bad science

No, Brain Waves Can’t Explain Politics

If you read Brian Stelter’s recent CNNMoney article, you learned some surprising things about the effects of Donald Trump on the human brain. Namely, that “Something about Trump — his face, his voice, his message — generates ‘increased brain engagement,’ SBB Research Group CEO Sam Barnett said in an interview with CNNMoney.”

Whoa! Maybe this, finally, can explain the Trump phenomenon. Let’s read on. According to Stelter, “Barnett, a hedge fund operator by day and neuroscience Ph.D. candidate by night, has been strapping electrode caps on focus group participants and showing them primary season debates.” CNN Money teamed up with him to gauge voters’ reactions to Trump, and found that:

The most intriguing result was that Trump generated more brain engagement than his rivals in almost every demographic category, even among Democrats and Independents,” Barnett said. “In each category, Trump outperformed the lowest ranked rival by a double-digit percentage.”
Barnett’s findings come from an electroencephalography study, or EEG study for short.

EEG systems measure electrical activity of the brain,” he explained.
The focus group participants might have been excited by Trump. Or they might have been repulsed. But one thing was for sure: they weren’t bored. This could be a testament to Trump’s considerable television performance skills. Or to his provocative nature. Topics like immigration, a staple of Trump’s speeches, triggered big surges in neural activity.

Trump led Ted Cruz, John Kasich and Marco Rubio among the women in CNN’s focus group of 12 Illinois viewers. Rubio led Trump slightly among the men.

While the focus group was relatively small, Barnett noted that his “neural engagement rankings” for the Republican viewers — Trump first, Cruz second, Kasich third, Rubio fourth — “exactly predicted the ranked outcome in the Illinois primary” the following week. [all emphasis mine]

If you read that closely, your eyebrows should be arched like the Cloud Gate by now. Let’s review the contradictory and questionable claims:

1. Barnett doesn’t know what it’s measuring, and maybe it’s measuring disgust, but it predicted the winner! Meaning maybe those were most disgusted by Trump … voted for him?

2. The “neural engagement rankings” exactly predicted the winners in the actual Illinois primary. Oh, except for men — Rubio elicited the most “engagement” among men, and then went on to rake in a grand total of 8.7 percent of the vote in Illinois.

3. The group of 12 voters was “relatively small.” Well, no: For a four-person race, that’s uselessly small. Take a situation in which each candidate gets three brain-scan “votes” from the participants. Now flip just one of those “votes” from Kasich to Trump. You’ve slightly altered one data point and completely changed the results from a four-way tie to a clear ordering with one tie in the middle. That is some fragile, fragile data.

Even the basic lingo here didn’t sit right to one brain expert. John Borghi, a cognitive neuroscientist whose tweet on the study first brought it to my attention, told me via DM that “brain engagement” doesn’t mean much in this context. “Usually ‘engagement’ means something like ‘involvement’ in this context,” he said. “For example, ‘the prefrontal cortex may be engaged in this process …’ But it seems like they are looking at a particular EEG measure and calling it ‘engagement,’ which isn’t appropriate.”

Borghi also said “neural engagement,” which comes up in both the article and in a graph label in the accompanying video (hooked-in voters’ “neural engagement” appears to spike at particularly Trump-arific moments), isn’t a term neurologists use in the way Barnett is. He checked Google Scholar to be doubly sure, then said, “Occasionally somebody will use it as a synonym for ‘brain activity.’ But the way they’re using it in this article is completely meaningless.” (In the graph section of the video, Barnett says of Trump to Stelter, “You can see he’s at 35.9, when these other candidate are in the twenties.” Thirty-five-point-nine what? Doesn’t say, though that could be the fault of CNN’s video editing.)

Barnett’s bio makes him out to be a pretty smart guy and a talented money manager. But an impressive investment career and a spot in a neuroscience Ph.D. program doesn’t mean he gets to claim to have developed a system to peer inside voters’ heads and use the resultant data to accurately predict the result of political contests. If he had, after all, that would be an astounding advance — and when you claim an astounding advance, you need to back up that claim with hard data.

Of course, at a certain level Barnett can make whatever claim he wants. What’s more disappointing is that Stelter let someone throw some fancy-sounding brain terms at him and then wrote up a credulous story. But this is, unfortunately, par for the course. Because neuroscience remains an area of such mystery and misunderstanding, people are particularly susceptible to believing and circulating less-than-rigorous claims about it. When those claims pertain to a hot-button subject like sex, or politics during a heated primary campaign, that susceptibility only gets exacerbated.

Anyway, my theory is that Stelter has an overactive overactive predorsal rear cingulate axolotl — a part of the brain I just invented that causes otherwise smart people to fall under the sway of neurobabble.

No, Brain Waves Can’t Explain Politics