One of the biggest areas of disagreement among statisticians is when to use parametric or non-parametric statistics. I got an inquiry about this today, and quite honestly, I don't think it matters all that much. One consideration is sample size--if your sample is large, then the assumption of normality is easily met because of the central limit theorem. I wrote up a brief page on the central limit theorem and linked to a few pages that show how the central limit theorem works.
You can find an earlier version of this page on my original website.