Norms and Practices in the Disclosure of Conflicts of Interest: Evidence from the Field of Economics

How to cite

Zingales, Luigi; Valletti, Tommaso; Barrios, John; Lancieri, Filippo; Levy, Joshua; Singh, Shashank. 2023. "Norms and Practices in the Disclosure of Conflicts of Interest: Evidence from the Field of Economics." AEA RCT Registry. AEARCTR-0012017. https://doi.org/10.1257/rct.12017-2.0

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

Luigi Zingales* , Tommaso Valletti , John Barrios , Filippo Lancieri , Joshua Levy , Shashank Singh

* Primary investigator

Registry ID
AEARCTR-0012017
DOI
https://doi.org/10.1257/rct.12017-2.0
First registered
Last updated
Status
On going
Trial dates
2023-09-18 – 2026-03-15
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/12017

Abstract

Our first-order interest lies in understanding the current attitudes of economists toward their own field’s standards and practices concerning the disclosure of conflicts of interest. To that end, we intend to conduct a survey of economists about these standards and present them with hypothetical scenarios of potentially inadequate disclosure. Additionally, we intend to compare these norms against norms and beliefs about adequate disclosure practices held by journalists and the general population.
We intend to use the responses of the survey described below to answer a number of questions about economists’ perceived practices and norms. Such questions include, but are not limited to, the following:

• Do economists believe that they, personally, adequately disclose their potential conflicts of interest?
• Do economists believe that their colleagues (i.e. the field as a whole) adequately disclose potential conflicts of interest?
• Do economists believe that certain kinds of relationships present greater potential for a conflict of interest than others?
• How do these beliefs (systematically) vary with individuals’ personal characteristics (past consulting affiliations, tenure-status, gender, nationality)?
• Do economists believe that existing policies impose a net cost or provide a benefit to the profession?
• Do economists’ beliefs on these issues diverge from those held by other groups?

Experimental design

We intend to disseminate our survey to five groups of individuals: “most-elite” economists, “elite” economists, economists generally, journalists, and the general population. We are interested in these five distinct groups for a number of reasons. The first three groups provide gradations of career concerns etc. among economists who are likely to be most familiar with the particular institutional dynamics that we are interested in. Second, journalists are of interest to us because they represent a sample of “watchdogs” who are often trained to investigate and root out various kinds of potential misbehavior. That is, their opinions about what constitutes conflicts of interest – and if there exist discrepancies between what economists and journalists think – are of particular interest. Finally, we are interested in the opinions of the general public, who serve as another barometer of potential conflicts of interest. Though we cannot guarantee full access to contact information, we intend to identify the following four groups. We define the “most elite” group of economists as those participating in the “Expert Surveys” conducted by the Kent A. Clark Center for Global Markets at the University of Chicago, Booth School of Business. The number of active panelists in this group varies over time but is approximately 130 at the time of writing. We define the “elite” group of economists as those who are affiliated with either the National Bureau of Economic Research (NBER) or the Center for Economic Policy Research (CEPR). Both are selective think tanks that require some kind of nomination and/or application for an economist to join. For the NBER, affiliates must have a primary appointment at a university, whereas CEPR affiliates may hold primary appointments at other institutions such as other think tanks, central banks, international organizations etc. Whereas the NBER membership is primarily American, CEPR membership is primarily European. We plan to construct the general sample of economists as a collection of economists in Javdani et al's database of economists (deduplicating the most elite, NBER and CEPR subgroups). We ensure that each economist respondent is a member of only one of these groups. The latter two groups are constructed as follows. We acquired the email IDs of journalists in the USA through presshunt.co who have accumulated one of the biggest journalist databases. We have about 13,000 email IDs of business and finance journalists and 5843 email IDs of journalists in government and politics. This sample of journalists is of interest to us because of their training/practice of journalism in addition to having some domain knowledge in economics/business. Finally, we intend to use YouGov’s services to identify a representative sample of the US population

Addendum (01/27/2026)
Following the original "within-subjects" experiment outline above, we intend to supplement our findings with a "between-subjects" design. The original manuscript employed a within-subject design where respondents: 1) read a description of a scientific finding; 2) reported their initial level of trust in the finding; 3) received additional information about the researcher's potential conflicts of interest; and 4) re-reported their revised level of trust. THe primary outcome was the "trust reduction" the difference between initial and revised trust levels.

The between-subjects design randomizes respondents to different information conditions rather than measures within-person revisions. Per vignette, respondents will be randomized into either the "Baseline" arm (respondents receive only the scientific finding, no CoI information); or one of the "Treatment" arms (respondents receive the scientific finding plus various CoI disclosures). The analog the trust reduction from the within-subject design in the between-subject design iis the difference between mean reported trust level reported by those respondents who where assigned to the Baseline arm and those who were assigned to the treatment arms.

In contrast to the within-subjects design where all respondents saw all vignettes (in a random order), respondents to the between-subjects instrument will be exposed to only a single vignette.

This between-subjects instrument will be disseminated to a representative sample of the United States general population using Prolific, and industry-standard online experiment platform.

Design details

Randomization method
Randomization done by qualtrics in the survey vignettes Addendum (01/27/2026) Randomization will be conducted by Qualtrics in each survey vignette with a quota/allocation rule that targets particular minimum detectable effects (MDEs) as outlined in the pre-analysis plan (PAP)
Randomization unit
Individual
Clusters
3 Major Clusters - Economists, Journalists, General public Addendum (01/27/2026) The between-subjects instrument will recruit a sample from the general population.
Observations
We expect 2000 economists, 2000 Journalists, 1500 general public Addendum (01/26/2026) The between-subjects instrument will target a total recruitment of 20,878 respondents across all vignettes. Addendum (01/27/2026) The between-subjects instrument will target a total recruitment of 20,878 respondents.
Arms
Each cluster has its own treatment arm (a separate survey for each), therefore 3. Economists, Journalists, General public Addendum (01/27/2026) Due to the large number of treatment arms and different targeted sample sizes (informed by pilot findings), please refer to the pre-analysis plan (PAP) for per-arm target sample sizes.
Primary outcomes
We intend to analyze the responses to the survey using various summary and analytical statistical methods. For example, questions of interest that could be answered include but are not limited to: • Conditional on having received tenure, do more respondents disclose engaging in consulting activities? • What are the changes in beliefs post disclosure treatment in the vignettes, how do these beliefs change vis-a-vis monetary, temporal and various other aspects • What are the differences amongst vignettes, are some conflicts more important than others? For example, are pure ideological conflicts more or less relevant than conflicts driven by incentives? • What is the effect of priming? Do the respondents react differently to economics vignettes when science/business vignettes are shown first? Are scientific conflicts more or less objective than economics conflicts? In addition to being able to ask questions about individual groups of respondents’ beliefs, our different samples permit us to compare beliefs between groups. For example, we could investigate how different groups differently (or similarly) perceive (the intensity of) certain kinds of conflicts of interest. In each of these within - and between - group analyses, we intend to exploit the random variation in the generated vignettes to understand how groups respond to varying degrees of conflict of interest.

Cited by

No citation data available for this registration yet.

Keywords