I want to perform multiple linear regression on my variables; since my dependent variable was negatively skewed, I reflected it and then I log transformed it. My independent variables was positively skewed, so I only log transformed them.

My first question is: after performing multiple linear regression on my variables, should I interpret results adversely?

I mean: if an independent variable has a positive beta, I say it has a negative effect on dependent variable; and if an independent variable has a negative beta, I say it has a positive effect on dependent variable?

My second question is: my adjusted $R^2$ is $0.35$, is that good enough to say that my model is acceptable?

  • 2
    $\begingroup$ 1. The data does not need to be normal, the residuals need to be. 2. What is a "good enough" $R^2$ varies by field. 3. Both these questions have been asked before (many times). $\endgroup$ – Peter Flom - Reinstate Monica Apr 20 '13 at 14:36
  • $\begingroup$ Thanks a lot for your comments. sorry if I asked duplicate question! I'll search in site more carefully next time. $\endgroup$ – PSS Apr 20 '13 at 14:41
  • 1
    $\begingroup$ You could search the site now, @PotentialScientist ;-). Re Peter's first point, you might try reading this thread: what-if-residuals-are-normally-distributed-but-y-is-not, eg. In addition, your question is not entirely clear: by "analyze", do you mean interpret? You refer to both sets of variables as "dependent variables", might one of them be independent? $\endgroup$ – gung - Reinstate Monica Apr 20 '13 at 15:22
  • $\begingroup$ yeah; thanks a lot; I edited my question. I'll take a look at the link. by the way, I checked the P-P plot of my variables and I saw that residuals are not normally distributed ; then what should I do now? do I need to transform my variables? $\endgroup$ – PSS Apr 20 '13 at 15:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.