Start with the data set you tabulated for the previous assignment – for US states cross-sectional data on poverty rate, state unemployment rate, and transfer payment. Raise the number of states at least to 35 for different issues. Add two variables state of inflation rate, state CPI or cost of living index is fine, and the percentage of college graduation completion rate in the state.
|Name of the State||Poverty rate, %||Transfer Payment Spending per Capita in dollars||Unemployment rate||Inflation rate – Cost of living Index||% of college graduation completion rate, Bachelor’s degree|
Answer: From the collected data, it is expected all independent variables to be proportional to the dependent variable, poverty rate, regardless of how big values one possesses. After plotting, the expectations which was also the hypothesis was found to be incorrect putting suggestions that the data is non-linear and had some multicollinearity.
Answer: The two regression have no strong correlation. The reason behind is because they have not proportional to the dependent variable.
After plotting the regression, the hypothesis was found to be true as seen in the values of R squared in the above graph.
Answer: From the regression curve for the four variable, there is multicollinearity of two variables which did not appear in the curve that is the inflation rate – cost of living index together with the unemployment rate. Besides, the coefficients of the two linear regression curves, the appearing one are have no strong correlation because their regression values, R, are less than 0.9. They are 0.5661 and 0.7493 for Transfer Payment Spending per Capita in dollars and % of college graduation completion rate, Bachelor’s degree respectively.
Answer: From the data and regression curves, there is no reason to believe that the original equation is understated. This is because there are two variables which did not appear in the graph as their values were too small and inconsistent which made them not to appear. Besides, this is a proof the equation with the four independent variables had some irrelevant variable.
Answer: From the scatter diagram, graph, above, the cost of living, unemployment rate and % of college graduation completion rate, Bachelor’s degree are not linear in connection with the poverty rate. Their lines cannot be traced graphically as they are have no single relationship. In this connection on Transfer Payment Spending per Capita in dollars can be linearly interpreted with poverty rate.
Goodwin, N., Harris, J. M., Nelson, J. A., Roach, B., & Torras, M. (2015). Macroeconomics in context. Routledge.