How do People Learn from Not Being Caught? An Experiment

How to cite

Zur, Tom. 2022. "How do People Learn from Not Being Caught? An Experiment." AEA RCT Registry. AEARCTR-0008992. https://doi.org/10.1257/rct.8992-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Tom Zur*

* Primary investigator

Registry ID
AEARCTR-0008992
DOI
https://doi.org/10.1257/rct.8992-1.0
First registered
Last updated
Status
In development
Trial dates
2022-02-19 – 2022-04-22
Countries
Private
JEL codes
Law and Economics; Illegal Behavior and the Enforcement of Law
Registry record
http://www.socialscienceregistry.org/trials/8992

Abstract

The law and economics literature has long posited that apprehending and imposing sanctions on individuals who commit offenses generates both general deterrence, aimed at the general population, and individual deterrence, aimed at the specific offender. While these general effects have been observed in various contexts, the empirical legal literature has failed to identify the mechanisms through which this learning process occurs. This paper will attempt to start filling this gap by examining, experimentally, whether potential repeat offenders (e.g., speeding, tax evasion, DUI) learn about the probability of detection from being caught at a different level of accuracy compared to learning from not being caught.

Experimental design

In each of the two stages of the experiment, the participants will be asked to perform a simple task of rolling a virtual six-sided fair dice after reporting their guess of its outcome. Participants' payment will be comprised of a flat participation fee, and bonus payment, based on their report on whether they guessed correctly, except for when they will be randomly audited, in which case falsely reporting a successful guess will lead to a lower payoff, compared to their payoff if truthfully reporting an unsuccessful guess.

Design details

Randomization method
Randomization will be done by the computer.
Randomization unit
Individual
Clusters
400~ Mturk workers.
Observations
400~ Mturk workers.
Arms
~200 control (audited) and ~200 treatment (not audited)
Primary outcomes
The change in the estimation of the probability of being audited between those previously audited and those not previously audited.
Secondary outcomes
The decision whether or not to report truthfully a guess of a dice roll's outcome

Cited by

No citation data available for this registration yet.

Keywords