Methodology and decision-making for others in voting experiments

Last registered on March 15, 2024

Pre-Trial

Trial Information

General Information

Title
Methodology and decision-making for others in voting experiments
RCT ID
AEARCTR-0009351
Initial registration date
March 11, 2024

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
March 15, 2024, 6:56 PM EDT

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

Locations

Primary Investigator

Affiliation
University of Portsmouth

Other Primary Investigator(s)

PI Affiliation
University of Portsmouth
PI Affiliation
University of Vienna

Additional Trial Information

Status
In development
Start date
2024-02-01
End date
2024-05-30
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
There is extensive research on optimal election rules, rationality on going to the polls and strategic voting (see, e.g., Kittel, et al., 2014) but little is known about the impact of being elected in itself on politicians’ decisions. Democratically elected leaders should represent the electorate’s interest, but we know that other motivations, such as budget maximization or increasing the probability of being re-elected might crowd out the electorate's best interest.
In this project we will study whether being elected rather than appointed leads politicians to act more in line with social preferences, giving more weight to the electorate’s preferences.

Our Hypotheses are based on a larger social closeness to the group affected by the decision if the decision maker has been elected, as well as reciprocity towards the voters.
As there might be a selection bias of pro-social decision makers via the election method, we will also investigate this possibility and two ways to circumvent this, as well as the possible influence on campaign slogans when the decision is taken before formulating the slogan.
External Link(s)

Registration Citation

Citation
Kittel, Bernhard, Wolfgang Luhan and Paul-Emile Mangin. 2024. "Methodology and decision-making for others in voting experiments." AEA RCT Registry. March 15. https://doi.org/10.1257/rct.9351-1.0
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2024-02-01
Intervention End Date
2024-05-30

Primary Outcomes

Primary Outcomes (end points)
H1. The choices of elected decision makers are closer to the stated preferences of the group affected by the decision, as compared to the choices of decision makers that are randomly determined.
Primary Outcomes (explanation)
Our primary hypothesis concerns the contrasting preferences of the decision maker and the group of people the decision is taken for. In our setting, the decision maker is not affected by the decision themselves, the decision only affects the welfare of others. The decision maker might still have an opinion on what the best decision will be that could differ from the preferences of the group they are deciding for and they might therefore paternalistically choose something other than what the affected group wants. If the decision maker is elected by the group, this will increase social closeness and might increase reciprocity, both should reduce paternalistic behaviour.
H1. The choices of elected decision makers are closer to the stated preferences of the group affected by the decision, as compared to the choices of decision makers that are randomly determined.

Secondary Outcomes

Secondary Outcomes (end points)
H2. Pro-social decision makers are more likely to be elected.
H2a. Participants deciding before being elected are more paternalistic (less affected by the election) than those who decide after being elected.
H2b. There are identifiable patterns in the slogans of elected decision makers.
H3. The content of the slogan will change depending on whether a decision has been taken or not.
Secondary Outcomes (explanation)
As described above a confounding factor might be that there is a selection bias leading to more social, less paternalistic people as the decision makers. In that case, the choices of elected decision makers would be closer to the affected group’s preferences, but not due to how the decisionmakers react to an election but purely because a specific group is elected. We will test this separately:
H2. Pro-social decision makers are more likely to be elected.
If that is not the case, yet we find that we cannot reject H1 (section 7.2) this would mean that it is the mere fact of being elected that changes people’s decisions. We will cross check this by systematically varying when the decision is made. Some participants decide before the election (with the decision being implemented if they are elected) and some only after they have actually been elected.
H2a. Participants deciding before being elected are more paternalistic (less affected by the election) than those who decide after being elected.
The only information available information at the point of the election is a short statement, a “slogan” that the candidates provide. If there is a selection bias towards more social decision makers, the people voting must somehow be able to identify this in the slogans (see, e.g., Milinski et al., 2016). We are aiming to qualitatively analyse these slogans to see whether there are discernible patterns
H2b. There are identifiable patterns in the slogans of elected decision makers.
We are going to control for this possibility with a variation of the order of tasks in the experiment. This has the effect that some participants will make a decision, as if they were elected, before formulating the slogan, while others will face the “normal” order of formulating a slogan, (possible) being elected and then making a decision. The former order (decision then slogan) could lead to a more topic-focussed slogan, that is less about personality and power and more about the decision and it’s benefits.
H3. The content of the slogan will change depending on whether a decision has been taken or not.

Experimental Design

Experimental Design
We will use a controlled lab experiment for data collection, implemented at the BaL BARE Laboratory.
Experimental Design Details
We will use a controlled lab experiment for data collection, implemented at the BaL BARE Laboratory.
As is the norm in economic experiments, students will be welcomed to the lab and seated at individual workstations separated by blinds. The instructions will be distributed and read aloud, allowing time for questions before the start of the experiment. All procedures detailed below are common knowledge and will be explained to all participants in full.
The experimental design consists of six elements:
A. social values test (Murphy et al.,2011)
B. decision for self
C. grouping
D. formulate a “slogan” for the election
E. selection of a decision maker
F. decision for group
G. questionnaire
The primary variation to study H1 is that D is either a random selection or an absolute majority vote.
In C, participants will be randomly allocated to groups of five participants. We employ a matching-groups protocol in which participants will repeat stages C-F task five times with newly determined groups without any overlap in group members (perfect stranger matching). In the election treatments, two of the five group members will be randomly determined to be decision maker “candidates” with the remaining group members casting votes. The decision in B and F is the investment level in a Gneezy and Potters (1997) investment game. The decision for the other group members will be implemented using the strategy method to allow the following change of orders.
To test hypotheses 2, 2a, 2b, and 3, we change the order of events to
A. social values test (Murphy et al.,2011)
B. decision for self
C. decision for group
D. formulate a “slogan” for the election
E. grouping
F. selection of a decision maker
G. automatic implementation of decision (from C).
H. questionnaire
This means that we will employ a 2 × 2 design, with two selection modes (random and election) and two orders.
Randomization Method
Randomisation is conducted at individual level, when registering to the experiment and attending the sessions.
Randomization Unit
Individual.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
None.
Sample size: planned number of observations
432
Sample size (or number of clusters) by treatment arms
108
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We require 108 independent observations per treatment, 432 in total. This number has been computed using G*Power, for a chi-squared test with alpha=5%, beta=20% and a medium effect size w of 0.3 for our H1.
IRB

Institutional Review Boards (IRBs)

IRB Name
Business and Law Faculty Ethics Committee - University of Portsmouth
IRB Approval Date
2024-01-24
IRB Approval Number
BAL/2023/67/LUHAN
Analysis Plan

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

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

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

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials