Lord Neuberger "The big problem, as it is everywhere, is with - - PowerPoint PPT Presentation

Cognitive Bias

Tom Stafford, University of Sheffield, 28/11/17 t.stafford@shef.ac.uk @tomstafford

Lord Neuberger

"The big problem, as it is everywhere, is with unconscious bias. I dare say that we all suffer from a degree of unconscious bias, and it can

• ccur in all sorts of manifestations. It is almost

by definition an unknown unknown, and therefore extraordinarily difficult to get rid of,

• r even to allow for. "

Fairness in the courts: the best we can do: Address to the Criminal Justice Alliance. Lord Neuberger, 10 April 2015 https://www.supremecourt.uk/docs/speech-150410.pdf

What is bias? Bias, in the psychological sense

Wason’s selection task

These four cards all have: a letter on one side a number on the other side. Rule: ‘All cards with a Vowel on one side have an Even number on the other side.’ Which cards would you have to turn over to decide whether this statement is true or false?

E X 1 6

‘Mental Contamination’

The 450 foot Capilano Suspension Bridge Image: Maarten Van Horenbeeck

Error Failure of rule following E.g. confirmation bias, errors in reasoning about probability, errors due to choice structure ‘Cognitive bias’ Inconsistency Unwanted influences E.g. “implicit bias”, contamination, priming due to group membership, stereotypes, etc ‘Social biases’

Mechanisms of implicit bias

Osgood & Luria (1954)

Stereotype accessibility Automatic associations Priming

The ‘Implicit Association Test’ (IAT) The ‘Implicit Association Test’ (IAT) Self-generating dynamic of biases

Ie target of bias underperforms due to the biases and/or a biased environment

Hehman, E., Flake, J. K., & Calanchini, J. (2017). Disproportionate Use of Lethal Force in Policing Is Associated With Regional Racial Biases of Residents. Social Psychological and Personality Science. https://doi.org/10.1177/1948550617711229

Discussion: hiring criteria case study

What advice would you give to chairs of future hiring committees so they could ensure a gender impartial hiring process?

Uhlmann, E. L., & Cohen, G. L. (2005). Constructed criteria redefining merit to justify

• discrimination. Psychological Science, 16(6), 474-480.

An applicant for the police chief job is described as highly educated, but not streetwise. The gender of the applicant affects which characteristics people judge are important for the job (but not which they are perceived to have).

http://www.independent.co.uk/arts-entertainment/films/news/idris-elba-is-too-str eet-to-play-007-says-james-bond-author-10480532.html

Do you worry about social bias? How? How do you address it in people who come to tribunals?

Anti-bias strategies

Why bias mitigation hard (1)

• Easier to see

biases in other people than

• urselves

Cartoon Reproduced with kind permission of Grizelda and Private Eye Magazine

Why bias mitigation hard (2)

• Source(s) of bias hard to locate
• Extent of bias hard to assess

What doesn’t work on its own

• Ignoring social identity (e.g. “colour blindness”)

Trying to be objective / suppression Awareness raising

• Duguid, M. M., & Thomas-Hunt, M. C. (2015). Condoning stereotyping? How awareness of stereotyping prevalence

impacts expression of stereotypes. The Journal of applied psychology, 100(2), 343-359.

Focussing on individuals

• Kalev, A., Dobbin, F., & Kelly, E. (2006). Best practices or best guesses? Assessing the efficacy of corporate affirmative

action and diversity policies. American sociological review, 71(4), 589-617.

What can judges do about bias - a framework

Effect on bias - Does your strategy...

Mitigate bias Insulate against bias Remove bias

Who it affects - Is your strategy...

Personal Interpersonal Institutional

Insulate Remove Personal Interpersonal Institutional

A 3x3 model

Mitigate Insulate Remove Personal Interpersonal Institutional

A 3x3 model

Mitigate

Avoid risk factors (hunger, fatigue), articulate reasoning, ‘imagine the opposite Cognitive training (e..g relearning associations) Remove information that activates bias

Insulate Remove Personal Interpersonal Institutional

A 3x3 model

Mitigate

Avoid risk factors (hunger, fatigue), articulate reasoning, ‘imagine the opposite Cognitive training (e..g relearning associations) Remove information that activates bias Identifying others’ biases is easier; challenging conversations Exposure to diversity (“Contact hypothesis”) Sub-divide tasks to insulate; independence of procedures

Insulate Remove Personal Interpersonal Institutional

A 3x3 model

Mitigate

Avoid risk factors (hunger, fatigue), articulate reasoning, ‘imagine the opposite’ Cognitive training (e.g. relearning associations) Remove information that activates bias Identifying others’ biases is easier; challenging conversations Exposure to diversity (“Contact hypothesis”) Subdivide tasks to ensure independence

• f procedures; reveal

identifying information last Tracking outcomes; predeclared criteria; recording process of decisions; norms of fairness Avoiding biased

• utcomes (e.g.

quotas / shortlisting requirements) Procedures that remove bias activating information;

Exercise: put your anti-bias strategies into the 3x2 framework

Addressing bias is like healthy eating

Image: By Evan-Amos (Own work) [CC0], via Wikimedia Commons

Need for collective approaches Possible future actions

Review practice using 3x3 framework Take an implicit association test https://implicit.harvard.edu/implicit/ Get a resource pack by emailing t.stafford@shef.ac.uk Other?

Exercise: future actions END

t.stafford@shef.ac.uk @tomstafford http://www.tomstafford.staff.shef.ac.uk/

Reserve slides (for discussion etc)

t.stafford@shef.ac.uk @tomstafford http://www.tomstafford.staff.shef.ac.uk/

examples of biased experts

Goldin, C., & Rouse, C. (2000). Orchestrating Impartiality: The Impact of" Blind" Auditions on Female Musicians. The American Economic Review, 90(4), 715-741. Schwitzgebel, E., & Cushman, F. (2015). Philosophers’ biased judgments persist despite training, expertise and

• reflection. Cognition, 141, 127-137.

Hoffrage, U., & Gigerenzer, G. (1998). Using natural frequencies to improve diagnostic inferences. Academic Medicine, 73(5), 538-40.

A 50-year-old woman, no symptoms, participates in routine mammography

• screening. She tests positive, is alarmed, and wants to know from you whether

she has breast cancer for certain or what the chances are. Apart from the screening results, you know nothing else about this woman. How many women who test positive actually have breast cancer? What is the best answer?

nine in 10 eight in 10

• ne in 10
• ne in 100

The probability that a woman has breast cancer is 1% ("prevalence") If a woman has breast cancer, the probability that she tests positive is 90% ("sensitivity") If a woman does not have breast cancer, the probability that she nevertheless tests positive is 9% ("false alarm rate")

50% 21%

Danziger, S., Levav, J., & Avnaim-Pesso, L. (2011). Extraneous factors in judicial decisions. Proceedings of the National Academy

• f Sciences, 108(17), 6889-6892.
• "97 percent of judges (thirty-five out
• f thirty-six) believed that they were in

the top quartile in “avoid[ing] racial prejudice in decisionmaking”"

Rachlinski, J. J., Johnson, S. L., Wistrich, A. J., & Guthrie, C. (2009). Does unconscious racial bias affect trial judges?. notre dame law review, 84(3), 09-11.