Hello, I am working on a graduate thesis about rate of return on Education. I ha

Question

Hello,

I am working on a graduate thesis about rate of return on Education.

I have generated a modified mincer equation with the following:

In(wage)= B0 + B1(education) + B2(Education)2 + B3(expereince) + B4(Experience)2 + B5(Gender) + B6( days of annual leave) + B7(city)

The regression I used is OLS regression but it doesnt seem to completely meet my professor requirements.

Is there any further calculation I can do other than (Oaxaca decomposition model and Heckman model)?

The analysis I have done showed that all of the variables of the regression were significant and does effect wages and it showed that men and women wages were not really equal, However the professor thinks just doing OLS regression isn't enough. I could do a comparison between two method to strengthen the thesis, but what could I use?

Explanation / Answer

OLS regression helps you to define a relationship between the dependent and independent variables and how much independent variables are statistically significant in affecting the dependent variable. The magnitude of the coefficient of the independent varibales define how much statistically significant the independent variables. In your equation, the independent variables are education, experience , gender , days of annual leave and city and the dependent variable is ln(wage). These independent variables may also be correlated with each other. For e.g. Education and Gender may be correlated. Education and City may be correlated. In many cases, Men have higher education than men. In Metro Cities, Education is high compared to small cities. To address these correlation problem among the independent variables (multicollinearity) and also reduce the number of independent variables (dimensionality reduction), you can use the methods Principal Component Analysis (PCA) or Factor Analysis which comes up with only few independent variables which explain most of the variance in the dependent variable wage. To check for multicollinearity, check Variance Inflation factor (VIF), and if VIF is high, you need to run PCA to eliminate highly correlated variables.

You can also perform ANOVA method , to check if there is any significant difference in wages among different genders, cities, experience.

You can also perofrm cluster analysis, which divides given data points into clusters of different wages on a plot and you can find what ranges of education values falls in different clusters.

Related Questions

Navigate

• Browse (All)
• Browse H
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.