在英留学时写的计量经济学论文(希望有帮助).docx

There are the main independent variables we research in this model. By testing this model, it can be concluded that which factors are significant for improving individual earnings, and know the relationships between them is positive correlation or negative correlation, or no linear correlation. Considering the validity of the model, the first step is to choose the more affect variables from the 22 variables. Referring to the research by others, the quality of male employee has far more influence on their income than other external factors from their wife or children ( ). So we select the following several variables: huseduc (years of schooling), husblck, husage, hushrs, hushisp, husexp, the independent variable is husearns.Some variables like the situation of economic development in 1991, external variables from wife and children, we are not able to quantify the size of influence, so we put them in the random fluctuations and fix them in heteroskedasticity test.Preliminary model: assumption: there is linear relationship between the dependent variable husearns and the independent variables huseduc, husage, hushrs, hushisp, husexp, husblck.Husearns=c+x1huseduc+x2husage+x3hushrs+x4hushisp+x5husexp+x6husblckUsing OLSDependent Variable: HUSEARNSMethod: Least SquaresDate: 01/01/11 Time: 16:01Sample: 1 5634Included observations: 5634VariableCoefficientStd. Errort-StatisticProb.??C-464.433941.94402-11.072710.0000HUSAGE31.761252683270.0000HUSEXP-31.509342.118723-14.871860.0000HUSBLCK-29.4199821.04346-1.3980580.1622HUSHISP23.9779420.873451.1487290.2507HUSHRS5.8247250845790.0000HUSEDUC6.4272352.3895892.6896830.0072R-squared0.153373?Mean dependent var453.5406Adjusted R-squared0.172492?S.D. dependent var406.9878S.E. of regression370.2265?Akaike info criterion14.66735Sum squared resid7.71E+08?Schwarz criterion14.67559Log likelihood-41310.92?F-statistic196.6973Durbin-Watson stat2.012364?Prob(F-statistic)0.000000You can see from the table, R^2=0.153373, which