Project 2: Regression Analysis to Predict Adult Smoking Prevalence Rank by State

Well folks, maybe the 3rd time will be the charm?!?

You see, here in MGS 8150 I ran into the same situation as I did with my regression model that I tried to build last year in MBA 7025 . . . what started out with three darn good independent variables (or so I thought) in the end only left me with one independent variable which was a paltry predictor of my dependent variable (R^2=.23). I think my lack of experience with analysis design is to blame, but the only way to get better at this is to keep on trying! :)

This time around, unexpected multicollinearity hindered my ability to properly (i.e. using more than one independent variable at once) run stepwise, backward, and forward regressions, and it turns out that ranked data doesn't lend itself well to regression analysis. So I live and learn. Again.

I am not a data person. I never have been as far as job experience or studies until now. But I am trying to be and I will persist and learn! :)

Anyway, please click here for my Regression Project and here for the accompanying Excel file.