Objective

This report presents a comprehensive descriptive analysis of the data collected to validate the adoption of the proposed methodology.

Data dictionary:

-Satisfaction: 1=Very dissatisfied, 2=Dissatisfied, 3=Neutral, 4=Satisfied, 5=Very satisfied -Usability: 1=Strongly disagree, 2=Disagree, 3=Neutral, 4=Agree, 5=Strongly agree

1.1 Data review

The following section aims to verify the variables contained in the database, where a total of 17 variables and 279 records were identified, including categorical and numerical variables. For further analysis, these variables will be divided into two large groups: the first related to characteristics of the level of satisfaction with the methodology, and the second related to usability.

Age: respondent age.

Gender:

-Male -Female -Non-binary/other -I prefer not to answer

Type of affiliation:

-Emeritus researcher -Senior Researcher -Associate Researcher -Junior Researcher -Researcher in training (graduate student) -Undergraduate Student -Affiliated Member

Research group:

-BRIAV32 -ESCAB -ESMIC -BAIDI -ESCOM -ESAVE

Satisfaction:

1. Satisfaction with methodology
   
2. Willingness to implement
3. Alignment with the Army
   
4. Clarity and structure
   
5. Improvement in innovation management

Usability:

1. Frequent use
2. Ease of understanding
   
3. Quick learning
   
4. Confidence in guidelines
5. Organized structure

## Rows: 279
## Columns: 17
## $ Age                                       <dbl> 34, 31, 33, 26, 32, 28, 30, …
## $ Gender                                    <chr> "Male", "Male", "Male", "I p…
## $ `Type of affiliation`                     <chr> "Researcher in training (gra…
## $ `Research group`                          <chr> "ESAVE", "ESAVE", "ESAVE", "…
## $ `1. Satisfaction with methodology`        <dbl> 4, 5, 3, 5, 5, 4, 4, 5, 4, 4…
## $ `2. Willingness to implement`             <dbl> 4, 5, 4, 4, 5, 4, 4, 5, 4, 4…
## $ `3. Alignment with the Army`              <dbl> 4, 5, 4, 4, 5, 5, 3, 5, 4, 5…
## $ `4. Clarity and structure`                <dbl> 3, 5, 4, 4, 5, 4, 4, 4, 4, 4…
## $ `5. Improvement in innovation management` <dbl> 3, 4, 4, 4, 5, 5, 4, 4, 3, 5…
## $ `1. Frequent use`                         <dbl> 4, 4, 5, 4, 5, 4, 4, 4, 4, 4…
## $ `2. Ease of understanding`                <dbl> 3, 4, 4, 4, 4, 3, 5, 4, 5, 4…
## $ `3. Quick learning`                       <dbl> 5, 2, 4, 5, 4, 4, 4, 3, 4, 3…
## $ `4. Confidence in guidelines`             <dbl> 4, 4, 5, 3, 3, 4, 4, 3, 5, 3…
## $ `5. Organized structure`                  <dbl> 4, 4, 5, 4, 4, 4, 5, 3, 3, 3…
## $ Index_Satisfaction                        <chr> "NO", "SI", "NO", "SI", "SI"…
## $ Index_Usability                           <chr> "SI", "NO", "SI", "SI", "SI"…
## $ Global_Score                              <chr> "NO", "SI", "SI", "SI", "SI"…

Variable type: character

Variable type: numeric

The above data enables an initial review and evaluation, allowing for the observation of a total of 279 records, which corresponds to the total number of surveys conducted. There are a total of 6 categorical data points and 11 numerical data points, with each record containing 17 columns of information provided by the personnel surveyed. According to the population census, a total of 354 samples were expected to be obtained, which is the total population.

For our analysis, a total of 279 samples were obtained from the total population N, which is a good sample, covering 78.8% of the population, with a margin of error of +-3.2% (better than the standard of +-5). This sample N has the capacity to detect subtle differences and effects between variables and greater statistical power to perform multivariate analyses, with a representativeness that significantly reduces sampling bias, meeting the statistical requirements for the results to be generalized to the entire population N.

1.2 Missing data

## Valores cero en edad: 7

According to the analysis carried out in section 1.1, no missing data were observed in the survey data. However, the statistical results for the numerical variables allow us to identify that there is age data that was filled in with a value of zero. Therefore, this data must be cleaned so that it does not affect the statistical measures.

Considering that age is the only variable affected due to incorrect completion, we will impute data for the seven identified respondents. The median will be used, as it is a measure of central tendency that is more resistant to extreme values and better represents the actual central value of the data.

## Median Age Calculated:32
## Imputed Values:7
## Imputation completed: ZEROS → NA → MEDIAN(32)

Once the data has been cleaned, we proceed to analyze it, proposing hypotheses that allow us to confirm or refute the respondents’ behavior toward accepting or rejecting the proposed methodology.

2.1 Hypothesis Quantitative Variables

Age: It is expected that the younger people the age, the greater the satisfaction/usability of the proposed methodology, taking into account that may more influenced younger people to follow a research methodology within their work area, as opposed to older people, who are expected to be less receptive to adopting a new proposed innovation methodology due to their experience, presenting resistance to change.

2.2 Categorical Variable Hypotheses

Gender: Men are expected to be more accepting of the methodology, considering that they represent the majority within the institution. Therefore, women are expected to be less accepting than men because they are a minority within the military institution.

Type of affiliation: It is expected that undergraduate students will be more receptive to the methodology, considering that they are new to the research groups, while senior staff may not be as receptive to adopting the innovation methodology.

Research group: It is expected that research groups at the central level will be more accepting of the methodology, while decentralized groups will not.

3.1 Numerical variable graphs

The above graph allows us to analyze the age distribution, where we can see that 75% of the data is below 34 years of age, the first quartile Q1 represents 25% of the data with ages below 29 years, and we find some outliers that are above the maximum value (41) and below the minimum value (22).

It can be concluded that there is moderate symmetry in the data due to its distribution, which is slightly skewed to the left, with 50% of the information falling between the ages of 29 and 34 (adults), meaning that the research groups’ workforce falls within this range and represents half of the data collected. It is also worth noting that there is little participation among people under the age of 23 and those over 36, with no obvious extreme outliers.

3.2 Categorical variable graphs

The gender distribution allows us to identify the presence of 77.4% male staff versus 19.7% female staff. This has a lot to do with the current workforce, where male staff are in the majority. Finally, 2.9% preferred not to respond.

The graph above shows the distribution according to type of employment, where undergraduate students have the highest participation (33%), while emeritus researchers only account for 0.7% of the sample n.

Finally, the distribution by research group is evident, with ESAVE (31.2%) and ESCOM (30.8%) having the highest participation rates compared to BRIAV32 (4.37%).

4.1 Satisfaction category

## null device 
##           1

## === GRÁFICO PRINCIPAL: DISTRIBUCIÓN DE RESPUESTAS ===

The graphs above show that, in general, respondents rated most questions as satisfactory to very satisfactory, with values of 4 and 5 representing more than 80% of all responses. This indicates that, overall, there is a high level of satisfaction with the methodology proposed by the surveyed staff.

The boxplot graph shows the consistency of the information, as it can be verified that the values 4 and 5 are predominant and are where most of the data is concentrated.

4.2 Usability category

## === GRÁFICO PRINCIPAL: DISTRIBUCIÓN DE RESPUESTAS ===

In the case of usability, most respondents lean towards the numerical value 4, that is, they agree with the usability of the methodology, and between 20 and 25% are totally in agreement. However, a large portion (between 20-26%) expressed a neutral position regarding the use of the methodology, which means they are undecided about whether to implement it.

The boxplot graph allows us to identify how the data is grouped. Unlike the satisfaction category, which ranged between 4 and 5, the usability category ranged between 3 and 5, with some responses even scoring 2 (disagree), suggesting that there is a group of people who disagree with the use of the methodology.

4.3 Category correlations (Satisfaction/Usability)

The correlation matrix indicates that the questions were correctly structured and divided into two categories. Considering that there is only a correlation between questions within the same category, we can see in the upper right part of the graph that we obtain values close to zero, indicating that there is no correlation between these variables when different categories are crossed. The opposite is true for the satisfaction category, where we can see that there is a moderate correlation between most of its variables (0.40- 0.59), and that there is a high correlation (0.60-0.79) between the variables “Willingness to implement” and “Clarity and structure.”

In the case of the usability category, the variables are weakly correlated (0.20-0.39) and there is a moderate correlation between the variables “ease of understanding” and “quick learning,” as well as with the variable “frequent use,” suggesting that these variables are associated but not causal.

4.4 Dendogram

The graph above allows us to identify how the categories of the study variables are being grouped hierarchically. The dendrogram shows a cluster, analysis of the variables based on their correlation patterns.

In summary, we can conclude that there are two distinct dimensions or clusters, and that within each cluster there is internal cohesion between the variables. Therefore, strategies to improve each group must be presented and proposed separately. For example, for the usability group, frequent use correlates with confidence in guidelines and organized structure, which means that variable 1 could be influenced by variables 4 and 5.

4.5 Cronbach’s coefficient

To establish a metric that allows us to determine the reliability of the proposed questionnaire, we will use Cronbach’s coefficient, a statistical measure used to assess the consistency of a data set. If the questions are related to each other and measure the same construct, we expect to obtain a high coefficient, indicating high reliability.

Cronbach’s alpha provides a way to ensure that the questions in our questionnaire measure the same thing, providing more reliable and accurate results.

## Cronbach's Alpha coefficient for the Satisfaction group: 0.857

## Cronbach's Alpha coefficient for the Usability group: 0.755

The above result allows us to identify that there are high Cronbach’s coefficients that guarantee that internally the questions in our questionnaire measure the same thing, providing more reliable and accurate results for our research, demonstrating the correlation of each group, and concluding that the constructs for usability and satisfaction are highly consistent and dependable.

5.1 Bivariate categorical variables

The initial hypothesis was that men would be more represented than women. However, this hypothesis was not validated, as we found that female staff had a higher acceptance of usability (69.1%) compared to male staff, suggesting that more women tend to be more receptive to the use of technology. Regarding the satisfaction category, we found that men are more satisfied (73.6%) than women (60%), indicating that a higher percentage of men are more willing to adopt technology.

It was expected that, under this hypothesis, undergraduate students would have greater adoption of the technology. However, this theory can be refuted, considering that for the usability category, students have divided opinions (50%), which could indicate that they are unclear about the impact of the methodology and tend not to know whether to use it or not. On the other hand, concerning the satisfaction category, a high percentage indicates a greater willingness to implement the methodology in terms of clarity, structure, alignment, and opportunities for improvement in management and innovation.

However, analyzing positions such as emeritus researcher, we can see that they are 100% aligned with the implementation of the methodology (low sample size, only two of the respondents).

According to this hypothesis, it was expected that the central-level research groups (ESCAB-ESAVE-BAIDI-ESMIC) would have greater acceptance. However, it can be seen that ESCAB has a 25% usability of the methodology, while external groups such as ESCOM (59.3%) and BRIAV (69.2%) showed better reception of the innovation methodology.

On the other hand, regarding the satisfaction category, it can be seen that BAIDI was the group with the lowest score (57.1%), while the other agencies appear to be more satisfied with the proposed methodology, with percentages ranging from 60% to 81.6%.

5.2 Correspondence analysis

Correspondence analysis will allow us to analyze the variables in order to determine their association. It is very important to define the objective of our CMA.

Objective: To understand how perceptions of satisfaction and usability (10 questions) vary according to gender, type of affiliation, and research group. To answer the question, “What do people think?” and then, “Who are the people who think that way?” In this particular case, the responses to the 10 questions (five on satisfaction and five on usability) were considered active variables. In contrast, gender, type of relationship, and research group were considered complementary variables. It is essential to set the objective, taking into account that MCA analyzes the relationships between all categories. We should add up the scores of the questions to create a single index. In that case, we are losing key information about which specific aspect of satisfaction or usability is influencing the relationship between variables.