Measuring evidence use under social pressure

Last registered on October 07, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
Measuring evidence use under social pressure
RCT ID
AEARCTR-0019845
Initial registration date
September 30, 2026

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
October 07, 2026, 10:21 AM EDT

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

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Ankhbayar Delgerchuluun
Affiliation
Center for Evaluation and Development
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Atika Pasha
PI Affiliation
Center for Evaluation and Development
Contact Investigator

Additional Trial Information

Status
On going
Start date
2026-02-16
End date
2026-12-14
Keywords
Behavior, Governance
Additional Keywords
Social desirability bias, evidence-informed policymaking, survey experiment, face-saving preambles, self-reported evidence use, COM-B framework, civil servants, Ethiopia
JEL code(s)
C83, D91, H83
Secondary IDs
Prior work
This trial is based on or builds upon one or more prior RCTs.
Abstract
Evidence-informed policymaking is often measured through self-reports, which may be affected by social desirability bias: public officials may feel pressure to appear more confident and committed to using research evidence. This survey experiment tests whether short face-saving preambles affect these reports among approximately 350–400 Ethiopian civil servants.

Participants are randomly assigned to answer either standard questions or identical questions preceded by statements that normalize challenges and differing views about evidence use. The questions measure capability, opportunity, and motivation to use evidence, drawing on the COM-B framework. We test whether the preambles lower self-reported scores and whether effects differ by seniority, professional role, and education. The findings will inform the design and interpretation of surveys used to assess evidence-informed policymaking, while distinguishing changes in reporting from changes in actual practice.
External Link(s)

Registration Citation

Citation
Delgerchuluun, Ankhbayar and Atika Pasha. 2026. "Measuring evidence use under social pressure." AEA RCT Registry. October 07. https://doi.org/10.1257/rct.19845-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
The intervention consists of short normalizing, or “face-saving,” preambles embedded in an online survey of Ethiopian civil servants. Participants are randomly assigned with equal probability to receive either three standard questions or the same questions preceded by brief statements acknowledging common challenges and differing views about using research evidence.

Each question measures one dimension of the COM-B framework: capability to interpret and apply quantitative evidence, opportunity to use evidence in daily work, and motivation regarding evidence’s contribution to policy quality. The preambles acknowledge differences in technical backgrounds, institutional constraints, and policy priorities, making it more acceptable to report limitations or less favourable views. Question wording and response scales are identical across arms; only the preambles differ.
Intervention Start Date
2026-05-18
Intervention End Date
2026-06-29

Primary Outcomes

Primary Outcomes (end points)
EIPM COM-B index combining self-reported capability, opportunity, and motivation to use research evidence in policymaking.
Primary Outcomes (explanation)
The index is the unweighted mean of three standardized survey items: confidence in interpreting and applying quantitative evidence; the extent to which the work environment enables research evidence use; and belief that research evidence improves the quality and credibility of policy decisions.

Each item is recorded on a 1–5 scale and standardized by subtracting the control-group mean and dividing by the control-group standard deviation. The three standardized items are then averaged. Higher index values indicate higher self-reported capability, opportunity, and motivation. All items are measured during the survey, immediately following the corresponding preambles in the treatment arm or without preambles in the control arm.

Secondary Outcomes

Secondary Outcomes (end points)
Capability: self-reported confidence in interpreting and applying quantitative evidence to inform policy decisions.
Opportunity: self-reported extent to which the work environment enables research evidence use in daily tasks.
Motivation: belief that research evidence improves the quality and credibility of policy decisions in the respondent’s directorate.
Secondary Outcomes (explanation)
The three components of the primary index are also analysed separately. Each is measured using a single survey item, with responses recorded as whole numbers from 1 to 5 on a slider with verbal endpoint labels. Capability ranges from “Not at all confident” to “Extremely confident.” Opportunity and motivation range from “Not at all” to “To a very great extent.” Higher scores indicate higher levels of the respective self-reported dimension. The wording and response scales are identical across treatment and control arms; only the presence of the preambles differs.

Experimental Design

Experimental Design
This is a two-arm, individually randomized survey experiment embedded in the registration questionnaire for the wider Ethiopia evidence-informed policymaking (EIPM) study. The planned sample is approximately 350–400 civil servants across about 70 directorates in Ethiopian federal ministries and agencies.

Respondents are assigned with equal probability to a direct-question control arm or a normalizing-preamble treatment arm. Both groups answer the same three questions measuring capability, opportunity, and motivation to use research evidence. Treatment-arm respondents receive a short face-saving preamble before each question. Each respondent sees only one version, and outcomes are measured within the same survey.
Experimental Design Details
Not available
Randomization Method
Computer-generated randomization within SurveyMonkey. Each respondent is independently assigned to the direct-question control arm or the normalizing-preamble treatment arm with probability 0.5 each. Randomization is conducted at the individual level without stratification.
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
350–400 individual randomization units (civil servants). Treatment is not assigned by cluster. Respondents are drawn from approximately 70 directorates.
Sample size: planned number of observations
Approximately 350–400 individual respondents.
Sample size (or number of clusters) by treatment arms
Approximately 175–200 respondents in the direct-question control arm and 175–200 respondents in the normalizing-preamble treatment arm. Each respondent is independently assigned with probability 0.5, so realized arm sizes may differ.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The power calculations in the pre-analysis plan report minimum detectable effects of approximately 0.38–0.43 standard deviations, assuming 80% power, a two-sided significance level of 5%, and equal allocation between the two arms. These calculations use effective analysis samples of 180–231 respondents under alternative sample-loss scenarios.
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
Ethiopian Soceity of Sociologists, Social Workers and Anthropologists
IRB Approval Date
2026-05-16
IRB Approval Number
ESSSWA/L/AA/05930/2026
Analysis Plan

Analysis Plan Documents

SDB_EIPM_PAP_v5.pdf

MD5: 590bce46d635e22f56e3cb766cfe49c6

SHA1: 521b771c41a2af0cf4f61a07a1ecd8e5f7df777b

Uploaded At: September 30, 2026