Hardware Reference
In-Depth Information
group of factors (Q 53.6, Q 5.6, Q 4.1, Q 5.2, Q 2.4, Q 6.1, Q 2.5, Q 3.5, Q 5.1, Q 6.2,
Q 2.3) have a greater proportion of support although an important percentage of peo-
ple are not convinced of their presence in professional environments. And finally, a
group of factors (Q 3.1, Q 3.2, Q 5.4, Q 1.1, Q 3.3, Q 3.4, Q 4.2, Q 5.5, Q 5.3) have a
confirmation percentage above 80% so they can be considered as real facts in the
software development world. As an additional action, for some of the respondents
(33), we collected information about testing training. Combining with the first survey
(Section 2), a global 36% of professionals have attended specific testing training.
4.2 Detailed Analysis of Results
A first objective is the validation of the questionnaire used to collect data analyzing
correlations between pairs of questions. First of all, reliability of the scale should be
tested through the Cronbach´s alpha coefficient. This measure helps to verify if the
questions are related between them, i.e., all the questions measure the same concept.
Then factorial analysis will be applied to verify which concept measures each group
of questions. This allows determining the structure of the scale [24].
4.2.1 Previous Analysis
Exploratory analysis enables the detection of possible errors during data collection as
well as the checking of feasibility of factorial analysis. Subsequently we examine
descriptive statistics (means, standard deviation, median, minimum and maximum
and absolute and relative frequencies) of all the variables in the study. Moreover, box
plots can help to determine data entry errors and the coefficient of variation can be
used to check the homogeneity of data. The correlation matrix gives information
about factorial analysis applicability: correlations higher than 0.30, significance levels
and determinants close to 0 shows there are correlated variables [25].
SPSS 16.0.1 and LISREL 8.80 statistical programs have been used to analyze the
collected data. A first visual inspection of correlation matrix showed us that there was
an essential number of correlations higher than 0.30; consequently, we concluded that
there were interrelated variables [25]. Moreover, as almost all significance levels are
close to zero, we had to reject the null hypothesis and concluded there was linear
relationship between the variables. The determinant is near zero too (9,06E-005): it
confirms these variables are highly correlated so factorial analysis is applicable.
4.2.2 Reliability Analysis
The first validation is the reliability analysis of the used factors. The reliability is the
degree in which the observed variable measures the real value and is free of error. The
reliability analysis includes the examination of corrected item-to-total correlations to
find out if each factor measures the same issue than the rest of the factors. We elimi-
nate specific items to improve reliability alpha coefficient. In this case, questions 1.1
(Q1.1) and 3.5 (Q3.5) have been eliminated. The final Cronbach´s alpha coefficient
value after these eliminations is 0.908 demonstrating high consistency and reliability
of the model
2
.
2
The conventional minimum for Cronbach's alpha coefficient is established in 0.7 [26].
