Estimated Results of the Probit Model
Results of the probit model are presented in
Table 4. The diagnostics tests such as Pseudo R
2
,
Likelihood Ratio, Chi-Square and estimated value
for the Log-likelihood functions are also reported in
the table. All the explanatory variables accounted
for about 58% of the variations in the probability
that a farmer would decide to adopt the alternative
pest management method. The overall fit,
expressed by the likelihood test, is high and
significant. This demonstrates that the variables
included in the model are relevant influences of the
adoption decisions of the sampled cocoa farmers
regarding the alternative pest management method
in the study area. The respondents’ years of
education, household size, participation in
community groups, farm size were all significantly
related to the probability of adoption by the
respondents (Table 4). In binary models, the
coefficients (βs) cannot be interpreted as the
marginal effects on the dependent variable (Greene,
1994; Gujarati, 2007). Therefore, marginal effects
were computed, as the percentage change in the
probability of adoption associated with a unit
change in an explanatory variable