Docstoc

Decision Making

Document Sample
Decision Making Powered By Docstoc
					Decision Making


Chapter 10
Decision Making

   Phases of Decision Making ►
   Basic Concepts of Probability ►
   Cognitive Illusions in Decision
    Making ►
   Utility Models of Decision Making ▶
   Descriptive Models of Decision
    Making ▶
       Phases of Decision Making

          Setting Goals ►
          Gathering Information ►
          Structuring the Decision ►
          Making a Final Choice ►
          Evaluating ►




Back              A Schematic View of these Phases
Phases of Decision Making




                            Back
Setting Goals

   The goals for a decision
       What are you going to accomplish?


   The goals influence decision making
    in various ways




                                      Back
Gathering Information

   Consider the decision of
       Buying a new computer
       Buying an apartment
       Choosing a college major




                                   Back
Structuring the Decision

   Decision structuring
       To manage various information
         When there are a great number of options
         When there are lots of considerations to
          be used in making the decision




                                          Back
Making a Final Choice

   Decide when to stop gathering
    information

   Decide which information is more
    relevant or reliable




                                Back
Evaluating

   To evaluate the whole process
          What went well?
          What didn’t go so well?

       To reflect and identify the areas of the
        process that can stand improvement
        and those that ought to be used again
        in future, similar decision




                                           Back
Basic Concepts of Probability

   The condition of uncertainty
       Probability
          A measurement of a degree of uncertainty
          Subjective probabilities
                 Probabilities are influenced by the estimators
                     Happy or sad; optimistic or pessimistic

            Objective probabilities
                 Not influenced by the estimators




                                                     Continue
Basic Concepts of Probability
   A 30-year-old woman discovers a lump in
    her breast and goes to her physician. The
    physician knows that only about 5 in 100
    women of the patient’s age and health
    have breast cancer. A mammogram
    (breast X-ray) is taken. It indicates
    cancer 80% of the time in women who
    have breast cancer but falsely indicates
    cancer in healthy patients 20% of the
    time. The mammogram comes out
    positive. What is the probability that the
    patient has cancer?
   5*80%÷[(100-5)*20%]=4/23=0.17

                                       Back
Cognitive Illusions in Decision Making

   How people gather and access the
    relevance of different pieces of
    information
   Cognitive illusions
       Certain systematic and common biases
        under most conditions but can lead to
        error when misapplied




                                   Continue
Cognitive Illusions in Decision Making

   Availability ►
   Representativeness ►
   Framing Effects ►
   Anchoring ►
   Sunk Cost Effects ►
   Illusory Correlation ►
   Hindsight Bias ►
   Confirmation Bias ►
   Overconfidence ►

                                 Back
Availability
   Ten students from a nearby college have
    indicated a willingness to serve on a
    curriculum committee. Their names are
    Ann, Bob, Dan, Elizabeth, Gary, Heidi,
    Jennifer, Laura, Terri, and Valerie.
       The dean wants to form a two-person
        committee. What is your estimate of the
        number of distinct committees that could be
        formed? (don’t use formulas; just respond
        intuitively)
       The dean wants to form an eight-person
        committee. What is your estimate of the
        number of distinct committees that could be
        formed? (don’t use formulas; just respond
        intuitively)
                                           Continue
Availability
   Consider the two structures shown below:




   A path in a structure is a line that connects one “x”
    from each row, starting with the top row and
    finishing at the bottom row. How many paths do
    you think each structure has?
                                             Continue
Availability

   Tversky & Kahneman, 1973
       When faced with the task of estimating
        probability, frequency, or numerosity,
        people rely on shortcuts or rules of
        thumb (heuristics) to help make
        judgments easier.
            Availability heuristic
                 Assessing the ease with which the relevant
                  mental operation of retrieval, construction,
                  or association can be carried out.
                 Formulas: 10!/{(x!)([10-x]!)} for problem 1
                              xy for problem 2
                                                      Back
Representativeness
   Two students, Linda and Joe, are having a
    boring Saturday afternoon in the student
    union. For lack of something better to do,
    they each begin flipping a quarter,
    keeping track of the way it lands over
    time. Then they compare results. Linda
    reports that her sequence of coin flips
    was heads, heads, heads, tails, tails, tails.
    Joe gets the following results: tails, tails,
    heads, tails, heads, heads.

   Which student has obtained a more
    statistically probable series of results?
                                       Continue
Representativeness

   Kahneman & Tversky, 1973
       Representativeness heuristic
            A belief that outcomes will always reflect
             characteristics of the process that
             generated them
       Conducted an experiment
            Using three conditions
                 Base rate ▶
                 Similarity ▶
                 Prediction ▶


                                               Continue
Representativeness
   Write down your best guesses about the
    percentage now enrolled in each of the
    following nine fields of specialization.
       Business administration
       Computer science
       Engineering
       Humanities and education
       Law
       Library science
       Medicine
       Physical and life sciences
       Social science and social work


                                         Back
Representativeness
   How similar Tom W. is to the typical graduate
    student in each of the following nine fields of
    graduate specialization.
       Tom W. is of high intelligence, although lacking in
        true creativity. He has a need for order and clarity,
        and for neat and tidy systems in which every detail
        finds its appropriate place. His writing is rather dull
        and mechanical, occasionally enlivened by
        somewhat corny puns and by flashes of imagination
        of the sci-fi type. He has a strong drive for
        competence. He seems to have little feel and little
        sympathy for other people and does not enjoy
        interacting with others. Self-centered, he
        nonetheless has a deep moral sense.


                                                    Back
Representativeness

   Participants were given the
    personality sketch but were told it
    was written several years ago,
    during Tom W.’s senior year of high
    school, based on his response to
    projective tests. They were asked to
    predict the likelihood for each field
    that Tom W. was currently a
    graduate student in it.

                                  Back
Representativeness




                     Gambler’s fallacy
Representativeness
   A random process will not always produce
    results that look random, especially in the
    short run.




                                       Continue
Representativeness

   Tversky & Kahneman, 1971
       Law of small numbers
         Misuse of representativeness
         “man who” argument (Nisbett & Ross,
          1980)
               I know a man who smoked three packs a day
                and lived to be 110.




                                               Back
Framing Effects
   You want to buy some fuels for your car,
    and you see two service stations, both
    advertising gasoline. Station A’s price is
    $1.00 per gallon; station B’s, $0.95.
    Station A’s sign also announces, “5
    cents/gallon discount for cash!” Station
    B’s sign announces, “5 cents/gallon
    surcharge for credit cards.” All other
    factors being equal, to which station
    would you choose to go?


                                      Continue
Framing Effects

   Tversky & Kahneman, 1981
       People evaluate outcomes as changes
        from a reference point--- their current
        state
            Depending on how their current state is
             described, they perceive certain outcomes
             as gains or losses.
                 The description “frame”s the decision or
                  provides a certain context
                 Context effect


                                                 Continue
Framing Effects

   Simply changing the description of a
    situation can lead people to adopt
    different reference points and
    therefore to see the same outcome
    as a gain in one situation and a loss
    in the other.




                                  Back
Anchoring

   Estimate the population of
    Philadelphia in April 2000
       Tim and Kim were given a starting
        value respectively: 1 million & 2 million
          Tim: 1.25 million
          Kim: 1.75 million

   The starting point have huge effects
    on their final estimates
       Correct value: 1.5 million

                                      Continue
Anchoring

   Definition
       A decision-making heuristic in which
        final estimates are heavily influenced
        by initial value estimates
   Estimate values
       8x7x6x5x4x3x2x1
       1x2x3x4x5x6x7x8
          2250
          512


                                         Back
Sunk Cost Effects
   A major educational initiative is begun in
    your hometown; $3 million is invested to
    help students stay away from cigarettes,
    liquor, and other drugs. In the third of
    four years, evidence begins to accumulate
    that the program is not working. A local
    legislator proposes ending funding to the
    program before the scheduled date.
    Howls of protest go up from some
    individuals, who claim that to stop a
    program after a large expenditure of
    funds has been spent would be a waste.


                                     Continue
Sunk Cost Effects

   A bias in decision making in which
    already “spent” costs unduly
    influence decision making to
    continue.




                                   Back
Illusory Correlation
   Hair twisting
       The person pinches a strand of hair between
        thumb and forefinger and proceeds to twist
        it around the forefinger.
   If you believe this behavior is especially
    likely in people undergoing a great deal
    of stress.
       Observe 150 students
   The results:





                                                Continue
Illusory Correlation

   People report seeing data
    associations that seem plausible
    even when associations are not
    present

   Illusory correlation
       An association between factors that is
        not supported by data but seems
        plausible

                                         Back
Hindsight Bias

   A tendency to exaggerate the
    certainty of what could have been
    anticipated ahead of time
       Once you know how a decision has
        turned out, you look back on the
        events leading up to the outcome as
        being more inevitable than they really
        were



                                       Back
Confirmation Bias

   The tendency to search only for
    information that will confirm one’s
    initial hunch or hypothesis, and to
    overlook or ignore any other
    information




                                  Back
Overconfidence
   Choose one answer for each question,
    and rate your confidence in your answer
    on a scale from .5 (just guessing) to 1.0
    (complete certain)
       Which magazine had the largest circulation in
        1978?
          Time                 Reader’s Digest
       Which city had the larger population in 1953?
          St. Paul, MN         New Orleans, LA
       Who was the 21st president of the United
        States?
          Arthur              Cleveland
       Who began the profession of nursing?
          Nightingale         Barton

                                                Back
Overconfidence

   Typical findings




                       Calibration
                       curve



                       Continue
Overconfidence

   An overly positive judgment of one’s
    own decision-making abilities and
    performance
       Confidence ratings are higher than
        actual accuracy




                                        Back
Utility Models of Decision Making
   What people are doing when they
    structure a decision and choose from
    alternatives?
       Normative models
          Ideal performance under ideal circumstances

       Prescriptive models
          How we ought to make decisions under non-
           ideal circumstances
       Descriptive models
          Detail what people actually do when they
           make decisions

                                            Continue
Utility Models of Decision Making
   Expected-utility theory
       A normative model of decision making in which
        the decision maker weights the personal
        importance and the probabilities of different
        outcomes in choosing among alternatives in
        order to maximize overall satisfaction of
        personal goals
       Expected value = ∑(Pi x Vi)
          P=probability of the ith outcome

            V=the monetary value of that outcome


                                         Continue
Utility Models of Decision Making

   Imaging a lottery with ten tickets
    numbered 1 through 10. If the
    ticket drawn is numbered 1, you
    win $10. If the ticket drawn is
    numbered 2, 3, or 4, you win $5.
    Any other numbers drawn are worth
    nothing. The EV of this lottery is
       (.1 x $10) + (.3 x $5) + (.6 x $0)
        =$1.60

                                     Continue
Utility Models of Decision Making
   Expected utility (EU) = ∑(pi x ui)




                                         Continue
Utility Models of Decision Making
   Multiattribute Utility Theory
       A normative model of decision making that
        provides a means of integrating different
        dimensions and goals of a complex decision.
   It involves six steps
       Breaking a decision down into independent
        dimensions
       Determining the relative weights of each of
        those dimensions
       Listing all the alternatives
       Ranking all the alternatives along the
        dimensions
       Multiplying the rankings by the weightings to
        determine a final value for each alternative
       Choosing the alternative with the highest value

                                            Continue
Utility Models of Decision Making
Weightings of five dimensions in the decision “choosing” a major.




                                                       Continue
Utility Models of Decision Making




                           Continue
Utility Models of Decision Making




                           Continue
Utility Models of Decision Making




                           Continue
Utility Models of Decision Making

   Payne, 1976
       People do not always spontaneously
        use MAUT.
            How people chose apartments when given
             different amounts of information about
             different numbers of alternatives.
                 Participants were presented with an
                  “information board” carrying a number of
                  cards. ▶
Utility Models of Decision Making




                   Continue
Utility Models of Decision Making
   Two factors were varied in the
    experiment
       The number of alternatives presented
       The number of factors of information
        available per alternative
   When participants had to decide
    among 6 to 12 apartments, they
    used another strategy.
       They eliminated some alternatives on the
        basis of only one or a few dimensions.
            Elimination by aspects
                 A descriptive model     Back
    Descriptive Models of Decision Making

   Image Theory ▶

   Recognition-Primed Decision Making ▶
Descriptive Models of Decision Making

   Image theory
       A descriptive theory of decision making
        that posits that the process consists of
        two stages
            A noncompensatory screening of options
             against the decision maker’s image of
             values and future in which the number of
             options is reduced to a very small set
                 Prechoice screening of options
            If necessary, a compensatory choice
             process


                                                   Continue
Descriptive Models of Decision Making

   Three images
       Value image
            Containing the decision maker’s values,
             morals, and principles
       Trajectory image
            Containing the decision maker’s goals and
             aspirations for the future
       Strategic image
            The way in which the decision maker
             plans to attain his or her goals

                                              Continue
Descriptive Models of Decision Making

   If there is more than one survivor of
    the prechoice screening phase, the
    decision maker may go on to use
    compensatory or other decision
    strategy to make the final choice




                                 Back
Descriptive Models of Decision Making

   Gary Klein, 1998
       Studied special experts
            Firefighters, intensive care pediatric
             nurses, military officers
       Experts most likely to rely on intuition,
        mental simulation, making metaphors
        or analogies, and recalling or creating
        stories



                                              Continue
Descriptive Models of Decision Making

   Recognition-Primed Decision Making
       A theory of expert decision making that
        holds that decision makers choose
        options based on analogy of a given
        situation with previously encountered
        situations




                                          Back