Forecasting the Results of Experiments: Piloting an Elicitation Strategy

How to cite

Otis, Nicholas; DellaVigna, Stefano; Vivalt, Eva. 2019. "Forecasting the Results of Experiments: Piloting an Elicitation Strategy." AEA RCT Registry. AEARCTR-0005211. https://doi.org/10.1257/rct.5211-2.0

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

Nicholas Otis* , Stefano DellaVigna , Eva Vivalt

* Primary investigator

Registry ID
AEARCTR-0005211
DOI
https://doi.org/10.1257/rct.5211-2.0
First registered
Last updated
Status
Completed
Trial dates
2019-12-18 – 2020-01-09
Countries
United States of America (N/A, Online Survey)
JEL codes
C81, C93, D03
Registry record
http://www.socialscienceregistry.org/trials/5211

Abstract

This study involves collecting forecasts of results of three field experiments. We randomly vary four features when eliciting forecasts: (1) small versus large reference values in an example preceding the predictions; (2) whether forecasts are in standard deviations or raw units; (3) slider versus text-entry responses; and (4) small versus large slider bounds.

Intervention

(1) small (0.1 SD) versus large (0.3 SD) reference values in an example preceding the predictions; (2) whether forecasts are in standard deviations or raw units; (3) slider versus text-entry responses; and (4) small (+-0.5 SD) versus large (+-1.0 SD) slider bounds.

Experimental design

We collect forecasts of the results of three experiments preliminarily accepted (before October 2019) that publicly posted their proposals, but that had not yet released any experimental results. We draw on a pool of academics, researchers, and practitioners. Consenting participants provide predictions for results of up to three studies. At the start of the survey, participants are randomized across the four conditions described in the “Interventions” section above.

Design details

Randomization method
Randomization takes place automatically when the survey is initiated.
Randomization unit
Randomization is at the individual level.
Clusters
We plan to collect forecasts from at least 80 individuals.
Observations
We plan to collect forecasts from at least 80 individuals.
Arms
Our main treatments will have N/2 participants, where N is our total sample size.
Primary outcomes
Forecasts of experimental results for three studies preliminarily accepted as Registered Reports to the Journal of Development Economics.

Materials

Papers from this trial

  1. DellaVigna, S., Otis, N. and Vivalt, E., 2020. Forecasting the Results of Experiments: Piloting an Elicitation Strategy. AEA Papers and Proceedings (Vol. 110, pp. 75-79). https://www.aeaweb.org/articles?id=10.1257/pandp.20201080

Cited by

No citation data available for this registration yet.

Keywords