Hallo,
I would like to output confidential intervals for the significant tests for my (pearson-)correlation-matrix.
I allready have a formula, but there must be a way to get SPSS to calculate this automaticly.
Thanks for your help!
Kind regards!
Jan
Here are three options for you using SPSS:
Do not use the method shown in the following video--it is not correct.
- https://www.youtube.com/watch?v=-dSoWqDyT4E
The method shown in that video yields a CI that is symmetrical regardless of the value of r. The correct method yields a symmetrical CI only when r = 0. Otherwise, the side of the interval toward 0 is wider--see the attached image.
HTH.
I no longer use SPSS (partly because it is a pain to do this sort of thing). It would be simpler to do this in R (or even in Excel). You can do this using compute commands (see link) and there may well be macros.
I attach a simple excel calculator I wrote for this a while back (though in practice I would use R).
Here are three options for you using SPSS:
Do not use the method shown in the following video--it is not correct.
- https://www.youtube.com/watch?v=-dSoWqDyT4E
The method shown in that video yields a CI that is symmetrical regardless of the value of r. The correct method yields a symmetrical CI only when r = 0. Otherwise, the side of the interval toward 0 is wider--see the attached image.
HTH.
The online calculator at http://vassarstats.net/rho.html does the job for me. The site does state that "the confidence interval of rho is symmetrical around the observed r only with large values of n". I did not face this problem with a sample size up to approx. 100. For greater sample sizes, the links provided by Bruce might be more appropriate.
All the best.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS