What does it mean if...
Posted: Tue Nov 18, 2014 9:54 pm
I created a model with life expectancy as dependent and obesity, salary, % that workout, %female
and I get results like
Coefficient
Salary 8.28
Obesity -.00689
Exercise .03343
Obese -.-689004
However if I include % of smokers as a variable, the signs of the coefficients change and don't make sense anymore. For example I would get
Salary -1.58
Obesisty7 .002334
Exercise .033434
Obese .002225
Smokers -0.17787
Which is basically saying that those who make more live less years, and if you are obese you will live longer
Why would adding that varible change the signs, and how can I resolve it
and I get results like
Coefficient
Salary 8.28
Obesity -.00689
Exercise .03343
Obese -.-689004
However if I include % of smokers as a variable, the signs of the coefficients change and don't make sense anymore. For example I would get
Salary -1.58
Obesisty7 .002334
Exercise .033434
Obese .002225
Smokers -0.17787
Which is basically saying that those who make more live less years, and if you are obese you will live longer
Why would adding that varible change the signs, and how can I resolve it