A Randomized Study to Abate Truancy and Violence in Grades 1-8 in Chicago Public Schools

How to cite

Guryan, Jonathan. 2016. "A Randomized Study to Abate Truancy and Violence in Grades 1-8 in Chicago Public Schools." AEA RCT Registry. AEARCTR-0000994. https://doi.org/10.1257/rct.994-1.0

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

Jonathan Guryan*

* Primary investigator

Registry ID
AEARCTR-0000994
DOI
https://doi.org/10.1257/rct.994-1.0
First registered
Last updated
Status
On going
Trial dates
2011-09-01 – 2035-08-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/994

Abstract

Together with the Chicago Public Schools (CPS), we are testing an intervention designed to reduce school truancy and dropout. Specifically, we are fielding a large-scale experiment to test one of the few interventions deemed by the U.S. Department of Education’s What Works Clearinghouse to have promising effects on student attendance and persistence in school – the Check & Connect (C&C) monitoring and mentoring program, which is designed to strengthen youth connections to school. Check & Connect mentors are full-time, in-school staff who regularly meet with frequently-absent students to monitor school engagement and progress, provide support services, and establish positive relationships with students and their families.The program was implemented in one third of eligible CPS elementary schools among two cohorts of students during the 2011-2012, 2012-2013, 2013-2014, and 2014-2015 school years. The remaining eligible schools served as comparison schools. Within Check & Connect schools, two grades between grades 1-8 were randomly selected to serve as control grades. Within the remaining grades, a randomly selected group of frequently-absent students received Check & Connect services. This design allows us to determine the student-level, school-level, and spillover effects of the intervention. We are studying the impacts of this intervention on academic, engagement, and attendance outcomes, as well as on long-term educational and labor outcomes. This document outlines the analysis plan for students among the second cohort (2013-2014, 2014-2015).

Intervention

Check & Connect is a monitoring and mentoring program designed to strengthen youth connections to school, and it is one of the few interventions that the U.S. Department of Education’s What Works Clearinghouse has found to have promising effects on preventing student dropout. Check & Connect mentors are full-time, in-school staff who regularly meet with frequently-absent students to monitor school engagement and progress, provide support services, and establish positive relationships with students and their families.

Experimental design

We will estimate ITT and TOT effects among students in cohort 2, including the following models:
• We will estimate models that allow the treatment effect to vary by age (initial grades 1-4 vs. 5-7; and grade-by-grade) in order to determine if Check & Connect’s impact on outcomes depends on students’ grade level.
• We will estimate models restricted to the treatment schools, comparing treatment and control kids within those schools, which will enable us to estimate variations in impacts at the school level.
• We will also estimate models that include students at the control schools to determine if there is a spillover effect of Check & Connect on control students at treatment schools.

Additionally, because cohort 2 contains a subset of students from cohort 1, we are able to compare treatment effects among students who received two vs. four years of Check & Connect. This question will be answered by estimating ITT and TOT effects for cohorts 1 and 2 together.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
School-level - Cohort 1: Of the 99 eligible CPS schools, 24 were randomly selected for treatment, with one school eventually opting out. The remaining served as control schools. - Cohort 2: Of the 23 schools in Cohort 1, 9 were selected for treatment. Grade-level: - Cohorts 1 & 2: Within the treatment schools, five grades between grades 1-7 were randomly selected for treatment. The remaining two grades served as control grades. Group-level: - Cohort 2: Treatment schools were allowed to nominated up to 10 students for treatment (half of which were actually assigned), 5 other groupings were created to test components of the program's structure (4 vs 2 years of treatment, latent effects that appear after treatment ends, etc.). Also, an additional sample of treatment and control students were added to the sample after the initial randomization to fill empty treatment slots,. Student-level: - Cohorts 1 & 2: Within the treatment grades, one third of frequently-absent students (students who missed between 10 and 30 or 35 days of school the year before) were assigned treatment. The remaining served as control students.
Clusters
Cohort 1 (2011-2012, 2012-2013): 23 schools were assigned treatment Cohort 2 (2013-2014, 2014-2015): 66 clusters to capture school and randomization blocks. 9 were assigned to control schools. Within each of the 9 schools assigned to treatment, there were up to 7 possible randomization blocks based on group-level randomization.
Observations
Cohort 1: 10,000 students total in both treatment and control schools. 450 students received treatment. Cohort 2: 5,000 students total in both treatment and control schools. 415 students received treatment.
Arms
Cohort 1: 23 treatment schools, 76 control schools. 933 assigned to treatment, 2985 assigned to control in treatment schools, ~6600 students in control schools. Cohort 2: 9 treatment schools, 9 control schools. 1030 assigned to treatment, 1111 assigned to control in treatment schools, ~3000 students in control schools.
Primary outcomes
PRIMARY ANALYSES: Attendance - Days present - Days absent - Membership days - Excused absences - Unexcused absences - Percent of days present Academic achievement - GPA - Math and reading grades - ISAT math and reading status - ISAT math and reading scores Behavior - Days suspended Transfer status (likelihood of transfer, number of transfers, transfer within vs. outside of CPS) SECONDARY ANALYSES: Number of absence streaks (or consecutive absences) Likelihood of attending summer school Attendance during summer school, given a student is required to attend Parent engagement (as measured by likelihood of having a parent portal account and showing up to parent report card pick-up days) Likelihood of engagement in after-school programming Likelihood of applying to and/or being accepted to a selective enrollment high school

Cited by

No citation data available for this registration yet.

Keywords