Environmental and Sustainability Indicators 12 (2021) 100140
5
respectively.
Fig. 2
provides a report of computed farmers’
climate
change awareness index.
3.2. Factors influencing farmers awareness of climate change
The results revealed variations in climate change awareness among
farmers. Results show that farmers do not equally perceive the changes
in climatic conditions. We hypothesize that the differences in perception
or awareness level will influence the adaptation responses of farmers. To
test this hypothesis, it is important to first find out the factors that drive
the differences in awareness level among the farmers. The results of this
analysis
has several policy relevance;
including identifying which
farmers to target with climate information. The findings is also relevant
for identifying gaps in existing policy efforts in improving the climate
change literacy rate of farmers; with the purpose of equipping farmers
with the relevant information. We compared the robustness of the OLS,
multinomial and ordinal logit regressions to determine the model of best
fit.
The information statistics suggested
that the multinomial logit
regression best fit the data; as it contained the lowest AIC and BICs [AIC
=
107.211; BIC = 173.149]; compared to the OLS [AIC = 237.046; BIC
=
270.016] and the ordinal logit model [AIC = 150.357; BIC =
190.353]. The log-likelihood ratio test of
joint significance also sug-
gested that the multinomial logistic model is valid [LR
χ
2
(20) = 161.33;
p-value = 0.001] and possesses a strong explanatory power (Nagelkerke
R
2
=
0.782). We therefore interrogated the
factors that account for
differences in climate change awareness using the multinomial logistic
regression. Ultimately, we assessed the factors that explain the proba-
bility of a farmer becoming a member of a climate change awareness
subgroup; and by extension, the predictors of climate change awareness
among farmers. We normalized subgroup 2
in the multinomial logit
estimation to identify the factors that predict low and high awareness
with community covariates fixed. Thus, subgroup 2 is the base category
and the multinomial logistic regression results are interpreted as the
average change in the probability of
gaining a high or low level
awareness relative to a moderate climate change awareness level due to
Do'stlaringiz bilan baham: