It's a phenomena called "motivated reasoning" in which your thinky forebrain bits shut down and the lizard brainstem kicks in. If you don't want to sit through the video, the meat and potatoes study is titled "motivated reasoning and enlightened self-government." They did a test in which they measured a subject's mathematical reasoning ability as a control, then they started describing similar mathematical questions with highly contentious political concepts, the result was that people started answering math questions either poorly or well depending on their political bias.
The paper seems to argue kind of the opposite of what you're saying as far as the reasoning, I think. It's arguing that more intelligent people (well, more numerate people, which is something highly correlated with intelligence and quantitative reasoning) are less likely to heavily scrutinize reasonable information when it aligns with their prior biases, not that it affects their intelligence or anything.
The examples in the study are designed in a way that doing some math is required as a quick, lazy glance will be reasonable, but wrong:
In the test, all the lazy look answers were wrong. The smart people caught it in the control questions on both aisles. When the lazy look answer was thought to align with their beliefs, smart people activated the lazy glance and incorrectly answered. When it didn't, they activated their slow, methodical analysis and correctly answered.
So basically, you're less likely to scrutinize reasonable information that aligns with your prior beliefs, but will stop being lazy when it doesn't. It's honestly not much of a crazy finding to me, but more of a "Yeah, no shit." That's actually how humans save time, by building on prior understanding of shit they've heavily scrutinized.
It's like someone sharing a realistic AI video of a tranny freaking out or perving in the women's bathroom or something. That's reasonable to believe since 354864 real ones exist so I'm less likely to research whether it's fake or not.
x 75 
