On the measurement and estimation of risk preferences: From the lab to the field

How to cite

Vieider, Ferdinand; Panin, Amma. 2017. "On the measurement and estimation of risk preferences: From the lab to the field." AEA RCT Registry. AEARCTR-0002349. https://doi.org/10.1257/rct.2349-1.0

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

Ferdinand Vieider* , Amma Panin

* Primary investigator

Registry ID
AEARCTR-0002349
DOI
https://doi.org/10.1257/rct.2349-1.0
First registered
Last updated
Status
In development
Trial dates
2017-07-24 – 2017-12-15
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/2349

Abstract

This document constitutes a pre-analysis plan for an experiment to be conducted in the field in India, starting in late July/early August 2017. Comparison samples are to be obtained with students in Oxford, and budget allowing, with students in Pune, India. While we tried to present a complete discussion, some of the hypotheses remain less elaborate than we had hoped for. This is due on the one hand to time pressures coming from project funding, which did not allow us to further delay the experiments. We thus had to leave this plan incomplete in order to be able to submit before seeing any signifiant amounts of data coming out of the experiment. On the other hand, this same time pressure did not allow us to fully develop our econometric approach, so that some hypotheses are formulated only verbally, but do not yet have a more precise mathematical formulation.

Intervention

There is no treatment intervention. The study consists of the experimental measurement of risk preferences using a variety of elicitation tasks.

Experimental design

We aim to measure risk preferences using a variety of measurement tasks. These include four `blocks' of comparable tasks, each one of which contains seven unique choice lists (plus repetitions and variations). In addition, we use three more traditional choice lists. Measurements are repeated over time.

Design details

Randomization method
There is no randomization involved.
Randomization unit
We target one whole village in southern India.
Clusters
275 households in the field; 100 subjects at each of the two universities.
Observations
For each subject, we will obtain 96*2 measurements, plus a variety of survey questions.
Arms
n/a
Primary outcomes
The main dimensions to be investigated are: measurement reliability; predictive validity; noise in measurement; modelling of preferences; endogenous reference points activated by task characteristics.

Cited by

No citation data available for this registration yet.

Keywords