Page 676 - 2015-37

Logistic regression Number of obs = 18
LR chi2 (6) = 60.73
Prob > chi2 = 0.0000
Log likelihood = -77.795559
Pseudo R2 = 0.2807

Results and Discussion

There are twelve (12) variables (11 dependent and one indepen-
dent) in the proposed model. Marital Status, Social group, improved
livestock breed, Sex and Agro Processing were dropped by the Stata
may be due to perfect predictability or less variation with the im-
proved crop production technology. They are one hundred and forty
(140) adopters (treated) of “Fadama” III, while forty-nine (49) were
non adopters. Because of this therefore, these variable were dropped
from the analysis.

The likelihood ratio chi- square (the goodness of it) measured by
the value of 60.73 with a p-value of 0.0000 showed that the choice
of explanatory variables included in the Logit model explained the
variation in decisions in improved crop production technology.

In the estimated model, “fadama” participant, Household size,
Status of Respondent and Primary occupation of the respondents
are found to have statistically signiicant and positive impact on im-
proved crop production technology. On the other hand, age and edu-
cation of the respondents are found to be statistically insigniicant on
the improved crop production technology.

The estimated odds ratio shows that the probability of adoption
of improved crop production technology increases with increase in
“Fadama” participation. The logit (odds ratio) goes on increasing by
3.218% units with a positive marginal effect of 0.16% holding other
factors ixed. This could be explained by the fact that as times goes
on there is all possible indication that one would be able to realize

- 11 -