Transformational Leadership

This theory was developed by Burns (1978) to delve into the task and people approach, which were called transactional and transformational leadership, respectively. This theory was studied in greater detail by Bass and Avolio (1990) who determined seven main factors of trans

formational leadership and unified the approaches of tasks and people.

These factors are detailed below:

are encouraged to solve organizational problems in different ways,

as well as to promote the development of creativity and innovation within the organization. It is also known under the name of crea

tive leadership and finds its foundation in the labor of Hersey and Blanchard (1969) and has been extensively developed by different

authors (Drucker, 1985; Vizcaya, 2015; Labelle & Réyes-Lez, 2018; Jiménez & Torrens, 2018; Dilts & Sempau, 2022; Otong & Tagadiad, 2023).

carry out their work as they see fit (Lewin, Lippitt, & White, 1939).

However, for this to happen the worker’s labor and psychological maturity is a must (Hersey & Blanchard, 1969). Its main origin

is the theories of motivational self-efficacy and implies that the

leader’s participation is only carried out when the tasks cannot be resolved by the collaborators (Wellman & LePine, 2017; Khan & Tidman, 2021; Lundmark, Richter, & Tafvelin, 2021; Achhnani, 2021; Ahsan & Khalid, 2023).

Due to the magnitude of this model in the academic literature of administration, it is one of the theories chosen in this research. As expressed by Dimitrov and Darova (2016) it is the leadership theory that has marked the end of the 20th century and the beginning of the 21st century.

1.3 Authentic Leadership

Due to the brevity of its postulation in the academic field, authentic leadership is little known by leadership scholars and practitioners (Mubashar & Fida, 2023). Although the objective of leadership is to take advantage of the efforts of collaborators, in some cases, the improper use of power by supervisors or leaders often causes organizational dehumanization (Caesens, Nguyen, & Stinglhamber, 2019). That is why some authors have chosen to assess the impact of authentic leadership on organizational dehumanization, finding an inversely proportional relationship between these variables (Sainz, Delgado, & Moriano, 2021), that is, authentic leadership promotes behaviors or conducts focused on respect for the human and labor dignity of workers.

Authentic leadership arises because of the reflection on the conduct of organizational leaders in the Enron, Tyco and some other cases, which caused considerable economic losses not only for the companies, but also severely affected the economies of the countries generating laws like Sarbanes-Oxley. Derived from these behaviors in organizational leaders is that Bill George (2003, 2007) creates the concept of authentic leadership. Authentic leadership has been defined as the product of real psychological abilities and capacities by leaders through selfaware states and self-regulated positive behaviors to promote reciprocal self-development at the axiological level, individually, and in groups in an advanced business environment (Luthans & Avolio, 2003; Avolio & Gardner, 2005). Thus, wecandefinethat authentic leadershipis thenext evolutionary step of transformational leadership by providing a more human and ethical perspective to leadership; by giving a sense and identity of service to the workers of the organization, which is possible thanks to the fact that “Based on the theoretical study of leadership styles and social learning theory, it is explained that employees acquire values, emotions, attitudes and behaviors from leaders” (Mehrad, Fernández-Castro, & Olmedo, 2020, p. 67).

Additionally, Walumbwa states that “it is a pattern of behavior that promotes and is inspired by both positive psychological capacities and a positive ethical climate, to foster greater self-awareness, internalized morality, balanced information processing, and transparency in the relationships between the leader and the followers” (Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008, p. 94).

This allows us to clarify what the dimensions of authentic leadership

are, which were defined through a study carried out by Walumbwa et

al. (2008):

a. Self-awareness: This dimension represents the self-knowledge of strengths and weaknesses of the leader or manager. It is based

on one of the five disciplines of Senge (1990), which establishes

the full knowledge of the mental models of the leader (Sebastian & Hn, 2023).

emotions or feelings to influence the rational decisions that the

leader must make.

d. Internalized morale: This dimension is the leaders’ ability to self-regulate their behavior according to their values and principles, without being influenced by group or social pressures. In the academic literature on authentic leadership, this dimension is extremely valuable, giving meaning to the remaining three dimensions, as it is the axiological basis of authentic leadership and ethical leadership (Ferrell & Ferrell, 2011; Langvardt, 2012; McCann & Holt, 2013; Skubinn & Herzog, 2014; Eisenbeiss, van Knippenberg, & Fahrbach, 2014).

The leadership through the transactional approach provides organizations with a way to perform tasks more efficiently and is also a way to prolong the effort and strengthen the organizational identity. Furthermore, leadership provides a sense of how to perform its tasks while respecting the dignity of the worker, and at the same time strengthening socially responsible behavior in the leader. However, how can this form of work bemaintainedover time?This becomes considerably moredifficult if we consider that leaders can also be removed or promoted in their responsibilities. For this reason, it is necessary for leadership to be reflected in the organizational culture, in such a way that, despite the changes that may arise in those who occupy positions of power in the organization, the nature of the company continues to work under the fundamentals of transformational leadership and authentic leadership.

1.4 Learning Leadership

The concept of “learning” organizations became popular towards the end of the 20th century with the publication “The Fifth Discipline” by Senge (1990), who presented a series of organizational practices that allowed the company to adapt and transmit knowledge for solving problems and everyday challenges (Elkjaer, 2020).

According to Buch (2020), organizational learning occurs when people learn as a community and acquire new ways of working. This results in innovationandhigheffectiveness sinceit provides organizations withthe tools for knowledge acquisition, information distribution and understanding of the company’s operation (Iqbal, Shah, Shabbir Ali, & Shahzadi, 2021). Gorzelany-Dziadkowiec (2015) states that the company must manage knowledge as the most valuable and strategic resource to modify procedures and adapt organizations to the environmental conditions in order to make them highly competitive. However, for this to happen, not only the workers’ participation is necessary, “but also, and above all, the managers’ “ (p.336). For these reasons, Yeung, Ulrich,

Nason and Von Glinow (1999) made a publication where they defined

12 characteristics of the leaders of learning organizations, which are categorized into Gez Romero (2016) three main groups:

c) Visionary: Motivates their followers in search of challenges and goals.

Now that these three important models of leadership have been presented, this works aims to develop model that involves these theories. The resulting unifying model of leadership will have the purpose the help MSMEs to exercise better leadership practices within their organizations.

2.1. Instruments

Three different instruments for assessing leadership were used aiming to stablish a new multidimensional model to better describe the main characteristics of a leader in the MSMEs context. The three instruments:

1) transformational leadership (Bass & Avolio, 1990), 2) authentic leadership (Walumbwa, Avolio, Gardner, Wernsing, & Peterson, 2008), and 3) learning leadership (Yeung, Ulrich, Nason, & Von Glinow, 1999) were selected for this study due to their reliability to assess leadership

aspects in business research from companies of different contexts and

cultures (Tejeda, Terri, & Rajnandini, 2001; Calder Mafud, Beltrán, & Pando-Moreno, 2020).

The MLQ-6S instrument (Bass & Avolio, 1990) was selected for the transformational leadership to assess this variable using 14 items divided in seven subconstructs. For the authentic leadership instrument (George, 2003) eight items were considered to assess four subconstructs of this variable. And for the learning leadership instrument (Gez Romero, 2016) nine items were used to assess three subconstructs. All these instruments use a 5-point Likert scale ranging from 1 to 5 to assess leadership.

The 31 items from the three instruments were combined aiming to develop a new model for assessing MSMEs leadership in a multidimensional way for Latin American populations.

2.2 Participants

Data was collected from 384 employees of MSMEs in a Northern city of Mexico. From this population, 204 participants were males and 174 females. The whole sample (n=384) was divided in two halves (n=192) for conducting different factorial analyses. The sample size of both halves was considered big enough for the factorial analyses since Cattell (1978) suggests a ratio sample size between three and six subjects per item, while Gorsuch (1983) suggests a minimum of five subjects per item for each factorial analysis. Both samples used for the factorial analyses were above these suggestions, with 192 participants to analyze 31 items for a ratio of 6.19 subjects per item. Additionally, the Kaiser-Meyer-Olkin (KMO) test was conducted to analyze the items correlation of the sample (Kaiser, 1970). The KMO of the first and second halves of the sample was 0.91 and 0.95 respectively, suggesting that these samples were adequate for conducting factorial analyses (Hutcheson & Sofroniou, 1999).

The normality of the samples was tested using the skewness, Kurtosis and Mardian tests expecting a normal distribution of the data that could facilitate the factorial analyses process (Curran et al., 1996; Muthent & Kaplan, 1992).

2.3 Exploratory Factor Analysis

The exploratory factor analysis (EFA) is a data driven test that generates a factorial model that determines if there is an underlying latent structure between the items of an instrument [75]. The EFA is usually conducted in social sciences research when researchers are designing or adapting new instruments (Lee et al., 2018).

The whole sample (n = 384) showed a non-normal distribution with skewness values ranging from -0.6542 to 0.4944, Kurtosis values from 1.9221 to 2.7037; and a Mardia’s coefficient p-value < 0.001 indicating a violation of the multivariate normality. Following these results an EFA was conducted using an ordinary least squares estimator to find the minimum residual solution. This is a maximum likelihood with robust standard errors EFA that fits non-normal data distributions (Revelle, 2014). A promax rotation was selected to conduct the EFA to allow correlations between factors during the process, aiming to obtain a more realistic interpretation of the multidimensional model (Fabrigar et al., 1999) For this EFAresults interpretation, a cutoff valueof 0.32 or higher was established for the correlation factors to consider an item as a part of one factor (Tabachnick & Fidell, 2001).

Before conducting the EFA, both a scree plot test, and a parallel analysis were conducted aiming to determine the most appropriate number of factors for conducting the EFA (Kaiser, 1970; Fabrigar et al., 1999). The results of these two tests were compared side by side resulting in an EFA with three factors, which were in line with the theories of business leadership that were used to develop this new multidimensional model: 1) transformational leadership, 2) authentic leadership, and 3) learning leadership.

2.4 Confirmatory Factor Analysis

Aconfirmatory factor analysis (CFA) was conducted with the second half of the sample (n=192) following the EFA results. The CFA was conducted with the items that grouped in one of the three subconstructs established by the EFA results with a factor loading of 0.32 or higher. The CFA explicitly test the model suggested by the researcher, testing if each item is loading in the latent factor suggested by the EFA results (Brown, 2015; Fernández et al., 2010). The items covariation was allowed during the CFA to continue with the promax rotation structure that was used for the EFA (Fabrigar et al., 1999).

Several fit indices were used toevaluate the CFAresults and the significance level of the items grouping within one of the three subconstructs of the multidimensional model of business leadership. These fit indices included chi-square,which should be non-significantata p-value > 0.05 [81]; Comparative Fit Index (CFI), expecting acceptable values above

0.95 (Brown, 2015; Hu & Bentler, 1999); Tucker Lewis Index (TLI), expecting acceptable values above 0.95 (Brown, 2015; Hu & Bentler, 1999);

and root mean square error of approximation (RMSEA), expecting values

less than 0.08 for moderatefit, 0.05 for a good fit, or 0.01 for excellent fit

(MacCallum et al., 1996). Due to the non-normal distribution of the data, the Satorra-Bentler correction was used during the CFA aiming to adjust

the fit indices and the chi-square with the goal of finding a more robust

solution for the CFA. During the CFA, the items with lower factor loadings

(slightly abovethe cutoff line of 0.32) from the EFAresults were deleted one by one to run an additional CFAaiming to improve the fit indices, and

this process was repeated several times until reaching the necessary

statistical significance (Batista-Foguet et al., 2004). The elimination of

items during both, the EFA and the CFA is a common practice during the validation process of new and adapted instruments (Lee et al., 2018; Morán-Soto & González Pe, 2022), and losing items would ultimately help to validate the data collected with the multidimensional MSMEs leadership instrument. Figure 1 shows the validation process of the new multidimensional leadership model describing the psychometric tests.

Figure 1. Psychometric validation tests designs for the multidimensional leadership model.

2.5 Internal Consistency

At the end of the validation process, the internal consistency of the 22

remaining items was tested using Cronbach’s alpha (α) tests for each

of the three subconstructs, expecting values of 0.8 or higher (Thorndike & Thorndike-Christ, 2010).

3.1 Exploratory Factor Analysis

The results of the EFA are shown in Table 1. This table does not

show the five items that were deleted for not reaching the minimum

factor loadings required during the EFA, three from the transformational leadership (e.g. My boss rewards those who share knowledge with others), one from the authentic leadership (The others have full

confidence in my boss and feel good in their company), and one from the learning leadership subconstructs (My boss helps us find meaning

in our work).

Table 1. EFA results showing all items that loaded to a factor

3.2 confirmatory Factor Analysis

The CFA was conducted with the 26 remaining items after the EFA

results (see Table 1). The fit indices with these 26 items did not reach

the accepted values, hence the items with the lower factor loadings

from the EFA (slightly above the cutoff line of 0.32) and the CFA were systematically deleted (see Table 1) aiming to strengthen the fit indices

of the new model as suggested by Brown (2015) and Hu and Bentler (1999). Four additional items were deleted following this methodology, two from the transformational leadership (e.g. My boss explains to the workers how to do things), one from the authentic leadership (My boss’s actions reflect their personal values), and one from the learning leadership subconstructs (My boss looks outside the company for people who

can provide different solutions to those used previously). Losing the

items with the weaker factor loadings resulted in a stronger model to assess leadership in MSMEs, which ultimately provided evidence that the remaining items could describe this type of leadership in a thorough way by establishing the subconstructs that better explain how leadership is perceived by MSMEs employees.

At the end, the final version of the instrument for assessing business

leadership kept 22 items (see Table 1). The CFA results for this model

were acceptable with fit indices of 0.957 for the Comparative Fit Index,

0.952 for the Tucker Lewis Index, and a root mean square error of appro

ximation value of 0.056. Although the chi-square statistic was significant

at a p-value < 0.001, these fit indices were considereda good model fit

due to the relatively small sample size of the second half of the sample

(n=192). All the correlations tested with the CFA were significant at a

p-value < 0.001, suggesting that this multidimensional model could be used to collect and analyze data about MSMEs leadership with similar populations.

3.3 Internal Consistency

The Cronbach’s alpha values of the three constructs resulting from the factor analyses showedacceptableinternalreliability coefficients, withα = 0.949 for the transformational leadership, α = 0.861 for the authentic leadership, and α = 0.803 for the learning leadership.

3.4 The Transformational Leadership Subconstruct

The transformational leadership subconstruct kept 13 items after the factor analyses, keeping most of their original items and absorbing some of the items that were supposed to load in the learning leadership subconstruct according to the original theory. An additional EFA was conducted with the 13 items that the transformational leadership subconstruct kept. This EFA was considered necessary to determine the number of subconstructs that the transformational leadership subconstruct should have, since this subconstruct took away the three items assessing the visionary leadership from the learning leadership subconstruct, and kept only one item from the individual consideration, the idealistic influence, the laissez faire, and the contingence rewards transformational leadership subconstructs (Fabrigar et al., 1999). The EFA results showed that the best way to classify the transformational leadership subconstruct was in two subconstructs of six and seven items, with items X8, X9, X10, X11, X12, and X14 for the first subconstruct (see Figure 1), and items X7, X15, X17, X18, X20, X21, and X23 for the second subconstruct (see Figure 1). It is important to highlight that the correlation factors for the 13 items analyzed with this additional EFA were all above 0.47, and no items were deleted from the transformational leadership subconstruct analysis.

Figure 1. Final version of the instrument to assess multidimensional business leadership.

Funding sources

Thisresearch did notreceive anyspecificgrantfromfunding agencies in the public, commercial, or not-for-profit sectors.