Presentation is loading. Please wait.

Presentation is loading. Please wait.

Day 3 Decision-Making Strategies. 2  Used for multi-attribute choices where the choices are comprised of multiple dimensions  People trade off low.

Similar presentations

Presentation on theme: "Day 3 Decision-Making Strategies. 2  Used for multi-attribute choices where the choices are comprised of multiple dimensions  People trade off low."— Presentation transcript:

1 Day 3 Decision-Making Strategies

2 2

3  Used for multi-attribute choices where the choices are comprised of multiple dimensions  People trade off low values on one dimension for high values on another dimension  Different types of compensatory strategies  Linear model  Additive Difference model  Ideal point model 3

4  Each dimension is weighted according to its importance  The weighted values are summed to form an overall index of value  This is one of the foundations of choice modeling  Based on expected utility theory 4

5  Each choice alternative is evaluated on all dimensions  The first step is looking at all alternatives for a single dimension  The differences are weighted and summed  It simplifies the choice (eliminates unnecessary information)  Tends to seem closer to how people make decisions  Based on satisficing theory 5

6  Comparison of each alternative to an “ideal”  Evaluate each one as to how close or far it is from the “ideal” 6

7  Best suited for fairly simple decisions with few dimensions  Well suited when the dimensions are known or knowable, and when an “ideal” choice can be envisioned 7

8 8

9  Do not allow trade-offs  Four major types  Conjunctive rule  Disjunctive rule  Lexicographic strategy  Elimination by aspects strategy 9

10  Eliminate any alternatives that fall outside certain pre-defined boundaries  Example of satisficing rather than optimizing 10

11  Each alternative is evaluated in terms of its best attribute, regardless of the rating on other dimensions 11

12  Identify the most important dimension and choose the most desirable alternatives for that dimension  If there is more than one “winner,” go to the next most important dimension and evaluate the remaining alternatives in the same way 12

13  Probabilistic variation of the lexicographic strategy  Each dimension selected with a probability proportional to its importance  Then alternatives are compared on that highest probability dimension and winners move to the next dimension 13

14  People pick the most important dimension of the alternatives and then select the highest rated alternative on that dimension  Pragmatic approach 14

15 15

16  Logical thinking  Prefrontal cortex  Metacognition – ability to reflect on one’s own mind and thus regulate (to a degree) the emotions  Monitors emotions and decides what to take seriously and what to ignore 16

17  Claude Steele  Stanford sophomores took the Graduate Record Exam (GRE)  White students performed significantly better than black students  Called the Achievement gap  When students told it was just a preparatory drill, no difference in scores 17

18  Power of prefrontal cortex to modulate most body signals, like pain  Fake pain-relieving cream provided relief  Electric shocks mitigated  Sobe Adrenaline Rush – lower price seen as producing less effective in problem solving 18

19  CalTech and Stanford wine tasting experiment  Three levels of wine - $5, $45, $90  With blind testing, respondents could sort them out fairly accurately  When asked to take a short survey about the wine characteristics, they became confused and selected incorrectly 19

20  Brian Wansink, Cornell  Bottomless bowl of soup  Whatever people see on their plate, they eat  They keep track by counting plates, or scoops of M&M’s, not actual food 20

21  Daniel Kahneman  Random number generated by roulette wheel and shown to respondents  Estimate the number of African countries in the United Nations  Those who saw higher roulette number guessed higher number of African countries, and those who saw lower roulette number guessed lower number of African countries 21

22  Plous, Ch. 19  Overconfidence is greatest when accuracy is near chance levels  Overconfidence diminishes as accuracy increases from 50 to 80 percent, and when it exceeds 80 percent, people become underconfident  Discrepancies between accuracy and confidence are not related to intelligence 22

23  The degree to which confidence matches accuracy 23

24 Estimate the answers at a 90% confidence levelLowHigh 1. Martin Luther King’s age at death 2. Length of the Nile River 3. Number of countries that are members of OPEC 4. Number of books in the Old Testament 5. Diameter of the moon in miles 6. Weight of an empty Boeing 747 in pounds 7. Year in which Wolfgang Amadeus Mozart was born 8. Gestation period (in days) of an Asian elephant 9. Air distance from London to Tokyo 10. Deepest (known) point in the ocean (in feet) 24

25 Estimate the answers at a 90% confidence levelAnswers 1. Martin Luther King’s age at death39 years 2. Length of the Nile River4187 miles 3. Number of countries that are members of OPEC13 countries 4. Number of books in the Old Testament39 books 5. Diameter of the moon in miles2160 miles 6. Weight of an empty Boeing 747 in pounds390,000 pounds 7. Year in which Wolfgang Amadeus Mozart was born1756 8. Gestation period (in days) of an Asian elephant645 days 9. Air distance from London to Tokyo5959 miles 10. Deepest (known) point in the ocean (in feet)36,198 feet 25

26  Confidence in decisions is comforting  That can lead to disastrously wrong decisions  Counter that by paying attention to the details that don’t fit the overall pattern  George Day’s “small voices” 26

27  Stop to consider reasons why your judgment might be wrong  We all have blind spots – reducing overconfidence is a matter of recognizing that those blind spots may exist and trying to examine alternatives 27

28  Plous, Ch. 21  Embarking on a promising course of action that subsequently becomes untenable  Taxonomy of traps  Time delay traps  Ignorance traps  Investment traps  Deterioration traps  Collective traps 28

29  Momentary gratification versus long term consequences – smoking or extra dessert  Or avoidance of momentary discomfort versus long term consequences – skipping a dentist appointment  Marketing implication – buy now, pay later plans 29

30  Walter Mischel, 1970’s  Four year olds  Eat one marshmallow or wait a few minutes to get two marshmallows  Most kids couldn’t resist for long  Kids who can’t resist tend to exhibit behavioral problems later in life  Tends to stabilize after the teen years 30

31  Negative consequences not understood by the decision-maker  Common when new life paths are taken and consequences are not forseeable  Example – college students getting MSMR degrees 31

32  When prior expenditures lead people to make bad choices  People sometimes ignore the “sunk cost” of investments 32

33  Costs and benefits change over time, leading to less desirable outcomes 33

34  Apply to groups of people  Sub-optimization by an individual leads to adverse consequences for the group 34

35 35 Prisoner A Prisoner B Confesses Doesn’t confess 5 years 10 years 0 years 1 years 0 years 10 years Above diagonal - Prisoner’s A sentence; below diagonal – Prisoner B’s sentence

36  Dilemma is that if both don’t confess, they get a much shorter (1 year) sentence, but are also exposed to a much longer (10 year) possibility  Both are better off confessing and taking a 5 year sentence  What would you do? 36

37  A dollar is auctioned to two bidders  Four rules: 1. No communication is allowed among bidders while the auction is taking place 2. Bids can be made only in multiples of 5 cents, beginning with a nickel 3. Bids must not exceed $50 4. The two highest bidders both have to pay what they bid, even though the dollar goes only to the highest bidder 37

38  When the two highest bids total $1.00, the auctioneer is assured of a profit  Still attractive from the bidder’s perspective  Second inflection point is when the first bidder reaches $1  Second bidder can’t quit, as they stand to lose 95 cents if they don’t go to $1.05  Bidding often reaches a few dollars 38

39  Explicitly consider withdrawal costs before beginning  Use different decision-makers for the initial and subsequent decisions – less personal involvement 39

40 40

41  Moral decisions tend to be regulated by emotions  Reason is invented as logical support for the emotional decision  Moral decisions require taking other people into account, not just oneself 41

42  Scenario 1: You are the driver of a runaway trolley. The brakes have failed. If you do nothing, five maintenance workers will die. If you swerve, one maintenance worker will die.  Scenario 2: You are standing on a footbridge over a trolley track. Unless the trolley can be stopped, five maintenance workers will die. Standing next to you is a large man, who if you push over the bridge, will fall on the track and stop the trolley, but the large man will die. 42

43  In scenario 1, when you are the driver, 95% of respondents agree it is better to swerve and save five men with one other man dying. This is a personal moral situation.  In scenario 2, almost nobody is willing to push the man over the edge, resulting in five people dying. This is an impersonal moral decision, and activates different parts of the brain. 43

44  For personal moral decisions, a rational moral decision process activates to generate an optimal decision – one death is better than five other deaths  For impersonal moral decisions, the area responsible for thinking about other people (superior temporal suculus, posterior cingulate, and medial frontal cortex) activate and produce confusion and a sub-optimal decision – one death is capital murder 44

45  The ultimatum game  Two respondents  One gets $10 and decides how to divide it  The other decides to accept the offer or reject, in which case both get nothing  Economists thought most people would offer a nominal amount like $1 and keep the rest  The logical response is to accept any offer  Most people rejected low offers as “unfair” and walked away with nothing  Proposers anticipated this response, and actually made “fair” offers in the area of $5 45

46  The desire to help others  The brain rewards altruism with a pleasurable feeling  Autism – people who can’t engage in or understand social interactions with others  Results in inability to sympathize with others  Mirror neurons aren’t developed 46

47  Decisions feel “unanimous” to us  However, most decisions are the result of weighing multiple conflicting factors  Stimulate the NAcc and pacify the insula  Prime with highly coveted items  Use promotional stickers to make the deal seem like a good deal  Credit cards are less like “real money,” therefore result in more purchasing 47

48  Brock and Balloun, late 1960’s  Two groups – regular churchgoers, committed atheists  Played tape recorded message attacking Christianity, with annoying static added  Listeners able to press button and remove static  Atheists removed the static, churchgoers did not - they each heard what they wanted 48

49  Jewish tailor story (Plous p. 22)  The gang members tried to upset the tailor; when he paid them, he said they were making him happy; when he reduced the payment, they couldn’t justify their behavior  People are motivated to reduce psychological inconsistencies 49

50  Cortex can only handle about seven data elements at once  Car buying involves dozens of features, options, etc.  Dijksterhuis categorized products with a complexity score  Simple things like simple kitchen tools (oven mitts) and home accessories (light bulbs) are easy  Complex things like furniture is very hard 50

51  With complex decisions, the longer people ponder them, the less satisfied they are with their decisions  The optimal strategy – use your rational mind to gather needed information, then don’t think about it – let your subconscious arrive at a good decision 51

52 52

53  Allport, 1954 and Zajonc, 1965  When observers are present, a person performs simple, well-rehearsed tasks better  For complex or unmastered skills, the performance is worse 53

54  Moede (1927)  People do not work as hard in groups as they work alone  When alone, individuals bear sole responsibility for the outcome  When in a group, responsibility is shared (diffused)  The larger the group, the more likely that social loafing will occur 54

55  Social comparison theory – people evaluate their ability levels and appropriateness of their opinions by comparison with others  People tend to compare themselves with those who are most similar to them  This can lead to conformity  Asch (1951) bar chart test 55

56  Consistent vocal minority can have an impact on a larger group of people  This can happen even when minority group is not particularly powerful or prestigious 56

57  Social pressures in a group to conform to the perceived consensus opinion of the group  Janis (1982) – eight common symptoms 1. Illusion of invulnerability leading to overoptimism 2. Collective efforts to rationalize 3. Unquestioned belief in group’s morality 4. Stereotyped views of adversaries 5. Pressure directed at any new member who dissents 6. Shared illusion of unanimity 7. Self-censorship of deviations from group consensus 8. Self-appointed “mind-guards” who protect the group from external, divergent information 57

58  Group discussion leads to advocating greater risk taking than individuals would advocate  However, when the initial inclination of the group is toward caution, group discussion can lead to advocating greater caution than individuals would advocate  Called “choice shift” 58

59  Slightly better than individual judgments  Three person groups  However, best individual in a group tends to outperform the overall group 59

60  Five types of group decision techniques  Consensus – face-to-face discussion leads to agreement  Dialectic – group members required to discuss factors that might bias their decisions  Dictator – one member (hopefully best) makes judgments for the group  Delphi – iteration through a series of rounds  Collective – no interaction among group members; decision is simple average 60

61  Class discussion 61

62  Aggregation is the worst type  Dictator works best IFF best individual is chosen to be the dictator  However, dictator tends to modify decisions based on group consensus 62

63 63

64  Simple problems require reason  Novel problems require reason  Embrace uncertainty  You know more than you know  Think about thinking 64

Download ppt "Day 3 Decision-Making Strategies. 2  Used for multi-attribute choices where the choices are comprised of multiple dimensions  People trade off low."

Similar presentations

Ads by Google