205
Ugandan context. However, the measures of risk-taking were not sufficiently well
fitted to this context. After minor adaptations to the initial measurement model,
1
the
index values were the following
2
: GFI = 0.941, AGFI = 0.9, RMSEA = 0.055,
CFI = 0.933; chi square p-value = 0.05. This shows that the measurement model of
entrepreneurial orientation fitted well with the Ugandan context.
After performing the CFA, linear regressions were used to analyze the impact of
entrepreneurial orientation on product, process and market innovations, in interaction with farm characteristics. Multiple regression models were run using different
interaction terms (e.g., the combined effect of innovativeness and education level,
or the combined effect of proactiveness and age) in order to: (i) understand whether,
when considering different control and interaction variables, the effects on agricultural innovations were stable; and (ii) assess whether the effect of entrepreneurial
orientation and farm characteristics vary in their impact on innovation in general as
well as on specific types of innovation.
17.3 Findings
Figure 17.1 shows the key tested relationships among the variables of interest:
entrepreneurial orientation, farm characteristics and farmer innovations. In the first
tested regression model (when farm characteristics and entrepreneurial orientation
are considered together with interaction terms), it was found that only education
1 Different combinations have been created between the different first-order latent constructs.
While running the analysis, problems emerge if risk-taking is included amongst the latent constructs. If risk-taking is included the values are represented as follows: GFI = 0.855, AGFI = 0.803,
RMSEA = 0.084, CFI = 0.661, and the chi square is significant (p value = 0.000). Problems also
arises whether a CFA is performed for the first-order latent construct risk-taking, when taken
alone, thus without any combinations with innovativeness, proactiveness and intentions. At the
same time, CFA was conducted for each of the first-order latent constructs, which did not register
any issues: innovativeness, proactiveness and intentions. The correlation values of each variable
with the latent construct were high and the model fit was good as well. Furthermore, one questionnaire item for the measurement of innovativeness and one questionnaire item (out of four total in
each) for proactiveness were excluded. It has been proven that even with three items for dimension,
the questionnaire can still maintain statistical authenticity (Cook et al. 1981). If risk-taking should
not be included in the questionnaire to measure entrepreneurial competences—even with a oneitem reduction each for innovativeness and proactiveness—the questionnaire is still statistically
authentic.
2 A value of the RMSEA of about 0.05 or less would indicate a close fit of the model in relation to
the degrees of freedom. The requirement of exact fit corresponds to RMSEA = 0.0. A value of
about 0.08 or less for the RMSEA would indicate a reasonable error of approximation, and one
would not want to employ a model with a RMSEA greater than 0.1. GFI is less than or equal to 1.
A value of 1 indicates a perfect fit. It is acceptable when GFI >0.9 The AGFI (adjusted goodness
of fit index) considers the degrees of freedom available for testing the model. The AGFI is bounded
above by one, which indicates a perfect fit. It is not, however, bounded below by zero, as the GFI
is. It is acceptable when AGFI >0.9. CFI falls in the range from 0 to 1. CFI values close to 1 indicate a very good fit.
17 The Role of Farmers’ Entrepreneurial Orientation on Agricultural Innovations…
Ugandan context. However, the measures of risk-taking were not sufficiently well
fitted to this context. After minor adaptations to the initial measurement model,
1
the
index values were the following
2
: GFI = 0.941, AGFI = 0.9, RMSEA = 0.055,
CFI = 0.933; chi square p-value = 0.05. This shows that the measurement model of
entrepreneurial orientation fitted well with the Ugandan context.
After performing the CFA, linear regressions were used to analyze the impact of
entrepreneurial orientation on product, process and market innovations, in interaction with farm characteristics. Multiple regression models were run using different
interaction terms (e.g., the combined effect of innovativeness and education level,
or the combined effect of proactiveness and age) in order to: (i) understand whether,
when considering different control and interaction variables, the effects on agricultural innovations were stable; and (ii) assess whether the effect of entrepreneurial
orientation and farm characteristics vary in their impact on innovation in general as
well as on specific types of innovation.
17.3 Findings
Figure 17.1 shows the key tested relationships among the variables of interest:
entrepreneurial orientation, farm characteristics and farmer innovations. In the first
tested regression model (when farm characteristics and entrepreneurial orientation
are considered together with interaction terms), it was found that only education
1 Different combinations have been created between the different first-order latent constructs.
While running the analysis, problems emerge if risk-taking is included amongst the latent constructs. If risk-taking is included the values are represented as follows: GFI = 0.855, AGFI = 0.803,
RMSEA = 0.084, CFI = 0.661, and the chi square is significant (p value = 0.000). Problems also
arises whether a CFA is performed for the first-order latent construct risk-taking, when taken
alone, thus without any combinations with innovativeness, proactiveness and intentions. At the
same time, CFA was conducted for each of the first-order latent constructs, which did not register
any issues: innovativeness, proactiveness and intentions. The correlation values of each variable
with the latent construct were high and the model fit was good as well. Furthermore, one questionnaire item for the measurement of innovativeness and one questionnaire item (out of four total in
each) for proactiveness were excluded. It has been proven that even with three items for dimension,
the questionnaire can still maintain statistical authenticity (Cook et al. 1981). If risk-taking should
not be included in the questionnaire to measure entrepreneurial competences—even with a oneitem reduction each for innovativeness and proactiveness—the questionnaire is still statistically
authentic.
2 A value of the RMSEA of about 0.05 or less would indicate a close fit of the model in relation to
the degrees of freedom. The requirement of exact fit corresponds to RMSEA = 0.0. A value of
about 0.08 or less for the RMSEA would indicate a reasonable error of approximation, and one
would not want to employ a model with a RMSEA greater than 0.1. GFI is less than or equal to 1.
A value of 1 indicates a perfect fit. It is acceptable when GFI >0.9 The AGFI (adjusted goodness
of fit index) considers the degrees of freedom available for testing the model. The AGFI is bounded
above by one, which indicates a perfect fit. It is not, however, bounded below by zero, as the GFI
is. It is acceptable when AGFI >0.9. CFI falls in the range from 0 to 1. CFI values close to 1 indicate a very good fit.
17 The Role of Farmers’ Entrepreneurial Orientation on Agricultural Innovations…