Cognitive biases make it difficult to reason about vaccines

0


When it comes to thinking about the likelihood of events in the world, humans are sorely missed. Indeed, when we reason about probabilities and probabilities, we are liable to make all kinds of mistakes. Psychologists call these susceptibilities cognitive biases. Cognitive biases can lead to bad thinking and bad decision making, sometimes with life or death consequences.

One type of cognitive bias at play these days on social media, and even in mainstream media, is related to the likelihood of getting sick from Covid-19 after vaccination. Psychologists call the source of this bias the base rate error.

To make the base rate fallacy a reality, consider the following situation:

Suppose you hear that as a result of a party attended by ten people, one vaccinated person and one unvaccinated person fell ill with Covid-19. You might be thinking, “Wow, one in two (50%) people who have been vaccinated against Covid have been vaccinated! So why bother getting vaccinated? “

This is basically the argument you see on social media every day. But what’s wrong with this way of thinking?

To see the problem, imagine I tell you that on that same evening with the two diseases of Covid-19, the two people who fell ill were right-handed. Would you then conclude that if you are left-handed, you are safe? Probably not. But why not? The answer, of course, is that because there are a lot more right-handed people in general (90 percent of us are right-handed), there will also be a lot more right-handed illness. In other words, the chances that someone who gets sick will be right-handed are already much higher just because right-handed people are much more common.

The important point here is the idea that the overall prevalence of something in the population, like right-handed versus left-handed people, matters when thinking about the likelihood that something is in these two groups. This overall prevalence is the base rate, and our tendency to ignore the overall prevalence is the error of the base rate. The base rate of being right-handed is very high, so there will be more of just about anything in right-handed people than in left-handed people, including Covid-19 disease, simply because of this difference in basic rate. More right-handed people than left-handed people eat burgers every day. But that doesn’t mean right-handed people like burgers more than left-handed people. It just means that there are more right-handed people.

Now, if we replace “right-handed” with “vaccinated,” we can begin to understand why the baseline error is important in any discussion of immunization. Just like with right-handed people, as the vaccination rate rises and there are more people vaccinated and fewer unvaccinated people in the population, the absolute number of vaccinated people who get sick will increase.

Suppose you now learn that on this holiday with two Covid-19 diseases, out of ten people, eight people were vaccinated and two were not. You might still be tempted to say that one in two sick people have been vaccinated, but this clearly misses the important point about baseline rates. The right way to think about this is that the only vaccinated person who got sick was one in eight vaccinated people (12.5%), but the only unvaccinated person who got sick was one in two unvaccinated people (50%). ). Very different, right? The conclusion you would draw from this example when the base rate is factored in is that you would have been about four times more likely to get sick at this holiday if you hadn’t been vaccinated. In fact, the real numbers are even more strongly in favor of vaccination. And we are all at this party.

It’s no surprise that most of us find it hard to think about these situations. Our brains are simply wired that way, with built-in cognitive biases. Fortunately, being aware of these biases can help us spot them when they arise and ultimately avoid them. Remember, when you think about the likelihood that something is in different groups, the base rate matters.


Leave A Reply

Your email address will not be published.