Why experts argue bias can be tackled — but rarely head-on

*Originally published on Feb. 3, 2022.

Bias is, by far, the number one complaint lodged at journalists reporting for the Canadian public broadcaster, says the CBC Ombudsman, Jack Nagler — and that's to be expected.

"Particularly now … people are frustrated when they see or hear stories that don't align with their views on the world or the frame through which they see things," Nagler told CBC IDEAS for its ongoing documentary series on handling biases.

"So they often accuse media organizations of bias, and that's hardly unique to CBC. That's commonplace everywhere." 

Nagler recommends that journalists admit that they have biases but endeavour not to let these "take them off the road to truth."

However, bias experts warn that the attempt to quieten one's own biases usually runs into logical and practical trouble.

"It's very difficult," said Olivier SIbony, author of You're About to Make a Terrible Mistake and most recently Noise, co-written with Daniel Kahneman and Cass Sunstein.

Sibony explains that it's rare in practice for any individual to have a true statistical base of data from which to determine that they are biased in a particular direction.

See Sibony in action discussing bias and the COVID pandemic:   

"If three of your friends tell you you're politically biased, you always see things in a particular way. Well, maybe it's actually these three friends of yours who have different political leanings," said Sibony.

The bias paradox

Science writer Jessica Nordell believes it's still worth fighting for "freedom from bias," despite the difficulty of tackling biases at an individual level.

She says it's often perceived as a task aimed at helping others, but her research for her new book, The End of Bias: A Beginning, left her feeling it needs to be recognized how deeply our biases damage our own experience of the world, by separating us from reality.

"I was interested in solving the problem, and not just sort of discussing the problem," said Nordell. "It seems like a paradox — how can we change something that we don't even know we're doing?"

Nordell worked with a computer scientist to create a computer simulation of a workplace, drawing from the empirical results of various 'bias studies' concerning how women fare in the workplace. Assuming just a 3 per cent average bias against the value of women employees' work, over a period of 10 years, the simulated company saw 87 per cent of its upper positions go to men.

It demonstrates that a member of a group may be almost egalitarian in assessing his or her colleagues, and yet, by holding a tiny but persistent bias that is widely shared, the resulting impact on other people's lives can be huge.

Bias anatomy

Olivier Sibony offers a taxonomy of cognitive biases, based on whether they belong to the family of "pattern-recognition biases" or, for instance, "action-oriented biases." Examples of the former include the well-known "confirmation bias" and the "attribution error."

The latter term refers to the common tendency of attributing a success or failure to the wrong cause, usually to the role of individuals at the expense of the role of chance or circumstance.

However, Sibony does not recommend using this taxonomy of 24 biases as a menu, from which to choose particular ones to combat in yourself.

"Yes, you probably have all those biases," said Sibony. "After the fact, you can explain any failure by cherry-picking from the very long list of biases one that explains the mistake you've made. What good does that do you at the time you need to make the decision? That's the problem."

For Nordell, the fight to overcome social biases nevertheless continues. She suggests it's fairly easy to spot the major sources of cultural and social bias in the country where you live — usually relating to gender, race, and age, plus some other common factors such as body weight and disability.

Nordell points to studies showing that populations can become less biased against particular groups by being exposed to more complicated ideas about those groups.

People often show a bias towards perceiving "outgroup homogeneity" — i.e. they instinctively imagine that the members of a rival or foreign group resemble each other. Challenging this habitual way of thinking may do much more to reduce harmful biases than a more straightforward campaign to boost the image of a particular group by showcasing its positive characteristics.

More generally, checking through a simple list of "biases to overcome" may not do much good for a person who is trying to become more fair-minded. A more promising, though indirect, method may be called for: getting to know people better. 
 

Guests in this episode:

Olivier Sibony is the author of Noise: A Flaw in Human Judgment and You're About to Make a Terrible Mistake. 

Jessica Nordell is the author of The End of Bias: A Beginning. 

Jimmy Calanchini is assistant professor of psychology at University of California, Riverside.  

Jack Nagler is the ombudsman at the Canadian Broadcasting Corporation.
 

* This episode was produced by Tom Howell.