Experience, bias and numerical illiteracy: humans are lousy at judging vaccine risks

Humans are terrible judges of risk. It’s why the gambling industry rakes in billions every year, why people are terrified of flying but not driving, and why another death possibly linked to the AstraZeneca vaccine may fuel vaccine hesitancy.

The 52-year-old woman died after developing a serious blood clot in her brain. Developing vaccine-associated blood clots is extremely rare — fatal cases are even rarer. Australia has administered 5.4 million COVID-19 vaccine doses, 3.6 million of which were AstraZeneca. There have been 48 confirmed and probable cases linked to the AstraZeneca vaccine and two deaths.

What we see v what we hear

UNSW school of psychology deputy head Professor Ben Newell tells Crikey much of how people perceive risk comes down to experience.

He conducted a study where people picked which simulated microworld they would choose to live in based on how prone the worlds were to natural disasters. The research found if people experienced the risk — that their simulated home experienced an earthquake — they were still likely to choose to live there because the risk was one they had experienced.

“People tend to overweight described risks and treat them as occurring more often than statistically,” he said.

“But if it’s an experienced risk, where you don’t tell a person how likely something is but over trials they learn there’s a 10% chance, then people tend to act as if they underweight prevalence — and see the risk as more like 5%.”

We saw this following the 9/11 terrorist attacks in the United States, where people were more likely to drive places than catch a flight — take an experienced, higher risk over a less experienced risk of flying.

But given Australians are vaccinated against more than a dozen diseases by the time they’re 16, wouldn’t that experience outweigh any vaccine hesitancy?

“I think the general idea is that [other vaccines] don’t translate into this situation because [this vaccine] is a new one,” Newel said.

At the mercy of biases

People’s perceptions are easily manipulated, Victoria University economist and climate change risk researcher Roger Jones tells Crikey.

“If we were purely rational, you would have a proportional reaction to risk based on probability,” he said. “But the problem is that if you’re unaware of something, the minute you take notice of it, it immediately becomes present.

“So even if something has a 0.0001% of a risk, you’re going to take more notice.”

Jones says there are many ways biases can be introduced into thinking — from the anchoring effect (wherein showing a higher number to a person before introducing data can lead people to perceive the data as lower), to the priming effect (being told something immediately before being told about the risk), to how familiar a person is with a certain risk.

“The fear of the unknown is a big factor, and if you don’t get immediate feedback on your decision it can prey on you,” he said.

Jones suspects the fact that blood clots can take some days to emerge is heightening people’s perception of risk due to the fear of the unknown.

Should we compare risks?

Research fellow at the Murdoch Children’s Research Institute’s vaccine uptake group Dr Jessica Kaufman tells Crikey there was some concern the news could fuel hesitancy, although officials had done well to highlight the rarity of the risk.

“It puts it more front of mind again for people as a potential risk, but doesn’t change the numbers that we’ve been sharing consistently throughout,” she said.

As more people get vaccinated and more data emerges, the rate of death or blood clots in Australia is decreasing, she says.

Even though risks are rare, comparing them with other risks — such as that of developing a blood clot from taking the contraceptive pill, or the risk of dying in a car crash — isn’t useful, she says.

“You don’t want to compare a completely arbitrary risk to something — like the risk of dying from smoking when people go, ‘Well that’s a choice that I’ve made.. ”

Part of this is due to a poor grasp of statistics, which Kaufman says can be overcome by visual depictions — showing one dot in 100,000 to represent a risk.

Perception of choice and control is also important: Kaufman says everyone should be able to choose which vaccine they get and when.