Page 141 - Hojnik, Jana. 2017. In Persuit of Eco-innovation. Drivers and Consequences of Eco-innovation at Firm Level. Koper: University of Primorska Press
P. 141
Methodology 141
- Factors loadings are simple correlations between the variables
and the factors.
- Factor matrix contains the factor loadings of all the variables on
all the factors extracted.
- Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is
an index use to examine the appropriateness of factor analysis.
High values (between 0.50 and 1) indicate that factor analysis
is appropriate. Values below 0.50 imply that factor analysis may
not be appropriate.
- Percentage of variance is the percentage of the total variance
attributed to each factor.
Exploratory factor analysis was performed using the Maximum Like-
lihood extraction method and Direct Oblimin rotation. Most research
using EFA has extracted factors that are orthogonal – that is, uncorrelat-
ed with or independent of one another (Maruyama 1998). In our study,
we used the oblique rotation (e.g., Direct Oblimin), which predicts that
factors are correlated with or dependent on one another and is in line
with what the structural equation approaches hypothesize; factors in
structural equation models usually will be hypothesized to correlate with
one another (Maruyama 1998). Moreover, Bartlett’s test of sphericity and
the Kaiser-Meyer-Olkin (KMO) test were used to determine whether
data were appropriate for factor analysis. KMO values of 0.80 or above
are excellent, 0.70 or above are middling, 0.60 or above are mediocre, 0.50
or above are poor, and below 0.50 are unacceptable (Hair et al. 1998). In
our study, all the KMO values were above 0.50, and the sig. value of Bart-
lett’s test of sphericity was less than 0.05, which means that our data jus-
tify the use of exploratory factor analysis.
After the exploratory factor analysis, we also conducted confirmato-
ry factor analysis. Evaluating the results of SEM involves theoretical cri-
teria, statistical criteria, and an assessment of fit. Although the issue of fit
is discussed in literature in greater detail than the other issues, it should
be remembered that fit is of no interest unless the results meet theoreti-
cal and statistical criteria. A model submitted to an SEM program should
be based as much as possible on “theory” in the sense of a systematic set
of relationships providing a consistent and comprehensive explanation of
a phenomenon. After the parameters of the models are estimated, they
should be assessed from a theoretical perspective (e.g., the signs and mag-
nitudes of the coefficients should be consistent with “theory”). Besides
the theoretical criteria mentioned above, there are also two major statis-
tical criteria. The first pertains to the identification status of the model
- Factors loadings are simple correlations between the variables
and the factors.
- Factor matrix contains the factor loadings of all the variables on
all the factors extracted.
- Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is
an index use to examine the appropriateness of factor analysis.
High values (between 0.50 and 1) indicate that factor analysis
is appropriate. Values below 0.50 imply that factor analysis may
not be appropriate.
- Percentage of variance is the percentage of the total variance
attributed to each factor.
Exploratory factor analysis was performed using the Maximum Like-
lihood extraction method and Direct Oblimin rotation. Most research
using EFA has extracted factors that are orthogonal – that is, uncorrelat-
ed with or independent of one another (Maruyama 1998). In our study,
we used the oblique rotation (e.g., Direct Oblimin), which predicts that
factors are correlated with or dependent on one another and is in line
with what the structural equation approaches hypothesize; factors in
structural equation models usually will be hypothesized to correlate with
one another (Maruyama 1998). Moreover, Bartlett’s test of sphericity and
the Kaiser-Meyer-Olkin (KMO) test were used to determine whether
data were appropriate for factor analysis. KMO values of 0.80 or above
are excellent, 0.70 or above are middling, 0.60 or above are mediocre, 0.50
or above are poor, and below 0.50 are unacceptable (Hair et al. 1998). In
our study, all the KMO values were above 0.50, and the sig. value of Bart-
lett’s test of sphericity was less than 0.05, which means that our data jus-
tify the use of exploratory factor analysis.
After the exploratory factor analysis, we also conducted confirmato-
ry factor analysis. Evaluating the results of SEM involves theoretical cri-
teria, statistical criteria, and an assessment of fit. Although the issue of fit
is discussed in literature in greater detail than the other issues, it should
be remembered that fit is of no interest unless the results meet theoreti-
cal and statistical criteria. A model submitted to an SEM program should
be based as much as possible on “theory” in the sense of a systematic set
of relationships providing a consistent and comprehensive explanation of
a phenomenon. After the parameters of the models are estimated, they
should be assessed from a theoretical perspective (e.g., the signs and mag-
nitudes of the coefficients should be consistent with “theory”). Besides
the theoretical criteria mentioned above, there are also two major statis-
tical criteria. The first pertains to the identification status of the model
