Jay Cizeski, Ph.D
Western Michigan University Department of Economics
Friendmann 5406
I am an Assistant Professor of Economics at Western Michigan University. My research uses controlled laboratory experiments in order to better understand how people think and make economic decisions.
My CV is available here.
In Fall 2026 I am teaching ECON 2010: Principles of Microeconomics. Current students can find the course website on WMU ELearning.
About Me
Job Market Paper
Research
Skeptical Exit in Games
Abstract:: People often face environments which they suspect may be manipulated. Deciding when to exit such environments due to skepticism is important, but difficult, requiring not only statistical inference but also strategic inference: when a manipulator may intentionally conceal their manipulation, the problem becomes a game, rather than a simple decision. With multiple potential sources of error, do people tend exit from these types of environments too soon, too late, or approximately when models predict, and which mechanisms matter? Using a laboratory experiment implementing a novel adversarial bandit design, I vary the possibility and source of manipulation, whether the extent of manipulation is revealed, and whether participants receive correct Bayesian posteriors to help make their exit decisions. This design isolates the effects of failures of statistical inference, strategic misprediction, and other regarding preferences in order to identify the sources of over- or under- exit, if they exist. I find evidence of neither over- nor under-exit: subjects’ exit decisions are largely consistent across treatments. This holds even though elicited beliefs substantially overestimate the likelihood of manipulation on average. These results contribute to literatures in biased statistical inference, games of hidden information, multi-armed bandits, and a growing social science of conspiratorial thinking.
Other Work
Messaging in Bargaining: To Name your Price or Not to Name Your Price?
Revise and Resubmit at Journal of Behavioral and Experimental Economics
Abstract: Using a laboratory experiment involving extended ultimatum bargaining games, I test how different forms of
messaging used by responders affect bargainer outcomes. In particular, I study whether responders perform
better by specifically claiming reserve prices (“explicit messages”) or rejecting offers without claiming a particular
reserve price (“implicit messages”). I find that while both forms of messaging and their combination
all significantly increase average responder payoffs relative to no messaging, there is no significant difference
in average responder payoffs between them - in short, it may not matter exactly how someone chooses to
negotiate through reserve price disclosure, so long as they signal at all. Analysis of subjects’ strategies reveals
that proposers increase their offers in response to the presence of implicit messaging and the size of explicit
messages (up to a point), indicating that proposers may treat messages with a limited degree of credibility.
Findings have implications for practical negotiation strategy and the design of bargaining institutions.
Paternalism: an Information Account
In Progress
Abstract: I consider the role of information quality differences between individuals on a person's decision to intervene in another's choice or not. I estimate and compare an “information premium” - a measure of how much stronger of a signal one needs in order to intervene, to examine conditions when people are more or less likely to do so. I find stark differences in information premia between decisions occurring before the potential paternalist knows their own and their target's signals compared to after, indicating a significant role of contingent thinking in this exercise.
Anchoring and Insufficient Adjustment Affects Most-Likely Event Estimation (with Jonathan Bogard)
In Progress
Abstract: In a series of experiments, we show that subjects fail to correctly answer a maximum-likelihood state-guessing exercise as a result of a spontaneous anchor on a salient central option. Subject responses are demonstrated to be in the majority incorrect, susceptible to alternative arbitrary anchors, and over-report the hypothesized arbitrary anchor in absence of a signal. Results have implications for belief elicitation methods.