Reference Dependence and Human-Robot Interaction
Last registered on August 09, 2018

Pre-Trial

Trial Information
General Information
Title
Reference Dependence and Human-Robot Interaction
RCT ID
AEARCTR-0003170
Initial registration date
August 07, 2018
Last updated
August 09, 2018 1:42 AM EDT
Location(s)

This section is unavailable to the public. Use the button below to request access to this information.

Request Information
Primary Investigator
Affiliation
Cornell University
Other Primary Investigator(s)
PI Affiliation
Cornell Univesitry
PI Affiliation
Cornell University
PI Affiliation
Hebrew University of Jerusalem
PI Affiliation
Hebrew University of Jerusalem
Additional Trial Information
Status
On going
Start date
2018-07-23
End date
2018-09-30
Secondary IDs
Abstract
We test experimentally whether agents are loss averse when they compete in a real-effort (real time/simultaneous) competition against a Robot that runs in randomly varying speeds. Human and Robot count the number of ‘G’ letters in their texts and place ‘Green’ block in the corresponding bin. Following Gill-Prowse (2012), The probability of winning a predetermined prize increases/decreases by 1% per unit difference between Human and Robot points. We hypothesize that an agent who is loss averse around her endogenous expectations-based reference point will respond negatively to the Robot’s performance level.
External Link(s)
Registration Citation
Citation
Dreyfuss, Bnaya et al. 2018. "Reference Dependence and Human-Robot Interaction." AEA RCT Registry. August 09. https://www.socialscienceregistry.org/trials/3170/history/32835
Experimental Details
Interventions
Intervention(s)
Intervention Start Date
2018-07-23
Intervention End Date
2018-09-30
Primary Outcomes
Primary Outcomes (end points)
Number of correct placements
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
number of mistakes, number of correct placement corrected for mistakes, etc
Secondary Outcomes (explanation)
We are still developing the exact measures.
Experimental Design
Experimental Design
Human and Robot count the number of ‘G’ letters in their texts and place ‘Green’ block in the corresponding bin. For each correct placement they get 1 point. For incorrect arrangement there is a penalty of 10 seconds during which they cannot earn points. The probability of winning the prize increases/decreases by 1% per unit difference in their points. Participants can click on ‘Stop’ to quit the round. If participants choose to stop, the robot’s final score will be equal to its projected final score, and the participants’ final score will be their current score.
Experimental Design Details
Not available
Randomization Method
Robot's speed is randomly chosen by the computer
Randomization Unit
Round
Was the treatment clustered?
No
Experiment Characteristics
Sample size: planned number of clusters
600 subject-rounds.
Sample size: planned number of observations
600 rounds (= 10 rounds X 60 subjects).
Sample size (or number of clusters) by treatment arms
600 in total; randomization is at the subject-round level, and the main specification includes the main regression from Gill-Prowse (2012) Table 2.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Cornell University, Office of Research Integrity and Assurance
IRB Approval Date
2018-06-20
IRB Approval Number
1806008045