17.08.2016 Views

RESPONSIBLE ENTREPRENEURSHIP VISION DEVELOPMENT AND ETHICS

2aO8o2F

2aO8o2F

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

176 <strong>RESPONSIBLE</strong> <strong>ENTREPRENEURSHIP</strong><br />

this study, Statistical Package for Social Sciences (SPSS) will be used, together with AMOS<br />

(Analysis of Moment Structures). In order to test the relationship between variables and in<br />

order to find the statistical significance of the hypothesized proposed model, structural equation<br />

modeling, more specifically confirmatory factor analysis will be used.<br />

Results and data analysis<br />

Exploratory factor analysis<br />

The Exploratory Factor Analysis (EFA) has the scope of discovering the structure of a<br />

relatively large set of variables and the initial assumption is that any indicator is associated<br />

with any factor. The factor analysis initially identifies seven factors. The value of KMO is<br />

0.939, very close to 1, so this means that the correlation models are relatively compact and<br />

the factor analysis will establish trustworthy and distinct factors. If this would not have been<br />

met, we could have said that distinct and reliable factors cannot be produced. Communalities<br />

table reveals values smaller than 0.30.<br />

Total Variance Explained table determines the number of significant factors. Only extracted<br />

rotated values are meaningful for interpretation. The factors are arranged in the descending<br />

order based on the most explained variance. We choose the rotation method Promax with Kaiser<br />

Normalization.<br />

The extraction refers to the process of determining how many factors best explain the<br />

observed covariation matrix within the data set. 71.518% of the total variance is explained by<br />

the seven factors, only the first factor is explaining 50.698% of the total variance, followed<br />

by 5.485%, 4.889%, 2.194%, 2.423%, 2.391% and 2.437% (all the remaining six factors).<br />

Pattern Matrix table indicates the factors which can be used in the study. After eliminating<br />

the questions for which the average value is less than 0.70, we obtain four factors and<br />

we start over the analysis. We apply Kaiser-Meyer-Olkin and Barlett tests. The KMO test<br />

value is 0.945, very close to 1, so the correlation models are relatively compact and the factor<br />

analysis will establish trustful and distinct factors. Sig. value is less than 0.01, so we can<br />

say that the factor analysis is adequate in this case. Analyzing the Communalities table, we<br />

observe that all the values in the second column are greater than 0.30, which is a good sign.<br />

Afterwards we analyze Total Variance Explained table. Total variance explained indicates how<br />

much of the variability in the data has been modeled by the extracted factors.<br />

The first four factors are meaningful since they have Eigenvalues greater than 1. The Total<br />

Variance Explained table shows the actual factors that were extracted. The section labeled<br />

“Rotation Sum of Squared Lodings” show us only those factors that met our cut-off criterion<br />

(the extraction method). In our case there were four factors with eigenvalues greater than<br />

1. The “% of Variance column tells us how much of the total variability (in all of the variables<br />

together) can be accounted for by each of the summary scale of factors. We observe<br />

that 69.053% of the total variance is explained by the four factors, only the first factor explaining<br />

58.372%, followed by 4.096%, 2.992% and 3.593% (the other three factors). The value<br />

69.053% is less than in the case of the seven factors, but in this case we have a better and<br />

cleaner Pattern Matrix, so we can continue with the study. In the second Pattern Matrix, we<br />

can see that variables group into factors; more precisely they load onto factors. We can observe<br />

a very clean factor structure in which convergent and discriminant validity are evident by the<br />

high loadings within factors, and no cross-loadings between factors.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!