Rice Genomics and Genetics 2015, Vol.7, No.1, 1-10
5
Table 3 Probit estimates of the determinants of awareness of improved rice varieties
Variables
Coefficient
T-stat
Marginal effects
Age
-0.013
-0.81
-0.006
Household size
-0.031
-0.55
-0.001
Membership in association
0.394
0.86
0.018
Access to media
1.133
2.30**
0.108**
Ownership of mobile phone
0.270
0.51
0.012
Years of formal education
0.038
0.77
0.002
Main occupation
0.975
1.34
0.096
Constant
0.690
0.55
Number of respondents
149
Prob>chi2
0.072
Pseudo R
2
0.230
Note: *, ** and *** Significant at 1%, 5% and 10% respectively
and the dependent variable was whether a farmer
used (Adopted) improved rice variety or not. The
second stage examined the intensity of adoption and
the dependent variable was the proportion of area
cultivated to the improved rice variety. To test for
sample selection bias, the relationship between the
residuals for the two stages, the selection and
outcome equation was examined. The Wald test of
independent equations rejects the null hypothesis of
no correlation (rho=0) between the two disturbance
terms (i.e. in the regression equation and selection
equation) at 1% level of significance. The negative
sign of rho shows that the unobservable factors that
reduce the probability of adoption of improved rice
variety increase the intensity of adoption and vice
versa. The result as presented in Table 4 also shows
that the sigma is statistically significant, implying
that the choice of explanatory variables included in
the heckman model explained the level of adoption
of improved rice varieties. The inverse mills ratio
(lambda) was significant implying that covariates
that condition the proportion of area grown to
improved rice variety is conditional on the probability
of adoption implying that there is no sample
selection bias problem.
The results of the analysis showed that access to
credit, access to media and agricultural income
significantly influence the probability of adoption.
Access to credit positively and significantly (p≤0.05)
influenced adoption. This implies that farmers that
had access to credit faces lessen cash constraint; this
enables them to purchase inputs such as improved
seeds. In line with a priori expectation, access to
media had a positive and significant (p≤0.05)
influence on the probability of adoption. This
implies that a high level of awareness was created
through media. The study also revealed that
agricultural income had a positive and significant
(p≤0.01) influence on the probability of adoption
meaning that farmers’ with higher agricultural
income have the likelihood of re-investing back into
the farm production activities by purchasing more
modern productive inputs like improved seeds.
Table A5 also revealed that gender, household size,
farm size and agricultural income significantly
influence the intensity of adoption of improved rice
variety. Gender of the household head was positive
and significant (p≤0.1) which implies proportion of
area cultivated to improved rice variety by the male
headed households increases by 52.6% as against
female headed households. The implication of this
could be due to the fact that female headed households
have poor access and control over resources in
general and have shortage of farm labour. Farm size
had a positive and significant (p≤0.01) influence on
intensity of adoption implying that the proportion of
area cultivated to improved rice variety increases as
farm size increases. If the farm size increases by one
hectare the area allocated to improved rice variety
will increase by 94.9%. This may be due to the fact
that farmers operating larger farms tend to have
greater financial resources and a high probability of
receiving credit to further enhance production than
those of small farm size.
