Linear regression and dichotomous variable

#1
Hello

I would like to do a linear regression (stepwise), and one of the variables I want to enter is sick/ not sick. The other variables are from ratio scale. Is this is possible?

In addition, in my experiment N=60, how many potential predictors I can enter to the model?

Thank you
Tal
 

Karabiner

TS Contributor
#2
I would like to do a linear regression (stepwise),
This could cause problems.
http://www.stata.com/support/faqs/statistics/stepwise-regression-problems/
http://andrewgelman.com/2014/06/02/hate-stepwise-regression/

and one of the variables I want to enter is sick/ not sick. The other variables are from ratio scale. Is this is possible?
http://www.psychstat.missouristate.edu/multibook/mlt08m.html
http://bfy.tw/3OuC

In addition, in my experiment N=60, how many potential predictors I can enter to the model?
It depends, amongst other things, on the magnitude of effects assumed.
You could perform a power analysis using G*power http://www.gpower.hhu.de/
I suppose that 3-5 predictor could be typically used.

With kind regards

K.