0

I am analyzing a survey result. I have 2 factors 1) a trainer factor 2) a participant benefited factor. These 2 factors are compared against a training rating and correlation is calculated. can i compare the (r)s? i.e., (r) of trainer factor to rating and (r) of participant factor to rating

sagarsmenon08
  • 1

1 Answers1

0

Yes, the difference between correlation coefficients can be tested for statistical significance. The test statistic depends on the two $r$ values, as well as on the sizes of the samples involved, $n_1$ and $n_2$. The exact formula depends on whether the correlation coefficients were calculated in the same group or not; and if they come from the same group, on whether they share a common variable or not. In $R$, the cocor package handles all these situations conveniently. See the related publication: Diedenhofen & Musch 2015

Ous
  • 448
  • 2
  • 6
  • Sample size is same.For example: (r)of sample x->A is .6 and (r)of sample y->A is .4; can i have a method to tell that Sample x is more correlated to A by a factor, as compared to sample y, and the significance of this deviation??? – sagarsmenon08 Jun 27 '19 at 13:30
  • I think you mean variables rather than factors. Can you specify a bit what these variables are? More specifically, are values of $x$ and $y$ paired somehow? (e.g. related to the same individual or item) – Ous Jun 28 '19 at 09:18