I do the survey every few years. It’s quite detailed and takes a long time to complete, so it’s not something to be done lightly.
That said, I’ve noticed no significant change in the survey, so probably the best use is to compare the thousands of responses year-over-year rather than as a static and meaningful number by itself. It’s certainly true that some reliability incidents are minor some are major, but across tens of thousands of surveys one might expect that to balance out, h year by year
I’m not at all sure about that, especially as different brands may appeal to different buyer types with different attitudes toward different types of issues.
One sort of example of this is need for recalls, which for one brand may mean always going to the dealer and for another brand may most often mean over the air with no impact on the owner.
All of the above and much more. The survey includes things like door locks and latches, paint, body panel fit, interior fit and finish, etc. Not all of those are mission critical or unique to EVs. EV manufacturers are either new to making EVs or new to making cars in general, so they tend to experience more problems like the trunk not closing properly than legacy manufacturers.
Just for the record, the CR survey includes the responses from 330,000 owners covering virtually every model of popular car. (I would not be surprised to find Lambo not included 
When you get a sample size that large the results should be remarkably stable, although I suppose some “over the air” recalls might get interlaced with those requiring a trip to the dealer. But mostly I’d guess owners would ignore those.
My point was, even if CR’s survey is “statistically valid”, which it may be, it’s still crappy “journalism” by CBC to get some eyes on their ads.
For those interested in owners’ experiences, I think it’s a useful heads-up by the CBC both by category and brand.
DB2
Take Tesla, the first cars to roll out of Fremont were pretty crappy and many are still on the road. They will lower Tesla’s rating but the new buyers are getting much better cars. This was reflected in the comments in the article but not in the scoring.
The Captain
“Stable” perhaps, but that doesn’t translate to necessarily meaningful. If one doesn’t ask the question the right way, the results may be meaningless.
E.g., if the question was “how many recalls did you experience?” because it used to be that going in to the dealer was the only option, once over the air recalls came to be, one really needs two questions, one each for in person and over the air. Indeed, perhaps three since aren’t there some recalls which have involved Tesla going to where the car is? Without that break down, one can no longer interpret the question sensibly.
For Tesla, most people won’t even notice the “recalls”. Software updates are just a regular part of the experience, as with other electronic devices. That the update also changes something the NHTSA wants changed typically makes no difference.
Here’s a good example:
A few people took advantage of Tesla’s silly feature. When the “recall” made it go away, some of those people were probably annoyed at the dumb regulations. But everybody else didn’t even notice.
What’s really important is that the typical owner experienced no inconvenience. So, yeah, counting “recalls” for such things is meaningless.
-IGU-
That headline is illustrative of the problem. It really should be a different word than recall.
Part of the confusion is the name of the survey “Consumer Reports Reliablity Survey.” If there is a gap in the door panel it is reported in the survey as a manufacturing flaw, but it doesn’t mean the car is unreliable. The CBC article extrapolated things like the latter to mean the car is in fact unreliable. Which is not quite accurate.
Likewise, many of the over the air updates are enhancements or fix a problem that the driver was not experiencing*.
*Other drivers experienced it, which is why it got fixed.