I think that's why Microsoft desperately wants to rebrand their engine (over and over) so that people will think it's something new that they should evaluate for the first time rather than the pre-Google product that they found inferior to Google.
Is this scientific? No. It is at least more objective, though. And I'm not trying for scientific rigor, just "does it do the task?" It's totally fair, too; that's how Google picked me up in the first place, and anyone else who meets that bar will get my business.
There's a case you're missing - that Google doesn't outright fail, but its performance is inferior to Yahoo's or Microsoft's. For example, Google may have the page you're looking for at the bottom of the first page of results (in which case you'd probably be satisfied and not try another search engine), but Yahoo or Microsoft might list it as the first result, which would be superior performance.
Go do some mechanical turking for us, please. :)
A search engine meant for use by billions of people is not a trivial thing.
It's a very weird effect. Ignoring everything else, Google certainly wins at cultivating brand loyalty. They're up there with Apple's brand loyalty in that respect.