Not quite your question, but -corrci- from the Stata Journal site
(-search- for exact location) supports confidence intervals for Pearson
correlations.
Thinking is not uniform on this across statistical science, but you only
have to study a few correlations, sample sizes, and calculated P-values
to see that -- unless your sample size is absurdly small -- almost any
correlation qualifies as significant at conventional levels.
Nick
[email protected]
Katja Schulze
Thank you, Martin!
You got it, I've got Vista. But now it works perfectly fine after having
updated to 10.1.
Do you also know how to add p-values (best in parantheses) below each
correlation coefficient?
I could imagine
qui spearman `varlist', pw stats(rho p)
could help, but how do I combine the rho-/ and p-matrices? And I do not
find an option for corr similar to the "sig"-option of pwcorr.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
