How School Matching Designs Affect Inequality: Experimental Evidence

Last registered on February 09, 2023

View Trial History

Pre-Trial

Trial Information

General Information

Title

How School Matching Designs Affect Inequality: Experimental Evidence

RCT ID

AEARCTR-0010302

Initial registration date

December 21, 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

January 03, 2023, 4:44 PM EST

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

Last updated

February 09, 2023, 4:08 PM EST

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country

United States of America

Region

Primary Investigator

Name

Sabrina Herzog

Affiliation

Düsseldorf Institute for Competition Economics - Heinrich-Heine-University Düsseldorf

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Henning Hermes

PI Affiliation

Düsseldorf Institute for Competition Economics - Heinrich-Heine-University Düsseldorf

Contact Investigator

Additional Trial Information

Status

In development

Start date

2022-01-01

End date

2023-12-31

Keywords

Behavior, Education, Welfare

Additional Keywords

Socio-economic status

JEL code(s)

D02, D47, D81, D91, I24

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Education has long-lasting effects in life outcomes of individuals. For some time now, market designers investigate matching mechanisms for school allocation, aiming to enhance fairness and ensure equal opportunities for children. Widely discussed are the Immediate-Acceptance (IA) and the Deferred-Acceptance (DA) mechanism, where the latter one has been agreed upon as more desirable due to its strategy-proofness which makes truthful preference submission a dominant strategy.
However, lab and field investigations have revealed that often applicants do not understand the strategy-proofness of the mechanism. This becomes especially problematic if applicants differ in submission strategies correlating with their socio-economic status (SES). The intended fairness might actually end up in an even higher socio-economic gap in educational outcomes.
In an online experiment, parents matched in groups of 16 with induced preferences over 4 schools have to choose a school ranking. After explaining them either the IA or DA mechanism we ask them to submit an application form for a school choice problem in which they compete over the 16 seats at the 4 schools. We pre-select parents by SES. In an additional treatment variation, we examine whether advice on the submission strategy (‘skip-the-middle school’ in IA and ‘submit the true preference order’ in DA) affects the socio-economic gap in school admissions. To gain a deeper understanding of what drives parents’ choices, we additionally elicit survey measures including, e.g., cognitive ability, beliefs about other parents’ behavior and returns to school quality, or trust in institutions.

External Link(s)

Registration Citation

Citation

Hermes, Henning and Sabrina Herzog. 2023. "How School Matching Designs Affect Inequality: Experimental Evidence." AEA RCT Registry. February 09. https://doi.org/10.1257/rct.10302-2.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Intervention Start Date

2022-12-20

Intervention End Date

2023-01-31

Primary Outcomes

Primary Outcomes (end points)

Truth-telling rates, skip-the-middle rates

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

We employ a between-subject design with four treatments. In each treatment group we sort parents of low and high socio-economic status (SES) in equal proportions.

Experimental Design Details

Subjects face a school choice problem, in which they and 15 other participants are asked to submit an application form to a centralized school admission authority. The problem consists of four schools with four seats each. Subjects are randomly matched in groups with an equal proportion of low and high SES parents in each group of 16. Preferences over schools are induced by the payoffs subjects receive when they get admitted at a school. There are four types of parents with different but correlated preference orderings over schools. We randomly assign two low and two high SES parents to each type.
We have two treatment variations: First, treatments vary by the school choice mechanism that is applied to find the admission of applicants to schools, the manipulable Immediate-Acceptance (IA) mechanism and the strategy-proof Deferred-Acceptance (DA) mechanism. As second treatment variation, we add a strategic advice on the optimal preference submission strategy. In the IA mechanism, we advise to skip-the-middle school in order to get a higher chance at the third-preferred school and not end up in the least preferred school. In the DA mechanism, we advise to submit preferences on the application form in the truly preferred order. Doing this, we aim to explore whether advise is able to help to decrease a socio-economic gap in school admissions, that we expect to find when comparing low and high SES parents in either one of the mechanisms.
Additional to the school choice problem, we ask subjects in all treatments to answer a questionnaire that elicits the following measures: subjects' strategies, loss-aversion, altruism, risk preferences, patience, trust in general and trust in institutions, level-k thinking, beliefs about returns to school quality, malleability of children, importance of school quality and importance of the decision in this study, and some demographics. Before subjects are faced with the school choice problem, we let them do a fluid IQ test (similar to Ravens) to measure cognitive ability. All measures serve to understand what drives parents' decisions on their preference submissions.

Randomization Method

computerized random assignment

Randomization Unit

socio-econcomic status, individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

see "Sample size"

Sample size: planned number of observations

960 parents in three treatments

Sample size (or number of clusters) by treatment arms

240 in Deferred-Acceptance without advice; 240 in Deferred-Acceptance with advice; 240 in Immediate-Acceptance without advice; 240 in Immediate-Acceptance with advice (always 120/240 low-SES parents and 120/240 high-SES parents)

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

With 960 parents as decision makers, that is, 240 decision makers per treatment with 120 having a low SES and 120 having a high SES, we can detect the following effect sizes at conventional levels of power (two-sided chi^2 test, significance level 5%, power 80%): For truth-telling rates as well as skip-the-middle rates, we can detect effect sizes of around 15-20 percentage points when comparing treatments, socio-economic status and advice variation.

Supporting Documents and Materials

IRB