A valid test would also have to be done on a fairly large collection of printers (or at least print heads) and inks. This is why I'm always concerned about drawing (jumping to) a conclusion based on just my experience with one printer.
I had a problem with a particular third party cartridge with a particular print head where it would band. If I switched to an OEM cartridge, the problem went away. Or if I put the third party cartridge in a different print head but in the same printer, the problem went away. This shows that a marginal cartridge can cause a problem in one situation but not another.
Imagine if you're the OEM and you have to design for millions of customers who might even be abusing their systems. A friend of mine who I refill for actually had one of his cartridges not snapped down all the way! Think of what Canon has to go through with people like that. This is one of the reasons, I believe, they went to the chipped cartridges. The printers without the chipped cartridges cannot tell if a cartridge is not snapped down; they cannot even tell if a cartridge is missing! When you're designing a system to be fool-proof, it's because there are actually fools out there that are going to be operating your product and blaming you when something goes wrong that is really their fault. A lot of people have no problem using third-party cartridges, but when you're refilling for a lot of other people like I am, or giving advice to lots of other people, I just don't see why you'd mess with third party cartridges.