Self-serving attribution biases economic actions

How to cite

Drobner, Christoph; Orhun, Yesim. 2023. "Self-serving attribution biases economic actions." AEA RCT Registry. AEARCTR-0010627. https://doi.org/10.1257/rct.10627-1.0

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

Christoph Drobner* , Yesim Orhun

* Primary investigator

Registry ID
AEARCTR-0010627
DOI
https://doi.org/10.1257/rct.10627-1.0
First registered
Last updated
Status
In development
Trial dates
2023-06-01 – 2024-12-31
Countries
Private
JEL codes
C91, D83, D84
Registry record
http://www.socialscienceregistry.org/trials/10627

Abstract

Empirical evidence shows that people attribute good outcomes to their own merits and bad outcomes to external factors that are beyond their control. We will test how this self-serving attribution bias shapes people's beliefs about meritocracy in society and affects downstream consequences for economic actions. We hypothesize that people who experience bad outcomes form more pessimistic beliefs about the meritocracy of society and as a consequence become less motivated to work hard. In contrast, overconfident people who experience good outcomes form more optimistic beliefs about the meritocracy of society and as a consequence become more motivated to work hard.

Intervention

The experiment consists of two parts and we randomly choose one part to determine payments.

In Part A, subjects perform a logic test with 10 questions and we elicit their beliefs about scoring in the top half in a group of 20 subjects, including themselves.

Subjects receive two payoffs in Part A. One payoff is provided by evaluator 1, and the other by evaluator 2. Exactly one of the two evaluators is a performance evaluator and exactly one of the two evaluators is a random evaluator. However, subjects are not aware of which evaluator corresponds to each role. The performance evaluator's payoff is determined based on subjects' performance in the logic quiz compared to a group of 20 subjects, including themselves. If their performance ranks in the top half of the group, the performance evaluator pays them $2. On the other hand, if their performance ranks in the bottom half, the performance evaluator provides no payment, resulting in a payoff of $0. The random evaluator determines subjects' payoff by tossing a coin. If the coin toss results in heads, the random evaluator pays them $2. Conversely, if the coin toss results in tails, the random evaluator provides no payment, resulting in a payoff of $0.

Subjects who receive only high payments or only low payments from both evaluators will be redirected to a different survey with a similar expected payoff. Subjects who receive mixed payments will proceed with Part B.

In Part B, subjects can solve up to 25 decoding tasks and they can decide whether they want to continue or stop working after each decoding task. Subjects are paid by the same evaluator 1 from Part A. If evaluator 1 is the performance evaluator, they receive $0.1 for each correctly solved decoding task. If evaluator 1 is the random evaluator, they receive no payoff. Before starting the decoding tasks, we elicit subjects' beliefs about the likelihood that evaluator 1 is the performance evaluator.

The experiment concludes with a survey measure of risk aversion.

Experimental design

The experiment consists of two parts and we randomly choose one part to determine payments.

In Part A, subjects perform a logic test with 10 questions and we elicit their beliefs about scoring in the top half in a group of 20 subjects, including themselves.

Subjects receive two payoffs in Part A. One payoff is provided by evaluator 1, and the other by evaluator 2. Exactly one of the two evaluators is a performance evaluator and exactly one of the two evaluators is a random evaluator. However, subjects are not aware of which evaluator corresponds to each role. The performance evaluator's payoff is determined based on subjects' performance in the logic quiz compared to a group of 20 subjects, including themselves. If their performance ranks in the top half of the group, the performance evaluator pays them $2. On the other hand, if their performance ranks in the bottom half, the performance evaluator provides no payment, resulting in a payoff of $0. The random evaluator determines subjects' payoff by tossing a coin. If the coin toss results in heads, the random evaluator pays them $2. Conversely, if the coin toss results in tails, the random evaluator provides no payment, resulting in a payoff of $0.

Subjects who receive only high payments or only low payments from both evaluators will be redirected to a different survey with a similar expected payoff. Subjects who receive mixed payments will proceed with Part B.

In Part B, subjects can solve up to 25 decoding tasks and they can decide whether they want to continue or stop working after each decoding task. Subjects are paid by the same evaluator 1 from Part A. If evaluator 1 is the performance evaluator, they receive $0.1 for each correctly solved decoding task. If evaluator 1 is the random evaluator, they receive no payoff. Before starting the decoding tasks, we elicit subjects' beliefs about the likelihood that evaluator 1 is the performance evaluator.

The experiment concludes with a survey measure of risk aversion.

Design details

Randomization method
Randomization dony by a computer.
Randomization unit
Individual level randomization.
Clusters
~1000
Observations
~1000
Arms
~250 (evaluator 1 payment: $2, evaluator 2 payment: $2) ~250 (evaluator 1 payment: $2, evaluator 2 payment: $0) ~250 (evaluator 1 payment: $0, evaluator 2 payment: $2) ~250 (evaluator 1 payment: $0, evaluator 2 payment: $0)
Primary outcomes
Performance beliefs Evaluator beliefs Decoding task performance score

Cited by

No citation data available for this registration yet.

Keywords