Using Dummy Variables in Improving the Simple Linear Regression Model for the Ratio of Consumption to the National Income in Iraq during the Period (1970-1994)

Abstract

This paper discuss the concept of dummy variables and its importance usge in statistical analysis by transforming the qualitative variables to measurable quantitative variables and applying it in analyzing the linear regression in both simple and multiple forms. A comparison has been made between using dummy variables and power transformation methodology. This comparison aims to show which one of the two methods is better in improving the linear regression model by applying them on the data of ratio of consumption to the national income in Iraq for the period of (1970-1994). Depending on the data available of that period the results showed that the dummy variables are more efficient than power transformation in improving the regression model of the consumption to national income. The dummy variables helped explaining almost 80% from the consumption ratio in the given period in Iraq by making the data to be more intelligible and more homogeneous in the model.