The Impact of Summer Youth Employment on Youth Outcomes

Last registered on January 24, 2020

View Trial History

Pre-Trial

Trial Information

General Information

Title

The Impact of Summer Youth Employment on Youth Outcomes

RCT ID

AEARCTR-0004185

Initial registration date

May 28, 2019

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

June 10, 2019, 4:48 PM EDT

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

Last updated

January 24, 2020, 11:39 AM EST

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

Locations

Country

United States of America

Region

Michigan

Primary Investigator

Name

Brian Jacob

Affiliation

University of Michigan

Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status

On going

Start date

2018-04-01

End date

2022-09-01

Keywords

Crime, Violence, & Conflict, Education, Labor

Additional Keywords

JEL code(s)

Secondary IDs

Abstract

This study examines the impact of participating in a summer youth employment program. The program runs for approximately 6 weeks, during which time participating youth work 20 hours per week in a variety of community service related capacities. To determine the causal impact of the program, we randomly assigned eligible applicants the opportunity to participate in the program in summer 2018. We plan to follow the youth for several years, collecting data on a variety of outcomes via a survey as well as through administrative sources.

External Link(s)

Registration Citation

Citation

Jacob, Brian. 2020. "The Impact of Summer Youth Employment on Youth Outcomes." AEA RCT Registry. January 24. https://doi.org/10.1257/rct.4185-1.1

Former Citation

Jacob, Brian. 2020. "The Impact of Summer Youth Employment on Youth Outcomes." AEA RCT Registry. January 24. https://www.socialscienceregistry.org/trials/4185/history/61346

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)

The intervention is a summer youth employment program. The program runs for approximately 6 weeks, during which time participating youth work 20 hours per week in a variety of community service related capacities.

Intervention Start Date

2018-07-09

Intervention End Date

2018-08-01

Primary Outcomes

Primary Outcomes (end points)

Work readiness
Performance in school

Primary Outcomes (explanation)

Work readiness will be a composite variable created from survey responses. Performance in school will include composite variables from survey responses as well as measures of academic achievement and attainment collected from administrative education data.

Secondary Outcomes

Secondary Outcomes (end points)

Criminal justice involvement
Perceptions of the police
Employment and earnings

Secondary Outcomes (explanation)

Criminal justice involvement and perceptions of the police will be composite variables constructed from youth survey responses. Employment and earnings will be created from UI wage record data.

Experimental Design

The program is regularly oversubscribed. That is, more eligible youth apply than there are spots available. For this reason, we worked with the public agency that administers the program to randomly select a subset of applicants and invited them to apply to the summer program.

Experimental Design Details

The sample included applicants to Grow Detroit's Young Talent (GDYT) Summer 2018 program who were eligible to participate in JPFC, which is defined as living in the city of Detroit and being 14-15 years of age as of July 1. The sample included roughly 3,900 youth.

We conducted a stratified and systematic random sample of youth to determine the treatment group. We utilized 8 explicit strata created as the interaction between three binary indicators for gender, geographic region of the city and self-reported GPA. We utilized three implicit strata: a binary indicator for attended private school, a binary indicator for attending high school (as opposed to middle school), and a continuous measure of the academic proficiency rate in the students 2017-18 school. The probability of selection was weighted to obtain a mix of participants with the geographic representation desired by the agency.

Randomly selected youth were invited to participate in the JPFC program. It was possible for control youth to participate in JPFC (or other programs within GDYT), and we will account for potential crossover in the analysis.

To collect outcome data, we matched applicants to state administrative data from the Department of Education and the Unemployment Insurance Agency. To collect other outcome, we conducted a survey of applicants in June 2019, roughly 10 months after the completion of the summer 2018 program.

Randomization Method

Researcher generated randomization done via computer, and then provided to agency staff, who invited the treatment youth.

Randomization Unit

individual youth

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

Roughly 3,900 eligible youth were included in the study; not clustered

Sample size: planned number of observations

Roughly 3,900.

Sample size (or number of clusters) by treatment arms

2,650 youth were assigned to the treatment group and 1,260 youth were assigned to the control group

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

The sample size (including the proportion of treatment youth) was constrained by the operation of the program. However, we did conduct a power calculation to reassure ourselves that the study would be able to detect meaningful size effects. To determine the minimum detectable effect size (MDES) size for our experiment, we assumed a sample size of 1,800 treatment students and 1,200 control students. Because we are uncertain about the proportion of invited students who will take-up the offer to participate in JPC, we calculated the MDES for three different scenarios: 50% take-up, 70% take-up and 90% take-up. As a conservative assumption, we assume the R-squared of 0.10 for models that include randomization strata fixed effects as well as other student and neighborhood covariates. Based on these assumptions, we estimate our MDES for administrative outcomes will be 0.08 SD and 0.21 SD for survey outcomes. In order to translate these effect sizes into a more useful metric, we made assumptions (informed by some data we have available) about the distribution of outcomes in our population. For example, with 75% compliance, we will be able to detect effects on student attendance as small as 2 days per year and a 5 percentage point impact on the likelihood of having any arrests.

Supporting Documents and Materials

IRB