Below are the results of two different linear regressions. The first only has an N of 10 while the second has an N of 43. The first has a very high association between the dependent variable and the independent variables overall (Adj R2 = 0.948, P = 0.003). The second one only has one variable in the model with a significant association with the dependent variable and a lower overall association (Adj R2 = 0.436, P < 0.001).

1st case

2nd Case

I am trying to understand some of the reasons for these two sets of results. The first seems spurious due to the small N. I am still having trouble understanding how the second could have an adjusted R2 as high as it does considering only one variable has a significant association with the dependent. What kinds of things might be going on here?

  • $\begingroup$ If the dependent variable is different you can not compare the R2. $\endgroup$ – Robert Jul 30 '16 at 19:09
  • $\begingroup$ The dependent variable is the same. The study area is different though. $\endgroup$ – SteveC Jul 30 '16 at 21:07
  • $\begingroup$ R^2 for the second model is not that high imho... it only "explains" (or better covers) ~50% of the variance, right ? $\endgroup$ – Drey Aug 1 '16 at 10:42

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.