Heuristics and Biases

# Heuristics and Biases

Télécharger la présentation

## Heuristics and Biases

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
##### Presentation Transcript

1. Heuristics and Biases Behavioral and Experimental Economics

2. Making Decisions under Uncertainty • Many decisions are based on beliefs concerning the likelihood of uncertain events… • Who will win the election? How will a \$ trade for a € tomorrow? Is the defendant guilty? • What determines such beliefs? • How do people assess the probability of an uncertain event or the value of an uncertain quantity?

3. We Use Heuristics We reduce the complex tasks of assessing probabilities and predicting values to simpler judgmental operations. Resembles the subjective assessment of physical quantities such as distance or size. Example of how we judge the distance of an object using clarity. The more sharply the object is seen, the closer it appears to be. This is mostly right but subject to systematic errors.

4. Introduction What is a heuristic? Why do humans use them? Do heuristics help or hurt human decision making?

5. Definition from Wikipedia Heuristic ( /hjʉˈrɪstɨk/; or heuristics; Greek: "Εὑρίσκω", "find" or "discover") refers to experience-based techniques for problem solving, learning, and discovery. Heuristic methods are used to speed up the process of finding a satisfactory solution, where an exhaustive search is impractical. Examples of this method include using a "rule of thumb", an educated guess, an intuitive judgment, or common sense. In more precise terms, heuristics are strategies using readily accessible, though loosely applicable, information to control problem solving in human beings and machines.[1]

6. Kahneman and Tversky http://nobelprize.org/nobel_prizes/economics/laureates/2002/kahneman-autobio.html http://www.dangoldstein.com/dsn/archives/2005/07/amos_tversky_1.html "Judgment under Uncertainty: Heuristics and Biases," Science,1974. James Monitier “Behaving Badly,” 2/2006. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=890563

7. Source of Bias • This seminal article by Tversky and Kahneman describes three Heuristics people rely on to assess probabilities and predict values: • Representativeness • Availability • Adjustment and Anchoring

8. Representativeness • Many of the probabilistic questions with which people are concerned belong to one of the following types: • 1. What is the probability that object A belongs to class B • OR • 2. What is the probability that process B will generate event A

9. To answer we rely on Representativeness Heuristic Probabilities are evaluated by the degree to which A is representative of B… In other words, the degree to which A resembles B Is A similar to B?

10. Example Linda is 31 years old. She is single, outspoken and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and equality. Which is more likely? 1) Linda works in a bank 2) Linda works in a bank and is active in the feminist movement

11. Representativeness It is more likely that Linda works in a bank. To argue that (2) is more likely is to commit a conjunction fallacy. Tversky & Kahneman (1983) Bankers 85% of professional fund managers chose (1) Feminist Bankers Feminists

12. Representativeness Source of the error? The Representative Heuristic (rule of thumb) People base judgments on how things appear rather than how statistically likely they are. People are driven by the narrative of the description rather than by the logic of the analysis.

13. Representativeness "The best explanation to date of the misperception of random sequences is offered by psychologists Daniel Kahneman and Amos Tversky, who attribute it to people’s tendency to be overly influenced by judgments of “representativeness.” Representativeness can be thought of as the reflexive tendency to assess the similarity of outcomes, instances, and categories on relatively salient and even superficial features, and then to use these assessments of similarity as a basis of judgment. We expect instances to look like the categories of which they are members; thus, we expect someone who is a librarian to resemble the prototypical librarian. We expect effects to look like their causes; thus we are more likely to attribute a case of heartburn to spicy rather than bland food, and we are more inclined to see jagged handwriting as a sign of a tense rather than a relaxed personality.“ Gilovich (1991), page 18

14. Example: Steve “Steve is very shy and withdrawn, invariably helpful, but with little interest in people, or in the world of reality. A meek and tidy soul, he has a need for order and structure, and a passion for detail.”

15. Order the following occupations in terms of the probability in which Steve is engaged in them Farmer Salesman Airline pilot Librarian physician

16. People order by probability and similarity in exactly the same way Problem: similarity, or representativeness, is not influenced by several factors that SHOULD affect judgments of probability

17. Our Insensitivities • 1. Insensitivity to prior probability or outcomes: what is the base-rate frequency of the outcomes? • The fact that there are many more farmers than librarians in the popn should enter into any reasonable estimate of the probability that Steve is a librarian rather than a farmer

18. Example Subjects shown brief personality descriptions of several individuals Subjects asked to assess, for each description, the prob that it belonged to an engineer rather than a lawyer In one experimental condition, subjects were told that the descriptions had been drawn from a sample of 70 engineers and 30 lawyers In another condition, 30 engineers and 70 lawyers

19. Bayes’ Rule The odds that any particular description belongs to an engineer rather than to a lawyer should be higher in this condition, where there is a majority of engineers! (.7/.3)^2 = 5.44 In violation of Bayes’ rule, the subjects in the two conditions produced essentially the same probability judgments. People ignored the math and looked only at the representativeness!

20. Evidence, No Evidence, Worthless Evidence When given no other information, subjects used the prior probabilities (evidence) Evidence ignored when given a description + evidence Evidence ALSO ignored when given a useless description

21. Useless Description “Dick is a 30 year old man. He is married with no children. A man of high ability and high motivation, he promises to be quite successful in his field. He is well liked by his colleagues.”

22. Insensitivity 2 2. Insensitivity to sample size Example: What is the average height of the following group of men? N = 1000 N = 100 N = 10 How do you make this assessment?

23. Sample Size People failed to appreciate the role of sample size. As n increases, variability decreases John Nunley and I both got decent student evaluation scores last year (him: 3.87 me: 3.88 I taught 3 classes and he taught 6 Who is the better teacher?

24. Example: • 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. • For a period of one year, each hospital recorded the days on which more than 60 percent of the babies born were boys. Which hospital do you think recorded such days?

25. Results from undergrad students The larger hospital (21) The smaller hospital (210 About the same (that is, within 5 percent of each other) (53) Discuss

26. Representativeness Posterior OddsUrn is 2/3 Red Joe’s Sample 8-to-1 Red Betty’s Sample 16-to-1 Red Who should be more confident that the urn contains 2/3 red balls and 1/3 white? Conclusion: People underestimate evidence Don’t consider sample size.

27. Explanation Intuitive judgments are dominated by the sample proportion and are essentially unaffected by the size of the sample.

28. Insensitivity 3 3. Misconceptions of chance (AKA the gambler’s fallacy) People expect that a sequence of events generated by a random process will represent the essential characteristics of that process even when the sequence is short. Example: People think that HTHTTH is more likely than HHHTTT or HHHTH

29. Chance We think that “chance” will be displayed “globally” and “locally” Imagine a roulette wheel at Vegas that has fallen on red for the last five spins… The next spin MUST be black… Right? RIGHT? We think black is “due” because it will look more like a representative sequence than if the wheel spins red.

30. More on Chance We think of chance as a self-correcting process in which a deviation in one direction induces a deviation in the opposite direction to restore the equilibrium. In fact, deviations are not “corrected”, rather as the chance process unfolds, they are “diluted.”

31. Representativeness: Gambler’s Fallacy Suppose an unbiased coin is flipped 3 times, and each time it lands on head. If you had to bet \$1,000 on the next toss, what side would you chose? Heads, Tails or no preference? 300 fund managers: no preference = 81% Of the 29 2010 BEE students: No preference = 17 of 23 (74%) Tails = 4 (17%) Heads = 2 (8%) What is going on here?

32. The Gambler’s Fallacy Source of the fallacy? The coin has no memory and each side is equally likely You are betting on a SINGLE flip, not an ENTIRE sequence

33. The Gambler’s Fallacy The Gambler’s Fallacy and the Hot Hand: Empirical Data from Casinos Rachel Croson and James Sundali Journal of Risk and Uncertainty (2005) Abstract Research on decision making under uncertainty demonstrates that intuitive ideas of randomness depart systematically from the laws of chance. Two such departures involving random sequences of events have been documented in the laboratory, the gambler’s fallacy and the hot hand. This study presents results from the field, using videotapes of patrons gambling in a casino, to examine the existence and extent of these biases in naturalistic settings. We find small but significant biases in our population, consistent with those observed in the lab.

34. Insensitivity 4 4. Insensitivity to predictability: if predictability is nil, the same prediction should be made in all cases. Examples: which company will be profitable? What is the future value of a stock? Who will win the football game? What info do you have to make this assessment?

35. Predictions are often made by Representativeness Three descriptions of a company: favorable, mediocre, poor. If mediocre description, mediocre prediction. The degree to which the description is favorable is unaffected by the reliability of that description or by the degree to which it permits accurate description. Predictions insensitive to reliability of evidence.

36. Insensitivity 5 5. Illusion of validity: The unwarranted confidence which is produced by a good fit between the predicted outcome and the input information People are more confident in their predictions if they perceive that the inputs look more like the outputs-no regard for limits of predictability

37. Example: Accuracy versus Confidence People express more confidence in predicting the final GPA of a student whose first-year record consists entirely of B’s than in predicting the GPA of a student whose first year record includes many A’s and C’s. Basic statistics tells us we can make better predictions if we have independent inputs rather than redundant or correlated inputs

38. Insensitivity 6 6. Misconception of regression: people don’t understand regression toward the mean. Consider two variables X and Y which have the same distribution. If one selects individuals whose average X score deviates from the mean of X by k units, then the average of their Y scores will usually deviate from the mean of Y by less than k units. Note: this was first documented by Galton more than 100 years ago.

39. Examples throughout life Comparison of height of fathers and sons Intelligence of husbands and wives Performance of individuals on consecutive examinations People do not develop correct intuitions about this phenomenon 1. They do not expect regression in many contexts 2. When they do recognize it they invent spurious causal explanations.

40. Pilot Training Example(Tversky and Kahneman (1974)) Pilot instructors note that… Praise for an exceptionally smooth landing is typically followed by a poor landing on next try. Criticism for a rough landing is typically followed by an improvement on next try. Conclusion: Verbal rewards are detrimental to learning, while verbal punishments are beneficial.

41. Misconception of Regression Probability What’s more likely? Improvement or an even worse performance? Mean Performance Bad

42. Availability People assess the frequency of a class or the probability of an event by the ease with which instances or occurrences can be brought to mind Example: one may assess the risk of a heart attack among middle-aged people by recalling such occurrences among one’s acquaintances

43. Predictable Biases from reliance on Availability • 1. Biases due to retrievability of the data: a class whose instances are easily retrieved will seem bigger than a similarly-sized class whose instances are less retrievable • Car crashes versus airplane crashes • Familiarity: example lists of famous people • Salience: seeing a house burning • Recent-ness

44. Bias of Availability 2 2. Biases due to the effectiveness of a search set: searching for words by first letter Example: Suppose one samples a word (of three letters or more) at random from an English text. Is it more likely that the word starts with r or that r is the third letter?

45. Bias of Availability 3 3. Biases of Imaginability

46. Bias of Availability 4 4. Illusory Correlation: Example mental patients with clinical diagnosis and drawings.

47. Anchoring & Adjustment 1. What are the last four digits of your telephone number? 2. Is the number of physicians in Wisconsin higher or lower than that number? 3. What is your best guess as to the number of physicians in Wisconsin?

48. Anchoring Mean guesses of number of physicians Last 4 digits of telephone # Sources: “Behaving Badly” (2006) and B.E. Class Survey

49. Anchoring: 474BE Class 2011

50. What is going on? Anchoring – the tendency to grab hold of irrelevant and subliminal inputs in the face of uncertainty. If people are “rational,” there should be no difference between those who happen to have high telephone numbers and those who have low ones