Simple Elicitation of Risk Attitudes, SERA

How to cite

Garagnani, Michele. 2020. "Simple Elicitation of Risk Attitudes, SERA." AEA RCT Registry. AEARCTR-0005763. https://doi.org/10.1257/rct.5763-2.0

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

Michele Garagnani*

* Primary investigator

Registry ID
AEARCTR-0005763
DOI
https://doi.org/10.1257/rct.5763-2.0
First registered
Last updated
Status
Completed
Trial dates
2020-05-10 – 2020-05-31
Countries
Switzerland (Zurich)
JEL codes
C91, D81
Secondary IDs
SERA
Registry record
http://www.socialscienceregistry.org/trials/5763

Abstract

Attitudes toward risk are one of the primitive concepts in economics. Moreover, measuring risk preferences is critical for economic analysis and policy prescriptions alike. Therefore, it is only natural that economists have developed a variety of experimental methodologies. However, “risk elicitation is a risky business” (Friedman et al., 2014) with some of the current established methods being hard to understand for participants, difficulty to implement, or not able to capture attitudes well enough. In this study we propose an improved, simple but effective method to elicit on-line individuals’ attitudes toward risk, SERA. This new method sits between a fixed design (i.e., Multiple Price List, Holt & Laury, 2002) and a dynamically optimal design (i.e., BOARD, Wang, Filiba, & Camerer 2019) using an adaptive three questions procedure. Hence, it strikes a compromise among quality of the data, simplicity of implementation (for experimenters), and easiness to understand (for participants).
In an empirical validation we perform a within-subjects comparison of SERA with the most commonly used risk elicitation procedures (the MPL and the investment task) using as a metric of judgement the proportion of out-of-sample correctly predicted choices.

Intervention

SERA, a new method to elicit attitudes toward risk, is compared to established alterantive methods.

Experimental design

Within-subjects comparison of different risk-elicitation methods (SERA, MPL, investment task, qualitative risk assessment).

Design details

Randomization method
The order of the 36 lotteries is randomized and fixed to be the same for each participant. One out of all decision situations will be randomily used to determine the earnings of subjects.
Randomization unit
Individuals are the independent level of analysis in this study.
Clusters
N=400.
Observations
N=400.
Arms
N=400 is the desired sample size. Subjects will be recrited until the desired size is reached. However, because the experiment is performed on-line, more subjects will be invited to participate in the experiment. Subjects exceeding the desired sample will be included in the analysed sample.
Primary outcomes
Proportion of correctly predicted choices out-of-sample across 3 different elicitation methods (SERA, MPL, investment task).
Secondary outcomes
Gender differences in risk attitudes. Correlation between the estimated risk attitudes across the 3 eliciation methods. Robustness of SERA to a within-subject repetition with re-scaled values.

Cited by

No citation data available for this registration yet.

Keywords