Behavioral Attenuation

How to cite

Oprea, Ryan; Enke, Benjamin; Graeber, Thomas; Yang, Jeffrey. 2024. "Behavioral Attenuation." AEA RCT Registry. AEARCTR-0013308. https://doi.org/10.1257/rct.13308-1.3

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

Ryan Oprea* , Benjamin Enke , Thomas Graeber , Jeffrey Yang

* Primary investigator

Registry ID
AEARCTR-0013308
DOI
https://doi.org/10.1257/rct.13308-1.3
First registered
Last updated
Status
In development
Trial dates
2024-06-18 – 2024-12-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/13308

Abstract

This research aims to study behavioral attenuation: the idea that, as a result of noisy information processing, people’s decisions are generally insufficiently elastic to variation in economic primitives. To do this, we run 30 parallel treatments, covering a broad range of economic decisions ranging from preference elicitations to belief updating to generic optimization problems, involving risk, time, consumption-savings, effort supply, taxes, fairness, giving, energy savings, information disclosure, prediction and inference, in both individual decisions and strategic games. In each of these domains (treatments), we vary one parameter, within-subject over a series of rounds, that has a predicted monotonic relationship to the subject’s decision. After each decision we elicit the subject’s “cognitive uncertainty” (CU) – the likelihood with which the subject believes they failed to make an optimal decision. Our main hypothesis is that (i) subjects with higher CU will make decisions that are less responsive to the parameter varied in the experiment and (ii) in tasks with objectively correct decisions, subjects will be less responsive than optimal. Our subjects are recruited from Prolific, restricting to subjects from the US and UK.

Intervention

Each subject is assigned one of 30 tasks (reviewed in the experimental design field below), given instructions and asked to make a decision that will influence some outcome over a series of rounds (specified in the pre-analysis plan). Within each task, the main intervention is that a parameter relevant to the decision is varied across rounds (reported for each task in the pre-analysis plan). Subjects are asked for an estimate of the likelihood with which they made an optimal decision (exact wording varies across tasks/treatments but is provided in each case in the supporting documents).

Experimental design

Each experiment is drawn from a large list of different behavioral decision paradigms listed in the Task List document. In each case, the participant completes up to 11 rounds of the same task. In some tasks, participants’ decisions affect their own or each other’s payments, as described in the Task List document. Across rounds, we vary the task parameters. An example is a savings decision, where the subject decides how much of their budget to consume or save as a function of the interest rate. Across rounds, the interest rate varies. After a subject makes their decision, we elicit (without incentives) the likelihood with which they believe they made an optimal decision. Subjects receive a participation payment (varied across tasks and reported in our pre-analysis plan) and in some tasks payment bonuses based on their and/or others’ decisions.

Design details

Randomization method
Randomization of payments are conducted by the experimental software where applicable.
Randomization unit
The order in which individual subjects are assigned parameters over the course of rounds is randomized at the subject level.
Clusters
6000 participants
Observations
130,000 (11 rounds per subject in 29 tasks and 6 rounds in one task; each round includes both a choice and a cognitive certainty elicitation)
Arms
Please see Sample Size document.
Primary outcomes
The primary outcome for each subject is a sequence of decisions, the nature of which varies across treatments/tasks and is described in detail for each treatment below.
Secondary outcomes
The secondary outcome for each subject and decision is the subject’s confidence: the subject’s own estimate of the likelihood with which they made an optimal decision (“cognitive uncertainty”).

Cited by

No citation data available for this registration yet.

Keywords