Every once in a while someone will resign from the panel and, when I ask why, they’ll respond that their car hasn’t been having any problems, so their participation hasn’t been of any use. No doubt many people think that cars with problems provide more valuable information than those without any problems. But this just isn’t true.

The primary goal of TrueDelta’s reliability research is to estimate the average number of repairs a model will require per year. For this estimate to be accurate, the ratio between cars with problems and those without problems in the panel must be as close as possible to this ratio in the entire car population. So every car is equally important, repairs or no repairs.

If people without repairs dropped out, TrueDelta’s results would overestimate repair rates. So it’s a good thing that once I explain to these people that they are providing useful data, they almost always continue to participate.