TrueDelta’s survey and analysis have been fine-tuned to yield reasonably accurate vehicle reliability stats even with a fairly small non-random sample. But, when you have 179 results, a few are bound to be off. Those I suspect are off are noted as such in the comment you’ll see if you hover over the result.

My initial reaction on seeing the initial result for the 2008.5 and 2009 Mazda3 was that it had to be one of these errant results. This was especially likely, since only 17 responses were received and this result is asterisked for a small sample size. Every previous model year of the Mazda3 has had a low or at least lowish repair rate (the 2004 has been drifting into average territory). So why is the slightly tweaked 2008.5 running about twice the average, with 105 successful repair trips per 100 cars per year?

Click on the model name on the results page, and you’ll be taken to the Mazda3 repair histories page, which displays all repairs reported for the car. It turns out that the revised audio system has often lost bass at high volume, and many cars have had the unit replaced.

So, no fluke. Even with a sample size of 17 cars the Vehicle Reliability Survey has uncovered a common problem with the car. Is it a major problem? No, it’s not. But then we provide this information as well.