What’s more effective in countering misinformation – facts of logic? Along with Emily Vraga, Sojung Kim and Leticia Bode, we’ve just published a new study, Testing the Effectiveness of Correction Placement and Type on Instagram, which explores this question. If you don’t have library access to the International Journal of Press/Politics, here is an open-access pre-press of the study.
In this research, we were interested in two approaches to countering misinformation: fact-based or logic-based corrections. Fact-based corrections show how misinformation is false by providing accurate information. Logic-based corrections explain how misinformation misleads by exposing its rhetorical techniques or logical fallacies. Most research into misinformation tends to focus on fact-based corrections, and most fact-checkers are, well, fact-based.
In my Cranky Uncle vs. Climate Change book, I combine both methods to counter climate myths. For example, here is a page from the book where I counter the “CO2 is plant food so it’s good to emit CO2” myth:
I use the Fact-Myth-Fallacy structure throughout the book. Each debunking begins by explaining a key fact about climate change. Then after introducing the myth, I explain the fallacy the myth uses to distort the facts. In the case of the “plant food myth”, the fallacy is oversimplification.
I also use cartoons to illustrate both fact and fallacy. To reinforce the fact that plants need the right conditions to flourish, I used the analogy of a prickly (pun intended) plant. To reinforce the fallacy explanation, I use the parallel argument of ice cream. Credit where credit is due, I borrowed (stole? homaged?) the ice cream analogy from Sarah Green’s Denial101x lecture on the plant food myth.
While we’re on the topic of the Denial101x MOOC, note that the Fact-Myth-Fallacy structure is used in all our video lectures. The fact/fallacy of every climate myth is summarized neatly in a single fact-myth-fallacy page – a one-stop-shop resource for anyone needing a quick response on climate misinformation.
But back to the research! For our experiment, we chose for our myth the argument “CO2 is plant food so emitting CO2 is good.” As there wasn’t much (any?) research into debunking misinformation on Instagram, we decided to test the myth and our corrections in the form of Instagram posts. Here’s what our misinformation post looked like:
In our experiment, our fact-based correction explained how plants need a variety of conditions – including a comfortable temperature range and steady water supply – to flourish. We reinforced the factual explanation with the prickly pool plant cartoon.
The logic-based correction explained the fallacy of oversimplification, ignoring the complex set of factors required for healthy plant growth. It’s like eating ice cream and saying… well, it’s all in the cartoon:
In our experiment, we divided the participants into five groups. One group was shown only the misinformation. The other groups were shown either the fact-based or logic-based correction. We also manipulated the order in which participants saw the misinformation and correction. They either saw the correction before the misinformation (prebunking) or after the misinformation (debunking). After the Instagram posts, we measured their belief in the plant food climate myth. The following bar chart shows each group’s belief in the myth (the lower the value, the better as this is a measurement of misperceptions):
The overall result is that logic outperformed facts. Go team logic! But it’s important to dig deeper into why. Whether the fact-based correction was effective depended on whether it came before or after the myth. If the facts came last, they were just as effective as logic in reducing misperceptions. But if the myth came last, it cancelled out any possible good effect that the facts might have had.
In contrast, logic was effective regardless of ordering. This indicates that explaining the rhetorical technique that a myth uses is potentially more robust in neutralizing misinformation, whereas the effectiveness of facts depend on what order people receive information.
This is not to say that countering misinformation with facts is unimportant. On the contrary, my approach whenever possible is to have my cake and eat it too, countering misinformation with both facts and logic. But this research does highlight how logic and critical thinking – an under-utilized and under-studied approach – is a powerful tool in neutralizing misinformation.
Again, you can read the full paper at http://sks.to/factvslogic.
A. Randomjack
I would say effectiveness depends in who you’re arguing with.
Evolution (nature + nurture) makes all sorts of minds, as well as creatures.
I’d recommend “The Blind Watchmaker”.
If you haven’t been totally awed by Evolution, you never understood it (paraphrasing Douglas Adams)
Thanks your work
Thomas H. Pritchett
While I agree that we should be debunking misinformation by using fact and logic based arguments, I am not sure, without seeing the sample size of your study and the actual value of your error bars, whether there is enough statistical power to claim that one approach works better than the other.
John Cook
All our stats are reported in the full paper at http://sks.to/factvslogic. The different results between the fact-based and logic-based approaches are statistically significant. However, this is just one study – we are planning follow-up studies as replication is a key element to the scientific method. Ideally, this approach should be tested across different issues and different types of fallacies.
Paul Hayes
In the meantime… https://statmodeling.stat.columbia.edu/2020/05/24/banishing-black-white-thinking-a-trio-of-teaching-tricks/
Claude T Harrell
Wondering how effective either method really is. I was once taught there are 2 types of arguments: Emotion based and Fact based. I was also taught that you cannot deal with an emotion based issue with fact unless at the same time you satisfy the emotional need of the person with whom you are having the argument. Similarly, you cannot solve a fact/logic issue no matter how much emotion/feeling you put into the argument – if you do not have facts to support you. I find most people these days are rather set in their positions and many are emotion based. Especially with deniers, if climate change was a dog, they would continue to deny its existence until it barked and bit them in the behind. Ok, a bit of an exaggeration but seriously it seems the worst effects will have to be on us before many will accept, and then only grudgingly, that man-caused climate change is real.
John Cook
That’s an important point and I have a few comments on it. First, this work is based on inoculation theory and there are two elements to an inoculating message. The first element is a threat warning, the second element is counter-arguments explaining how the misinformation misleads. The threat warning has an emotional element to it – it works on a system 1, fast thinking level – while the counter-arguments are the more reasoned rational aspect – the system 2, slow thinking part of the inoculation. Research into inoculation has shown that it’s the threat warning that does most of the heavy lifting. So there is an emotional element to inoculation – warning people of the threat of being misled.
Second, it’s often implicit when I talk about inoculation but its based on the principle that you stop misinformation from spreading not by changing the minds of deniers but by stopping the rest of the population from being misled by their misinformation. Only 10% of the U.S. public are dismissive of climate science but they are a vocal minority. 90% of the population are open to evidence but also vulnerable to misinformation. Inoculating messages have their greatest impact on the 90% who aren’t dismissive and that’s where we should focus our communication efforts.