- #1
- 7,237
- 10,984
Hi all,
Hi all, hope this is not too simple; please feel free to give me a reference if this is so. I want to know how to address having "insignificant" coefficients in my regression:
I just did a multilinear regression returning ## y=a_1x_1+...+a_kx_k ##
In the resulting analysis, two of the coefficients , say ## a_1, a_2 ##, came out to be insignificant (at a 95% level) . Still, when I remove them and do a new regression ## y=a_3x_3+a_4x_4+...+a_k x_k ## the adjusted R^2 drops very slightly. Is it standard practice to drop "insignificant" variables?
Hi all, hope this is not too simple; please feel free to give me a reference if this is so. I want to know how to address having "insignificant" coefficients in my regression:
I just did a multilinear regression returning ## y=a_1x_1+...+a_kx_k ##
In the resulting analysis, two of the coefficients , say ## a_1, a_2 ##, came out to be insignificant (at a 95% level) . Still, when I remove them and do a new regression ## y=a_3x_3+a_4x_4+...+a_k x_k ## the adjusted R^2 drops very slightly. Is it standard practice to drop "insignificant" variables?