Statistical Identification of Factors Responsible for Rice Yield: Regression Analysis Approach

Abstract

The importance of agriculture in our economy is underscored by the fact that it provides essential food supplies to meet the needs of our growing population. In Pakistan, agriculture is a vital contributor to our economic growth, Accounting for 22.7% of the GDP and employing approximately 37.4% of the Labor force. The rice ranked at 3rd place with 5.03 billion metric tons production worldwide. Pakistan came at 9th place all over the world in producing milled rice with a production of 9.32 million metric tons, rice contributes 2.4% to the value addition and 0.5% to the GDP. In this study we have collected data of Rice yield from Agriculture Department with various factors that contribute in the yield of rice. Then we use regression approach to identify the factors that are contributing in the rice yield significantly and the factors that contributing insignificantly. From the multiple linear regression, we got the R2 = 58.1% and Adjusted R Squared = .553. And then we see that, some predictor variables were found to be statistically significant while others were not as they are important for the rice yield. We suspected multicollinearity and applied factor analysis to address it. Factor analysis is applied to the independent variables, generating factors that were then used in a multiple regression analysis. The results of this analysis showed an improvement over the previous model, by making factors from all the independent variables show that all the factors are now contributing significantly in the model and the R2 has also improved from 58.1% to 68.7%.