Research Summary
Research Summary
Self-environment relationship and its effect on decisions under risk and uncertainty
Description
My research seek to better understand the main cognitive and social abilities that guide our judgments, and the ways they interact with aspects of the situation to shape humans' decisions. It is currently comprised of three related subjects:
- The effect of losses on risk taking.
Previous research shows that many interesting deviations from rational choice can be explained with the assertion that losses loom larger than gains (Kahneman & Tversky, 1979). In one series of studies, my colleagues and I try to clarify the conditions that give rise to this "loss aversion" bias. Our results suggest that loss aversion is less general than assumed by some of the popular applications. For example, decision makers do not exhibit loss aversion when they rely on personal experience (Erev, Ert, & Yechiam, 2008), nor are they loss averse while choosing among low stakes lotteries based on precise description of the payoff distributions (Ert, & Erev, 2008, 2009). Loss aversion in these settings appear to emerge in two conditions: (1) When the participants are asked to make many decisions (more than 10) without feedback, and (2) when the safe prospect is explicitly framed as the status quo.
We currently explore potential accounts for the differences in behavior across situations. According to one explanation, choice among decision rules is a product of case-based reasoning, and some tasks (and/or frames) are more likely to trigger past experiences in which risk taking has been counter-productive. - Decision Making in Social Contexts
Many situations involve the possibility of learning from other people's behavior even when no communication is taking place. For example, exposure to other people's decisions in facilitates informational cascades when uncertainty is high (Raz & Ert, 2008), and facilitate risk taking when risks are associated with rare losses (Yechiam, Druyan, & Ert, 2008). In other situations, effective interaction with other people requires that their motives should be taken into account. For example, people might have a tendency to reject attractive prospects when they are suggested by another person, but prefer these offers when they are presented as an abstract choice tasks (Ert & Erev, 2008). This tendency might reflect people's attempt to protect themselves from "lemon" products (Ackerlof 1970). - Self-environment relation in decisions from experience (and from description)
A recent and constructive distinction in decision-making under risk and uncertainty relates to the difference between decisions that are based on descriptions of the potential outcomes and their corresponding probabilities (e.g., as while playing a roulette) and decisions that are based on experienced feedback (e.g., playing a slot machine). Previous studies show robust differences between these two domains (e.g., Hertwig et al., 2004). My research explores both consistent constructs of individual risk taking in each domain (e.g., Ert & Yechiam, 2010; Ert, Yechiam, & Arshavsly, 2009), as well as the potential effect of environmental factors that are likely to be relevant in natural settings. Examples include choice between multiple outcomes (Ert & Erev, 2007), and choice in dynamic settings (Biele, Erev, & Ert, 2009). - The Value of Quantitative Models and Aggregation Methods in Predicting Choice Behavior
This research venue explores the value of quantitative models and/or aggregation methods (e.g., polls, prediction markets) in predicting behavioral data. In a recent research we (Ido Erev, Al Roth, and myself) organized an open choice prediction competition aimed to evaluate the value of descriptive models in predicting risky choices. We suggest the competition method as a paradigm that can help in facilitating clear and useful descriptive models.
The details of the individual prediction competition are available here: http://tx.technion.ac.il/~eyalert/Comp.html.
And here is a link to a new choice prediction competition in Market Entry games: http://sites.google.com/site/gpredcomp/home. This competition is currently open for registration.