I'm asked this question often. It's a good one.

I do one thing differently from every other reliability study I'm aware of: with the exception of the first month, I only collect data on problems going forward. So, unlike with other surveys, people do not know what they will be reporting at the time they decide to participate.

For the typical model in the panel, 70-80% of panel members do not report a single repair trip in their first six months of responses. For some models the percentage is over 90.

So I don't appear to mostly be getting people with problems to report.

Some people do want to report their repair histories. So I created a separate survey for this. This second survey is not required of members, it's purely voluntary. And I would expect its results to be biased.