Choice Architecture for Messaging Corruption

How to cite

Torres, Felipe; Duch, Raymond. 2021. "Choice Architecture for Messaging Corruption." AEA RCT Registry. AEARCTR-0007233. https://doi.org/10.1257/rct.7233-1.0

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

Felipe Torres* , Raymond Duch

* Primary investigator

Registry ID
AEARCTR-0007233
DOI
https://doi.org/10.1257/rct.7233-1.0
First registered
Last updated
Status
In development
Trial dates
2021-03-01 – 2021-03-31
Countries
Chile ()
JEL codes
P48; D73
Secondary IDs
FONDECYT - Chile
Registry record
http://www.socialscienceregistry.org/trials/7233

Abstract

We study to what information and benchmark voters update beliefs regarding corruption when they are presented with ``objective'' information about malfeasance. We conducted and mixed within-between subjects design, where respondents will receive different information about corruption using different benchmarks. The treatment consists of 6 short videos with different information about malfeasance.

For all video treatments, we will measure beliefs about the prevalence of corruption using different metrics 1) beliefs about the prevalence of corruption using a 0 - 10 scale; 2) beliefs about the prevalence of corrupt behavior by civil servants in that local government; and 3) beliefs about the economic impacts of corruption.

Intervention

We will conduct a factorial design, where respondents will receive different information about corruption using different benchmarks. The treatment consists of 6 short videos with different information about malfeasance.

For all video treatments, we will measure beliefs about the prevalence of corruption using different metrics 1) beliefs about the prevalence of corruption using a 0 - 10 scale; 2) beliefs about the prevalence of corrupt behaviour by civil servants in that local government; and 3)beliefs about the economic impacts of corruption. Further details can be found in our pre-analysis plan.

Experimental design

We will implement a factorial design that generates 24 treatment conditions. Respondents will receive the information using six different metric dimensions: standard, severity, resources total, resources individual, resources foregone losses, and program; two benchmarking dimensions: spatial and temporal; and two valence dimensions: positive and negative performance. Hence we have a 6 X 2 X 2 factorial design. There are a total of 24 factorial cells. Further details can be found in our pre-analysis plan.

Design details

Randomization method
Simple randomization from a web service: random.org
Randomization unit
Individuals
Observations
9000 number of observations collected from at least 1500 respondents
Arms
At least 1500
Primary outcomes
1) Beliefs about corruption using a 0 to 10 scale 2) beliefs about the prevalence of corrupt behaviour by civil servants in that local government, and 3) beliefs about the economic impacts of corruption 4) Beliefs about the location o
Secondary outcomes
No secondary outcomes

Cited by

No citation data available for this registration yet.

Keywords