This Pump Sucks: Testing Transitivity with Individual Data Michael H. Birnbaum and Jeffrey P. Bahra California State University, Fullerton.

Slides:



Advertisements
Similar presentations
Experiments and Variables
Advertisements

New Paradoxes of Risky Decision Making that Refute Prospect Theories Michael H. Birnbaum Fullerton, California, USA.
Among those who cycle most have no regrets Michael H. Birnbaum Decision Research Center, Fullerton.
Science of JDM as an Efficient Game of Mastermind Michael H. Birnbaum California State University, Fullerton Bonn, July 26, 2013.
1 Upper Cumulative Independence Michael H. Birnbaum California State University, Fullerton.
1 Lower Distribution Independence Michael H. Birnbaum California State University, Fullerton.
True and Error Models of Response Variation in Judgment and Decision Tasks Michael H. Birnbaum.
Evaluating Non-EU Models Michael H. Birnbaum Fullerton, California, USA.
Who are these People Who Violate Stochastic Dominance, Anyway? What, if anything, are they thinking? Michael H. Birnbaum California State University, Fullerton.
Certainty Equivalent and Stochastic Preferences June 2006 FUR 2006, Rome Pavlo Blavatskyy Wolfgang Köhler IEW, University of Zürich.
Testing Lexicographic Semi- Order Models: Generalizing the Priority Heuristic Michael H. Birnbaum California State University, Fullerton.
Testing Heuristic Models of Risky Decision Making Michael H. Birnbaum California State University, Fullerton.
Some New Approaches to Old Problems: Behavioral Models of Preference Michael H. Birnbaum California State University, Fullerton.
1 Distribution Independence Michael H. Birnbaum California State University, Fullerton.
1 Upper Tail Independence Michael H. Birnbaum California State University, Fullerton.
Decision-making II choosing between gambles neural basis of decision-making.
Testing Models of Stochastic Dominance Violations Michael H. Birnbaum Decision Research Center California State University, Fullerton.
1 Upper Distribution Independence Michael H. Birnbaum California State University, Fullerton.
Ten “New Paradoxes” Refute Cumulative Prospect Theory of Risky Decision Making Michael H. Birnbaum Decision Research Center California State University,
Violations of Stochastic Dominance Michael H. Birnbaum California State University, Fullerton.
Do we always make the best possible decisions?
Testing Critical Properties of Models of Risky Decision Making Michael H. Birnbaum Fullerton, California, USA Sept. 13, 2007 Luxembourg.
Ten “New Paradoxes” Refute Cumulative Prospect Theory of Risky Decision Making Michael H. Birnbaum Decision Research Center California State University,
New Paradoxes of Risky Decision Making that Refute Prospect Theories Michael H. Birnbaum Fullerton, California, USA.
1 The Case Against Prospect Theories of Risky Decision Making Michael H. Birnbaum California State University, Fullerton.
Testing Transitivity (and other Properties) Using a True and Error Model Michael H. Birnbaum.
Assessing cognitive models What is the aim of cognitive modelling? To try and reproduce, using equations or similar, the mechanism that people are using.
Web-Based Program of Research on Risky Decision Making Michael H. Birnbaum California State University, Fullerton.
Web-Based Program of Research on Risky Decision Making Michael H. Birnbaum California State University, Fullerton.
1 A Brief History of Descriptive Theories of Decision Making: Lecture 2: SWU and PT Kiel, June 10, 2005 Michael H. Birnbaum California State University,
1 Gain-Loss Separability and Reflection In memory of Ward Edwards Michael H. Birnbaum California State University, Fullerton.
I’m not overweight It just needs redistribution Michael H. Birnbaum California State University, Fullerton.
1 Ten “New Paradoxes” of Risky Decision Making Michael H. Birnbaum Decision Research Center California State University, Fullerton.
Basic Business Statistics, 10e © 2006 Prentice-Hall, Inc. Chap 9-1 Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests Basic Business Statistics.
1 Gain-Loss Separability Michael H. Birnbaum California State University, Fullerton.
Is there Some Format in Which CPT Violations are Attenuated? Michael H. Birnbaum Decision Research Center California State University, Fullerton.
1 Lower Cumulative Independence Michael H. Birnbaum California State University, Fullerton.
Stochastic Dominance Michael H. Birnbaum Decision Research Center California State University, Fullerton.
Web-Based Program of Research on Risky Decision Making Michael H. Birnbaum California State University, Fullerton.
Testing Transitivity with Individual Data Michael H. Birnbaum and Jeffrey P. Bahra California State University, Fullerton.
1 Restricted Branch Independence Michael H. Birnbaum California State University, Fullerton.
Probability Population:
Grether and Plott: Economic Theory of Choice and the Preference Reversal Phenomenon Economics 328 Spring 2004.
Presidential Address: A Program of Web-Based Research on Decision Making Michael H. Birnbaum SCiP, St. Louis, MO November 18, 2010.
Business Statistics, A First Course (4e) © 2006 Prentice-Hall, Inc. Chap 9-1 Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests Business Statistics,
Behavior in the loss domain : an experiment using the probability trade-off consistency condition Olivier L’Haridon GRID, ESTP-ENSAM.
Chapter 10 Hypothesis Testing
PARAMETRIC STATISTICAL INFERENCE
Educational Research: Competencies for Analysis and Application, 9 th edition. Gay, Mills, & Airasian © 2009 Pearson Education, Inc. All rights reserved.
Decision making behavior Why do people make the choices they do? Reason-based choice Regret theory Effort-accuracy Choice and judgment heuristics.
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Unit 5: Hypothesis Testing.
Statistics (cont.) Psych 231: Research Methods in Psychology.
Time series Model assessment. Tourist arrivals to NZ Period is quarterly.
Lecture 15 – Decision making 1 Decision making occurs when you have several alternatives and you choose among them. There are two characteristics of good.
A Stochastic Expected Utility Theory Pavlo R. Blavatskyy June 2007.
Ellsberg’s paradoxes: Problems for rank- dependent utility explanations Cherng-Horng Lan & Nigel Harvey Department of Psychology University College London.
CROSS-VALIDATION AND MODEL SELECTION Many Slides are from: Dr. Thomas Jensen -Expedia.com and Prof. Olga Veksler - CS Learning and Computer Vision.
Testing Transitivity with a True and Error Model Michael H. Birnbaum California State University, Fullerton.
CHAPTER 15: Tests of Significance The Basics ESSENTIAL STATISTICS Second Edition David S. Moore, William I. Notz, and Michael A. Fligner Lecture Presentation.
Axiomatic Theory of Probabilistic Decision Making under Risk Pavlo R. Blavatskyy University of Zurich April 21st, 2007.
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Unit 5: Hypothesis Testing.
Can a Dominatrix Make My Pump Work? Michael H. Birnbaum CSUF Decision Research Center.
CHAPTER 15: Tests of Significance The Basics ESSENTIAL STATISTICS Second Edition David S. Moore, William I. Notz, and Michael A. Fligner Lecture Presentation.
Statistics (cont.) Psych 231: Research Methods in Psychology.
Inferential Statistics Psych 231: Research Methods in Psychology.
Unit 5: Hypothesis Testing
Significance Tests: The Basics
Significance Tests: The Basics
New Paradoxes of Risky Decision Making that Refute Prospect Theories
Presentation transcript:

This Pump Sucks: Testing Transitivity with Individual Data Michael H. Birnbaum and Jeffrey P. Bahra California State University, Fullerton

Transitivity of Preference If A > B and B > C then A > C. Satisfy it or become a money pump. But transitivity may not hold if data contain “error.” And different people might have different “true” preferences.

Tversky (1969) Tversky (1969) reported that selected subjects showed a pattern of intransitive data consistent with a lexicographic semi-order. Tversky tested Weak Stochastic Transitivity: If P(A>B) > 1/2 and P(B>C) > 1/2 then P(A>C) > 1/2.

Issues Iverson & Falmagne (1985) argued that Tversky’s statistical analysis was incorrect of WST. Tversky went on to publish transitive theories of preference (e.g., CPT).

Renewed Interest in Intransitive Preference New analytical methods for analysis of transitivity (Iverson, Myung, & Karabatsos; Regenwetter & Stober, et al); Error models (Sopher & Gigliotti, ‘93; Birnbaum, ‘04; others). Priority Heuristic (Brandstaetter, et al., 2006); stochastic difference model (González-Vallejo, 2002; similarity judgments, Leland, 1994; majority rule, Zhang, Hsee, Xiao, 2006). Renewed interest in Fishburn, as well as in Regret Theory.

Lexicographic Semi-order G = (x, p; y, 1 - p). F = (x’, q; y’, 1 - q). If y - y’ ≥  L choose G (  L = $10) If y’ - y ≥  L choose F If p - q ≥  P choose G (  P = 0.1) If q - p ≥  P choose F If x > x’ choose G; if x’ > x choose F; Otherwise, choose randomly.

Priority Heuristic “Aspiration level” is 10% of largest prize, rounded to nearest prominent number. Compare gambles by lowest consequences. If difference exceeds the aspiration level, choose by lowest consequence. If not, compare probabilities; choose by probability if difference ≥ 0.1 Compare largest consequences; choose by largest consequences.

New Studies of Transitivity Work currently under way testing transitivity using same procedures as used in other decision research. Participants view choices via the WWW, click button beside the gamble they would prefer to play. Today’s talk: Single-S data.

Studies with Roman Gutierez Four studies used Tversky’s 5 gambles, formatted with tickets or with pie charts. Studies with n = 417 and n = 327 with small or large prizes ($4.50 or $450) No pre-selection of participants. Participants served in other risky DM studies, prior to testing (~1 hr).

Three of Tversky’s (1969) Gambles A = ($5.00, 0.29; $0, 0.79) C = ($4.50, 0.38; $0, 0.62) E = ($4.00, 0.46; $0, 0.54) Priority Heurisitc Predicts: A preferred to C; C preferred to E, and E preferred to A.

Findings Results were surprisingly transitive, unlike Tversky’s data (est. 95% transitive). Of those 115 who were perfectly reliable, 93 perfectly consistent with EV (p), 8 with opposite ($), and only 1 intransitive. Differences: no pre-test; Probability represented by # of tickets (100 per urn), rather than by pies; Participants have practice with variety of gambles, & choices;Tested via Computer.

Pie Chart Format

Pies: with or without Numerical probabilities 321 participants randomly assigned conditions with probabilities displayed as pies (spinner), either with numerical probabilities displayed or without. Of 105 who were perfectly reliable, 84 were perfectly consistent with EV (prob), 13 with the opposite order ($); 1 consistent with LS.

Findings Priority Heuristic predicted violations of transitivity were rare and rarely repeated when probability and prize information presented numerically. Violations of transitivity are still rare but more frequent when probability information presented only graphically. Evidence of Dimension Interaction violates PH and additive Difference models.

Response to Birnbaum- Gutierrez Perhaps the intransitivity only develops in longer studies. Tversky used 20 replications of each choice. Perhaps consequences of Tversky’s gambles diminished since 1969 due to inflation. Perhaps transitivity occurs because those prizes are too small.

Birnbaum & Bahra Collected up to 40 choices/pair per person. (20 reps). 2 Sessions, 1.5 hrs, 1 week apart. Cash prizes up to $ participants, of whom 10 to win the prize of one of their chosen gambles. 3 5 x 5 Designs to test transitivity vs. Priority heuristic predictions

Notation-Two-branch Gambles G = (x, p; y, 1 - p); x > y ≥ 0 L = Lower Consequence P = Probability to win higher prize H = Higher consequence

LH Design A = ($84,.50; $24) B = ($88,.50; $20) C = ($92,.50; $16) D = ($96,.50; $12) E = ($100,.50; $8)

LP Design A = ($100,.50; $24) B = ($100,.54; $20) C = ($100,.58; $16) D = ($100,.62; $12) E = ($100,.66; $8)

PH Design A = ($100,.50; $0) B = ($96,.54; $0) C = ($92,.58; $0) D = ($88,.62; $0) E = ($84,.66; $0)

Priority Heuristic Predictions LH Design: E > D > C > B > A, but A > E LP Design: A ~ B ~ C ~ D ~ E, but A > E PH Design: A > B > C > D > E but E > A

One Rep = 2 choices/pair

Analysis Each replication of each design has 20 choices; hence 1,048,576 possible data patterns (2 20 ) per rep. There are 1024 possible consistent patterns (R ij = 2 iff R ji = 1, all i, j). There are 120 (5!) possible transitive patterns.

Within-Rep Consistency Count the number of consistent choices in a replicate of 20 choices (10 x 2). If a person always chose the same button, consistency = 0. If a person was perfectly consistent, consistency = 10. Randomly choosing between 1 and 2 produces expected consistency of 5.

Intransitive and Consistent

Within-Replicate Consistency The average rate of agreement was 8.63 (86% self-agreement). 46.4% of all replicates were scored 10; an additional 19.9% were scored 9.

LH Design: Overall Proportions Choosing Second Gamble

LP Design: Overall Proportions Choosing Second Gamble

PH Design: Overall Proportions Choosing Second Gamble

Majority Data WST LH Design A>B>C>D>E LP Design A>B>C>D>E PH Design E>D>C>B>A Patterns consistent with special TAX with “prior” parameters. But this analysis hides individual diffs

Individual Data Choice proportions calculated for each individual in each design. These were further broken down within each person by replication.

S# 8328 C = 9.6 Rep = 20

S# 8328 C = 9.8 Rep = 20

S# 8328 C = 9.9 Rep = 20

S# 6176 C = 9.8 Rep = 20; started with this pattern, then switched to perfectly consistent with the opposite pattern for 4 replicates at the end of the first day; back to this pattern for 10 reps on day 2.

S# 684 C = 8.1 Rep = 14; an intransitive pattern opposite that predicted by priority heuristic.

S# 7663 C = 6.3 Rep = 10; an intransitive pattern consistent with priority heuristic,  P = Few reps and low self-consistency in this case.

Data Summary For n = 51, there are 153 matrices. Of these, 90% were perfectly consistent with WST: P(A,B) ≥ 1/2 & P(B,C) ≥ 1/2 then P(A,C) ≥ 1/2. 29 people had all three arrays fitting WST; no one had all three arrays with intransitive patterns.

Summary of WST Individuals

29 People with 3 Perfectly WST Patterns

Within-Person Changes in Preference Pattern Criterion: Person must show perfect consistency (10 out of 10) to one pattern in one replication, and perfect consistency to another pattern on another replication. 15 Such cases were found (10%). There may be other cases where the data are less consistent.

Summary Recent studies fail to confirm systematic violations of transitivity predicted by priority heuristic. Adds to growing case against this descriptive model. Individual data are mostly transitive. Next Q: From individual data, can we predict, for example, from these data to other kinds of choices by same person, e. g., tests of SD?