Misinformation, the Brain and Tricks for Analyzing Data Accurately

Michela Marinelli, an associate professor of neuroscience and neurology, teaches a class particularly relevant as widespread misinformation becomes a hot topic in society.

Photo by Matt Wright-Steel/Alcalde.

​Marinelli won the 2020 Texas Ten award from The Alcalde, which recognizes professors who have made a difference in the lives of their students. 

This month, she is participating in the Texas Science Festival and is the featured speaker for UT Brainstorms for "The Fooled Brain: A Conversation on How We Are Easily Misled When Processing Information and Making Decisions." Her research seeks to understand the neurobiological basis of drug addiction.

You teach the Analytical Skepticism (NEU 377) course. Could you describe the content of the course and your experiences teaching it?

I teach students how to navigate data, which is difficult. We have irrational ways of judging data, making it difficult to interpret it, and we are often misled by data displayed poorly, whether on purpose or by accident. I teach some tricks of the trade on how to know if the relationships you are seeing in the data are real.

Analytical Skepticism is an extremely hard class, but it is very rewarding. Students often say it is the hardest class they have ever taken, but it changed the way they see the world. Sometimes I get emails from people who find that they reference their class notes in their job.

What are some common threads in the instances in which people are misled?

Our beliefs color the way we judge new information. So, if we believe one thing and we hear something different, we will find flaws in what we heard. We tend to see what we are looking for. Other flaws are related to data. If data are not portrayed the proper way, you might derive very different conclusions from it. Additionally, scientists will run an experiment without a control group and see a change over time. They attribute that change to an intervention, but because there was no control, they do not know if the change would have occurred without the intervention.

What are some examples in which data has been misinterpreted in the past?

A typical example is the J-shaped curve for the relationship between alcohol and health, which shows that people who do not drink any alcohol have worse health than those who have one drink a day, and as drinks per day increases beyond that, people start becoming more unhealthy. It turned out that those who did not drink any alcohol were a biased population: they were people who were sick so they could not drink.

Another example is a colleague of mine who did a study that found that people who take antidepressants spend less time in the hospital when they have a heart attack. They thought antidepressants could be a good medication for the heart, but it turned out that the women in the study were more likely to take antidepressants than men and tended to spend less time in the hospital. Adjusting for those differences, taking antidepressants wasn't the key to lowering recovery time, being a woman was.

How can we guard ourselves against being misled, especially on social media, when we have to analyze data or the news to make personal decisions?

It is important to refer to reputable sources, like educational or government sites. It is also helpful to know some tricks to see data differently. Knowing those tricks will not make you immune to being misled, though. Sensational news gets sent out more than bland news on social media. Anything that has a big finding definitely dwarfs anything where there is no finding, even when the findings are wrong. Additionally, social groups tend to reflect biases back to themselves on social media, which can make you more close-minded because you will not see the posts that have an opinion that differs with yours.

How should we approach conversations with family and friends who have been misled?

I understand that the only way to change someone's point of view or at least make them open their eyes to another one is to find a common ground to begin with. This will allow you to gradually show them the correct point of view, rather than getting immediately shut out. Additionally, in a leadership course I took, the teacher guiding me said, "You have to use your emotions to convince people." Telling someone that thousands of people are dying will not do as much as showing them the picture of a single person who is suffering.