2.4. Statistical Analyses

After linguistic validation, the statistical validation of the questionnaire was achieved, following the procedure described in Figure 1.

Validation algorithm.

Basic descriptive statistical tools were used for the exploratory analysis of the data. Additionally, item analysis was performed on two levels: item–item correlations and item–total correlations. Item analysis can be applied to samples of over 100 participants [22]. To perform item analysis, Pearson correlation coefficients were calculated, which measure the association between two variables according to the magnitude of the absolute value [28,29,30]. If 0.00 < r < 0.10 the association is very weak, if 0.10 ≤ r < 0.30 the association is weak, if 0.30 ≤ r < 0.50 the association is moderate, if 0.50 ≤ r < 0.70 the association is strong and if 0.70 ≤ r < 1.00 the association is very strong. For r = 0 there is no association, and for r = 1 the association is perfect.

The reliability of the scales for each of the seven independent dimensions considered was evaluated through the calculation of the Cronbach’s alpha (α), which measures the internal consistency of the different statements evaluated within a certain group [31]. The values of Cronbach’s alpha (α) range from 0.0 to 1.0. Higher scores indicate a more reliable, homogenous scale in which the individual items in each domain of the questionnaire reliably measure the domain core concept [32]. According to Hill and Hill [22], the alpha can be interpreted as follows: α < 0.6—unacceptable internal reliability; 0.60 ≤ α < 0.70—weak internal reliability, 0.70 ≤ α < 0.80—acceptable internal reliability, 0.80 ≤ α < 0.90—good internal reliability and α ≥ 0.90—excellent internal reliability.

In a complementary stage of the data analysis, factor analysis (FA) was also used, considering all 64 items. Firstly, the suitability of the data for this kind of analysis was tested through evaluation of the correlation matrix and the values of MSA (measure of sampling adequacy) in the anti-image matrix, the Kaiser–Meyer–Olkin measure of adequacy of the sample (KMO) and Bartlett’s test [24,33]. The solution was obtained through extraction with the principal component analysis (PCA) method with varimax rotation. The Kaiser criterion was used to stipulate the number of components to retain, which means considering eigenvalues ≥1. Communalities indicated the percentage of variance explained (VE) by the factors extracted [31], and they were to be equal to 0.5 or higher [30,34]. To determine the internal consistency in each factor, we again used Cronbach’s alpha (α) [31,35].

The analysis of the data used SPSS software from IBM Inc. (version 26, Armonk, NY, USA).