3.2. Statistical results from the regression analysis using SPSS
Table 2 summarizes all the statistical characteristics of the three cereals. Both population and precipitation positively correlate (≈ 0.70) with yields of all the three cereals (greater than 0.3); therefore, there is no multicollinearity.
Population is positively correlated with yields of sorghum (≈0.94) and maize (≈0.90) and strongly correlates with that of rice (0.85). R2 indicates that population accounts for about 88% of the variability of sorghum’s production, 80% of the variability of maize’s production, and 72% of the variability of rice’s production. Although positive, the correlation of precipitation with the cereals is lower (sorghum: ≈ 0.73; maize: ≈ 0.72; and rice: ≈ 0.78). Precipitation explains about 52% of the variability of sorghum’s production, 50% of the variability of maize’s production, and 60% of the variability of rice’s production. The two independent variables are both positively correlated with the crops, but population influences the variability to a greater extent than precipitation.
R2 values show that the two predictive variables (population and precipitation) explained the production of sorghum at about 90%, maize at 80%, and rice at 78%. Only about 10% of sorghum’s yield, 20% of maize’s yield, and 23% of rice’s yield are explained by other factors. Sig in ANOVA yielded 0.000 (<0.5) as the value for all three dependent variables. Thus, the model is statistically significant, and it accurately predicts the three crops. The standardized coefficients demonstrates that population explains about 83%, 75.80%, and 60% of the change in the production of sorghum, maize, and rice, respectively. The contribution of precipitation to explaining changes in the crops’ yields is about 16%, 19.60%, and 36.20%, respectively.
The population with Coefficient Sig value of 0.000 for all the cereals makes a unique and statistically significant contribution to the prediction of sorghum, maize, and rice production. Likewise, precipitation with a Sig value of 0.077 for sorghum, 0.084 for maize, and 0.005 for rice makes a statistically unique contribution to the prediction.
Table 2. Least squares characteristics associated with the regression analysis for crops using SPSS
Population accounted for about 88% of the variability in the production of sorghum, 80% of the variability of maize, and 72% of the variability of yields in rice. Precipitation explained approximately 52% of the variability in production of sorghum, 50% of maize, and 60% of rice.