But what isn't shown in the poll is that of the 1/3 that own/have owned. a Taurus with a problem, how many of the "own" guns are still owned because the problem was fixed, and the owner is now satisfied with the gun.
That is correct. To get that information the poll would have had to be designed differently.
However, even without that information, we can still get a pretty decent picture of what the situation is in terms of how likely one is to get a Taurus with problems, assuming that we've gotten reasonably accurate responses to the poll questions.