How would you represent a variable with values of 1 or 0 in a multiple regression using the StepwiseMulReg class? If I set the value to 0, then the entire set of observations is thrown out. Along the same lines, if a value really is zero (say a percentage) then how would you be able to include it in the regression.
Are you saying that you have an independent variable, that has a value of either 0, or 1 (no fractions)? And when you run a stepwise regression the variable is always deemed insignificant and left out of the final equation?
Is the binary variable significant? How do you know? If you run a standard multiple regression (NOT stepwise) on the data, including all variables, what is the F-STAT value for the binary independent variable.
binary variable using MulReg
Posted - 09 Sep 2009 : 15:49:36
