Document Type



In this thesis we explored some topics in regression analysis. In particular, we studied what linear regression is from a matrix theory perspective, and applied analysis of variance in a setting with two factors and unbalanced sample sizes. In addition, we applied Box-Cox variable transformation as a solution when the regression model violated the normality and equal variance (also called homoscedasticity) assumption. Our main goal is to use these theories to construct models and investigate questions related to lifetime earnings of people living in America by using real data. In doing so, we used the statistical software R to perform calculation involved in variable selection models, to identify and quantify relationships between variables as well as to test hypotheses.



Thesis Comittee

Dr. Kevin Rion, Thesis Advisor

Dr. Irina Seceleanu, Committee Member

Dr. Wanchunzi Yu, Committee Member

Copyright and Permissions

Original document was submitted as an Honors Program requirement. Copyright is held by the author.

Included in

Mathematics Commons