Decision Making, Noticing, and Decision Making, Noticing, and the Use of Behavioral Insightsthe Use of Behavioral Insights

Max H. Bazerman Max H. Bazerman

Harvard UniversityHarvard University

Decision BiasesDecision Biases

Daniel Kahneman’s Thinking, Fast and Slow Daniel Kahneman’s Thinking, Fast and Slow (over a million copies sold in the UK alone)(over a million copies sold in the UK alone)

Asian Disease Problem Asian Disease Problem (from Tversky & Kahneman, 1981)(from Tversky & Kahneman, 1981)

PROBLEM 1 (N=152):PROBLEM 1 (N=152): Imagine that the US is preparing for the outbreak of an unusual Asian Imagine that the US is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact to combat the disease have been proposed. Assume that the exact scientific estimates of the consequences of the programs are as follows. scientific estimates of the consequences of the programs are as follows.

Program A: If Program A is adopted, 200 people will be saved [72%]. Program A: If Program A is adopted, 200 people will be saved [72%].

Program B: If Program B is adopted, there is 1/3 probability that 600 Program B: If Program B is adopted, there is 1/3 probability that 600 people will be saved, and 2/3 probability thatpeople will be saved, and 2/3 probability that no people will be no people will be saved [28%]. saved [28%].

Which of the two programs would you favor?Which of the two programs would you favor?

Asian Disease Problem Asian Disease Problem (from Tversky & Kahneman, 1981)(from Tversky & Kahneman, 1981)

PROBLEM 2 (N=155):PROBLEM 2 (N=155): Imagine that the US is preparing for the outbreak of an unusual Asian Imagine that the US is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact to combat the disease have been proposed. Assume that the exact scientific estimates of the consequences of the programs are as follows. scientific estimates of the consequences of the programs are as follows.

Program C: If Program C is adopted 400 people will die [22%]. Program C: If Program C is adopted 400 people will die [22%].

Program D: If Program D is adopted there is 1/3 probability that Program D: If Program D is adopted there is 1/3 probability that nobody will die, and 2/3 probability that 600 people will die nobody will die, and 2/3 probability that 600 people will die [78%]. [78%].

Which of the two programs would you favor?Which of the two programs would you favor?

Other BiasesOther Biases

FramingFramingAnchoringAnchoringEgocentrismEgocentrismDiscounting the futureDiscounting the futureVividnessVividnessIgnoring Base RatesIgnoring Base RatesMythical Fixed PieMythical Fixed PieEscalation of CommitmentEscalation of CommitmentArbitrary Mental AccountingArbitrary Mental AccountingIgnoring the Decision of OthersIgnoring the Decision of Others

The U.S. versus TobaccoThe U.S. versus Tobacco

1999 – The US DOJ files suit under RICO1999 – The US DOJ files suit under RICO

a)a) FraudFraudb)b) ConspiracyConspiracy

The FraudsThe Frauds1)1) Selling addictive substances while denying doing so.Selling addictive substances while denying doing so.2)2) Manipulating nicotine while denying doing so.Manipulating nicotine while denying doing so.3)3) Leading the public to make the false inference that Leading the public to make the false inference that

light/low tar products are less hazardous than full light/low tar products are less hazardous than full flavored brands.flavored brands.

4)4) Denying the effects of second hand smoke despite Denying the effects of second hand smoke despite having data to the contrary.having data to the contrary.

5)5) Denying the primary health effects of smoking Denying the primary health effects of smoking despite having internal studies that clarify these despite having internal studies that clarify these effects.effects.

6)6) Making false claims about conducting independent Making false claims about conducting independent research.research.

7)7) Suppressing evidence on health effects of smoking. Suppressing evidence on health effects of smoking.

RecommendationsRecommendations

My testimony concluded with the argument My testimony concluded with the argument that absent significant court intervention, that absent significant court intervention, misconduct would occur in the future. My misconduct would occur in the future. My recommendation to the court, based on the recommendation to the court, based on the assumption that D.O.J. had proven liability, assumption that D.O.J. had proven liability, was that the court should appoint monitors to was that the court should appoint monitors to the tobacco companies to consider appropriate the tobacco companies to consider appropriate structural changes, not limited to, but structural changes, not limited to, but including removal of senior management. including removal of senior management.

Timing of My WorkTiming of My Work

1) Hired: 3/10/051) Hired: 3/10/05

2) Filed expert report: 3/21/052) Filed expert report: 3/21/05

3) Deposed: 4/10/05 (7 hours, 10 attorneys on the other side)3) Deposed: 4/10/05 (7 hours, 10 attorneys on the other side)

4) Mid-April: Submitted written direct testimony 4) Mid-April: Submitted written direct testimony

5) 4/30/05: At the beginning of a prep session for 5/4/05 5) 4/30/05: At the beginning of a prep session for 5/4/05 court appearance, I was asked to amend my testimonycourt appearance, I was asked to amend my testimony

6) 5/4/05: Testified in court6) 5/4/05: Testified in court

7) 5/4-6/05: Prepared for Daubert challenge7) 5/4-6/05: Prepared for Daubert challenge

8) 5/6-13/05: Helped DOJ prepare for Fischel (my opposite) 8) 5/6-13/05: Helped DOJ prepare for Fischel (my opposite) testimonytestimony

Total: 214 hours Total: 214 hours

11

12

9/119/11

The U.S. government knew that terrorists were willing to become martyrs for their cause, and that their hatred toward the United States was increasing.

1993: Terrorists had previously bombed the World Trade Center.

1994: Terrorists hijacked an Air France airplane and made an aborted attempt to turn the airplane into a missile aimed at the Eiffel Tower.

1994: Terrorists had also failed in an attempt to simultaneously hijack twelve U.S. commercial airplanes in Asia.

Airline passengers knew how simple it was to board an airplane with items, such as small knives, that could be used as weapons.

EnronEnron

How could Arthur Andersen vouch for the financial health of Enron, concealing billions of dollars in debt from its shareholders?

At the heart of the Enron debacle is a conflict of interest that experts have been warning about for years. In 2000, multiple scholars testified before the SEC that the rapid growth of the Big 5 accounting firms’ consulting divisions have made impartial financial audits almost impossible.

Psychologists have known for a long time that individuals with a vested self-interest, even honest ones, are incapable of unbiased (independent) judgment.

a) Auditors have made tremendous profit from selling consulting

services to their audit clients.

b) Auditors want to be rehired.

c) The personnel on the audit often take jobs with the client firm.

June 17, 2005June 17, 2005

1) Waking up in London at 5 AM and reading about 1) Waking up in London at 5 AM and reading about Matthew Myers in the New York Times.Matthew Myers in the New York Times.

2) I realized that I should have acted six weeks 2) I realized that I should have acted six weeks earlier, and it was time to act now.earlier, and it was time to act now.

3) My story was covered by the Washington Post, and 3) My story was covered by the Washington Post, and led to congressional letters demanding an led to congressional letters demanding an investigation.investigation.

4) But, I recognized my earlier failure, and 4) But, I recognized my earlier failure, and appreciated what Matthew Myers had done.appreciated what Matthew Myers had done.

What Can You Do More Broadly?What Can You Do More Broadly?1) Put it on your agenda to be a first-class noticer (Warren 1) Put it on your agenda to be a first-class noticer (Warren Bennis). Focusing is good, but sometimes you need to take a Bennis). Focusing is good, but sometimes you need to take a look around.look around.

2) Take an outsider’s view.2) Take an outsider’s view.

3) Audit your organization. Identify changes that will help to 3) Audit your organization. Identify changes that will help to create a noticing organization.create a noticing organization.

4) When something is wrong, figure out what it is (tobacco 4) When something is wrong, figure out what it is (tobacco story).story).

The Role of NoticingThe Role of Noticing

What problems exist that do not make What problems exist that do not make sense?sense?

What are the barriers to fixing these What are the barriers to fixing these problems?problems?

Are there viable solutions?Are there viable solutions?

Using Behavioral Insights to Using Behavioral Insights to Improve GovernmentImprove Government

Thaler and Sunstein’s NudgeThaler and Sunstein’s NudgeThe Behavioral Insights Team – led by David The Behavioral Insights Team – led by David HalpernHalpernThe UK’s What Works ProjectThe UK’s What Works ProjectNew Demonstrations throughout the WorldNew Demonstrations throughout the World

2003 Organ Donation Rates2003 Organ Donation Rates(Johnson and Goldstein, 2003)(Johnson and Goldstein, 2003)

Gender Equality Nudge: Gender Equality Nudge: Comparative EvaluationComparative Evaluation

“Today, 94 of the

FTSE 100 companies

count women on

their boards as do

over two thirds of

all FTSE 350

companies.”

Two Types of Promises Two Types of Promises (Shu, Mazur, Gino, Ariely, and Bazerman)(Shu, Mazur, Gino, Ariely, and Bazerman)

Math PuzzlesMath Puzzles

Design

Two Types of Promises Two Types of Promises (Shu, Mazur, Gino, Ariely, and Bazerman)(Shu, Mazur, Gino, Ariely, and Bazerman)

1) It works in the lab – with replication2) It works on reporting mileage to an insurance company3) We believe it will work on collecting taxes4) Two large trials ongoing

Research Going on Harvard’s Research Going on Harvard’s Behavioral Insights GroupBehavioral Insights Group

1)1) Employees at a business-process outsourcing company based in Bangalore, when Employees at a business-process outsourcing company based in Bangalore, when asked to reflect for the last 15 minutes of each day on the lessons they had learned asked to reflect for the last 15 minutes of each day on the lessons they had learned that day, increased performance by 22.8 percent (Francesca Gino, HBS)that day, increased performance by 22.8 percent (Francesca Gino, HBS)

2)2) Sending a letter to eight million US households comparing household energy use to Sending a letter to eight million US households comparing household energy use to that of their neighbors can reduce energy consumption equivalent to that used by that of their neighbors can reduce energy consumption equivalent to that used by St. Louis and Portland (Todd Rogers, HKS)St. Louis and Portland (Todd Rogers, HKS)

3)3) Opt-out defaults greatly increase savings rates at every stage of the savings lifecycle, Opt-out defaults greatly increase savings rates at every stage of the savings lifecycle, including savings plan participation, contributions, asset allocation, and rollovers including savings plan participation, contributions, asset allocation, and rollovers (Brigitte Madrian, HKS, David Laibson, FAS, and John Beshears, HBS)(Brigitte Madrian, HKS, David Laibson, FAS, and John Beshears, HBS)

4)4) Plan-making and stressing high turnout increases the likelihood that an individual Plan-making and stressing high turnout increases the likelihood that an individual will vote (Todd Rogers, HKS)will vote (Todd Rogers, HKS)

5)5) When tax preparers at H&R Block helped some low- and moderate income clients When tax preparers at H&R Block helped some low- and moderate income clients fill out the FAFSA using existing financial data from the IRS, there was a significant fill out the FAFSA using existing financial data from the IRS, there was a significant boost in the rate at which participants in the program applied for financial aid, the boost in the rate at which participants in the program applied for financial aid, the number who received aid and the amount they got, and the proportion of number who received aid and the amount they got, and the proportion of participants who enrolled in college (Bridget Terry Long, HGSE)participants who enrolled in college (Bridget Terry Long, HGSE)