Why are we bad at risk calculations?

‘Magical thinking’ can cause us to make mistakes

Humans have an odd relationship with risk. When threat is real and present, we respond adaptively and almost instantly — long before there’s time for any careful consideration. We have just one aim in mind: to save our own life and/or those of our loved ones.

However, when danger is only a possibility and we have time to consider the extent of the risk, we often make mistakes. Why?

First, the character of an event can skew our judgment. When something unusual and terrible happens, our immediate fright causes us to ignore logic altogether and react as if the event is highly likely to happen again, this time to us. That’s why, in the immediate aftermath of

9/11, many people cancelled flight reservations and took to the roads instead, apparently ignoring the fact that statistically, it’s much safer to fly than drive long distances.

Second, the location of the event influences our judgment. When we’re in unfamiliar places we tend to overestimate risk, whereas in familiar surroundings we underestimate it. That’s why most people overestimate the chance of being mugged when they’re in a foreign city, even if — statistically — their current location is safer than their hometown.

Third, David Ropeik, an international consultant on risk perception, has shown we overestimate risk when we don’t feel in control of what’s happening. For example, if you’re a passenger in a car, you’ll probably believe there’s a greater chance of a collision than if you’re the driver, even if you’re taking the same route and even if you know the person driving is competent.

Another reason we miscalculate risk is that we often engage in “magical thinking.” That is, we assume a causal connection between an action and a particular outcome, even though there’s no evidence to back that up. A good example is the belief that if we fasten our seat belt, we’re less likely to have a car crash. True, we may come to less harm if we’re involved in a crash, but the risk of collision doesn’t change just because we remember to buckle up.

Finally, we often allow anecdotal information and/ or personal biases to colour our perception of risk, as Lyle Brenner and colleagues at the University of Florida demonstrated. They told participants there was currently a 30 per cent chance of rain in both Phoenix (a desert town) and Seattle (a coastal town), then asked them to estimate how likely it was that they would need an umbrella if they visited either city that day. Participants were more likely to say they needed an umbrella in Seattle than Phoenix, even though they’d just been told the chance of rain was the same in both cities.

The lesson?

Whenever you estimate personal risk, start by asking: Is that information based on reliable scientific sources? If not, find out what you can from such sources while disregarding anecdotes and purely personal opinions. Only then make your estimation.