Discussion: View Thread

Since so many of you were kind enough to respond to our initial question, I thought it would be appropriate to send out a follow-up. As I'd said before, the client came to us after her initial attempt at simple correlation was non-significant. I do believe that she was not "fishing", but honestly trying to figure out if that analysis was the correct one to do under the circumstances. (We have had clients who wanted to "fish" before, and we work hard to educate them and not to cast the net with them.)

In the end, the client would not show me the data. All we could do was to say that replicating the nurse scores 10 times for each patient score was not appropriate, offer to help if she brought the data in, and offer to help with the design of future studies so that this problem could be avoided.