Spring 2019

Choosing for others: The influence of incentives and similarity on the decision process

Research on surrogate decision making has established a robust and consistent finding: people are not very good at choosing what another person would choose for themselves. However, if surrogates are given the chance to learn another’s preferences, can they do so? Moreover, why do surrogates choose the “wrong” options for their recipients? Across two studies, subjects observe the choices made by others and choose for them. We find that people are able to learn the preferences of others, but whether they actually choose in line with their recipient’s preferences is dependent on the incentives and their similarity to the recipient.
 

Ian Adams and Nicole Sintov (School of Environment and Natural Resources)

Carbon footprint feedback framing influences the cognitive and emotional pathways to consumption behavior

Although feedback (e.g., carbon footprint) can be effective in reducing consumption, little research examines the cognitive and emotional processes between feedback reception and pro-environmental behavior (PEB). Attribution theory can illuminate these processes. 397 adults received a positively-, neutrally-, or negatively-framed bogus feedback message. Frame impacted the attributions people made for feedback, which had consequences for experienced emotions. Pride promoted PEB following positive feedback. Guilt promoted PEB following neutral, but not negative feedback, the latter of which activated self-serving biases. Findings clarify the circumstances by which emotions influence PEB and can be applied to improve feedback interventions.

Shivangi Bhardwaj, Michael Silverstein, Pär Bjälkebring, and Ellen Peters (Psychology)

Numeracy and the Interpretation of Medical Information: How to Avoid Bias in Low Numerate populations

This study investigates whether individual differences in numeric abilities moderate the effect of certainty or pseudocertainty (e.g., 100% effectiveness). Participants (N=434) read a statement from the Center for Disease Control and Prevention about a cancer-causing virus. They were randomized to one of two conditions (percentages only or percentages and net-effectiveness) and asked about the effectiveness of five potential vaccines. After each vaccine, participants rated their preference for the vaccine on a 7-point scale. Strong pseudocertainty biases emerged. Our results suggest a way to remedy this bias in vaccine decisions by doing the math for participants. However, it mainly benefitted the highly numerate.

Pär Bjälkebring and Ellen Peters (Psychology)

Negative Affect Influence Objective Numeracy

This project investigated whether negative affect was related to temporary decreases in objective numeracy. 500 mTurk participants were pre-tested on objective numeracy and answered questions about their current affective state (positive and negative affect), and also completed a 5-item numeracy scale. Participants who felt more negative affect on the day of testing scored lower compared to those with less negative affect. Positive affect had no relationship to numeracy scores. Further, we examined whether education, subjective numeracy or cognitive ability could protect from the impact of negative affect on objective numeracy. Implications for decision making are discussed.

Breanne Butters (Department of Design)

Debunking myths about contraceptive safety among women in Kingston, Jamaica: Randomized controlled trial

Myths about contraception side effects and safety are stopping women in Kingston, Jamaica from using contraception, specifically the IUD and implant. We designed and produced an animated video to debunk these myths. During the design phase, we conducted focus groups with women in Kingston to design with and for them to better understand their responses to the information we are designing. We conducted focus groups through Skype where the stakeholders reviewed and gave feedback on drawn storyboards and narration. Later I conducted an in person focus group in Kingston to get feedback on revised storyboards and some finished animation.

Ben Casner (Economics)

Learning While Shopping: An Experimental Investigation Into the Effect of Learning on Consumer Search

In many consumer search environments searchers do not know the distribution of prices in the market before they search. I conduct an experiment exploring the implications of search problems in an environment with learning about the distribution of payoffs. My results support theoretical predictions that learning results in declining reservation values, providing evidence that learning may be an explanation for declining reservation utility and recall seen in field data. However, while the theory predicts a myopic reservation value strategy, many subjects instead choose to set a high reservation value to learn about the distribution before adjusting based on their observations.

Rachael Gwinn (Psychology), Deniz Ozkardas (Psychology), Julie Golomb (Psychology), and Ian Krajbich (Psychology and Economics)

The Role of Attention in Value-Based Auditory Choices

We make visual, preference-based decisions, like what to wear, every day. However, we don’t live in a purely visual world. We frequently make auditory decisions, like what song to listen to, yet these decision processes are not as well understood. To investigate such processes, including the role of attention, we had subjects complete a binary choice task. On each trial they controlled which of two songs played at any moment and ultimately decided which song they preferred. We found striking similarities between vision-based and auditory-based decisions, namely that attention played an important role in determining which song was chosen.

Ritesh Jain and Kirby Nielsen (Economics)

A Systematic Test of the Independence Axiom

We investigate the independence axiom, a central tenet of expected utility theory. We design a laboratory experiment to test this axiom on the entire probability simplex. While the existing literature primarily has shown violations of Independence where individuals favor certainty (the "certainty effect"), we find that the reverse is broadly more common---violations of Independence are most common when individuals first prefer risk over certainty. This result runs contrary to commonly-held "folk wisdom" regarding Independence violations and suggests a new direction for theoretical models.

Osub Kwon (Economics)

Sequential Ascending Clock Auction and the Declining Price Anomaly: An Experimental Study

This paper experimentally investigates the bidders’ behavior in two-stage sequential ascending clock auction (i.e., English auction) design. The equilibrium given by Milgrom and Weber (2000) predicts that the prices of the two goods sold are exactly the same regardless of the value realization. We, in addition, show that this equilibrium can be obtained in our setting by two rounds of iterated deletion of weakly dominated strategies. Despite such strong prediction, we find little support for equilibrium-like behavior. Instead we find “on average” the same prices in three-bidder auctions and the declining prices in five-bidder auctions.

Amanda Montoya, Stephanie Smith, and Pär Bjälkebring (Psychology)

Numeracy-Related Responding to Numeric Self-Report Scales

Self-report scales are used throughout judgment and decision making research. However, response style research suggests that there are individual differences in how respondents use response scales. Our research investigates how numeracy (an individual’s ability to understand and use math) predicts how individuals use numeric scales. Using Mturk samples, we explore how numeracy predicts response style, and how number of response options and response labels moderate the relationship between numeracy and response style. We investigate patterns of responding based on numeracy, and discuss the potential impact this could have on the conclusions of judgment and decision making research.

Scott Schwenter and Kendra Dickinson (Department of Spanish and Portuguese)

(Not) Choosing Negative Concord

Negative concord, the use of multiple negative words in one sentence to express only one negation, is common across languages, albeit stigmatized in English (e.g. I don't have nothing). In Brazilian Portuguese (BP), neither concord, which is normative (e.g. Eu não tenho nada), nor lack of concord (e.g. Eu tenho nada), are stigmatized. How do BP speaker/writers choose between them? We present evidence from three sources: acceptability judgments, a forced-choice survey, and qualitative analysis of naturally-occurring examples. We show that BP lack of concord is constrained by the negative word in question, the verb, and the person/number of the subject.

Brittany Shoots-Reinhard and Ellen Peters (Psychology)

Numeric ability and intuitive understanding of probabilistic risk

Higher numeric ability generally predicts better decision making, but some probabilistic concepts (e.g., cumulative probability) are beyond even those high in numeric ability. In this study, participants answered questions about distracted driving and crash risks. The highly numerate were more likely to give answers that were logically consistent. For example, when given risk information that driving distracted doubles 1-year crash risk, the highly numerate were more likely than the less numerate to give cumulative risk estimates with the same ratio, even though their estimates were incorrect. Thus, the highly numerate may have better intuitive understanding of risk.

Michael Silverstein, Par Bjalkebring, and Ellen Peters (Psychology)

Developing an Adaptive Latent Trait Measure for Objective Numeracy

The ability to incorporate numeric information, called numeracy, has shown to be important for judgments and decision making. Existing measures of numeracy vary greatly in difficulty and length making them more appropriate for measuring different levels of numeric ability. The goal of this study is to address the need for a new measure by creating an adaptive scale that accurately captures numeracy across the range of abilities.  Eighty-four new numeracy items are calibrated using a two parameter logistic model that estimates the difficulty and discriminability of each question, and are then formed into an adaptive test based on those parameters.

Hayley Svensson, Brittany Shoots-Reinhard, Tyler MacDonald, and Ellen Peters (Psychology)

Numeracy and the Strength of Monetary versus Non-Monetary Incentives on Task Performance

While incentives have traditionally been used to motivate people, recent research suggests small monetary incentives may be de-motivating (Heyman & Ariely, 2004). To test whether numeric ability would influence the extent to which small monetary incentives reduce motivation, we examined the effects of measured numeric level, incentive type (monetary or non-monetary) and level (small or large) on effort exerted to perform a simple, clerical-like task. Our results suggest that high amounts of a monetary incentive may actually be de-motivating to the receiver, and this did not depend on level of numeric ability.