Illusion of Control in Mental Models

Last registered on August 03, 2022

View Trial History

Pre-Trial

Trial Information

General Information

Title

Illusion of Control in Mental Models

RCT ID

AEARCTR-0009831

Initial registration date

August 01, 2022

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

August 03, 2022, 3:05 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country

United States of America

Region

Primary Investigator

Name

Qiaofeng (Tony) Fan

Affiliation

Stanford University

Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status

In development

Start date

2022-08-01

End date

2022-08-31

Keywords

Behavior

Additional Keywords

mental models, belief formation

JEL code(s)

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

I test for an illusion of control effect using direct elicitations of experimental participants' mental models and identify targeted learning as an economic mechanism driving the bias. Participants learn from a multivariate dataset before taking actions involving one variable and stating their beliefs about the data-generating process (i.e., relationships between variables). Illusion of control refers to a tendency to believe one's own action variable as outcome-relevant. Different treatments allow me to disentangle targeted learning about the action variable as a potentially important mechanism that organically generates the effect.

External Link(s)

Registration Citation

Citation

Fan, Qiaofeng (Tony). 2022. "Illusion of Control in Mental Models." AEA RCT Registry. August 03. https://doi.org/10.1257/rct.9831-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Participants learn from a multivariate dataset before taking actions involving one randomly selected variable and stating their beliefs about the data-generating process (i.e., relationships between variables).

Intervention Start Date

2022-08-01

Intervention End Date

2022-08-31

Primary Outcomes

Primary Outcomes (end points)

Propensity to state different variables (attributes) as outcome-relevant, in multiple choice or elicited model (function) forms; probabilistic beliefs about the chances that different variables (attributes) are outcome-relevant.

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Measures of attention to different variables (attributes) during learning

Secondary Outcomes (explanation)

Attention is measured by mouse clicks on different variables (attributes) during learning.

Experimental Design

See intervention

Experimental Design Details

Randomization Method

Randomization done by a computer

Randomization Unit

Individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

No cluster

Sample size: planned number of observations

2,000 participants

Sample size (or number of clusters) by treatment arms

1,000 participants assigned to the Targeted Learning treatment, in which the action variable is known to the participants before the learning stage. This further breaks down to 500 participants randomly assigned an action variable that is in fact outcome-relevant and 500 participants randomly assigned an action variable that is in fact outcome-irrelevant.

1,000 participants assigned to the Untargeted Learning treatment, in which the action variable is unknown to the participants during the learning stage and only revealed immediately before the action task. This further breaks down to 500 participants randomly assigned an action variable that is in fact outcome-relevant and 500 participants randomly assigned an action variable that is in fact outcome-irrelevant.

During the belief elicitation stage, half of the 2,000 participants are cross-randomized to state the probabilistic beliefs first, and the other half are cross-randomized to guess the exact function first.

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

~10 percentage points for the binary outcomes (state a variable as outcome-relevant or not, in different forms); ~6 percentage points for the probabilistic beliefs (about the chance that a variable is outcome-relevant).

Supporting Documents and Materials

IRB