Page 161 - Hojnik, Jana. 2017. In Persuit of Eco-innovation. Drivers and Consequences of Eco-innovation at Firm Level. Koper: University of Primorska Press
P. 161
Results 161
comprises nine items. All the coefficients were found to be positive, high
and significant; these are indicated in Table 29 and Figure 7.
Statistical information of the construct expected benefits, pertain-
ing to reliability (reliability coefficients) and convergence (goodness-of-
fit model indexes) based on the overall sample (N= 223), is as follows. The
construct expected benefits showed good reliability (Cronbach’s alpha=
0.911). The goodness-of-fit indexes are as follows: NFI = 0.889; NNFI =
0.879; CFI = 0.909; SRMR = 0.058; RMSEA = 0.13. CFI showed good fit
(over the threshold of 0.90), while other goodness-of-fit indexes showed
slightly worse fit.
From Table 29, we can see that three items have lower standardized
coefficients (approximately 0.60); these are: “To reduce costs (energy, ma-
terial, etc.)”, “To improve firm reputation” and “Adjustment to EU”. In
addition, the goodness-of-fit model indexes are also low. Therefore, we
decided to conduct another exploratory factor analysis, in which we elim-
inated these three items due to their low correlations with other items.
For instance, the item “To reduce costs (energy, material, etc.)” had cor-
relations with other items ranging between 0.310 and 0.571, followed by
the item “To improve firm reputation”, which had correlations with other
items ranging between 0.363 and 0.666 and “Adjustment to EU,” which
had correlations with other items ranging between 0.310 and 0.488.
Moreover, communalities of those items are as follows: “To reduce costs
(energy, material, etc.)” = 0.335, “To improve firm reputation” = 0.383
and “Adjustment to EU” = 0,283. After eliminating those three items,
we conducted exploratory factor analysis once more, and the value of the
Kaiser-Meyer-Olkin measure for sampling adequacy was 0.896. Bartlett’s
test of sphericity also showed a statistically significant value (chi-square =
865.338; df = 15; p = 0.000), meaning that the correlation matrix has sig-
nificant correlations. The communality index shown good communal-
ities for almost all items (the lowest communality after extraction was
0.546), while variance explained was estimated at 64.096%. We can see
that with fewer items (six instead of nine items), we are able to explain
more variance; therefore, we conducted the confirmatory factor analysis
again to check whether the goodness-of-fit indexes are any better.
A confirmatory factor analysis was conducted in order to validate the
findings of the exploratory factor analysis, which resulted in one factor
composed of six items. This has also been confirmed by the confirmatory
factor analysis. The eco-innovation determinant expected benefits com-
prises six items. All the coefficients were found to be positive, high and
significant. These are indicated in Table 30 and Figure 7.
comprises nine items. All the coefficients were found to be positive, high
and significant; these are indicated in Table 29 and Figure 7.
Statistical information of the construct expected benefits, pertain-
ing to reliability (reliability coefficients) and convergence (goodness-of-
fit model indexes) based on the overall sample (N= 223), is as follows. The
construct expected benefits showed good reliability (Cronbach’s alpha=
0.911). The goodness-of-fit indexes are as follows: NFI = 0.889; NNFI =
0.879; CFI = 0.909; SRMR = 0.058; RMSEA = 0.13. CFI showed good fit
(over the threshold of 0.90), while other goodness-of-fit indexes showed
slightly worse fit.
From Table 29, we can see that three items have lower standardized
coefficients (approximately 0.60); these are: “To reduce costs (energy, ma-
terial, etc.)”, “To improve firm reputation” and “Adjustment to EU”. In
addition, the goodness-of-fit model indexes are also low. Therefore, we
decided to conduct another exploratory factor analysis, in which we elim-
inated these three items due to their low correlations with other items.
For instance, the item “To reduce costs (energy, material, etc.)” had cor-
relations with other items ranging between 0.310 and 0.571, followed by
the item “To improve firm reputation”, which had correlations with other
items ranging between 0.363 and 0.666 and “Adjustment to EU,” which
had correlations with other items ranging between 0.310 and 0.488.
Moreover, communalities of those items are as follows: “To reduce costs
(energy, material, etc.)” = 0.335, “To improve firm reputation” = 0.383
and “Adjustment to EU” = 0,283. After eliminating those three items,
we conducted exploratory factor analysis once more, and the value of the
Kaiser-Meyer-Olkin measure for sampling adequacy was 0.896. Bartlett’s
test of sphericity also showed a statistically significant value (chi-square =
865.338; df = 15; p = 0.000), meaning that the correlation matrix has sig-
nificant correlations. The communality index shown good communal-
ities for almost all items (the lowest communality after extraction was
0.546), while variance explained was estimated at 64.096%. We can see
that with fewer items (six instead of nine items), we are able to explain
more variance; therefore, we conducted the confirmatory factor analysis
again to check whether the goodness-of-fit indexes are any better.
A confirmatory factor analysis was conducted in order to validate the
findings of the exploratory factor analysis, which resulted in one factor
composed of six items. This has also been confirmed by the confirmatory
factor analysis. The eco-innovation determinant expected benefits com-
prises six items. All the coefficients were found to be positive, high and
significant. These are indicated in Table 30 and Figure 7.
