A fucking large pile of expensive broken shit that just packed in is not anecdotal - it's known as evidence.
I wonder how often these studies are taken- annually? Quarterly? It'd be neat if they were done frequently, rather like polls during campaign season. Maybe a month or two after major product launches to gauge if satisfaction has changed.
But rationally it's quite clear that it is better to not need repairs (=waste time) in the first place.
It's the same as in many jobs - saving the day in a disaster will get more praise than preventing a disaster from happening; but, of course, prevention is far more valuable.