Making Good Decisions

1
Making Good Decisions

• Douglas M. Stewart, Ph.D.
• Anderson Schools of Management
• University of New Mexico

2
Lecture Overview

• Common Decision Making Biases
• Dealing with Uncertainty
• Motivation
• Escalation of Commitment
• Note This lecture and many of the examples are
  drawn from Bazerman, Max, Judgment in Managerial
  Decision Making, 5th Ed., John Wiley Sons New
  York, 2002, and Hammond, Keeney and Raiffa, Smart
  Choices A Practical Guide to Making Better Life
  Decisions, Broadway books, New York, 1999.

3
Human Beings are Quasi-Rational

• Bounded rationality
• Bounded willpower
• We place greater weight on present concerns than
  on future concerns
• Bounded self-interest
• We care about the outcomes of others
• Sowe use heuristics.

4
Common Biases

• Availability
• Representativeness
• Anchoring and Adjustment
• Confirmation Bias
• Hindsight and the Curse of Knowledge

5
Availability Example 1

• Which of the following lists was the cause of
  more premature deaths in the United States in
  1999?
• (a) Tobacco use, obesity/inactivity, and alcohol
• (b) Cancer, heart disease, and auto accidents

6
Availability What this Means

• List (a) are the 3 leading causes. Auto
  accidents are 8th.
• Vividness and recency (ease of recall of
  instances of an event) tend to make us
  overestimate its likelihood.

7
Availability Example 2

• In four pages of a novel (about 2,000) words),
  how many words would you expect to find that have
  the form _ _ _ _ _ n _ (seven-letter words that
  have n in the 6th position)?
• 0
• 1-2
• 3-4
• 5-7
• 8-7
• 11-15
• 16

8
Availability Example 2 (cont.)

• In four pages of a novel (about 2,000) words),
  how many words would you expect to find that have
  the form _ _ _ _ i n g (seven-letter words that
  end with ing)?
• 0
• 1-2
• 3-4
• 5-7
• 8-7
• 11-15
• 16

9
Availability (2) What this Means

• People tend to have higher estimates for _ _ _
  _ i n gthan for _ _ _ _ _ n _
• Memory structures make some information easier to
  recall.
• Why are multiple gas stations at the same
  intersections?
• Why do upscale retailers want to be in the same
  mall?
• Why are most of the restaurants in a city often
  located within a few blocks of each other?

10
Availability Example 3

• Is marijuana use related to delinquency?
• Are couples who get married under age 25 more
  likely to have bigger families?

11
Availability (3) What this Means

• People typically try to recall instances of
  delinquent marijuana users (or couples who
  married young and had large families), and if you
  know a lot such people you assume that the
  relationship is stronger than it may actually be.
• Availability of perceived co-occurring instances
  in our minds leads to higher probability of
  future co-occurrence.
• You actually need to consider 4 groups in
  assessing any two dichotomous events
• Marijuana users who are delinquents
• Marijuana users who are not delinquents
• Delinquents who do not use marijuana
• Non-delinquents who do not use marijuana

12
Representativeness Example 1

• Mark is finishing his MBA at a prestigious
  university. He is very interested in the arts
  and at one time considered a career as a
  musician. Is he more likely to take a job
• (a) in arts management
• (b) with a technology company

13
Representativeness (1) What this Means

• Did you ask yourself How likely is it that a
  person working in the management of the arts
  would fit Marks description?
• You should have asked How likely is it that
  someone fitting Marks description will choose
  arts management?
• But what about the base rate?
• People use base-rate data correctly when no other
  information is provided.
• Life savings and the base-rate of business
  failures.
• Pre-nuptials and the divorce base-rate
• Why are do my neighbors with the smaller houses
  drive all of the nice cars?

14
Representativeness Example 2

• A certain town is served by two hospitals. In
  the larger hospital about 45 babies are born each
  day and in the smaller hospital about 15 babies
  are born each day. As you know, about 50 percent
  of all babies are boys. However, the exact
  percentage varies from day to day. Sometimes it
  may be higher than 50 percent, sometimes lower.
  For a period of 1 year, each hospital recorded
  the days in which more than 60 percent of the
  babies born were boys. Which hospital do you
  think recorded more such days?
• (a) The larger hospital
• (b) The smaller hospital
• (c) About the same (that is within 5 of each
  other)

15
Representativeness (2) What this Means

• Most people choose C.
• Think about it
• What is more likely getting 6 heads on 10 flips
  of a coin, or
• Getting 6,000 heads on 10,000 flips of a coin?
• Sample size is rarely part of our intuition.
  People think how representative it would be for
  60 of babies to be born boys in a random event.
• Consider in advertising Four out of five
  dentists surveyed recommend sugarless gum for
  their patients who chew gum.

16
Representativeness Example 3

• You have started buying stocks on the internet,
  beginning with five different stocks. Each stock
  goes down soon after your purchase. As you
  prepare to make a sixth purchase, you reason that
  it should be more successful since the last five
  were lemons. After all, the odds favor making
  at least one successful pick in the six
  decisions. The thinking is
• (a) Correct
• (b) Incorrect

17
Representativeness (3) What this Means

• The sixth stock is independent of the performance
  of the first five.
• The Gamblers Falicy - We have an inappropriate
  tendency to assume that random and non-random
  events will balance out.
• After holding bad cards on ten hands of poker,
  the player believes he is due for a good hand.
• After winning 1,000 on the New Mexico Lottery, a
  woman changes her regular number (after all, how
  likely is it that that number will come up
  twice.)
• Play a slot machine that someone has put a lot of
  money into without winning any big payouts

18
Representativeness Example 4

• You are a sales forecaster for a department store
  chain with nine locations.
• The chain depends on you for quality projections
  of future sales in order to make decisions on
  staffing, advertising, information system
  developments, purchasing, renovations, and the
  like.
• All stores are similar in size and merchandise
  selection. The main difference in their sales
  occur because of location and random
  fluctuations.
• Sale for 2004 are show on the next slide.

19
Example 4 Continued

• Your economic forecasting service has convinced
  you that the best estimate of total sales
  increase between 2004 and 2005 is 10 (to
  99,000).
• Your task is to predict 2005 sales for each
  store. Since your manager believes strongly in
  the economic forecasting service, it is
  imperative that your total sales equal 99,000.

20
Example 4 Continued
21
Representativeness (4) What this Means

• Most people increase the sales at each store by
  10.
• Are 2004 sales perfectly correlated with 2005
  sales? No.
• People tend to neglect regression to the mean.
• Forecast should be somewhere between a perfect
  correlation forecast and a un-correlated forecast
  of 11,000 for each store.

22
Representativeness Example 5

• Linda is 31 years old, single, outspoken, and
  very bright. She majored in philosophy. As a
  student, she was deeply concerned with issues of
  discrimination and social justice, and she
  participated in antinuclear demonstrations.
• Rank order the following eight descriptions in
  terms of the probability (likelihood) that they
  describe Linda

23
Example 5 continued

• (a) Linda is a teacher in an elementary school
• (b) Linda works in a bookstore and takes yoga
  classes
• (c) Linda is active in the feminist movement.
• (d) Linda is a psychiatric social worker.
• (e) Linda is a member of the League of Women
  Voters.
• (f) Linda is a bank teller
• (g) Linda is an insurance salesperson.
• (h) Linda is a bank teller who is active in the
  feminist movement

24
Representativeness (5) What this means

• Look at where you ranked these issues
• (c) Linda is active in the feminist movement.
• (f) Linda is a bank teller
• (h) Linda is a bank teller who is active in the
  feminist movement
• Most people rank h f c (which is strong on
  representativeness.
• This is a conjunction fallacy.
• So why are we willing to buy insurance for death
  by airline crash, at a higher cost than term life?

25
Anchoring Example 1

• A new internet company recently made its initial
  public offering, becoming publicly traded. At
  its opening, the stock sold for 20/share. The
  companys closest competitor went public one year
  ago, also at a price of 20/share. The
  competitors stock is now priced at 100/share.
• What will the new firm be worth one year from now?

26
Anchoring (1) What this means

• How would you have responded if the other firm
  was now worth only 10/share?
• We tend to develop estimates by starting from an
  initial anchor, and adjusting from there to yield
  a final answer. This adjustment is usually
  insufficient.
• Different starting points yield different
  answers, even if the anchor is ignored as being
  rediculous.
• A child is tracked by a school system that may
  categorize them at a certain level of performance
  at an early age.
• Employers often ask applicants to reveal their
  current salaries. Why?
• What are the implications for negotiation?

27
Anchoring Example 2

• Which of the following appears most likely?
  Which appears second most likely?
• (a) Drawing a red marble from a bag containing 50
  percent red and 50 white.
• (b) Drawing a red marble seven times in
  succession (with replacement) from a bag
  containing 90 red and 10 white.
• (c) Drawing at least one red marble in seven
  tries (with replacement) from a bag containing
  10 red and 90 white.

28
Anchoring (2) What this means

• The most common ordering is ba-c
• The correct order is c (52), a (50), b (48).
• We tend to overestimate the probability of
  conjunctive (AND) events, and underestimate the
  probability of disjunctive (OR) events.
• This is a real problem with multi-stage planning.
• Home remodeling always takes longer than
  expected.
• After 3 years, Doctoral students dramatically
  overestimate their ability to complete their
  dissertation in 1 year.

29
Anchoring Example 3

• Listed on the next slide are 10 uncertain
  quantities. Write down your best estimate of the
  quantity. Next put a lower and upper bound
  around your estimate, so that you are 98
  confident that your range surrounds the actual
  quantity.

30
Example 3 continued
31
Anchoring (3) What this means

• Answers
• Wal-marts revenue 405.61 Billion
• Targets revenue 64.95 Billion
• Fords Market Value 8.27 Billion
• MSFT Market Value 166.88 Billion
• CSCO Market Value 97.3 Billion
• Households 77,873,000
• Extinct mammals 22
• Did you have more than 6 of 7 right?
• People tend to be overconfident in the
  correctness of their answers. Moreover, the
  overconfidence is greater when asked about areas
  we know less.

32
General Bias Example 1

• The following sequence of numbers follows a rule.
  What is the rule?
• 2 4 6
• You may present other sequences of numbers and I
  will tell you if they follow the rule or not.

33
General Bias (1) What this means

• Did you try submitting numbers that do not
  conform to your rule?
• Answering the questions requires that you attempt
  to falsify hypotheses not confirm them.
• This is the confirmation trap.

34
General Bias Example 2

• You are traveling by car in an unfamiliar area,
  and your spouse is driving. When you approach an
  unmarked fork in the road, your spouse decides to
  go to the right. Four miles and fifteen minutes
  later, it is clear that you are lost. Your blurt
  out. I knew you should have turned left at the
  fork.
• A manager who works for you hired a new
  supervisor last year. You were well aware of the
  choices he had at the time and allowed him to
  choose the new employee on his own. You have
  just received production data on every
  supervisor. The data on the new supervisor are
  terrible. You call in the manager and claim,
  There was plenty of evidence that he was the
  wrong man for the job.

35
General Bais (2) What this means

• Sound familiar?
• This is hindsight bias.
• It can keep us from learning from the past and
  evaluating decisions objectively.

36
Dealing with Uncertainty

• Framing of choices
• Anchors matter
• Pseudocertainty
• Insurance premium vs. Guaranteed loss
• Transactional utility

37
Framing of choices Example

• You face the following concurrent sets of
  choices
• 1. Choose between
• (a) A sure gain of 240
• (b) A 25 chance to gain 1,000 and a 75 chance
  to gain nothing.
• 2. Choose between
• (c) A sure loss of 760
• (d) A 75 chance to loose 1,000 and a 25
  chance to lose nothing.

38
Example Continued

• Now choose between
• (e) A 25 chance to win 240, and a 75 chance to
  lose 760.
• (f) A 25 chance to win 250 and a 75 chance to
  lose 750

39
Framing of Choices What this means

• Did you chose (a) and (d)? Why not?
• (a)(d)(f) and (b)(c)(e).
• Framing the same problem differently can reverse
  the choices.

40
Anchors matter Example

• What makes you feel worse
• (a) You receive a letter from the IRS saying that
  you made a minor arithmetic mistake in your tax
  return and must send them 100. You receive a
  similar letter the same day from the New Mexico
  Treasurer saying you owe 50 for a similar
  mistake. There are no other repercussions from
  either mistake.
• (b) You receive a letter from the IRS saying that
  you made a minor arithmetic mistake in your tax
  return and must send them 150. There are no
  other repercussions from the mistake.

41
Anchors matter What this means

• Most people find the two small losses to be worse
  than the single larger loss.
• The extra 50 loss is less bothersome from the
  point on the utility curve where you have lost
  100 than from where you have lost nothing.

42
Pseudocertainty - Example

• Which of the following do you prefer
• (a) A sure win of 30
• (b) An 80 chance to win 45

43
Example continued

• Consider the following 2-stage game. In the
  first stage there is a 75 chance to end the game
  without winning anything, and a 25 chance of
  moving into the second stage. You must choose
  (a) or (b) prior to the start of the game.
• (a) A sure win of 30
• (b) An 80 chance to win 45

44
Example continued

• Which of the following do you prefer
• (a) A 40 to win 30
• (b) An 20 chance to win 45

45
Psuedocertainty What this means

• Did you choose the same option in all three
  choices (first or second)?
• If not, then you have reacted to certainty and
  psuedocertainty effects.
• The reduction in the probability of an outcome is
  more important when the initial state is certain
  rather than merely probable
• The change in game 2 is due only to the
  introduction of uncertainty of reaching the
  second state (psuedocertainty.)

46
Insurance premium vs. Guaranteed loss

• People are more willing to accept a risk if
  certain cost is phrased as an insurance premium
  rather than a guaranteed loss.

47
Transactional utility Example

• Imagine you are about to purchase a cable modem
  for 50. The salesperson informs you that same
  product is available on sale at the stores other
  branch, located 20 minutes away. What is the
  highest sales prices that the product could have
  for which you would drive to the other store to
  make the purchase.

48
Example continued

• Imagine you are about to purchase a television
  for 500. The salesperson informs you that same
  product is available on sale at the stores other
  branch, located 20 minutes away. What is the
  highest sales prices that the product could have
  for which you would drive to the other store to
  make the purchase.

49
Transaction Utility

• Did you demand a greater discount in absolute
  dollars for the television? Why?

50
Dealing with Uncertainty - Summary

• Prospect Theory
• People evaluate rewards and losses relative to a
  neutral reference point.
• People think about potential outcomes as gains or
  losses relative to this fixed, neutral reference
  point.
• People form their choices based on the resulting
  change in asset position as assessed by an
  S-shaped value function

51
Motivation

• Egocentrism Perceptions and expectations are
  biased in a self-serving manner.
• If someone sues you and you win the case, should
  he pay your legal costs?
• (85 say yes)
• If you sue someone and lose the case, should you
  pay his costs?
• (44 say yes)
• Regret Avoidance We are biased towards
  decisions that avoid the opportunity for regret.
• We will choose options that shield us from
  feedback of foregone alternatives
• We will make choices that are likely to compare
  favorably to foregone alternatives

52
Escalation of Commitment Example 1

• You accept a position with a prestigious
  consulting firm, believing that the job offers an
  excellent career opportunity in a firm that you
  can grow with.
• Two years later, you have not progressed as
  rapidly as you had expected. Anxious to
  demonstrate your worth the company you decide to
  invest large amounts of unpaid overtime to get
  ahead. Still you fail to get the recognition you
  think you deserve.
• By now you have been with the organization for
  several years and would lose numerous benefits,
  including vested interest in the companys
  pension plan, if you decide to leave. You are in
  your late 30s and feel that you have invested
  your best years with the company.
• Do you quit?

53
Escalation of Commitment (1) What this means

• Do you escalate?
• It appears that the mechanism underlying
  escalation is self-justification.
• Conditions that lead to escalation
• Failure can be explained away with causes
  unrelated to initial decision.
• Individuals are more likely to escalate than
  groups.

54
Auction

• Auctioning a 20 bill.
• Rules
• Minimum 2 independent bidders
• Highest offer gets item at strike price
• Second highest offer must also pay their last bid
  price

55
Discussion

• Do you see implications
• Negotiations
• Investments
• Professional Services

56
Bias Strategies

• Availability
• Examine your assumptions so that you are not
  being swayed by memorable distortions.
• Try to get statistics. Dont rely on your memory
  if you dont have to.
• If you dont have direct statistics, try to build
  up an estimate with indirect statistics and other
  data.
• Representativeness
• Dont ignore relevant data, consider base rates
  explicitly.
• Dont confuse one type of probability with
  another. (Probability that an arts manager would
  be like Mark with the probability that someone
  like Mark would be an arts manager.)
• Anchoring
• Always view the problem from different
  perspectives, using alternate starting points and
  approaches. Then reconcile the differences.
• Think about the problem on your own first (to
  avoid being anchored by their perceptions.)
• Seek opinions from a wide variety of people to
  push your mind in new directions. (Be careful not
  to share your own estimates lest you anchor them.)

57
Bias Strategies (2)

• Confirmation trap
• Get someone you respect to play devils advocate
  to argue against the decision you are making (or
  build these arguments yourself)
• Be honest about your motives. Are you really
  gathering evidence to inform your choice or
  justify it.
• Expose yourself to conflicting information.
  Dont go soft on the disconfirming evidence.
• When seeking advice from others, dont ask
  leading questions.
• Overconfidence
• Try to challenge your own extreme figures. Try
  hard to imagine how they might be exceeded.
• Challenge an experts estimates in a similar
  manner.
• Substitute facts for opinion whenever possible.

58
Bias Strategies (3)

• Escalation of Commitment
• Never think of the status quo as your only
  alternative. Identify other options to use for
  comparison.
• Ask yourself if you would choose the status quo
  if it werent already the status quo.
• Avoid exaggerating the cost of switching from the
  status quo.
• Be sure to compare how the status quo WILL BE not
  just how it is now. Things change.
• If several alternatives are clearly better than
  the status quo, dont default to the status quo
  simply because you cannot choose among them.
• Seek out and listen to the arguments of people
  who werent involved in the earlier decision that
  got you here.
• Examine why admitting to an earlier mistake
  distresses you.
• Remember that sunk costs are sunk.

59
Thar be process failures ahead

• Most common serious process errors in decision
  making
• Working on the wrong problem.
• Failing to identify your key objectives.
• Failing to develop a range of good alternatives.
• Overlooking crucial consequences of the
  alternatives.
• Giving inadequate thought to tradeoffs.
• Disregarding uncertainty.
• Failing to account for risk tolerance.
• Failing to plan ahead when decisions are linked
  over time.

60
The Decision Making Process

• Define the Problem
• What triggered this decision? Why am I
  considering it?
• Question the constraints assumed in your problem
  statement.
• Identify the essential elements of the problem.
• What other decisions hinge on this decision?
• Should related decisions be included? Should the
  problem be narrowed?
• Ask others how they see the problem.
• Take your time here. A good solution to a well
  defined problem is better than an excellent
  solution to a poorly defined one.

61
Decision Making Process (2)

• Identify your objectives
• Make sure you include the intangible and
  subjective.
• Write down all of the concerns you hope to
  address through your decision.
• Convert these concerns to succinct objectives
  (verb and object).
• Separate the means from the ends. (the 5-Whys.)
  The means can stimulate alternatives. The ends
  are used to evaluate alternatives.
• Clarify what you mean by each objective.
• Test the objectives. Do they seem to lead to
  decisions you can live with? Could you use them
  to justify decision to others? Or have you
  overlooked something important.
• Practical advice
• Objectives are personal.
• Different objectives will suit different problems
  (dont recycle.)
• Butwell thought out objectives for similar
  problems should remain relatively stable.
• Objectives should not be limited by what data can
  be obtained easily.
• If a decision sits uncomfortably in your mind,
  you may have overlooked an important objective.

62
Decision Making Process (3)

• Alternatives
• Dont box yourself in with limited alternatives.
• Default option
• First solution
• Alternatives presented by others
• Use your objectives to ask How?
• Challenge constraints (assumed vs. real)
• Set high asperations
• Do your own thinking first
• Learn from history, but dont be constrained by
  it.
• Ask others for suggestions
• Give your subconscious time to operate
• Generate first, evaluate later
• Four categories of alternatives
• Process alternatives
• Win-win alternatives
• Information gathering alternatives
• Time-buying alternatives

63
Decision Making Process (4)

• Describe the consequences
• Mentally put yourself into the future (when the
  consequences will be revealed)
• Create a description of the consequences of each
  alternative. Compare against objectives. Do
  descriptions cover all objectives? Have
  objectives been overlooked?
• Eliminate any clearly inferior objectives.
• Organize descriptions of the remaining into a
  consequences table.
• Goal
• Enough precision to make a smart choice, but not
  exhausting detail.

64
Decision Making Process (5)

• Tradeoffs
• Find and eliminate dominated and effectively
  dominated alternatives.
• Make tradeoffs using even swaps.
• Make the easiest swaps first.
• Concentrate on the amount of the swap, not the
  perceived importance of the objective.
• Value an incremental change based on what you
  start with.
• Make consistent swaps (if AgtB and BgtC then be
  sure AgtC.)
• Seek out information to make informed swaps.

65
Decision Making Process (6)

• When faced with linked choices and uncertainty
• Understand your risk tolerance
• Estimate the probabilities
• Build decision trees.

66
Summary

• We are all quasi-rational
• We have in place strong biases in our
  decision-making that can cause errors
• Awareness, quantitative decision aids, and
  counter-strategies can help minimize these errors