We know that scanning is really a test of how drives read disks.
Not so, unfortunately. Perhaps you mis-used the term "read" here. "Reading" is what goes on in transfer rate tests and file tests. Scanning forces a drive to lock speed and never re-read to correct, so it eliminates error correction in the reading process. The disc is being read, but the data is not.
There are a number of reasons to "eliminate" some drives as error scanners, mostly the way they report errors and what they actually DO report as PI/PIF. In this regard, they are all different in some ways, and IMHO some of them are not usefull for my own purposes. Most, if not all, of the arguments offered for one drive over another as a scanner are just rationalizations in an attempt to justify using one's favorite drive.
Personally, I will not use a drive for scanning that reports PIF in anything other than 1ECC blocks. There's just no reason to accept 8 ECC or higher reporting when 1 ECC is available. Again, my own preference.
It's always been true that comparing scans from different drives is foolish, but that is equally true whether they are 2 different drives or even 2 drives of the same model.
People never seem to get a grip on the idea that the purpose of scanning is not to compare drives, but to compare DISCS.