I don't think they are saying that we are predominantly logical with a few exceptions around controversial topics...
I think they are saying something much, much deeper and maybe this is just my confirmation bias kicking in here but it really supports my observations on a position I've held for a couple of years now.
I think what they are alluding to is that we are not logical and not rational in how we choose all of our beliefs as we do not try to seek out what is true (except for one edge case), instead we try to seek consensus within some political sphere (a group of any kind).
The so called controversial topics are where it's easiest to spot and easiest to study. But there is a stunningly deep implication that just because there is an uncontroversial consensus around something it doesn't mean the conclusion was arrived at by a sound process. So you actually really have to question everything.
For a while now I've been saying people don't argue for what is true, they argue for political affiliation.
It's close to impossible for humans to tell what's true. Often it would take trillions of dollars, millions of people or thousands of years to actually verify something as being the case. We take 99.99% of our existence on faith because it's simply not possible for us to fact check it. And if it's not possible for us to fact check what is true except for in the most basic of cases why would it ever be important to us as a social behavior? So that's why we see what this article points out. At least in my view.
We do do some degree of fact checking but it's usually the minimum viable fact checking to coordinate with our peers.
The interesting edge case they point out where some people do reason consistently is also signaling political affiliation. They are attempting to signal "I'm on team evidence-based". But the interesting thing about team "evidence-based" is they tend not to question or be skeptical so much of the fraud and other social phenomenon (replication crisis etc) present in science. They take it on faith that "this is the best answer I have whether it is true or not, but it's most likely true therefore it's true". Even team evidence-based isn't purely rational. And for good reason. They can't be. It's simply not viable to personally try to reproduce every study you read.
We are political animals that care about our survival and forming consensus. Not truth.