Nobel Prize-Winning Research on Risky Decision Making
For this, the ninth nomination in the Top Ten Psychology Studies, it's Nobel Prize-winning research. Daniel Kahneman and Amos Tversky were interested in the apparently strange way in which people make decisions in risky situations. One example is: would you bet £10 on the flip of a coin if you stood to win £20? So you've got a 50% chance of losing £10 and a 50% chance of winning £20. This seems like a good bet to take and yet studies show that people tend not to take it. Why?
Changes in wealth
Before Kahneman and Tversky (1979) published their ground-breaking research, risky decisions were usually analysed by thinking about the total wealth involved. When you look at this bet in the context of the total wealth it makes sense to gamble. It's obvious you've got more to gain than you have to lose. So, why do people tend not to?
"It is actually the changes in wealth on which people base their decision-making calculations."What Kahneman and Tversky suggested was that, in fact people think about small gambles like this in terms of losses, gains and neutral outcomes. It is actually the changes in wealth on which people base their decision-making calculations. But that doesn't completely explain why people don't take the bet. There's a further piece to the puzzle.
It turns out that at low levels of risk, such as this coin flip situation, people are more averse to the loss of £10 than they are attracted by the chance of winning the £20. Studies have shown that people actually need the chance of winning £30 before they'll consider risking their own £10.
Just as people show illogical risk aversion in some circumstances, they also show risk seeking behaviour in other circumstances.
Imagine you have to choose between these two options. The first is that you have an 85% chance of losing £1,000 along with a 15% chance of losing nothing. The second is a 100% chance of losing £800. Not much of a choice, right!? You're between a rock and hard place. Still, sometimes we have to cut our losses.
"When the potential for loss is there, suddenly people prefer to take a risk."According to the maths you should choose the sure loss of £800, but most people don't. Most people choose to gamble. So when the potential for loss is there, suddenly people prefer to take a risk. They've become risk seekers. Yet, when there's the potential for gains, people are often risk averse.
Framing bias
This way of thinking about how people behave in risky situations, which Kahneman and Tversky called Prospect Theory, has a second major insight that follows on from the risk aversion and risk seeking described above.
What they realised was that people behaved in different ways depending on how the risky situation was presented. Remember that if a risk is presented in terms of losses, people will be more risk seeking, and if it's expressed in terms of gains, people will be more risk averse.
Their classic example involves this fictional situation:
"Imagine your country is preparing for the outbreak of a disease expected to kill 600 people. If program A is adopted, exactly 200 people will be saved. If program B is adopted there is a 1/3 probability that 600 people will be saved and a 2/3 probability that no people will be saved."
Here, the risk is presented in terms of gains so people tend to choose option A (72%), which is, in fact, worse. Here's the same problem but this time presented in terms of losses:
"Imagine your country is preparing for the outbreak of a disease expected to kill 600 people. If program A is adopted, exactly 400 people will die. If program B is adopted there is a 1/3 probability that no one will die and a 2/3 probability that 600 people will die."
Now most people (78%) choose B because the problem is presented in terms of losses. People suddenly prefer to take a risk. In fact, if you look at both the situations you'll see that, mathematically, they're identical and yet people's decision is heavily influenced by the way the problem is framed. This effect has been termed preference reversal.
Now back to the real world
After considering these sorts of problems for a few minutes, it's easy to wonder what all of this abstract reasoning has to do with the real world. Quite a lot argue Kahneman and Tversky. The Nobel Prize committee agreed.
"Everyday life involves endless 'gambles'."Everyday life involves endless 'gambles' and betting examples are just one of the easiest ways to understand how humans make decisions in risky situations. Certainly Kahneman and Tversky's work has plenty to say about some of the apparently strange decisions people make in everyday life.
So, next time you're agonising over a decision in terms of losses, try this simple trick. Re-imagine the whole decision in terms of gains. I can't promise it will help you make your decision, but at least you'll better understand Kahneman and Tversky's insightful research. Humans are not as rational as we would like to think.
Now Vote!
All the nominations for the top ten studies in psychology are now in. It's time for you to vote for your favourite. Which one most captures your imagination? You can recap the runners and riders here, where you can also vote.
References
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decisions under risk. Econometrica, 47, 313-327.
Join 21938 readers
Cool. I learn something new every day. Thanks.
You're welcome!
Are you sure you have the right numbers in you fictional 'outbreak of a disease' scenario? As far as I can tell both choices are mathematically equivalent. Both programs seem to have an Expectation value of 200 persons saved.
I read this blog and never comment (even though I should, boourns me), but this entry was fascinating, great work =)
Anonymous, the numbers are definitely right. But I was implying the two programs are not equivalent in expectation value, which they are. I've edited that. Thanks for pointing this out.
Scarf, thank you!
The 2 options in the 'outbreak' scenario aren't identical at all. Option 1 gives you the certainty that 200 will be saved. Option 2 has two possible outcomes: 600 will be saved (a 1 Iin 3 chance) OR no one will be saved (a 2 in 3 chance).
Or didn't I get it?
Anonymous, the idea is that they're identical in terms of probabilities. If these scenarios were repeated over an over again, the average number of people dying would be the same whichever option you chose:
Imagine you repeat the scenerio three times. When you choose program A, 3 x 200 = 600 people die.
When you choose program B 3 times, the probability is that twice no one dies and one time 600 people die.
So the total number of people dying, on average, is the same - which means the 'expectation value' is the same.
'If these scenarios were repeated over and over again'...
But that is not what the premise implies: 'Imagine your country is preparing for the outbreak of a disease expected to kill 600 people.' That looks like an absolute 600 to me. No hint of repetition in that.
It should have read something like: '600 people of every (e.g.) 10.000...' and: '200 of every 10.000 will be saved".
Sorry to be such a bore, but my point is this: how can you draw conclusions from a test that can be (mis)interpreted in more ways than the 2 it intended to test?
No need to apologise! You're right - it can be interpreted in different ways. My previous comment was explaining how they are equivalent on probabalistic terms, but as you say, human interpretation is, as we now know, a different matter altogether. This study is designed to show two different types of these interpretations. Although, as you say, there are others.
On the first scenario there are actually three options.
1.) Have £10 (is this a newly gained £10? An existing £10 out of the person's own pocket? How does this influence the person's state of mind).
2.) 50% chance of losing £10
3). 50% chance of gaining £20
It would be interesting to look at the starting point - ie does it make a difference if they feel they have already "won" by being given £10 or is it money they've "hard earned", as this is a part of option 1).
I found this article really interesting and so I'm using Kahneman's and Tversky's experiment for a psychology project at school. However, I can't seem to find any sources that can lead me to their actual experiment (since I have to replicate it), so could you please help me out here? Thanks : )
A great article. I am a Risk management Coach and I am often asked to explain what is meant by Risk. I tend to use the example below to illustrate my point.
Risk is often seen as the likelihood of an event occurring. i.e. what is the mathematical probability of a particular outcome. However, we humans rarely make our decisions based on mathematics but n emotions. And emotions are subjective.
I placed a thick wooden beam 3 meters long (10 foot) and 30 cm square section between two dining chairs. On one end I put a £20 note. Everything was solid and the beam was inflexible. I asked for a volunteer to walk the beam and if they did they could keep the £20 note. Everyone volunteered. So rather than select I suggested we raise the height. When I reset the height to 2 meters 80% of the class dropped out. However, I still didn’t want to make a selection so I raised it to 6 meters. I was left with one volunteer. All the rest thought the risk was too great. Yet the risk of falling was exactly the same. It was the perceived severity of harm that a fall would cause that was the significant influential factor in the attendees acceptance of the risk. Risk is not as much about mathematics as it is about the individual’s perception between loss and reward. However, their position wasn’t fixed as when I increased the amount of money on offer the class again volunteered on mass. I even suggested that I put the beam between two buildings 200 meters above the ground and as long as the reward was significant I had volunteers.
Wilf Archer