The Science of Rationality
Are you irrational? I’m not talking about irrational by the dictionary definition of not having reason or lacking normal mental clarity. I’m talking about whether you’re irrational in the scientific sense. And the answer is, yes, you are irrational, and so am I. Today, we’re going to explore what that means and why it matters, and whether it’s worth striving to be more rational. Many years ago, I ran a survey about rationality to find out what people thought about it. It turns out
People are in favor of it. They think it should be taught to children. They thought their political opponents needed it desperately. But there is one group that people think doesn’t need much help with rationality, and that’s themselves. That being said, most people will acknowledge that they’re not perfectly rational all the time. On our website, personalitymap.io, where we map out over a million human correlations, we have data on the extent to which people agree with the statement, “I am a completely rational
Person.” And while the vast majority of people agree with this statement, only 12% agree to the maximum extent, acknowledging some level of non-rationality. But how do we decide what’s rational scientifically? Well, to do so, we need a gold standard of what rational behavior looks like. Once we have a model of that, we can compare real behavior against that standard to see how irrational humans are being. There are two broad domains of irrationality to look at, beliefs and
Behaviors. Let’s start with beliefs. To evaluate how irrational or rational our beliefs are, we can ask the question, “To what extent do our beliefs actually match the evidence that we have?” We define irrational beliefs as beliefs that don’t line up with the evidence. But can we actually say what a belief should be? Do we have a rational standard to compare human beliefs against? It turns out there is a rigorous way to look at this, which we
Can break down with a simple example. Suppose that I have two coins. There’s a real coin that has heads on one side, tails on the other, and when I flip it, there’s a 50% chance of landing on heads and a 50% chance of landing on tails. The second coin is a fake coin. It has heads on both sides. So, if I flip it, there’s a 100% chance that it lands on heads. Now, suppose that I put both
Coins in a bag and I draw one out completely at random. The first question to ask is, “What’s the chance that I’m holding the real coin versus the fake coin?” Well, clearly, because I drew one out at random, there’s a 50% chance I’m holding the real one and a 50% chance I’m holding the fake one. Any other answer in this scenario is irrational. It doesn’t match what we actually know. Now, of course, in a real-life scenario,
There could be other considerations. Maybe slight of hand was used to put the coins into the bag, or maybe the coins weren’t really drawn out with 50% chance each. But let’s keep the example simple. Let’s assume for now that we know all of the information and there’s no misinformation. Suppose now that I flip the coin and it lands on heads. How does that adjust your view on whether it’s the real or fake coin? Intuitively, it seems like this has to increase the
Probability that the coin is fake. If the coin had been real, there’s a chance it would have landed on tails, but it didn’t land on tails. It landed on heads, which is more consistent with the fake coin than the real coin. So, what’s the probability that the coin is actually heads? Intuitively, we should be able to agree that rational people will now all think that there’s more than a 50% chance that the coin is fake. But you might feel that they can still
Disagree with each other, that it might be reasonable for a rational person to think it’s a 70% chance of being fake or a 55% chance of being fake. All we can really say is that the probability has increased. But that’s actually not true. You can prove mathematically that there’s only one answer to this question. There’s an exact probability now that the coin is fake, and any other answer is wrong. That is, it doesn’t conform to rational thought. I won’t
Bore you with all the details of the calculation, but you can prove mathematically what the answer is. You can write out the probability of the coin being fake given that it landed on heads. You can use rules of probability to rewrite this equation in a different form, and then you can actually use that new form to do the calculation, and you find an exact answer. The probability that the coin is fake is 2/3. Every other answer is wrong. This is something
That a lot of people find surprising. They think, “Yeah, when you get evidence, you should change your mind.” But they think of that as a subjective judgment. The amount you change your mind, they don’t think of as an objective fact. But if you know all of the details, all the facts of the situation, the amount you should change your mind is actually a provable mathematical fact. If you’ve ever heard of Bayes’ rule, it’s that rule that
Tells us how much we should change our mind in the light of new evidence. Although Bayes’ rule is a precise mathematical statement, there’s an English language version of it that you can use, which I call the question of evidence. The question of evidence is, “How many times more likely am I to see this evidence if my hypothesis is true compared to if my hypothesis is false?” Let’s apply the question of evidence to the prior example. How many times more
Likely am I to see this evidence, that is, the coin landing on heads, if the hypothesis is true, that is, the coin is fake, compared to if the hypothesis is false, that is, the coin is a fair coin? Well, if my hypothesis is true, that the coin is fake, there’s a 100% chance that it lands on heads. So, that’s the first part of the equation. If my hypothesis is false, and it’s actually a fair coin, there’s a 50% chance it lands on heads.
So, how many times more likely am I to see this evidence if the coin is fake compared to real? Well, it’s simply 1 / 1/2, which is 2. That’s the strength of this evidence. We now know the strength of the evidence. The last thing we need to know is how to combine it with our prior knowledge. Remember, originally, there was a 50% chance of the coin being real versus fake. In other words, our prior was 1:1 odds. We now simply
Multiply those initial odds by the strength of the evidence, two, to get our final odds. So, two times one gives us two. So, now the odds that the coin is fake are 2:1, which is a 2/3 probability of being fake. Now, here’s the big takeaway. Suppose we have some belief about the world. We think something is true with a certain likelihood. Now, we get some new evidence against that thing. Well, how confident should we be? It turns out there’s only one
Mathematically correct answer to this, and that’s encapsulated in the question of evidence. This gives us an objective standard to evaluate human beliefs against to see whether they’re rational. We can see if they respond to evidence in the way that probability theory requires. So, how well do humans do by this objective standard? The answer is, not well. We see so many cases where our beliefs don’t conform to the actual evidence. Anchoring is an amazing and fascinating example of this. In one of
Their experiments, Kahneman and Tversky, who are two of the most famous rationality researchers, asked participants to spin a wheel of fortune to pick a random number that the participants knew was random. They were then asked, “Is the percentage of African nations in the UN above this random number or below it?” Finally, they were asked to estimate what they believe is the actual percentage of African nations in the UN. Remarkably, the random number from the roulette wheel influenced what their final
Estimate was. Those that had a higher random number tended to estimate a higher percentage of African nations in the UN, whereas if the initial random number was lower, they tended to make a lower estimate. Of course, the random number has no bearing whatsoever on the true answer. So, essentially, they were updating their view based on invalid evidence. What about cases where we overreact to valid evidence? That is, the evidence actually matters, but we adapt to it too much. Imagine there’s a
Rare disease that only one in 10,000 people have. And suppose that you have a 99% accurate test for this disease. That means that if the test is given to someone who does have the disease, there’s a 99% chance it will say they have it. And if the test is given to someone who doesn’t have the disease, there’s a 99% chance it will say they don’t have it. Now, suppose that in a completely random screening, someone is
Tested for this disease using this test. The test says that they do have it. So, what’s the probability that they actually have the disease? Take a moment to think about it. Many people will say that there’s a 99% chance they have it because the test is 99% accurate. But this is not accurate. It’s a massive overreaction to the evidence given. To see why, remember that we assumed that only one in 10,000 people have the disease. So, if we’re randomly screening
10,000 people, we only expect one to have it. When we apply that test that’s 99% accurate to the 10,000 people, it’s going to make a mistake about 1% of the time. That means if you consider all the people that tested positive, roughly one of them will have the disease and about 100 of them won’t. So, there’s only about a 1% chance that you actually have the disease if you test positive. A 1% chance is a massive difference from a
99% chance, which is what most people, even licensed physicians, think the answer is. Remarkably, when social scientist Gerd Gigerenzer gave a similar question to many doctors, only 20% of them got the answer right. And non-doctors are even worse. This is an illustration of what’s known as the base rate fallacy. That’s when we don’t properly take into account the prior information we had before getting this new evidence. In other words, we form our judgment on the new evidence, but we
Don’t incorporate what we knew before we got that evidence. The third issue with evidence is probably the most relatable. This is where we don’t change our mind as much as we should when we’re given valid evidence, or we dismiss that valid evidence outright and don’t change our mind at all. Now, while I’m sure that you always update on valid evidence when your political opponent points out why you’re wrong about your beliefs, you immediately change your mind, I’m sure
You’ve noticed that your political opponent doesn’t do the same. When you point out why they’re wrong using valid evidence, they don’t seem to change their mind one iota. We’ve talked about what it means to have irrational beliefs, but that’s just one of the two types of irrationality. The other is having irrational behaviors. But what does it mean to have irrational behaviors? It means that you’re not acting to produce the things that you fundamentally value as much as you could
Be. In other words, if you took a certain action, you’d produce more of what you value than if you took another action, and for some reason, you’re taking the other action instead of the first one. So, for example, suppose you value money, and that option A gives you 1. And further, suppose that in all other ways, the two options are identical. In that case, it would be irrational to take option B because it gets you less of
What you fundamentally want. Of course, in practice, humans value many things, not just money. So, acting rationally is, based on the evidence you have, choosing the actions that tend to get you more of what you value rather than less. We can think of three different types of ways that we can act irrationally. The first is pursuing things that we don’t actually value. The second is pursuing things that we value less over things we value more. And the
Third is when we value things in an inconsistent or self-contradictory way. Let’s start with the third type. In a famous experiment, Kahneman and Tversky, the same rationality researchers we talked about before, asked people what they would do if they were a policy maker when there was a disease outbreak. risks the lives of 600 people. You have to decide, as a policy maker, what to do. With option A, exactly 200 lives will be saved. With
Option B, there’s a 1/3 chance that all 600 people’s lives will be saved and a 2/3 chance that nobody will be saved. Take a moment to consider this. Which of the two options would you choose? Now, I want you to consider a different scenario. Again, 600 people’s lives are at risk from a disease outbreak. And again, you have two options. Option C, exactly 400 people will die. Option D is that there’s a 1/3 chance that nobody
Will die and a 2/3 chance that all 600 will die. Take a moment to consider this. Do you choose option C or option D? Well, this was all actually a trick question. The two scenarios are actually identical. The only difference is framing. In the first scenario, we framed it in terms of lives saved. In the second scenario, we framed it in terms of lives lost. But otherwise, the two scenarios were identical and in fact, the options were identical. Option
A is exactly the same as option C and option B is exactly the same as option D. In the first scenario, most people choose option A. They want the guaranteed lives saved. And yet, in the second scenario, most people don’t choose option C because they don’t want the guaranteed lives lost. But of course, those are identical options. 200 lives saved means 400 lives lost in the scenario. The only difference is whether we’re framing it in terms of lives saved
Or lives lost. People’s choice is inconsistent. Irrelevant details about how something is described can change our actual behavior. In other words, one form of irrationality is being inconsistent with ourself. A second type of irrationality is when we pursue things that are not valuable to us. This often happens from us picking up on what other people value and accidentally pursuing it instead of what we value. An example where this happened in my own life was one day where a friend came to
Me and said, “I’m really unhappy with my boyfriend, but I’m confused. I don’t understand why because he’s such a great guy.” I had her go through an exercise where she wrote down a list of the things that she fundamentally values. I then had her evaluate how her boyfriend does on those things that she values. He did not fare well on this list. I then asked her to list what her parents value and to evaluate him on that list. On
That list, he did really well. In other words, she was dating someone who matched what her parents value, not someone who matches what she values. And that’s why she was so unhappy. We see a ton of this happen on social media where people adopt from the people around them what they think they’re supposed to value and pursue things that are not actually valuable to them. The third type of irrational behavior is when we pursue something that we value less over
Something that we value more. A common example where we see this is with anxiety. People with high anxiety often avoid things that are really valuable because of the anxiety. For instance, someone with social anxiety may avoid going to a party where they really value the interactions merely because they’re trying to avoid the discomfort. Or for someone with public speaking anxiety, they might avoid public speaking even though doing so prevents them from advancing in their career. In those
Cases, they may very much know that they value their career more or value the social connection of the party more than they do avoiding things that make them anxious. And yet, when the anxiety strikes, they act based on what the anxiety tells them to do rather than based on what they fundamentally value. Another example that most of us can relate to is when we get stuck on social media scrolling and scrolling and not even enjoying ourself
Very much when we could do something that’s much more valuable to us like spending time with a loved one or working on a personal project. The issue here is that there are many things that can drive our behavior other than what we fundamentally value such as fear, addiction, or instant gratification. Taking a step back, we’ve seen now that our beliefs are often formed in irrational ways. In other words, we’re not reacting to evidence in the way that
Rationality requires. We’ve also seen that our behaviors can often be out of sync with what we fundamentally value. In other words, we’re not taking actions to produce as much value as we could according to what we care about. No human is perfectly rational. Rationality is an ideal that with work, we can get closer and closer to, but we’ll never actually get there. That being said, I think the endeavor is well worthwhile. It’s worth striving to be rational. By
Working to have more rational behaviors, we can produce more of what we fundamentally value. If you want to learn more about your own irrationality, take our quick how rational are you test. It’s a fun quiz that gets you to reflect on ways that you could be more rational. If you found this interesting and would like to see more things like it, I’d love it if you’d subscribe to our channel. We regularly put out videos about human psychology and even do a lot
Of our own studies that we report on.