I have also noticed the same sort of variation with my new PX-716a (I'm going try to get a replacement at Fry's today to see if it's the drive's fault). Perhaps it is more representative of actual burn quality to take an average of, say, 5-6 PIE scans. In this respect, the more discrete TA test may be more useful. It's faster to scan (so you can repeat the test several times to get a better overall assessment) and moreover it has been much more consistent for me. That is, PIE results can vary widely while the TA test gives pretty much the same results time after time.
It might be helpful, in the scan thread, for people to run the same disc through, say three times over, and report the average and standard deviation of the results. Sometimes I wonder if all these "best scan" posts are genuine or really just luck-of-the-draw one-offs or the best result out of several scans.