Remember, this is a European survey. Mazda and the others sell a bunch of cars over there that they wouldn't even think of selling in the US.
My biggest problem with this kind of survey is that they are actually a bit biased against luxury cars. 1st of all, a Luxury car like a BMW 3-series or an Audi A4 will usually have more toys and features then your average Honda Civic or Mazda3, like nav, heated seats, or traction control. The presence of more gizmos means that there are more things to go wrong. Imangine it as shooting at a 1 foot dartboard versus shooting at a 1 year dartboard.
This survey doesn't tell us the magnitude or cost of the issues that arise. For all we know, 50% of Saabs might have a faulty power window, but impeccable drivetrains, while 1 in 10 Mazdas' engines explode, but their electronics are second to none. Surely the engine exploding is a bigger problem, but it would count for just as much as the faulty window.
There is also an issue of expectations. We expect a $35,000 Audi to be next to perfect while we would expect more issues from a Kia. If an interior light on your A4 is flickering, you would be more likely to report the problem then if it was in a Kia Rio.
On top of all of this you have to ask, can you really put all cars from one company under one banner? I would argue that from company to company, certain cars are more reliable then others. A BMW 328i would probably be more reliable then a 530i with IDrive which would be more reliable then an M5. Since we buy individual cars and not a whole company's lineup, information about individual models would be much more telling.
Finally, there has to be a hundred of these surveys and they all get slightly different information. I tend to look more at trends between surveys then at one survey. What I have found is that Honda, Toyota, and Lexus are always at or near the top, American car makers are usually in the middle, Land Rovers are unreliable, and that luxury cars usually do worse then regular cars.