You sound fine to me, and I get your point.

The key thing to realize, and that often escapes people at first glance, is that this isn't an opinion poll. With an opinion poll it is critical to have as representative a sample as possible.

In this case, it's important to ask how the people who are being sampled are likely to differ from the general population, AND how this might affect the results.

Ways forum visitors might differ from the general population:
1. Drive more aggressively
2. More likely to mod
3. Might maintain their cars better
4. Might notice a problem with their car only after hearing of it on a forum
5. Came to the forum because they have been having problems

Of these, I'm most concerned about #3 and #5. But looking at the details of the repairs that are reported, the great majority, including the many electrical problems people report:

1. Could not have been caused by how the car was driven
2. Could not have been caused by a mod (and respondents are asked if their cars are heavily modded--few are)
3. Could not have been prevented by maintenance--there's no way to maintain a window regulator, for example; that said, maintainable parts do fail more often with older cars, so this becomes more of a factor with those
4. Are problems that anyone would notice
5. Occurred after the person learned of the research and signed up

Yes, there might be some distortion from the above factors, but not enough to render the results misleading unless someone is trying to split finer hairs than I'd suggest. This isn't a presidential election poll, where a result that is off by more than a couple percentage points would be misleading.