Decision processes brown bag, spring 1999

12-1:30 PM, Mondays, room D9-10 Graduate School of Education
Note new room.

11-Jan, Chip Heath, Duke
On the Social Psychology of Agency Relationships: Lay Theories of Motivation, Overemphasize Extrinsic Rewards

18-Jan, Steve Hoch, Wharton
The Variety of an Assortment (with Eric Bradlow and Brian Wansink)

25-Jan, Dave Schmittlein, Wharton
A Live Baby or Your Money Back: The Marketing of in vitro Fertilization Procedures (with Donald G. Morrison (UCLA)

1-Feb, Andy Schotter, NYU
Coordination Conventions in Inter-Generational Games: An Experiment in Lamarackian Evolutionary Dynamics

8-Feb, Lola Lopes, Iowa
Attention and Aspiration in Risky Choice: A Theoretical Account

15-Feb, Horward Kunreuther, Wharton (with C. Hee)
The affection effect in insurance decisions

22-Feb, Teck Ho, Wharton
Adaptive and Sophisticated Experience-Weighted Attraction Learning in Normal-form Games

1-Mar, Hal Arkes, Ohio U.
Does Articulating the Basis of A Decision Improve or Impair the Decision?

8-Mar, Spring Break

15-Mar, Ken Hammond, Colorado
Distinguishing Two Varieties of Cognitive Competence Will Improve Forecasts and Judgments Under Stress

22-Mar, Tom Schelling, Maryland (4:30-6, 351 SH-DH)
Rational choice and some of its alternatives

29-Mar, Ravi Dhar, Yale
The effect of forced choice on choice

5-Apr, Richard Zeckhauser, Harvard
Earnings management to exceed thresholds
This is a co-sponsored event by Applied Economics and Decision Processes.

12-Apr, Cancelled

19-Apr, John Payne, Duke
Difficult Valuations: Assigning a Dollar Amount to an Environmental Good or to a Punitive Damage Award

26-Apr, Craig Fox, Duke
Familiarity Bias and Belief Reversal in Relative Likelihood Judgment


Three laboratory studies and one field study show that people generally hold lay theories which contain an extrinsic incentive bias -- people predict that others are more motivated than themselves by extrinsic incentives (job security, pay) and less motivated by intrinsic incentives (learning new things). The extrinsic incentives bias can be separated from a self-serving bias and it provides an empirical counterexample to the traditional actor-observer effect in social psychology (although its theoretical explanation is similar). This kind of bias may hinder organizations from organizing because people who act as principals may use improper lay theories to offer inappropriate deals to agents.


How do people perceive the amount of variety contained in an assortment and how do these perceptions influence satisfaction and choice? Specifically, we consider the variety question in the context of assortments of multi-attribute categorical objects of fixed number n. We develop a simple and flexible metric for characterizing the informational structure of any categorical assortment and show that it is a special case of a more general ordinal metric. The metric is calculated by considering the dissimilarity between all {n choose 2} pairs of objects in the assortment. Aggregating across all pairs in the set, the variety of each assortment can be represented as a function of the number of pairs that differ by zero attributes (duplicate objects), one attribute, ..., up to M attributes (completely distinct objects). The metric is comparable to a generalized Hamming distance, a measure developed to study coding transmission errors in WWII. It is related to work in the psychology of similarity and classification and concept learning, a difference being that we focus on the overall psychological distance contained in an assortment of objects rather than the single object focus of previous work. We generalize a simple version of our metric to accommodate differences in attribute salience and importance and also allow for differential weighting of pairwise differences depending on whether comparison involve local (adjacent objects) or non-local (non-proximate objects) information structure.

We then use the metric to develop assortments that vary widely in terms of their pattern of pairwise distinctions between objects in each set. In the empirical work, we consider the influence of three factors on variety perceptions: the informational structure of each assortment (i.e., the pattern of pairwise distinctions); the level of organization of the objects (organized vs. random displays); and the task orientation (analytic, intentional processing vs. holistic, incidental processing). First, we find that informational structure matters a lot. Distinctions of one have the biggest marginal influence (.10) on variety perceptions, followed by distinctions of two (.04), and distinctions of three (.03); people easily pick up the presence of duplicate objects but there are diminishing returns to higher-order uniqueness. We also find that people are more influenced by local information structure (directly adjacent objects) than by non-local information structure. Proximity matters. Organization of the display can either increase or decrease variety perceptions. When processing is holistic, random displays are seen as more varied. When people engage in analytic processing, however, organized displays appear to offer more variety and variety perceptions are more sensitive to the underlying information structure. Analysis of individual level variety perception models reveals substantial heterogeneity and four distinct groups of subjects: one group that engages in holistic processing and sees organized displays as much less varied; another group that engages in detailed processing of local information and sees organized displays as much more varied; and two other groups that fall in between.

We also find that both variety perception and organization are key drivers of satisfaction and choice. People are more satisfied with and likely to choose those assortments that are perceived as offering high variety and that are displayed in an organized rather than random manner. Variety perceptions mediate the effect of information structure on satisfaction and choice. We end the paper by considering other applications for our variety metric including: choosing stimuli for concept testing; product extensions and positioning; and experimental design for meta-analysis and conjoint measurement.

A Live Baby or Your Money Back: The Marketing of in vitro Fertilization Procedures
David Schmittlein (Wharton) and Donald G. Morrison (UCLA)

A large number of clinics that offer in vitro fertilization (IVF) have begun to aggressively market the following options to couples seeking to have a genetically related baby: (1) an a la carte program where the couple pays $7,500 per attempt regardless of the outcome; or (2) a money-back-guarantee program where the couple pays a $15,000 up-front fee that covers up to three attempts - however, if after three cycles there is no live birth delivery, then the full $15,000 is refunded.

If the couple contemplating these two choices knew the success probability for each attempt, then a simple analysis would show whether the a la carte or money-back-guarantee program is better. Unfortunately, it is difficult for couples to gather the relevant data and even more daunting to adjust the aggregate data to their own situation.

In this article we assemble the most recent available data and develop a model that allows patients, clinics, and public policy advocates to assess the a la carte vs. money-back-guarantee programs. The most surprising result of our analysis is that the money-back-guarantee program appears (for the patients) to be "too good to be true." That is, with reasonable projections from the most recent data, the money-back guarantee yields a substantial negative expected profit per couple for the clinics. More importantly from the patients' perspective, the money-back-guarantee turns out to be the better option for all couples with less than 0.5 success probability per cycle. (The breakeven probability is even higher if risk aversion is considered.) Virtually all traditional IVF patients can be considered to have per-cycle success probabilities well below 0.5. Can it be that clinics are offering money-back-guarantees that both lose money for the clinics and give the patients a deal that is far better for them than the traditional a la carte payment method?

After a detailed analysis of the key variables - i.e., success rate per attempt, heterogeneity of couples' base rates of success, individual couples' "learning" on successive attempts, and cost to the clinic per attempt - nothing makes these money-back-guarantees profitable for the clinics. Since presumably clinics are not in business to lose money, the analysis must be missing something major. Based on the kind of aggressive marketing (e.g., mass media) for the money-back guarantees, we believe it is bringing in younger and less infertile patients than those who were in IVF clinics prior to 1996. In other words, the marketing of money-back guarantees may be inducing couples who would previously have used - successfully - other less invasive procedures with fewer potential side effects and less risk of multiple births to decide, "Why wait: let's jump to IVF now and we'll get our money back if it doesn't work." We show that under this scenario the money-back-guarantees can be profitable for the clinics.

The implications of earlier use of IVF are then considered for patients, clinic managers, and from an overall public policy point of view. Although the clinics that make profits and the couples who either receive a baby or their money-back are unlikely to complain, there are some significant downsides to the marketing of IVF money-back-guarantees that need to be understood.

Lola Lopes
Attention and Aspiration in Risky Choice: A Theoretical Account

It has long been known that expected utility theory provides a poor account of human decision making under risk. Recently, however, both theoreticians and experimentalists have looked to a rank-dependent variant of EU to provide the foundation for a better account. This talk develops the psychological basis for preferring a rank-dependent account but shows that rank-dependence alone is insufficient. A theory (SP/A Theory) is described in which an attentional factor (Security- Potential) is joined with a stochastic control factor (Aspiration) to more accurately describe risky choice.

Christopher K. Hsee
Graduate School of Business, University of Chicago

Howard C. Kunreuther
The Wharton School, University of Pennsylvania
National Bureau of Economic Research


This research studies individual insurance behavior, and documents a robust, normatively-irrelevant "affection effect" in both the decision to purchase coverage and to invest the time and energy in making an insurance claim after suffering a loss. People who had greater affection toward an article they were shipping were more willing to buy an insurance policy covering damage to the item than those who had less affection, holding the amount of coverage and all other normatively-relevant factors constant. In case the article was damaged, those who had greater affection for the article were also more willing to go through the trouble of collecting the compensation. We explain these findings by a "consolation hypothesis," according to which, people perceive insurance compensation not just as a monetary payment, but as a token of consolation. The implication of this research goes beyond insurance behavior. It suggests (a) that decisions can be influenced heavily by affective factors that are often ignored in normative analyses, and (b) that the same amount of money can evoke different feelings and have different utilities depending on its source.

Teck Ho
"Adaptive and Sophisticated Experience-Weighted Attraction Learning in Normal-form Games"

In strategic situations, people care about what others are likely to do. In the last few decades a large body of mathematical `game theory' has developed about how people will make choices in these situations. However, these theories generally assume that people know a lot about how other people in the strategic situation are likely to behave, or they can figure it out. In fact, people usually figure out what others are likely to do by learning from experience.

The first part of the talk describes a theory of how this learning occurs. This theory combines two very different forces-- `reinforcement', which means that successful strategies will be repeated, and `belief learning', which means that players keep track of what other people have done to figure out what those people will do in the future, then they choose strategies which will give the biggest payoff if their guesses are right. These two learning theories were thought to be different for about 50 years. It will be shown that they are actually special kinds of a single kind of learning, `experience weighted attraction' learning (EWA 1.0).

The second part of this talk establishes EWA 1.0's potential for prescriptive use by estimating its economic value. Since its economic value is positive, EWA 1.0 fails the so-called "publishability test." We fix this problem along with several other "user complaints" and develop an extended version of the theory called EWA 2.0. The EWA 2.0 model allows for both adaptive and sophisticated players and includes Nash equilibrium and Quantal Response Equilibrium as special cases. We will also briefly discuss how EWA 2.0 can be extended to explain strategic teaching/manipulation and reputation phenomena in practice.

Hal Arkes
Does Articulating the Basis of A Decision Improve or Impair the Decision?

Wilson, Schooler, and their colleagues have shown that considering the bases of one's decision may result in poorer decisions than if one merely chooses without any forethought. This phenomenon has been demonstrated whether one uses experts' opinions as the gold standard or the decision maker's own satisfaction with the choice. I will report on a naturalistic field study using high school students' college choice as the testbed for the Wilson-Schooler hypothesis. I will try to reconcile my surprising results with those of Wilson and Schooler.

Ken Hammond
Distinguishing Two Varieties of Cognitive Competence Will Improve Forecasts and Judgments Under Stress

I briefly indicate the strong advantages of the explicit recognition of the metatheories of correspondence and coherence for the field of judgment and decision making (explained in detail in "Human Judgment and Social Policy", Hammond, 1996). Failure to recognize them has led to researchers finding themselves in separate camps and the consequent obstruction of progress because two varieties of cognitive activity become confused. When recognized, however, the complementary nature of these metatheories will lead to a constructive convergence that will advance everyone's interests. Moreover, when recognized,the distinction can be used to improve forecasts, advance our study of judgments under stress, and even help us to understand the political judgments induced by the constitutional separation of powers.

Ravi Dhar
The Effect of Forced Choice on Choice

Whereas most academic and industry studies of consumer decision making involve forced choice (i.e., participants are told to choose one of the presented product/service alternatives), buyers usually have the option not to select any alternative, and this no-choice option has by far the greatest share. An implicit assumption in the experimental practice of forcing choice is that the no-choice option draws proportionately from the various available alternatives, such that the qualitative conclusions are unaffected. However, we propose that the no-choice options competes most directly with alternatives that buyers tend to select when they are uncertain about their preferences but are forced to make a choice. Building on this general proposition, we show that the introduction of the no-choice option strengthens the attraction effect, weakens the compromise effect, and decreases the share of an option that is "average" on all dimensions. We also examine the mechanisms underlying the impact of having the option not to choose. These results are consistent with the notion that the no-choice option provides an alternative way of resolving difficult choices that is not available when subjects are forced to choose. We discuss the theoretical and practical implications of this research.

Richard Zeckhauser
Earnings management to exceed thresholds

Earnings provide important information for investment decisions. Thus executives - who are montiored by investors, directors, customers, and suppliers - acting in self interest and at times for shareholders, have strong incentives to manage earnings. We introduce behavioral thresholds for earnings management. A model shows how thresholds induce specific types of earnings managemnet. Empirical explorations identify earnings management to exceed each of three thresholds: report positive profits, sustain recent performance, and meet analysts' expectations. The positive profits threshold proves predominant. The future performance of firms suspect for boosting earnings just across a threshold is poorer than that of control group firms.

Craig Fox
Familiarity Bias and Belief Reversal in Relative Likelihood Judgment

People are often called on to make an assessment of the relative likelihood of events (e.g., which of two investments is more likely to perform at least as well as the market?). The ordering of beliefs over events can also be established by assessing the relative likelihood of their complements (e.g., which of the two investments is more likely to underperform the market?). Probability theory and most descriptive models of judged probability, including support theory (Tversky & Koehler, 1994; Rottenstreich & Tversky, 1997), assume that belief orderings over events and their complements should mirror each other (e.g., P(A) * P(B) iff P(not-A) * P(not-B)). This principle is violated in several surveys in which we asked people to assess the relative likelihood of familiar versus unfamiliar events. In particular, respondents are biased to view more familiar events (and their complements) as more likely than less familiar events (and their complements). Our data show that the proportion of subjects judging a familiar event as more likely than an unfamiliar event exceeds the proportion of subjects rating the complement of the unfamiliar event more likely than the complement of the familiar event. An identical pattern was observed in studies in which subjects indicated which of two events they thought was less likely. Further studies suggest that the familiarity bias is less pronounced among subjects who are asked to judge the probability of each event rather than which event is more likely, and a greater proportion of subjects rate the more familiar event as more likely than assign a higher probability to that event. These patterns can be construed as belief reversals, analogous to the preference reversal phenomenon in decision making. The data are consistent with a contingent weighting model in which the process of judging relative likelihood biases attention toward evidence supporting focal hypotheses (and away from evidence supporting complementary hypotheses). Because it is easier to recruit evidence supporting familiar events than unfamiliar events, this skewed attention causes both familiar events and their complements to be judged more likely, on average, than unfamiliar events and their complements.

Some talks given last semester

11/2: Maurice Schweitzer, Wharton School, OPIM
The Impact of Alcohol on Negotiator Behavior: Experimental Evidence

11/9: Max Bazerman, Northwestern University
Explaining How Preferences Change Across Joint Versus Separate Evaluation

11/16: Chris Hsee, University of Chicago
The Evaluability Hypothesis and its Implications for Consumer Behavior

11/23: JDM conference

11/30: Thanksgiving

12/7: Christian Schade
Consumers' Purchasing Decisions on Warranties: Experimental Results on Framing Effects and Personality Differences

Maurice Schweitzer
The Impact of Alcohol on Negotiator Behavior: Experimental Evidence

Alcohol influences many of the same decision processes required for negotiations. This work investigated effects of a moderate amount of alcohol on negotiator behavior and negotiated outcomes. We conducted two experiments involving a negotiation exercise that paid M.B.A. subjects in proportion to the number of points that they earned. Results from study 1 revealed that inebriated dyads reached less integrative agreements than sober dyads. Study 2 investigated process differences between inebriated and sober negotiators. In this study inebriated negotiators were paired with sober negotiators and their negotiations were tape recorded, transcribed, and coded. Results form this study revealed that inebriated negotiators used more aggressive tactics and made more mistakes. Further, across both studies we found that inebriated negotiators were unaware that alcohol had affected their negotiation.

Max Bazerman
Explaining How Preferences Change Across Joint Versus Separate Evaluation

This paper examines how preferences for outcomes change across joint versus separate evaluation of alternatives. In joint evaluation, two (or more) options are presented and evaluated simultaneously. In separate evaluation, each option is presented and evaluated separately. We review a growing body of evidence demonstrating this type of preference shift and discuss how it is different from existing biases and preference reversals documented in the literature. We then review and integrate three competing explanations for this type of preferential inconsistency.