Nudging Chicago Youth Away from Dangerous Defaults: Evaluating the ChiPlan Mobile Application

Last registered on December 15, 2017

Pre-Trial

Trial Information

General Information

Title
Nudging Chicago Youth Away from Dangerous Defaults: Evaluating the ChiPlan Mobile Application
RCT ID
AEARCTR-0002203
Initial registration date
December 14, 2017

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
December 15, 2017, 5:39 PM EST

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

Locations

Primary Investigator

Affiliation
University of Chicago Crime Lab

Other Primary Investigator(s)

PI Affiliation
University of Pennsylvania
PI Affiliation
University of Chicago Booth School of Business

Additional Trial Information

Status
Withdrawn
Start date
2017-06-01
End date
2017-12-14
Secondary IDs
Abstract
Interventions to reduce crime often focus on increasing the likelihood and severity of punishment. Deterrence works if offenders carefully deliberate their actions, and so are considering the potential costs of their actions. But, we suggest that many offenses stem from more automatic behavior. Potential offenders often default into risky habitual activities, which might lead to crime and violence (even if that was not the intention). Here, we propose to evaluate an intervention that uses a plan-making app to introduce active choice (and behaviorally informed choice architecture) into everyday moments. This app lists safe activities for at-risk youth and uses behavioral levers to increase the likelihood that youths participate in these safe alternatives.
External Link(s)

Registration Citation

Citation
Diop, Binta Zahra, Aurelie Ouss and Anuj K. Shah. 2017. "Nudging Chicago Youth Away from Dangerous Defaults: Evaluating the ChiPlan Mobile Application ." AEA RCT Registry. December 15. https://doi.org/10.1257/rct.2203-1.0
Former Citation
Diop, Binta Zahra, Aurelie Ouss and Anuj K. Shah. 2017. "Nudging Chicago Youth Away from Dangerous Defaults: Evaluating the ChiPlan Mobile Application ." AEA RCT Registry. December 15. https://www.socialscienceregistry.org/trials/2203/history/24023
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
2017-06-30
Intervention End Date
2017-09-15

Primary Outcomes

Primary Outcomes (end points)
The key outcomes are:
- Arrests
- Victimization
Primary Outcomes (explanation)
Arrests:
We will use administrative datasets provided by the Chicago Police Department (CPD) to see whether using ChiPlan decrease the number of arrests, the likelihood or arrests and if it changes the type of offenses youth are arrested for.

Victimization:
We will look at the victimization reports compiled by CPD and see if the use of ChiPlan has any impact on community level and individual level victimization. The City of Chicago also administer surveys to One Summer Chicago applicants. A section of this survey collects information on victimization. We are also hoping to look at self-reported victimization information collected through the survey.

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Clustered Randomized Controlled Trial
Experimental Design Details
Randomized Controlled Trial:
The Department of Family & Support Services (DFSS), a key partner for us in this work, directly served approximately 8,000 youth through a summer job program over the summer. Those summer jobs recipients are assigned to 44 service providers (sites). Service providers are for the most part located on the South Side of Chicago, which, along with Chicago’s West Side, are the two most violent areas of the city. DFSS serves youth who are typically at high risk of victimization or involvement in crime: roughly 15% of participants have been arrested in the past. The service providers (with roughly 200 youth per site) train youth and place them in jobs.

Randomization Method
Randomization will be conducted at the grounds of the Crime Lab using statistical softwares such as Stata and R
Randomization Unit
With the summer 2017 cohort of an anticipated 8,000 youth, we will randomize access to the ChiPlan app by service provider. We will randomize the 44 sites to be either “treatment” or “control” sites. All youth at the treatment sites will be given a code to log on to the app and the concept of using the app will be presented to them during their program. We chose clustered randomization instead of individual randomization to ensure equity among youth assigned to the same site and buy-in from the service providers we will ask to encourage youth to use the app. We will be comparing outcomes of youth who were offered access to the app to those of youth who weren’t.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
Among the 44 original service providers, 2 pairs of providers are providing services from offices located in the same building. To minimize spillovers, we bundle each of those pairs into one unit. As a result, we will randomize 42 units.
Sample size: planned number of observations
Approximately 8,000 summer job recipients
Sample size (or number of clusters) by treatment arms
Approximately 4,000 summer job recipients in the treatment group and 4,000 summer job recipients on the control group
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Using data from 2015 on arrests among One Summer Chicago (OSC) participants, we estimate that we would be able to detect an 19% drop in the number of arrests, from a baseline of 0.53 arrests per youth. However, it is also important to note that in the past year, arrests in Chicago have decreased substantially. This contrasts with the historically high levels of violence that Chicago is currently experiencing: 757 people have been killed in Chicago so far this year, which represents a 58% increase since last year. The decrease in arrests is therefore not reflective of a reduction in violence, but rather may stem from a change in police behaviors. That is, it reflects a measurement problem rather than an indication that there is no room for improvement on the challenge of youth violence. In light of these drastic changes in the environment of arrests and violence in Chicago, we think that the minimum detectable effect (MDE) we mention above is the lower bound of the MDE for this project. As a result, we are looking for alternative measures of victimization and violence including Chicago Police Department (CPD) victimization reports, hospital admission data, and a survey administered to all One Summer Chicago participants. We are confident that our strong relationships and years of experience with city agencies will allow us to find the most accurate and relevant way to measure victimization and crime. Baseline mean for number of arrests 0.531293 Standard deviation of arrests 1.731572 Minimum detectable effect for a one-sided test 0.1058607
IRB

Institutional Review Boards (IRBs)

IRB Name
University of Chicago IRBs
IRB Approval Date
2016-06-30
IRB Approval Number
IRB16-0813

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?
Yes
Intervention Completion Date
September 15, 2017, 12:00 +00:00
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