Personal Growth

Your Binary Bias Makes It Hard to See in Shades of Gray

It's time to order lunch for the office, and seven people want sandwiches while four people want soup. Sorry, soup lovers, the majority has spoken: Sandwiches for all! Except ... there's not really any reason why you couldn't order seven sandwiches and four cups of soup, is there? Your brain loves to set up false dichotomies for you to choose between, and it can sometimes hurt you when it doesn't need to.

Coping With Contradiction

Picture this: You're reading through the reports of the effects of a certain drug, and the results you're seeing don't all agree with each other. Some test groups are hungry four times more frequently, some aren't hungry any more frequently, and some are hungry about half as often. Now it's your job to gauge how the drug affects hunger levels overall. It's a tough call — seems like it makes the most sense to split the difference between the extreme effects and say that the drug makes you hungrier about twice as often. The only problem is, that's completely wrong.

In a new study by Matthew Fisher and Frank Keil, participants were given that exact job, with the results they viewed being a spread including some combination of "strong positive" (4 times more likely to be hungry), "weak positive" (2 times more likely), "does not change likelihood" (no effect), "weak negative" (half as likely), and "strong negative" (one quarter as likely). After seeing all of the results, they were asked how much the medication changed the likelihood of being hungry. The trick? Every set of data had the same average number, but some appeared to have more results on the higher end or the lower end. Most people tended to judge the likelihood of hunger based on how many high results there were compared to low results, not what the average result was. There was one other factor that shifted people's estimated answers: the first number they saw.

Related Video: How to Beat Bias Online

The Extremes Aren't Everything

The researchers found similar results when they had participants guess which of two menus came from a cheaper restaurant. One menu would have two $12 items, five $14 items, one $17 item, one $19 item, and one $20 item; the other, two $10 items, six $16 items, and two $17 items. The second menu looks cheaper, right? Wrong — both average out to $15 per entree. The effect was even more strongly pronounced when people had the data visualized for them in a chart.

This bar graph, for example, shows the seating availability of two different restaurants. Five percent of the tables at the first seat one diner, 45 percent seat two, 10 percent seat three, 25 percent seat four, and 15 percent seat five. At the second restaurant, 25 percent seat one, 15 percent seat two, zero percent seat three, 55 percent seat four, and five percent seat five. Eyeball the charts and you're likely to guess that the first restaurant skews closer to two seats per table and the second restaurant closer to four seats per table — really, the tables at both seat an average of three.

This tendency to favor the extremes goes far beyond the way we think about numbers. It's easy, for example, to paint those we disagree with politically with the same brush as the people on the absolute extreme end of the spectrum. Remember, there's almost always a lot of middle ground to cover.

Get stories like this one in your inbox or your headphones: sign up for our daily email and subscribe to the Curiosity Daily podcast.

If you're working on experiencing the world the way it really is instead of the way your brain thinks it should be, we recommend Howard Ross' "Everyday Bias: Identifying and Navigating Unconscious Judgments in Our Daily Lives." We handpick reading recommendations we think you may like. If you choose to make a purchase through that link, Curiosity will get a share of the sale.

Written by Reuben Westmaas November 20, 2018

Curiosity uses cookies to improve site performance, for analytics and for advertising. By continuing to use our site, you accept our use of cookies, our Privacy Policy and Terms of Use.