ECON 360 Lecture Notes - Lecture 7: Homoscedasticity, Multicollinearity, Autocorrelation
21 Feb 2018
School
Department
Course
Professor
Document Summary
The regression regress pgdp birth fem1549pop urbangrowth femurban femedu ubranedu lpgdp policy_dummy pgdpxpolicydummy. The f- statistics shows that all the variables are significant in estimating this model. However the p values for birth rate, fem1549, femurban, femedu, ubranedu and lgdp suggest that they are not statistically significant in estimating the per capita gdp. Of the four diagnostic plots, we asses one of the graphs(residual vs fitted plot: gauss markov assumptions. As for this paper, the predicted model is diagnosed using the gauss markov assumptions. These assumptions are helpful in checking whether biasness exists in the predicted model. The gauss markov assumptions dictate that the model parameters must be linear, with the variables having been acquired through random sampling, and no multicollinearity should exists between the variables. In addition, the conditional mean of errors must be equal to zero, and lastly, homoscedasticity should always occur. The plot show a slight deviation at around 0. 5(fitted values)
