Consumer Reports is a subscription based service. Everything they report publicly is designed to give you just enough to get you interested so you will pay for a subscription. In short, yes it is click bait, but it's backed by real data which you can pay to access if you so choose.
The data might be real, but it's heavily biased.
They survey only their subscribers, not the general public. So in the world of scientific polling, their data is fundamentally biased, and thus would not be considered scientific.
In some respects it's a self fulfilling prophecy. If a someone picks up the magazine on a grocery store news stand and reads that the car they just purchased, and love, is one that consumer reports praises, they then are more likely to subscribe to the magazine, and later report their experiences in the annual survey. Now, if they are having trouble with that new car, and regret their purchase, then they would conclude that the magazine is BS, and therefore unlikely to subscribe, and then have no option to report their problems. Similarly if consumer reports pans the car they just bought, they are likely to subscribe if they agree, and not subscribe if they disagree.
As an example of readership population bias, In the 2016 Auto issue they listed the 2015 Ford Fiesta as having insufficient data to report. Meaning very few owners of that car reported their experiences. But this is a volume production car, with US sales over 60,000 units, but is a car Consumer Reports does not like. So apparently not many Fiesta owners subscribe. Prior years showed poor reliability reported by those that do subscribe. The Nissan Senta is another example, this car sells over 200,000 units in the US, but yet consumer reports shows insufficient data. And it is a car they don't like.
But the for the Lexus LX, which they love, they did have sufficient data, even though this car much lower has US sales, less than 8,000 units. There are other examples.
In polling this is a serious flaw. For example if you were to poll Donald Trumps rally attendees on who they like, you will get a very biased result which means nothing.
I imagine if we were to poll the readers of this forum, we might get a very different result than consumer reports
I have been a loyal subscriber to consumer reports for more than 30 years.
But I can say that my experience with reliability of the cars I've owned to be quite contradictory to the data they publish.
One example was the Toyota Prius, which consumer reports loves, I found to be the most unreliable car I ever owned.
However when I bought my FiST, it took the precaution of leasing it for 36 months, the length of the bumper to bumper warranty.
