# The Effects of Information on the Impact of Development Interventions in Fragile States

Last registered on August 03, 2022

## Pre-Trial

Trial Information

### General Information

Title
The Effects of Information on the Impact of Development Interventions in Fragile States
RCT ID
AEARCTR-0009818
Initial registration date
July 29, 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, 2:26 PM EDT

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

Region
Region

### Primary Investigator

Affiliation
Chr. Michelsen Institute

### Other Primary Investigator(s)

PI Affiliation
University of Osnabrueck
PI Affiliation
University of Bergen

Status
In development
Start date
2022-08-07
End date
2022-12-01
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Prior research shows that the beneficiaries of foreign aid often lack basic informa- tion on development projects in their communities—in terms of the initiators of the interventions or the rationale of aid distribution. This lack of transparency can undermine citizens’ political efficacy and nurture perceptions of unfair aid distribution. Customized aid information campaigns may avert these effects and impact positively on state-society and inter-group relations. We aim to investi- gate these potential effects of information in a randomized controlled trial among 10,000 respondents of a three-wave panel survey in Mali and Niger. We focus on 20 community-based infrastructure projects per country. In the baseline wave, mem- bers of the treatment group will receive customized information on ongoing aid projects in their vicinity (e.g., type, volume, and formal rules of project selection). In the mid-line, we repeat this information. In addition, half of the treatment group will have the opportunity to provide feedback on the project to the government and responsible aid agency. In the end line, we estimate the effects of these two elements of the information campaign on peoples’ evaluation of political efficacy, distribu- tional fairness, engagement with political institutions and inter-group relations. This pre-analysis plan describes our hypotheses, research design and planned data collection.

### Registration Citation

Citation
De Juan, Alexander, Paul Hofman and Carlo Koos. 2022. "The Effects of Information on the Impact of Development Interventions in Fragile States." AEA RCT Registry. August 03. https://doi.org/10.1257/rct.9818
Experimental Details

### Interventions

Intervention(s)
In this condition, respondents receive information on an aid project in their vicinity. This information focuses on several key defining features of the process of project selection as well as the project output itself. In terms of project selection, we provide information on (1) how funding volumes are allocated as across communes, (2) provisions for the identification of development needs, and (3) procedures for the selection of concrete aid projects. In terms of the projects themselves, we focus on (1) who is responsible for the funding program, (2) the type of project, (3) the costs and (4) the expected duration of the project. Note that we conceive of these individual pieces of information as elements of one single compound treatment; i.e., we are interested in the effects of being informed (or not) about key characteristics of aid projects rather in the (potentially heterogeneous) effects of any specific pieces of information.
Intervention Start Date
2022-08-07
Intervention End Date
2022-12-01

### Primary Outcomes

Primary Outcomes (end points)
State-Society Relations (Index)
Intergroup relations (index)
Primary Outcomes (explanation)
State-Society Relations: Rather than explicitly asking respondents about their “trust” or “confidence” in institutions, we gauge their willingness to engage (or not) with state institutions. Three considerations inform this decision: first, we expect that concrete activities are easier to assess for respondents than abstract notions like “trust”. Second, from an aid-policy perspective, (intended) behavior seems more relevant than general attitudes. Thus, we rely on four survey items related to voting, taxation, contacting and complaining. Respondents can answer to each item on a four-option Likert scale ranging between very likely'' (4), likely'' (3), unlikely'' (2), or very unlikely'' (1). Responses to each of the four item will be standardized, combined into an index and standardized again (See PAP for detailed questions).
Intergroup relations: To measure our second main outcome, we rely on vignettes that confront respondents with hypothetical distributional decisions in zero-sum situations. Again, our main motive for relying on vignettes rather than standard questions related to “trust in others” is to make survey items as concrete as possible. Specifically, we rely on two types of scenarios: the first one focuses on the inter-village distribution of a resource surplus, the second one focuses on political support for discriminatory political campaigns. In both cases, we ask for respondents' own distributional preferences as well as their expectations of other villages’ preferences. Responses to each of the four items will be standardized, combined into an index and standardized again. (See PAP)

### Secondary Outcomes

Secondary Outcomes (end points)
Civic engagement (index)
Secondary Outcomes (explanation)
In addition to our two main outcomes, we register a civil engagement index for an exploratory analysis. We understand civil engagement as people's voluntary contribution to a public good in terms of investing some of their \textit{time} and \textit{money} to upkeep the service and infrastructure. We believe that customized information and feedback opportunities can potentially increase civic engagement. The provision of information signals to the respondents that they are taken seriously as beneficiaries and stakeholders of the respective projects. This may, in turn, increase people's awareness and commitment to contribute to sustainability of the service. We rely on two items that ask respondents about the likelihood to voluntary contribute some of their time and money. Respondents can
answer to each item on a four-option Likert scale ranging between “very likely” (4), “likely” (3), “unlikely” (2), or “very unlikely” (1). Responses to each of the four item will be standardized,
combined into an index and standardized again
(See PAP for questions)

### Experimental Design

Experimental Design
In wave 1 respondents will be randomly assigned to one of three treatment conditions and will keep their treatment status throughout the panel survey; the information and feedback treatments are administered across the first two waves of the panel.
Experimental Design Details
Not available
Randomization Method
Randomization assigned on tablet during survey.
Randomization Unit
Individual
Was the treatment clustered?
No

### Experiment Characteristics

Sample size: planned number of clusters
-
Sample size: planned number of observations
1500 Individuals (in final wave, accounting for expected attrition)
Sample size (or number of clusters) by treatment arms
500 control, 500 T1, 500 T2
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
See PAP for details
IRB

### Institutional Review Boards (IRBs)

IRB Name
Institutional Review Board (Ethikkommission) of the University of Osnabrück
IRB Approval Date
2021-09-21
IRB Approval Number
471043.5
Analysis Plan

### Analysis Plan Documents

KfW_Information_Mali_Niger_PAP.pdf

MD5: 0f1db97ec453fd6525b791681cc7172a

SHA1: fcfbabfce315631cbdb73b9a73b204288ef2dcd2