Skills Measurement Strategic Leadership Based on Knowledge Analytics Management through the Design of an Instrument for Business Managers of Chilean Companies

Abstract

The growth of business intelligence and analytics (BI&A) and technological advancement is having an impact on business dynamics, implying that executives need to adjust their management skills to create value and a sustainable competitive advantage in agile environments. In this research study, a model (LDM–BI&A) consisting of knowledge and skills domains for BI&A leadership has been proposed to measure strategic leadership skills for BI&A business managers in Chilean SMEs. The skills were studied according to their demand (role requirements) and supply (university curricula in Chile). Factorial validity was determined using Confirmatory Factor Analysis, showing that the model satisfactorily fit the data. The instrument proves reliable, showing solid psychometric properties, and contributes to the literature as it has the design and empirical validation of an instrument that measures skills linked to business managers in Chilean SMEs. This will provide support to Human Resources managers on how BI&A leadership skills can be diagnosed, maintained, and developed.

1. Introduction

Business intelligence and analytics (BI&A) is often understood to be the techniques, technologies, systems, practices, methodologies, and applications with which to analyze critical business data toward improving the understanding of a company’s business, the market, and how to make timely business decisions. BI&A has become increasingly important to both the academic and business communities around the world [1]. Today BI&A is one of the fastest-growing industries and has become a strategic component of many organizations because competitive advantages can be created and so can ways to differentiate one’s wares from those of the competition [2].

There exists continuous and strong pressure from the competition in surrounding industries upon companies, which are offering similar products using similar technologies, pushing them to the limits of the possibilities of differentiation. In this competitive environment, analytical competitors extract every last drop of value from these processes as they seek to be effective in their strategic purposes [3]. Therefore, large and small companies should be on a continuous path of evaluating and improving the ways data are used, so that value and competitiveness can be engendered [4]. With competency-based knowledge analytics management, the traditional job concept is blurred, organizational charts are simplified, and the different tasks can be compared with each other so that the possibilities for mobility, promotion, career, replacement, rotation, etc., can be organized more rationally. Competency-based knowledge analytics management has also affected decision-making processes, which have become more decentralized. Although autonomy has been encouraged and each person assumes their own responsibility, increasingly, joint projects and work teams need to combine the competences of their members to achieve their objectives; hence, there must be transparency and a greater exchange of information and knowledge between all.

The steady increase in data has led to a growing need for more real-time and predictive information, paving the way for a new generation of BI&A, with a greater focus on data intelligence and with analytics becoming more forward-looking [5,6].

In this sense, access to information by managers—understanding managers as professionals that primarily direct and guide the work of others, making sure that things are completed—means that different scenarios can be evaluated immediately and different possible actions can be assessed that can be tested and compared in parallel, which could affect approaches to the decision-making process [7].

In digital BI&A environments, leaders must adapt or improve their acting skills, anticipate markets and trends, make decisions, and change or make adjustments to planning, while relying on technology in the face of changes in the environment and the market [8]. Faced with these new demands from contemporary society, manager leadership must manage, with the appropriate skills, to enable reconfigurable competitive advantages [9] for the organization and be able to deal with attitudes, values, knowledge, and skills [10], to effectively articulate change in the organization under different scenarios arising between agile environments and changing markets.

The COVID-19 crisis has accelerated the digital transformation. However, the lack of clarity about the specific skills the workers of tomorrow will need to function in the world of work is telling [11]. A review of studies on SMEs and large firms shows that research on BI&A has mainly focused on large firms [12]. The limited financial resources of SMEs have implications for BI&A investment strategies. SMEs constitute 99.5% of companies in the region (almost 9 out of 10 are microenterprises) and generate 60% of formal productive employment. However, Latin American SMEs have a particularly significant productivity gap, accounting for only a quarter of the region’s total value of output [13]. In this context, it is important to advance the studies that focus on SMEs to identify the specific benefits and barriers they face when embarking on BI&A initiatives.

McKinsey & Company [11] developed a study illustrating the demand for technological, social, and emotional cognitive skills as poised to be increasingly relevant and the needs for physical, manual, and basic cognitive skills as becoming less required. In this fast-paced context of technological advancement, effects on business dynamics, which demand that executives adjust their managerial skills, exist [14]. This presents challenges for business schools, which have to introduce changes in content and teaching in technology, aiming for a comprehensive look at business analysis and an appropriate pedagogical design to develop students’ professional skills [15].

Studies suggest a severe shortage of BA talent needed to meet current and future market demand [16,17]. However, progress in addressing these voids through education and training programs will not be easy because of the different approaches and emphases on determining what skills analytics professionals need, which leads to difficulties when designing training programs [15]. While there are adequate BI&A curricula in the US, there is no such clarity in other countries. Success in developing new leaders depends on the development of fundamental managerial skills, such as cognitive, social, and emotional intelligence [18]. However, in the supply of training programs for these skills, some key concepts and topics are not covered in sufficient depth as required by the market [19], which is a problem considering that skill development is becoming strategic [20].

From the numerous research works carried out on skills, four main approaches are depicted [21]: the Behaviorist approach, the Generic approach, the Cognitive approach, and the Constructivist approach. The Behaviorist approach (promoted by McClelland) proposes that job performance is about skills, and it is understood as being the person’s characteristics that relate to effective job performance and may be common in other situations [22,23]. The Generic approach identifies the most effective people [21], and through statistical analysis, the principal and generic characteristics of the most effective performers are defined. The Cognitive approach includes all the mental resources that individuals use to perform important tasks, to acquire knowledge, and to achieve good performance [24]. Finally, the Constructivist approach (on which this research is based) focuses on learning processes, the product of interactions that people make in a given context when talking or discussing meaning of reality. The value of one’s existing knowledge is processed and renegotiated in the interaction, and, thus, the new skill (the integrated knowledge) is the added value for the individual and for the whole learning community. According to this perspective, the training process takes place mainly through and by interaction with educational and social communication, which is constructed as a meaningful experience for individuals [25]. Consequently, interaction and collaborative social communication entail extending the concept of skill–performance to encompass social or emotional skills. This is the social–constructive approach [26], which emphasizes the similarity between the skills needed for successful performance in society (such as learning, cooperation, problem-solving, information processing, dealing with uncertainty, decision-making based on incomplete information, and risk assessment) and the development of collaborative skill (as a synonym for social–constructive learning). Therefore, it seems appropriate for the focus of this study to be advancing the identification of essential business leadership skills to manage with BI&A [27]. This research study seeks to advance knowledge of essential BI&A skills in managerial leadership, considering the evidence of obstacles faced by universities when creating and delivering appropriate curricula to meet the demand for skills in organizations, given the accelerated adoption of advanced technologies, such as BI&A and AI, while understanding the constantly changing nature of BI&A and the need for updating skills and competences [15,16,28,29].

Marzano and Kendall proposed a hierarchical taxonomy of learning domains, based on levels of awareness that increase as one advances through the levels of the taxonomy. In this sense, retrieval processes can be automatic, demanding a low level of awareness; analysis or utilization of knowledge requires thinking with a higher degree of awareness, as well as when establishing plans by the metacognitive system and much higher still, in the internal system (self), for example, for decision-making. Our model is based on the hierarchical taxonomy of learning domains [25,30]. Figure 1 shows the LDM–BI&A model, in which the Competency Model Domains for BI&A Leadership Development for Business Managers, consisting of four domains of BI&A leadership, are observed.

Professional Capability Development for BI&A business managers: Refers to the set of technical and theoretical skills and knowledge on business intelligence and analytics acquired in formal education processes, complemented by professional experience in organizations. They include mastery of concepts, characteristics, and management of terms related to business intelligence and analytics, as well as mastery of data analysis techniques, statistical techniques (descriptive, bivariate, inferential, etc.), platforms, and applications associated with business intelligence or Business Intelligence (BI).

Proficiency in the BI&A learning environment: Refers to the active promotion of a coexistence environment based on interpersonal relationships of trust, acceptance, equality, and respect among the different work teams, in order to establish an adequate work environment that allows peer support, identification, dissemination of good practices, and active learning of technologies associated with business intelligence and analytics or BI&A.

BI&A Strategic Vision Proficiency: Refers to the ability to think creatively about the future, emerging contexts, trends, and key aspects and imagine different future scenarios in order to determine their implications and possible outcomes in a global and comprehensive perspective. A person with strategic vision is capable of analyzing the effects of external factors or variables (political, social, cultural, and economic aspects of the country) and internal (the organization) on the performance of the organization of which they are a part.

Integrates BI&A skills into expert work (Habits of Mind): Refers to the ability to put into practice, autonomously, the skills and knowledge acquired in Business Intelligence and Analytics in order to solve problems that arise in the organization and, thus, achieve the attainment of organizational objectives. A person with Habits of Mind is able to set goals and make decisions about what information is needed and which tool associated with BI&A will be adequate to achieve the business objectives.

Table 1. Table of instrument specifications to measure skills linked to BI&A.

Dimensions	Items	Total
Professional Capability Development for BI&A business managers	48 to 67	20
Proficiency in the BI&A learning environment	10 to 29	20
BI&A Strategic Vision Proficiency	1 to 9	9
Integrates BI&A skills into expert work (Habits of Mind)	30 to 47	18
	Total	67

below defines each dimension of the proposed LDM–BI&A model.

In summary, the purpose of this research is to validate a standardized measurement instrument to measure skills associated with BI&A in professionals working in BI areas of small and medium-sized companies operating in different regions of Chile. This instrument will allow Human Resources areas to determine the level of professionals through the skills it measures, allowing the creation of development plans in case skills that are not well-consolidated are detected.

2. Methodology

This is a nonexperimental, descriptive investigation with a cross-sectional design [27]. A sample of 262 professionals from different regions of Chile was selected, in particular managers or executives in the role of boss, manager, and related administrators working in the area of BI. The procedure used for the selection of the sample was non-probabilistic purposive sampling [31].

Capability domains for the development of BI&A leadership skills. Each domain describes the fundamental practices, personal resources, skills, and knowledge that guide the development of business managers in BI&A leadership (compiled by the authors).

The instrument used and validated in this research study has the measurement of skills in BI&A as its purpose, and it is made up of 67 items distributed in four main dimensions. These are shown in Table 1. Forty-seven items have a 5-point Likert frequency scale that ranges from Never to Always, while the remaining twenty items, from the dimension Development of Professional Capabilities for BI&A business managers, have a binary rating, given that they only have four options and one of them is the correct option.

2.1. Procedure

To validate the instrument in order to measure skills associated with BI&A, a series of stages were carried out, which are described below.

2.1.1. Preparatory Stage

At this stage, we used the Web Data Scraping procedure to collect, from social networks such as Indeed and LinkedIn, the competency terms linked to BI&A, using the following query string: “business intelligence & analytics” or “business intelligence and analytics” or “business intelligence analytics” or * “Analytics” or “Business” or * “BI” or * “BI” or * “BI&A” or “Business Intelligence” or “Business Intelligence and Analytics” or “Analytics”. In addition, we added a restriction related to the location of job advertisements and academic study programs: Chile. From the terms collected, we selected those most conceptually related to the dimensions of the instrument and then used these terms as a reference for the wording of the items for each of the four dimensions. Finally, the preliminary version of the instrument, with its items and respective response options, was assembled.

2.1.2. Exploratory Stage

At the first instance, various scientific journal articles and books were consulted to support the instrument’s content validity. The instrument’s structure is based on the taxonomy of Marzano and Kendall [25] and a set of practical dimensions for good management and leadership established by the Chilean Ministry of Education [32].

Second, five experts in the BI&A area were consulted to assess the congruence of each item of the instrument along with the dimension it forms a part of.

Third, a quantitative analysis of the instrument’s items was carried out to examine the quality of each item according to a series of statistics. With the quantitative analysis of the items, those that did not contribute to the measurement process of the dimensions were discarded.

2.1.3. Definitive Stage

At this stage, the final version of the instrument was available after the quantitative analysis of the items.

First, the validity study was carried out through the factorial validity procedure, using Confirmatory Factor Analysis (CFA) [33] statistical technique, to confirm the instrument’s four-dimensional structure.

Second, the reliability of the measurements of the instrument dimensions as well as the instrument as a whole was estimated. For the dimensions with multi-item components, such as Mastering the Strategic Vision of BI&A, Mastering the BI&A Learning Environment, and Integrating BI&A Skills into expert work, Cronbach’s Alpha coefficient was used [34], and for the dimension Professional Capability Development for BI&A business managers, the items of which are binary rated, the Kuder Richardson KR-20 formula was used. Raju’s Beta Coefficient [35] was used to estimate the instrument measurements as a whole.

Third, norm-related scores were calculated, in particular, standardized typical scores (Zn- and T-scores). Furthermore, tests such as Kolmogorov–Smirnov and Student’s t-tests for two independent groups were used. To assess the significance of the results, a confidence level of 95% was defined, which implies a significance level equal to 0.05. The first was carried out with the aim of assessing whether the data behaved in a normal way, and the second was carried out in order to find out whether statistically significant differences existed between men and women according to their scores on each of the instrument’s dimensions.

3. Results

To start writing the items for each of the instrument dimensions, a Web Data Scraping technique was used, as explained above.

Table 2. Resource and number of scraped URLs.

Resource	Number of Scraped URLs	How
Indeed	247	Collected using Indeed search
LinkedIn	362	Collected using LinkedIn search
University study programs	172	Using a pre-collected list of URLs
Total	781

shows the sources and number of URLs scraped. From this process, a set of terms was compiled and divided into two groups, with the first and second groupings being those associated with soft skills and technical skills, respectively.

The classification consisted of evaluating each term against the definitions of the four dimensions, whereby a term related to a definition was automatically entered into the dimension, but a term not linked to any definition was excluded, as in the case of the term “Management”.

Once classification of each term had been carried out, the items of the instrument were edited on the basis of the terms classified in the four dimensions established in the domains of leadership–BI&A model (Figure 1).

The number of terms located by dimension can be seen in

Table 3. Number of terms located in the proposed dimensions.

Dimensions	Number of Terms
BI&A Strategic Vision Proficiency	10
Integrates BI&A skills into expert work (Habits of Mind)	11
Proficiency in the BI&A learning environment	2
Professional Capability Development for BI&A business managers	51

; as observed, not enough terms were found that associated with the dimension “mastering the enabling environment for BI&A learning”, with this being the dimension with the lowest number of terms. Meanwhile “developing professional capabilities for BI&A business managers” had the highest number of terms.

For the wording of the instrument’s items, the number of items was established as reflected in the specifications table shown in

Table 4. Preliminary table of instrument specifications.

Dimensions	Items
Professional Capability Development for BI&A business managers	61 to 90
Proficiency in the BI&A learning environment	21 to 40
BI&A Strategic Vision Proficiency	1 to 20
Integrates BI&A skills into expert work (Habits of Mind)	41 to 60
Total	90

.

The number of items established for the dimensions was considered sufficient to cover the characteristics of each of them. After defining the number of items to be written per dimension or domain, the content of each item was written, taking as a reference point the terms found by the Web Data scraping technique. Note that the content of most of the items is a reflection of each of the terms found by the scraping; however, some items existed that were drafted on skills based upon the bibliography [32,36], as the number of terms found was not sufficient for the number of items established per domain.

3.1. Content Validity

Initially, to guarantee the instrument’s content validity by verifying that its items covered all the characteristics of the variable to be measured [37], several articles from scientific journals, graduate papers, and books were consulted in order to provide support to the instrument’s content (items). The instrument’s structure is based on the taxonomy of Marzano and Kendall [25] and on a set of practical dimensions for good management and educational leadership established by the Chilean Ministry of Education [32].

Second, a preliminary table of instrument specifications was devised, as shown in Table 4, which portrays the dimensions that constitute the instrument and the items that measure them.

In the third and last instance, a qualitative analysis of the instrument’s items was conducted to determine whether there was congruence between these items and the dimensions they claimed to measure. To this end, five experts in the area of BI&A were consulted, all male; three were civil engineers in computer science, one was a computer engineer, and one was an industrial engineer, with ages ranging between 38 and 51 years. Three of them had master’s degrees, and the other two had doctoral degrees.

After consulting the five experts, 100% agreement was reached for eight items, which meant that all five experts agreed that these items measured the dimension they claim to measure. On the other hand, it was documented that 80% agreement was reached for 16 items, which meant that four out of the five experts agreed that these items measured the dimension that they claimed to measure. Then, 60% agreement was obtained for 18 items, where three out of the five experts consulted agreed that these items measured the dimension they claimed to measure. It should be noted that 48 items obtained percentages of agreement lower than 50%; adjustments were made to the wording of these items so that they would be linked to the characteristics of the dimension they formed a part of.

Regarding the wording of the items, it was observed that all the experts agreed that more than half of the items, specifically 59, were suitably worded. In addition, it was found that for 26 items, four out of the five experts agreed with the suitability of their wording. In the case of four items, only three of the consulted experts agreed they were well written. Finally, there was just 40% agreement that one item was well written.

Based on the aforementioned discussion, as stated, adjustments were made to the wording of items with a percentage of agreement lower than 50% so that they would be linked to the dimension’s characteristics that they were intended to measure.

3.2. Quantitative Analysis of Items

The purpose of item analysis is to examine item properties associated with instrument properties that are in development and to determine which of these are relevant, discarding those that do not aid the variables’ measurement process that the instrument is intended to measure [37].

Sample Description

To conduct analysis of the items, a non-probabilistic sample of 262 professionals from different regions of Chile was collected, of which 100 were women, representing 38.2% of the total sample, while 162 were men, representing the remaining 61.8%. The average age of the sample was 37.52 years, with a standard deviation equal to 8.52 years, making it a relatively heterogeneous sample in terms of age. The largest proportion of participants came from the city of Santiago de Chile (30.2%), followed by Valparaíso (17.2%) and Viña del Mar (15.6%).

Regarding education level, most participants were university-educated, specifically 128 cases, representing 48.85% of the total sample, followed by 113 cases (43.13%) with a master’s degree, 7.63% with at least one specialization, and only 0.38% with a doctorate.

In relation to participant occupation, the largest proportion worked in accounting, specifically 52 cases, representing 19.85%, followed by a group of 49 cases, who were in the field of engineering, representing 18.70% of the total sample.

Regarding the quantitative analysis of the items, three statistics were selected: arithmetic mean, standard deviation, and Pearson’s product–moment correlation coefficient.

Table 5. Scale for multipoint items.

Statistics	Criteria	Score	Decision
Arithmetic Mean	Less than 2 or greater than 3	0	Rejected: Total score less than 2
	Between 2 and 3	1
Standard Deviation	Less than 1.10	0
	Greater or equal to 1.10	1	Accepted: Rating greater than or equal to 2
Item-Dimension Correlation	Less than 0.70	0
	Greater or equal to 0.70	1

shows the item score that is a frequency score ranging from Never to Always (Multipoint); within this score are the statistics chosen and the criteria established for each value an item yields in each statistic. An item will be accepted when its arithmetic mean ranges between 2 and 3, its standard deviation is greater than or equal to 1.10, and its correlation coefficient is equal to or greater than 0.70. Values within the aforementioned ranges are assigned one point, adding up to a total of three points; therefore, if the total score of an item is greater than or equal to it, it is accepted and included in the final version of the instrument.

Table 6. Scale for binary score items.

Statistics	Criteria	Score	Decision
Mean	Less than 0.30 or greater than 0.60	0	Rejected: Total score less than 2
	Between 0.30 and 0.60	1
Standard Deviation	Less than 0.25	0
	Greater than or equal to 0.25	1	Accepted: Rating greater than or equal to 2
Item-Dimension Correlation	Less than 0.30	0
	Greater than or equal to 0.30	1

shows the scale for items with a correct answer; i.e., those scored as binary (0 and 1). For this scale, an item is accepted when its arithmetic mean ranges between 0.30 and 0.60, its standard deviation is greater than or equal to 0.25, and its correlation coefficient is greater than or equal to 0.30.

Of the 90 items, 67 were accepted (74.44% out of the total number of items), while 23 were rejected (25.56%). The 67 accepted items had arithmetic means located in the central values, were able to discriminate between people who selected the different options that these items offered, and had the ability to differentiate between people who obtained high and low scores. The rejected items had extreme arithmetic means in that they did not adequately discriminate between people because they assumed relatively homogeneous behaviors, with little dispersion with respect to the mean.

From the item analysis, a new table of specifications was derived, which can be seen in Table 1, showing the BI&A Strategic Vision Domain as the dimension with the lowest number of items; however, this number is considered sufficient.

3.3. Construct Validity

A factorial validity study was conducted to demonstrate the association of the scores yielded by the instrument with the four skills of the LDM–BI&A model [38].

Factorial Validity

This was done by means of the CFA multivariate statistical technique, which allows comparison of a model constructed in advance, in which the researcher, basing it on a theory, already knows how the variables involved are related [39].

Figure 2 shows the factorial structure of the instrument that aimed to measure skills associated with BI&A, with factor 1 being the Development of Professional Capabilities for BI&A business managers, factor 2 being Proficiency of the BI&A learning environment, factor 3 being Proficiency of BI&A Strategic Vision, and factor 4 being Integrates BI&A skills into expert work (Habits of Mind) dimensions, most items had low-to-high correlations with their respective factors. Note that the items of the first factor have the lowest factor loadings (correlations), with item 60 being the one with the smallest factor loading, specifically 0.09.

After the EFA, the measurements of the model’s goodness of fit were obtained in terms of the chi-squared likelihood ratio gathered, $\chi$² (2217, N = 262) = 3323.90, p < 0.001, which indicates that the result was statistically significant at the 0.05 level and designates statistical significance to the differences between the observed data matrix and the matrix estimated by the model. Therefore, the proposed factorial model does not fit the data [40]. Note that the chi-squared likelihood ratio is sensitive to sample size; hence, the calculation of other measures of goodness of fit is necessary. The standardized chi-squared test obtained $\chi$²/gl = 1.57, a result within the threshold, which is usually a range from 2 to 3 [41].

The Normative Fit Index yielded the Comparative Fit Index (CFI) = 0.918, a result above the threshold (0.90), meaning that the estimated model differs to a greater degree from the null model (model where the variables are not related), so the variables are related to their factors.

In relation to the root mean square error of approximation, RMSEA = 0.047 was obtained, with a 90% confidence interval ranging between 0.44 and 0.50; hence, there is an adequate fit of the model to the data, as the value obtained is below the threshold (0.050), which is an acceptable value. The value obtained expresses the amount of variance the factorial model cannot explain per degree of freedom.

A Bentler–Bonet Normative Fit Index (NFI) of 0.804 was obtained, which is below the threshold of 0.90. However, it can be accepted as it is close to the recommended value; therefore, the estimated factorial model is better than the null or independent model.

A Tucker–Lewis Index (TLI) of 0.915 was obtained, above the threshold of 0.90. This result allows us to affirm that the estimated factorial model is better than the null model, serving also as additional evidence to accept the proposed model or factorial structure.

3.4. Reliability

The internal consistency method was used to evaluate the consistency of the instrument’s measurements when evaluated at different times [42].

Table 7. Cronbach’s Alpha coefficients.

Dimension	Cronbach’s Alpha	No. of Items
BI&A Strategic Vision Proficiency	0.94	9
Proficiency in the BI&A Learning Environment	0.98	20
Integrates BI&A Skills into Expert Work (Habits of Mind)	0.98	18

shows the internal consistency coefficients obtained through Cronbach’s alpha. The dimensions BI&A Strategic Vision Domain, BI&A Learning Environment Domain, and Integrate BI&A skills into expert work, obtained high coefficients; therefore, the items that encompass them are strongly related to each other. Hence, they measure the dimensions they claim to measure very homogenously and with the same direction, degree, and intensity.

Table 8. KR₂₀ formula coefficient.

Dimension	KR20	No. of Items
Professional Capability Development for BI&A business managers	0.68	20

shows the coefficient calculated using the KR₂₀ formula for the Professional Capability Development for BI&A business managers dimension as equal to KR₂₀ = 0.68, which is moderate but acceptable.

Table 9. Raju’s beta coefficient.

β	No. of Items
0.80	67

shows Raju’s beta coefficient, ideal for estimating the reliability of the measurements of an instrument of an instrument with dimensions that are independent, so it is not possible to obtain a single score for the instrument. This instrument obtained a coefficient equal to β = 0.80, which indicates that the items measure with the same direction, degree, and intensity the variables they claim to measure.

3.5. Scores Related to Standards

Normative scores, or norms, represent the behavior of a sample in the instrument, giving insight into the typical behavior of a group of people in the instrument [43].

There are several norm-related scores; however, each may or may not require the gross scores to behave normally. To evaluate whether the original scores of the instrument behave according to a normal distribution, the non-parametric Kolmogorov–Smirnov test was used for a sample, the results of which are shown in

Table 10. Kolmogorov–Smirnov test for one sample.

	BI&A Strategic Vision Proficiency	Proficiency in the BI&A Learning Environment	Integrate BI&A Skills into Expert Work	Professional Skill Development for BI&A Business Managers
Test statistic	0.079	0.076	0.075	0.076
p-value	0.000	0.001	0.001	0.001

Note: the theoretical distribution established is the normal distribution.

. As can be seen, all the dimensions obtained statistically significant values; therefore, the scores of each of them behave in a non-normal manner.

Based on the results presented, derived normalized typical scores are calculated for all dimensions, specifically T-scores, which are appropriate for distributions that behave non-normally.

To know whether normative scores are to be calculated for men and women, Student’s t-test was calculated for two independent samples for each dimension, according to gender, the results of which are shown in

Table 11. Student’s t-test of two independent groups for gender.

	t	gl	p-Value
BI&A Strategic Vision Proficiency	−0.433	260	0.665
Proficiency in the BI&A Learning Environment	0.124	260	0.901
Integrate BI&A Skills into Expert Work	−0.455	260	0.649
Professional Capability Development for BI&A Business Managers	−0.702	239.509	0.483

. As can be seen, all the values for each dimension were not statistically significant, which indicates no significant differences in the scores according to gender; that is to say, women tended to score similarly to men in the four dimensions that constitute the instrument; hence, it was not necessary to calculate normative scores by gender.

Finally, based on the results obtained, normalized typical scores, namely, Zn- and T-scores, were calculated. These scores report the distance of a person evaluated with respect to the average of the distribution. It is important to add that Zn-scores have a mean of 0 and a standard deviation of 1, while T-scores have a mean of 50 and a standard deviation of 10.

4. Discussion

Research on the development of leadership competencies in BI&A has been mainly of an exploratory nature to identify these competencies, oriented to case studies or recent studies related to identifying the hard and soft competencies of supply and demand in the market, that is, in higher education programs (supply) associated with BI&A or on the demand side, BI&A competencies required in job positions, as well as other studies associated with identifying BI&A competencies for professionals for the coming years.

For its part, research related to BI&A curricula in higher education is scarce and has been rather focused on the United States [11,15,44,45,46].

Consequently, this research responds precisely to the lack of an articulated model of the knowledge domains and skill sets necessary for leadership in BI&A, with this purpose a standardized instrument based on LMD–BI&A, Figure 1 is designed to measure the strategic leadership competencies linked to BI&A in business managers of small and medium-sized Chilean companies, to support Human Resources managers on how to diagnose, maintain, and develop leadership competencies in BI&A.

A Web Scraping procedure was carried out in order to collect terms associated with BI&A skills from social networks, such as Indeed and LinkedIn. This allowed us to choose terms mostly conceptually related to the dimensions of the instrument, which were used as a reference for the wording of the items.

The content validity of the instrument was studied by consulting five experts in the BI area. Consequently, modifications were introduced in the wording of the items, so that they would be conceptually linked to their respective dimensions. Next, a quantitative analysis of the items was performed to determine the properties of the items directly associated with the properties of the instrument [37], thus obtaining a new version of the instrument with 67 items and discarding the items that showed little discriminatory capacity.

Construct validity was assessed by means of factorial validity. The CFA was used, which obtained results that confirmed the hypothesized four-dimensional factor structure, given that the satisfactory fit of the model to the data such as the CFI, NFI, and other measures of the quality of the fit yielded acceptable indices, recommending acceptance of the fit of an EFA model [47]. The model fit was also good when considering the residuals, as the RMSEA index value was <0.047 [40] and was even lower than the more restrictive threshold of 0.06, recommended by Hu and Bentler [48].

The reliability of the instrument as determined by Cronbach’s Alpha, the KR₂₀ formula, and Raju’s beta coefficient yielded coefficients that account for a reliable instrument according to Shahirah S. and Moi N. [49].

Finally, norm-related scores were calculated, which, although they do not have psychometric properties, allow standardizing the measures of the instrument, thus facilitating the interpretation of the results in the person evaluated. Statistical tests such as the Kolmogorov–Smirnov test and Student’s t-test for two independent samples were used. The Kolmogorov–Smirnov test revealed that the distribution of the scores of each dimension of the instrument behaves in a non-normal way. This, according to Lezama [43], is usually due to sampling errors, so it was pertinent to calculate the normalized scores to convert the non-normal distribution into a normal one, which allows us to contrast the scores under equal conditions. Thus, the Student’s t-test provided evidence of the inexistence of differences in the way women and men scored in the four dimensions comprising the instrument, which affirms that both men and women tend to behave or score similarly in the four dimensions of the instrument.

5. Conclusions

In summary, with respect to construct validity, it can be affirmed that the scores of the BI&A competency instrument are indicators of the four established dimensions, having valid measures, given that the model was a satisfactory fit to the data and that the items achieved high factor loadings with their factors.

Regarding the reliability, the scores obtained tend to be consistent over time, in other words, the instrument yields reliable measurements.

Thus, the instrument that measures the skills associated with BI&A assumes robust psychometric properties, given that its measures are valid and reliable. Hence, it can be used for the purposes it is intended to serve, contributing to the literature with the design and empirical validation of a standardized instrument used to measure skills linked to business managers in small and medium-sized Chilean companies and opening a new avenue for future research. In addition, this study informs Human Resources managers on how to diagnose, maintain, and develop leadership skills in BI&A for the updating of relevant skills in demand as well as the skills required in this area for professional development.

6. Limitations and Suggestions

This research study showed that the instrument yields valid and reliable measures although it is not free of limitations, which are noted below with some recommendations for future research.

One limitation of this research study lies in restricting the emphasis to social values, economic conditions, and historical moments, as these are the decisive elements that explain how these phenomena are created because of their complexity and psychological characteristics that reflect the broad structures of a given society.

Considering that the instrument presents valid and reliable measures, a revaluation of its properties in larger samples is also proposed, by running pilot tests to assess whether the examinees understand the instructions and content of the items.

Advancing research studies in BI&A, by applying underlying concepts and theories to other disciplines, such as psychology, is recommended [45].