Keep your eyes on the prize: How to make penalty contracts work 2

Last registered on October 17, 2023

View Trial History

Pre-Trial

Trial Information

General Information

Title

Keep your eyes on the prize: How to make penalty contracts work 2

RCT ID

AEARCTR-0007650

Initial registration date

May 10, 2021

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

May 10, 2021, 11:45 AM EDT

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

Last updated

October 17, 2023, 8:41 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country

Switzerland

Region

Country

United States of America

Region

Primary Investigator

Name

Andrea Essl

Affiliation

University of Bern

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Frauke von Bieberstein

PI Affiliation

University of Bern

Contact Investigator

PI Name

Stefanie Jaussi

PI Affiliation

University of Bern

Contact Investigator

Additional Trial Information

Status

Completed

Start date

2021-05-01

End date

2022-06-30

Keywords

Behavior, Labor

Additional Keywords

Online experiment, Contract framing, Penalty, Bonus, Loss aversion, Salience

JEL code(s)

Secondary IDs

Prior work

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

Abstract

This online experiment examines the influence of payment modalities on productivity when working under framed incentive contracts. We investigate the effect of payment visualization to make the payment situation more salient to the participants. We employ a 2x2 design with four different treatments varying the incentive contract (bonus vs. penalty) and the visualization (visualized vs. description only). Participants work on a real effort task. At the end, one task is randomly selected for payment. If the selected task is solved correctly, participants in the bonus treatment receive an additional payment. In the penalty treatment, participants receive a higher payment upfront, but can lose part of it, if the selected task is solved incorrectly. Both contracts are payoff-equivalent. We vary whether the payment options are visualized on screen with dollar bills and coins, or are only explained using a verbal description.

External Link(s)

Registration Citation

Citation

Essl, Andrea, Stefanie Jaussi and Frauke von Bieberstein. 2023. "Keep your eyes on the prize: How to make penalty contracts work 2." AEA RCT Registry. October 17. https://doi.org/10.1257/rct.7650-1.2

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Online experiment on Amazon’s Mechanical Turk (MTurk). 2x2 between-subject design. Participants are randomized at the individual level. In the main stage of the Bonus treatments the payment specifies a base pay plus a bonus if the individual solves the randomly selected task correctly. In the main stage of the Penalty treatments the payment specifies a higher base pay including the incentive pay, which has to be paid back (penalty) if the individual solves the selected task incorrectly. Both contracts are payoff-equivalent. In the Visualization treatments images of USD bills / coins are shown on screen. In the Verbal treatments a verbal explanation without visualization of the possible outcomes is used.

Intervention Start Date

2021-05-17

Intervention End Date

2021-12-30

Primary Outcomes

Primary Outcomes (end points)

Productivity measured as the number of correct answers in the real-effort task.

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Skipping behavior measured as the number of skipped tasks in Part 3; number of incorrectly answers in the real-effort task

Secondary Outcomes (explanation)

Experimental Design

The experiment consists of three parts and a questionnaire. In Part 1, we assess loss aversion based on a lottery task. In Part 2, participants receive a fixed wage and work on a real effort encryption task (Encryption task with Double Randomization). Participants have to encode letters into numbers and can proceed with the next word only if they encoded all letters correctly. Effort is measured as the number of solved words. In Part 3, the main stage of the experiment, participants work on the same encryption task. This time, payment is performance based and tied to a randomly selected task. Depending on the treatment, the payment is framed as a bonus or penalty and the payment is visualized or not visualized (see above).

Experimental Design Details

Randomization Method

computer (online experiment)

Randomization Unit

individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

The number of participants planned for the experiment is 1440 people completing the study. We advertise for 1600 participants to take into account individuals excluded due to the restrictions below.

We will exclude subjects who do not complete the MTurk study within 60 minutes of starting and who are not approved for any other reason (e.g. not having a valid MTurk ID);

We will analyze the data both, for all participants and for those participants, who skip all or most of the tasks solve at least one task correctly.

In addition, and in line with previous research, we will analyze the data once excluding those who make inconsistent decisions in the loss aversion test and once including these subjects.

The study will be kept open on MTurk until 1600 subjects have completed the assignment.

Sample size: planned number of observations

The number of participants planned for the experiment is 1440 people completing the study. We advertise for 1600 participants to take into account individuals excluded due to the restrictions below.

Sample size (or number of clusters) by treatment arms

About 400 participants per treatment (about 360 participants per treatment after excluding participants as described above)

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

Based on a two-sided Wilcoxon-Mann-Whitney test, an error probability of 0.05, and a power of 0.80, we require about 360 participants per treatment to detect an effect of Cohen’s d of 0.21. This is comparable to the treatment effect identified in a pilot study.

Supporting Documents and Materials

IRB