Using Behavioral Science
to Improve Criminal Justice
Outcomes
Preventing Failures to Appear in Court
Brice Cooke
Binta Zahra Diop
Alissa Fishbane
Jonathan Hayes
Aurelie Ouss
Anuj Shah
November 2017
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 2
Acknowledgments
We are grateful to the Laura and John Arnold Foundation, the John D. and Catherine T. MacArthur
Foundation, and the Abdul Latif Jameel Poverty Action Lab (J-PAL) for their early and continued
support of this work. We are also greatly appreciative for the support and close partnership with
the New York City Mayor’s Oce of Criminal Justice, in particular Elizabeth Glazer, Alex Crohn, Allie
Meizlish, and Angela LaScala-Gruenewald; the New York City Police Department, in particular Deputy
Commissioner Susan Herman, Detective Kenneth Rice, and Lieutenant Denis O’Hanlon; and the New
York State Unified Court System Oce of Court Administration, in particular Justin Barry, Jason Hill,
Karen Kane, Carolyn Cadoret, and Zac Bedell.
We also would like to thank our colleagues at ideas42 and the University of Chicago Crime Lab and
especially Roseanna Ander, Katy Brodsky Falco, Chelsea Hanlock, Zachary Honoro, Christina Leon,
and Jens Ludwig for their invaluable assistance, Christina Avellan, Hannah Furstenberg-Beckman,
and Jessica Leifer for their contributions to the diagnosis and designs of the text message reminders,
and Ethan Fletcher, Jaclyn Leowitz, David Munguia Gomez, and Allison Yates-Berg for their contri-
butions to the summons form re-design.
The views expressed in this report are solely those of the authors and do not necessarily reflect those
of any funders or data providers.
Contact: Alissa Fishbane (alissa@ideas42.org) or Aurelie Ouss (aouss@sas.upenn.edu)
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 3
About ideas42
We’re a leader in our field with unique expertise and experience at the forefront of behavioral science.
We use this to innovate, drive social change, and improve millions of lives. We create fresh solutions
to tough issues based on behavioral insights that can be scaled up for the greatest impact. ideas42
also educates leaders and helps institutions improve existing programs and policies.
Our work spans 30 countries and encompasses consumer finance, economic mobility, education,
energy and the environment, health, international development, and safety and justice. As a global
nonprofit organization, our partners include governments, foundations, companies, NGOs, and many
other institutions.
At its core, behavioral science helps us understand human behavior and why people make the
decisions they do. It teaches us that context matters, that asking the right questions is critical, and
that simple solutions are often available, but frequently overlooked or dismissed. We work to identify
the subtle but important contextual details that can have a disproportionate impact on outcomes.
Visit ideas42.org and follow @ideas42 on Twitter for more.
About the University of Chicago Crime Lab
The U.S. has the highest rate of homicide among any developed nation in the world. The U.S. also
has by far the highest rate of incarceration among any high-income nation, with over 2.2 million
people currently incarcerated nationwide. Both of these problems disproportionately aect our most
economically disadvantaged and socially marginalized communities.
Taken together, all levels of government in the U.S. spend well over $200 billion per year on the
criminal justice system (including police, courts, and corrections). Yet we have made little long-term
progress on these problems. The homicide rate in America today is about the same as it was in 1950,
or even 1900. This stands in stark contrast to the enormous progress the U.S. has made toward
reducing mortality rates from almost every other leading cause of death. One key reason we have not
made more progress on these problems is a striking lack of rigorous evidence about what actually
works, for whom, and why.
The University of Chicago Crime Lab and sister organization Crime Lab New York aim to change
this by doing the most rigorous research possible in close collaboration with city government and
non-profits. Using randomized controlled trials, insights from behavioral economics, and predictive
analytics, the Crime Lab partners with government agencies and frontline practitioners to design
and test promising ways to prevent violence and reduce the social harms of the criminal justice
system, with the ultimate goal of helping the public sector deploy its resources more eectively (and
humanely) to improve lives.
Building on the model of the Crime Lab, the University of Chicago launched Urban Labs in 2015 to
help cities identify and test the policies and programs with the greatest potential to improve human
lives at scale. Under the direction of leading social scientists, Urban Labs utilizes this approach across
five labs that tackle urban challenges in the crime, education, energy & environment, health, and
poverty domains.
Visit urbanlabs.uchicago.edu/labs/crime
Executive Summary
I
n 2014, nearly 41% of the approximately 320,000 cases from tickets issued to people for low-level
oense in New York City (NYC) had recipients who did not appear in court or resolve their summons
by mail. This represents approximately 130,000 missed court dates for these oenses. Regardless of
the oense severity (summonses are issued for oenses ranging from things like riding a bicycle on
the sidewalk to drinking in public), failure to appear in court automatically results in the issuance of an
arrest warrant. Because warrants are costly and burdensome for both the criminal justice system and
recipients, the NYC Mayor’s Oce of Criminal Justice—in partnership with the New York City Police
Department and New York State Unified Court System Oce of Court Administration—asked ideas42
and the University of Chicago Crime Lab to design and implement inexpensive, scalable solutions to
reduce the failure to appear (FTA) rate.
We tackled this problem using a two-sided approach. First, we re-designed the NYC summons form
to make the most relevant information stand out, making it easier for people to respond appropriately.
In the new form, important information about one’s court date and location is moved to the top, the
negative consequence of failing to act is boldly displayed, and clear language encourages recipients
to show up to court or plead by mail.
Second, we created text message reminders. We identified behavioral barriers leading many to
miss their court dates: people forget, they have mistaken beliefs about how often other people skip
court, they see a mismatch between minor oenses and the obligation to appear in court, and they
overweigh the immediate hassles of attending court and ignore the downstream consequences. We
then designed dierent reminders targeted at helping recipients overcome these barriers.
From March 2016 to September 2017 we implemented and evaluated our interventions, and showed
that both have significant and positive eects on appearance rates. We found that behavioral
re-design of the form reduced FTA by 13%. has already been scaled system-wide to all criminal
court summonses, and, based on 2014 figures, translates to preventing roughly 17,000 arrest warrants
per year.
Using a randomized controlled trial, we found that the most eective reminder messaging reduced
FTA by 26%, relative to receiving no messages. Looking 30 days after the court date, the most
eective messaging reduced open warrants by 32% relative to receiving no messages. This stems
from both reducing FTA on the scheduled court date as well as court appearances after the FTA
to clear the resulting warrant. These results are in addition to the gains already realized from the
summons form re-design. The most eective messaging combined information on the consequences
of not showing up to court, what to expect at court, and plan-making elements.
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 5Executive Summary
Traditionally, criminal justice policy is informed by the assumption that people make an explicit decision
to oend, and so most approaches aim to make crime less worthwhile. But our interventions are built
on the view that people who miss their court date do not necessarily make an active choice to skip it.
Rather, they may have failed to consider the decision at all due to a number of obstacles. The results
indicate that crime policies that focus on behavioral barriers can oer humane approaches to reduce
negative consequences for both citizens and the criminal justice system, without resorting to the
traditional lever of increasing enforcement.
WITH OLD FORM: 41%
WITH MOST EFFECTIVE
TEXT MESSAGES:
Estimates for summons recipients who provide a phone number
Improvements in timely court appearance
FTA Rates
13%
DECREASE
36%
DECREASE
26%
DECREASE
WITH NEW FORM: 36%
26%
Introduction
T
o bring about behavior change and crime prevention, policymakers within the criminal justice
system have traditionally focused on deterrence. For example, longer prison sentences are often
used to discourage crime by making crimes more costly for oenders.
However, these policies will only be eective if people carefully consider the costs and benefits of
their actions. Yet a growing body of literature in the behavioral sciences suggests that people often
do not think systematically about costs and benefits before acting. Instead, people often base their
decisions on intuitive or automatic processes that falter in predictable ways. Fortunately, the predict-
ability of these processes opens up additional levers for generating behavior change. For example,
behavioral science has shown people will reduce their energy consumption if told how much energy
they use relative to their neighbors
1
or that medical adherence can be boosted with simple reminders
to reduce forgetting.
2
However, insights from behavioral science have yet to be methodically applied
to criminal justice, where they hold promise for making the system fairer and more ecient.
To illustrate this, we focus on one problem: failures to appear in court (FTA). The criminal justice
system cannot work if people fail to appear in court, which is why the system places great weight
on ensuring that people attend required hearings and enforces prescribed responses if they fail to
do so. Nationally, the FTA rate is approximately 21-24% for felony cases.
3
FTA rates for misdemeanor
and low-level oenses are even higher: historically around 40% for summons cases in New York City
(NYC), which in 2014 represented about 130,000 missed court dates. In many jurisdictions, failing to
appear can result in an arrest warrant; in NYC this is the default response in accordance with state
law.
To reduce FTAs, a traditional policy approach would propose stricter enforcement of arrest warrants,
based on the assumption that people skip court because they weren’t deterred by existing penalties.
However, a behavioral science perspective suggests many other factors could lead people to miss
court. For example, they may not have paid close attention to information about their court date when
they got it, they may have simply forgotten, or they may not have planned for taking time o from work
in order to attend their court date. If these behavioral barriers account for some instances of FTA, then
behavioral interventions may help courts reduce FTA rates without resorting to stricter enforcement.
1
Allcott, Hunt, and Sendhil Mullainathan. "Behavior and energy policy." Science 327, no. 5970 (2010): 1204-1205.
http://science.sciencemag.org/content/327/5970/1204
2
Dai, Hengchen, Katherine L. Milkman, John Beshears, James J. Choi, David Laibson, and Brigitte C. Madrian. "Planning prompts
as a means of increasing rates of immunization and preventive screening." Public Policy & Aging Report 22, no. 4 (2012): 16-19.
http://nber.org/aging/roybalcenter/planning_prompts.pdf
3
Cohen, T. H. (2010). Pretrial release of felony defendants in state courts: State court processing statistics, 1990-2004.
https://www.bjs.gov/content/pub/pdf/prfdsc.pdf
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 7Introduction
This policy brief outlines the process and results of a joint project with ideas42 and the University
of Chicago Crime Lab, in partnership with the New York City Mayor’s Oce of Criminal Justice
(MOCJ), New York City Police Department (NYPD), and the New York State Unified Court System
Oce of Courts Administration (OCA). The project’s aim was to develop and test two behavioral
approaches to addressing the common issue of FTA, which plagues court systems across the country.
Instead of applying traditional approaches to increase compliance with court summonses (via stier
enforcement), we looked for opportunities to address contextual factors that were contributing to
missed appearances in NYC courts.
In the following sections, we outline the extent of the FTA problem in NYC, the contextual factors we
identified as contributing to the problem, and two simple, cost-eective solutions we designed and
tested to address it. After presenting results of each intervention, we conclude with thoughts and
recommendations for moving forward.
What is behavioral science?
Behavioral science is the study of how people make decisions and act within a
complex and textured world where details matter. It draws from decades of research
in the social sciences to create a more realistic framework for understanding people.
The standard approach to predicting human behavior suggests that we consider
all available information, weigh the pros and cons of each option, make the best
choice, and then act on it. The behavioral approach suggests something different.
We make decisions with imperfect information and do not always choose what’s
best for us. Seemingly small and inconsequential details undermine our intentions to
act. Behavioral science has been used across a variety of fields to realign policies,
programs, and products with how we really behave in order to improve outcomes.
Behavioral Reasons People
Fail to Appear in Court
C
ourt appearance tickets are issued for low-level oenses, which range from public consumption
of alcohol and public urination to riding a bicycle on the sidewalk and spitting. Among summonses
requiring an in-person court appearance (and were not resolved through plea by mail
4
), historically,
around 40% end in FTA.
Who Receives Summonses?
Descriptive Statistics of Summons Recipients
5
Summons recipients between January 2016 and June 2017
Borough Oense Past Summons Recipients
24%
29%
20%
22%
5%
8%
6%
8%
25%
10%
34%
9%
32%
68%
First summons
(since January 2012)
Received a prior summons
(since January 2012)
Bronx
Brooklyn
Manhattan
Queens
Staten Island
Alcohol
Park Trespassing Violation
Marijuana
Disorderly Conduct
Public Urination
Motor Vehicle
Other
Gender Breakdown
88%
12%
Male Female
4
The plea by mail option is available for two oenses: public consumption of alcohol and public urination.
5
Source: New York State Unified Court System data
34
years old
Average age
of summons
recipients
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 9Behavioral Reasons People Miss Court
In NYC, hundreds of thousands of arrest warrants are currently open due to FTA, which is problematic
for both civilians (who can be taken into custody upon future interaction with police ocers) and for
law enforcement (whose time and resources are spent taking individuals into custody who might
otherwise walk away from interactions with the police). The negative consequences for recipients
could include time in police custody, potential immigration issues, and disruptions to work and family
life—not to mention the psychological costs of worrying about being picked up on an open warrant.
Dealing with these warrants also places a burden on the police and court systems.
To uncover the psychological and contextual features contributing to FTAs, ideas42 and the Crime
Lab conducted quantitative and qualitative research using our behavioral diagnosis methodology.
We uncovered four main barriers contributing to FTAs.
Mental
Models
First, some recipients believe that receiving a ticket for a minor
oense and having to attend court for it is unfair. The crime feels
misaligned with the punishment. That is, having to go to court
for a seemingly minor oense (e.g., being in a park after hours)
doesn’t match with people’s “mental model” of what necessi-
tates a court appearance, much less an arrest warrant.
Present
Bias
Second, the immediate costs of attending court, such as taking
time o work and/or arranging childcare, outweigh the (often
unknown) consequences of not appearing. Many people we
interviewed weren’t aware that a warrant was a consequence
of FTA, but even among those who were aware of the warrant,
some still reported missing court because immediate costs of
going loomed larger than the risk of getting arrested in the future.
This focus on immediate costs over future ones, even when those
future costs are objectively larger, is known as “present bias.”
Social
Norms
Third, there is a misperception about court attendance. A majority
of interviewees hold the misperception that most people do not
attend their court dates, which (consciously or unconsciously)
may influence their own decision to attend or not. Prior work from
behavioral science shows that the perceived behavior of peers
(“social norms”) can have a strong influence on our decisions
and actions.
Inattention
Fourth, the long lag time between receiving the summons and
attending court leads many to forget. In NYC, the court date is
typically 60 to 90 days after the ticket was issued, which is plenty
of time for people to forget about their court date or the summons
altogether. This forgetting can be attributed to “inattention.”
Behavioral Interventions
to Reduce FTA
W
ith our understanding of the contextual and psychological barriers influencing court atten-
dance, we designed two simple, low-cost, scalable solutions to increase appearances.
6
Our
first touch point was the summons form itself, which is the recipients’ main source of information
regarding where and when they must attend court. One reason for FTA could be that people do not
take the time to carefully read the form. We redesigned it to limit the attention needed to acquire the
most important information by putting the essential details near the top of the form and clearly stating
the consequences of missing court.
Comparing the old and new summons form
We made several changes to the recipient copy of the summons form. Some of the main changes of
the front page of the form are described in the call out boxes on the next page.
7
6
These behavioral interventions complement policymaker and judicial system actions to reduce the stock of older open warrants.
Most recently, in August 2017, the Manhattan District Attorney cleared 644,494 warrants dating back at least 10 years.
7
See idea42's website for more details on the form redesign: http://www.ideas42.org/summons
CRC-3206 (5/12)
Name (Last, First, MI)
Street Address
City
ID/License Number
Date of Birth (mm/dd/yy)
Reg State
Time 24 Hour (hh:mm)
Place of Occurrence
In Violation of
Section
VTL
Subsection Admin
Code
Penal
Law
Park
Rules
Other
The Person Described Above is Charged as Follows:
Precinct
Date of Offense (mm/dd/yy) County
Plate Type Veh Type Make Year ColorExpires (mm/dd/yy)
Ht Wt Eyes Hair Plate/Reg
State SexType/Class Expires (mm/dd/yy)
State Zip Code
Apt. No.
Complaint/Information
The People of the State of New York vs.
NYC Pink
Copy
Title of Offense:
Defendant stated in my presence (in substance):
I personally observed the commission of the offense charged herein. False statements made herein are punishable
as a Class A Misdemeanor pursuant to section 210.45 of the Penal Law. Affi rmed under penalty of law.
Complainant’s Full Name Printed Rank/Full Signature of Complainant Date Affi rmed
(mm/dd/yy)
Agency Tax Registry #
The person described above is summoned to appear at NYC Criminal Court
located at:
Command Code
Summons Part County
Date of Appearance (mm/dd/yy)
At 9:30 a.m.
Bronx Criminal Court - 215 E 161
st
Street, Bronx, NY 10451
Kings Criminal Court - 346 Broadway, New York, NY 10013
Redhook Community Justice Center - 88-94 Visitation Place, Brooklyn, NY 11231
New York Criminal Court - 346 Broadway, New York, NY 10013
Midtown Community Court - 314 W 54
th
Street, New York, NY 10019
Queens Criminal Court - 120-55 Queens Boulevard, Kew Gardens, NY 11415
Richmond Criminal Court - 67 Targee Street, Staten Island, NY 10304
DEFENDANT’S COPY
1452220.indd 3 11/25/14 9:29 AM
GLUE LINE
DEFENDANT’S COPY
Criminal Court Appearance Ticket
**To avoid a warrant for your arrest, you must show up to court.**
At court, you may plead guilty or not guilty.
Please see back for exceptions for Public Consumption of Alcohol and Public Urination offenses.
You are Charged as Follows:
Title of Offense:
Time 24 Hour (hh:mm)
Other
VTL
Admin
Code
Penal
Law
Park
Rules
For Additional Information and Questions:
Visit the website or call the number below for additional information about your court
appearance and translation of this document.
www.mysummons.nyc
OR
Call 646-760-3010
County
Date of Offense (mm/dd/yy)
Place of Occurrence
In Violation of Subsection
Section
Show up to court on:
Cell Phone Number (where court may contact you)
Home Phone Number (where court may contact you)
Complainant’s Full Name Printed Rank/Full Signature of Complainant
Date Affirmed
(mm/dd/yy)
Command Code
Defendant stated in my presence (in substance):
(mm/dd/yy)
Court Appearance Date (mm/dd/yy):
at: 9:30 a.m.
Your court appearance location:
Bronx Criminal Court ……………………...……………….. 215 E 161
st
Street, Bronx, NY 10451
Kings & New York Criminal Court ……............ 1 Centre Street, 16
th
Floor, New York, NY 10007
Redhook Community Justice Center ….………......... 88-94 Visitation Place, Brooklyn, NY 11231
Midtown Community Court ………...…………….……. 314 W 54
th
Street, New York, NY 10019
Queens Criminal Court ……………….……. 120-55 Queens Boulevard, Kew Gardens, NY 11415
Richmond Criminal Court …………………………. . sland, NY 1030
Court Locations: You must appear at the court location identified above.
Redhook
Community Justice Center
Bronx
Criminal Court
Kings & New York
Criminal Court
Richmond
Criminal Court
Midtown
Community Court
Other (specify) ______________________________________________
Name (Last, First, MI)
Date of Birth
()
(
)
.
.
.
26 Central Ave, State nI 1..
CRC-3206 (1/16)
Queens
Criminal Court
I personally observed the commission of the offense charged herein. False statements made herein are punishable as a
Class A Misdemeanor pursuant to section 210.45 of the Penal Law. Affirmed under penalty of law.
Precinct
Agency Tax Registry #
2035619.indd 3 1/7/16 10:21 AM
OLD
NEW
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 11Behavioral Interventions to Reduce FTA
CRC-3206 (5/12)
Name (Last, First, MI)
Street Address
City
ID/License Number
Date of Birth (mm/dd/yy)
Reg State
Time 24 Hour (hh:mm)
Place of Occurrence
In Violation of
Section
VTL
Subsection Admin
Code
Penal
Law
Park
Rules
Other
The Person Described Above is Charged as Follows:
Precinct
Date of Offense (mm/dd/yy) County
Plate Type Veh Type Make Year ColorExpires (mm/dd/yy)
Ht Wt Eyes Hair Plate/Reg
State SexType/Class Expires (mm/dd/yy)
State Zip Code
Apt. No.
Complaint/Information
The People of the State of New York vs.
NYC Pink
Copy
Title of Offense:
Defendant stated in my presence (in substance):
I personally observed the commission of the offense charged herein. False statements made herein are punishable
as a Class A Misdemeanor pursuant to section 210.45 of the Penal Law. Affi rmed under penalty of law.
Complainant’s Full Name Printed Rank/Full Signature of Complainant Date Affi rmed
(mm/dd/yy)
Agency Tax Registry #
The person described above is summoned to appear at NYC Criminal Court
located at:
Command Code
Summons Part County
Date of Appearance (mm/dd/yy)
At 9:30 a.m.
Bronx Criminal Court - 215 E 161
st
Street, Bronx, NY 10451
Kings Criminal Court - 346 Broadway, New York, NY 10013
Redhook Community Justice Center - 88-94 Visitation Place, Brooklyn, NY 11231
New York Criminal Court - 346 Broadway, New York, NY 10013
Midtown Community Court - 314 W 54
th
Street, New York, NY 10019
Queens Criminal Court - 120-55 Queens Boulevard, Kew Gardens, NY 11415
Richmond Criminal Court - 67 Targee Street, Staten Island, NY 10304
DEFENDANT’S COPY
1452220.indd 3 11/25/14 9:29 AM
GLUE LINE
DEFENDANT’S COPY
Criminal Court Appearance Ticket
**To avoid a warrant for your arrest, you must show up to court.**
At court, you may plead guilty or not guilty.
Please see back for exceptions for Public Consumption of Alcohol and Public Urination offenses.
You are Charged as Follows:
Title of Offense:
Time 24 Hour (hh:mm)
Other
VTL
Admin
Code
Penal
Law
Park
Rules
For Additional Information and Questions:
Visit the website or call the number below for additional information about your court
appearance and translation of this document.
www.mysummons.nyc
OR
Call 646-760-3010
County
Date of Offense (mm/dd/yy)
Place of Occurrence
In Violation of Subsection
Section
Show up to court on:
Cell Phone Number (where court may contact you)
Home Phone Number (where court may contact you)
Complainant’s Full Name Printed Rank/Full Signature of Complainant
Date Affirmed
(mm/dd/yy)
Command Code
Defendant stated in my presence (in substance):
(mm/dd/yy)
Court Appearance Date (mm/dd/yy):
at: 9:30 a.m.
Your court appearance location:
Bronx Criminal Court ……………………...……………….. 215 E 161
st
Street, Bronx, NY 10451
Kings & New York Criminal Court ……............ 1 Centre Street, 16
th
Floor, New York, NY 10007
Redhook Community Justice Center ….………......... 88-94 Visitation Place, Brooklyn, NY 11231
Midtown Community Court ………...…………….……. 314 W 54
th
Street, New York, NY 10019
Queens Criminal Court ……………….……. 120-55 Queens Boulevard, Kew Gardens, NY 11415
Richmond Criminal Court …………………………. . sland, NY 1030
Court Locations:
You must appear at the court location identified above.
Redhook
Community Justice Center
Bronx
Criminal Court
Kings & New York
Criminal Court
Richmond
Criminal Court
Midtown
Community Court
Other (specify) ______________________________________________
Name (Last, First, MI)
Date of Birth
()
(
)
.
.
.
26 Central Ave, State nI 1..
CRC-3206 (1/16)
Queens
Criminal Court
I personally observed the commission of the offense charged herein. False statements made herein are punishable as a
Class A Misdemeanor pursuant to section 210.45 of the Penal Law. Affirmed under penalty of law.
Precinct
Agency Tax Registry #
2035619.indd 3 1/7/16 10:21 AM
1
Clear title describes the purpose and required action.
2
The date, time, and location of the appearance is moved from the bottom to the top, where it is more
likely to be read.
3
The consequence of missing is clearly articulated and framed to spur loss aversion, the human tendency
to feel losses more severely than equivalent gains.
OLD
NEW
1
2
2
3
1
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 12Behavioral Interventions to Reduce FTA
The second touch point addressed the lag time between receipt of the summons and the court
date. We designed text message reminders tailored to address the bottlenecks described above.
Compared to other forms of reminders, such as letters or robo-calls, text messages are inexpensive,
and information is easily received and retrievable later.
We designed multiple sets of text messages to determine which messaging is most eective at
reducing FTA. Some were sent before a person’s scheduled court date (pre-court messages) and
some were only sent if they had missed their court date (post-FTA messages). In order to test which
messages were most impactful on FTA rates, recipients were randomly assigned to receive some
combination of pre-court and/or post-FTA messages, or no message at all.
The pre-court message sets consist of three dierent texts, sent seven, three, and one day before
the scheduled court date. This schedule was chosen in order to prompt recipients to take preemptive
action for attending court (i.e. scheduling time away from work or securing childcare) without reminding
them too early, which could lead to procrastination.
Some pre-court messages emphasized the consequences of failing to appear and provided infor-
mation about what to expect at court (“consequences”), while others focused on helping people
develop concrete plans for appearing in court (“plan-making”). A third set combined consequences
and plan-making messages. All messages helped to address inattention or forgetting the court date.
Pre-Court Messages
CONSEQUENCES MESSAGES
7 days before court 3 days before court 1 day before court
Helpful reminder: go to
court Mon Jun 03 9:30AM.
We'll text to help you
remember. [Show up to
avoid an arrest warrant.]
Reply STOP to end texts.
www.mysummons.nyc
Remember, you have
court on Mon Jun 03 at
346 Broadway Manhattan.
[Tickets could be dismissed
or end in a fine (60 days to
pay).] [Missing can lead to
your arrest.]
At court tomorrow at
9:30AM [a public defender
will help you through the
process.] [Resolve your
summons (ID##########)
to avoid an arrest warrant.]
1
Makes the costs of FTA more salient to overcome present bias.
2
Reduces the ambiguity and perceived costs of attending court.
3
Highlights penalties to overcome present bias and the mental model that you don’t need to go
to court for minor violations.
4
Repeats the consequence to keep the cost of missing court top-of mind, reinforcing that despite
the mismatch between crime and punishment, you must attend to avoid a warrant.
1
3
2
2
4
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 13Behavioral Interventions to Reduce FTA
PLANMAKING MESSAGES
7 days before court 3 days before court 1 day before court
Helpful reminder: go to
court on Mon Jun 03
9:30AM. [Mark the date on
your calendar and set an
alarm on your phone.] Reply
STOP to end messages.
www.mysummons.nyc
You have court on Mon
Jun 03 at 346 Broadway
Manhattan. [What time
should you leave to get
there by 9:30AM? Any
other arrangements to
make? Write out your plan.]
You have court
tomorrow for summons
ID##########. [Did you
look up directions to 346
Broadway Manhattan?]
Know how you're getting
there? Please arrive by
9:30AM.
1
Encourages people to set reminders to help them remember.
2
Aids people to think ahead and overcome potential barriers (or costs) to showing up to court.
3
Helps plan how to get there and makes the act of going more concrete.
COMBINATION MESSAGES
7 days before court 3 days before court 1 day before court
Helpful reminder: go to
court Mon Jun 03 9:30AM.
We'll text to help you
remember. Show up to
avoid an arrest warrant.
Reply STOP to end texts.
www.mysummons.nyc
You have court on Mon
Jun 03 at 346 Broadway
Manhattan. What time
should you leave to get
there by 9:30AM? Any
other arrangements to
make? Write out your plan.
Remember, you have court
tomorrow at 9:30AM.
Tickets could be dismissed
or end in a fine (60 days
to pay). Missing court for
########## can lead to
your arrest.
These messages, combining elements from both sets above, address present bias,
mental models, and plan-making as previously described.
3
2
1
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 14Behavioral Interventions to Reduce FTA
In addition to the pre-court reminders, we developed two types of messages sent only if a person
had missed the court appearance and a warrant had been issued. The first type focused on conse-
quences, letting recipients know that a warrant was issued, but that they wouldn’t be arrested if they
clear it at the court. The second type relied on the power of social norms and informed recipients that
most people actually had attended their court date. Again, both addressed inattention or forgetting.
Post-FTA Messages
CONSEQUENCE MESSAGE
[Since you missed court on
Jun 03 (ID##########),
a warrant was issued.]
[You won’t be arrested
for it if you clear it at 346
Broadway Manhattan.]
www.mysummons.nyc
Sent when a warrant is triggered by an FTA
1
Notifies of the serious consequence that has occurred.
2
Encourages action to resolve the open warrant.
SOCIAL NORMS MESSAGE
[Most people show up to
clear their tickets but records
show you missed court for
yours (ID##########).]
Go to court at 346
Broadway Manhattan.
www.mysummons.nyc
Sent when a warrant is triggered by an FTA
1
Provides feedback that their behavior goes against
the norm.
1
1
2
Results
Solution 1: Summons Form Behavioral Redesign
The redesigned summons form was first introduced to replace old forms in March 2016 and univer-
sally adopted by July 2016. The rollout period culminated in a rapid adoption of the new form across
NYC between June and July 2016. Once the new form was issued citywide, the old forms were
revoked and collected for destruction.
In order to isolate the impact of the redesigned summons form from other contributing factors to
FTA, we compared outcomes between people issued an old form and a new form using a quasi-
experimental approach called a regression discontinuity design. We focused on the narrow time-
window around new form adoption, comparing people who received summonses just before and just
after their issuing ocer switched to the new form. The intuition behind this research design is that
within a few weeks of the switch, the form version a recipient received was as good as random: they
happened to get whichever form the ocer was using at that time. This means that any change in FTA
is likely caused by the new forms.
8
Those who happened to receive the new summons form have an FTA rate that is 13%, or 6.4
percentage points lower than those who happened to receive the old summons form because
their issuing ocer had not switched yet. As the key variable between these two similar groups of
summons recipients, we can determine that the new forms caused this reduction in FTA.
Solution 2: Behavioral Text Messages to Reduce FTA
We evaluated the eect of behavioral text messages using a randomized controlled trial. Anyone in
NYC who was issued a summons and provided their cell phone number was eligible to receive text
message reminders. Summons recipients were randomized to receive one of the pre-court or post-FTA
message sets, or no messages (the “comparison group”). All eects seen here are in addition to the
gains in court attendance already realized through the behavioral summons form redesign.
8
In fact, the characteristics of summons recipients were very similar just before and just after ocers switched forms, in terms of
the kinds of oenses they received summonses for, their age and gender composition, and their likelihood of having received
summonses in the past. Thus, any dierence in FTA rate between those who received the old and new forms would suggest that
the new forms were responsible for the change.
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 16Results
TEXT MESSAGE SETS
PRE-COURT MESSAGES
Combination
Messages
Consequences
Messages
Plan-making
Messages
Comparison Group
(No Messages)
1
2
3
1
2
3
1
2
3
If FTA at initial summons court date
POST-FTA MESSAGES
Group A
Consequences
Group B
Consequences
Group C
No Message
Group D
Consequences
Group E
No Message
Group F
Consequences
Group G
Social Norms
Comparison
No Messages
We found that receiving any pre-court message reduces FTA on the court date by 21%. The combi-
nation messages, using elements of both the consequences and plan-making sets, were the most
eective, reducing FTA by 26% (from 38% to 28%). This 26% FTA reduction is measured on the
court date, and comes after receiving the sequence of three pre-court messages.
We also looked at the impact 30 days after the court date, as some summons recipients show up to
court to clear their warrants after their scheduled court date. For the combination messages, when
an individual fails to appear in court, they also receive a post-FTA message. Relative to receiving no
text message, we find a 32% reduction in open warrants for people who received a combination
message set and a post-FTA message (from 24% to 17%). This reflects both the change in FTA
on the court date, as well as subsequent court appearances to clear warrants within 30 days of the
scheduled court date.
There is also a question of whether timing of messages matters for reducing FTA—are messages more
eective when they are sent before missing a court date or after? We find that post-FTA messages
alone are helpful, leading to a 15% reduction in failures to return to court within 30 days, but not
as helpful as pre-court messages. Among post-FTA messages, the consequences message (16%
reduction) was more eective than the social norms message (13.6% reduction).
9
9
We also compared sending just the pre-court messages vs. pre-court plus post-FTA messages. Here, we find that for people who
received pre-FTA messages the eect of receiving an additional post-FTA message is encouragingly in the right direction, but not
yet statistically significant at the typical 5% level.
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 17Results
16%
DECREASE
37.8%
31.8%
29%
28.1%
The dierence in FTA rates between the comparison group
and any treatment arm is significant at the 1% level (p<0.01)
0
5%
10%
15%
20%
25%
30%
35%
40%
Comparison Group (No Messages)
Plan-Making
Consequences
Combination
24%
DECREASE
26%
DECREASE
CONTROL GROUP A GROUP C GROUP E
19.8%
Pre-Court Plan-Making +
No Post-FTA message
32%
DECREASE
24.3%
The dierence in FTA rates between the comparison group
and any treatment arm is significant at the 1% level (p<0.01)
16.6%
0
5%
10%
15%
20%
25%
Comparison Group (No Messages)
Pre-Court Combination +
Post-FTA Consequences
CONTROL GROUP A
20.5%
No Pre-Court Message +
Post-FTA Consequences
GROUP F
17.8%
Pre-Court Consequence +
Post-FTA Consequences
GROUP B
20%
Pre-Court Plan Making +
Post-FTA Consequences
GROUP E
19.2%
Pre-Court Consequence +
No Post-FTA message
GROUP C GROUP D
21%
No Pre-Court Message +
Post-FTA Social Norms
GROUP G
FTA Rate by Type of Message
Open Warrant Rate 30 Days After Court Dates
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 18Results
Cost-Benefit Analysis
Both the redesign and text message interventions are inexpensive and scalable. Using the re-designed
form has exactly the same cost as using the old form, and the only cost is incurred during a one-time
change. The text messages are also inexpensive, at less than one cent ($0.0075) per message. For
example, sending all 2014 summons recipients three messages would have cost less than $7,500.
By contrast, the costs of failing to appear in court are much higher. Entry into the criminal justice
system—as would be the case if a person was arrested for having an FTA warrant—can have major
adverse impacts on people’s lives, regardless of the severity of the initial oense. Failures to appear
in court also divert time and resources in both courts and policing. The benefits could be even
larger if these kinds of messages also reduce FTA for more severe oenses, since this could result
in a lesser use of pre-trial detention. By reducing FTA rates,
10
behavioral interventions might make it
possible to allow more people to await trial outside of jail—an important goal for NYC and other U.S.
jurisdictions that are concerned about the racial and social disparities of pre-trial detention. Because
they are so inexpensive and easy to replicate, both interventions could easily be adapted for other
locations and for other types of courts and oenses.
10
Prior failures to appear in court are among the strongest predictors of future missed court appearances, and similarly the most
heavily weighted factors in considering pre-trial release https://www.bjs.gov/content/pub/pdf/prfdsc.pdf
http://www.nycja.org/lwdcms/doc-view.php?module=reports&module_id=678&doc_name=doc
WARRANTS THAT
COULD HAVE BEEN
AVOIDED IN 2014
20,800-31,300
Total Warrants Avoided Per Year (APPROXIMATE)
3,700-14,300
WARRANTS
AVOIDED
WITH
NO TEXTS
114,100
FTAs
WITH
COMBO TEXTS
110,400
FTAs
13% PHONE COLLECTION
(ACTUAL)
WITH
COMBO TEXTS
99,900
FTAs
50% PHONE COLLECTION
(HYPOTHETICAL)
320,000
scheduled arraignments
41%
FTA RATE
WITH
OLD FORM
131,200
FTAs
WITH
NEW FORM
114,100
FTAs
17,100
WARRANTS
AVOIDED
Next Steps
T
he interventions described here are among the first applications of behavioral science to criminal
justice policy.
11
Promisingly, not only are these solutions impactful, their eects are as large or
larger than some of the most successful similar behavioral interventions in other domains. We see
these interventions as an encouraging first step toward incorporating insights from behavioral science
into criminal justice reform.
An immediate next step is to build o of the results and continue to scale the most eective interven-
tions to reach more people. As a measure of the potential for future growth, a recent survey found that
87% of adults nationwide own a cell phone, with ownership reaching nearly 96% in NYC.
12
While text
messages are very eective, only about 13% of summons recipients in NYC currently provide a cell
phone number, which represents a significant opportunity to expand reach. Enabling more recipients
to get these messages would increase the potential impact of this intervention.
Another promising avenue we are exploring is “personalized reminders.” The usual approach in
behavioral science is to identify the intervention with the largest average eect and administer the
same “nudge” to everyone. We might achieve larger gains by tailoring reminders to individuals, so
that a given individual receives messages specific to the barriers that they are experiencing. For
instance, busy people may be particularly responsive to plan-making messages, while first-time
summons recipients may be more responsive to consequences messages.
Our findings have the potential for impact beyond low-level oenses and beyond NYC. Another
aim is to scale both the redesign of other complex forms that recipients receive and text message
reminders across dierent court systems and cities. Future work could specifically investigate the
gains to behavioral enhancements at criminal courts that handle more serious misdemeanors and
felonies in jurisdictions across the country.
The work we describe here represents an early success in using behavioral science to improve the
criminal justice system. Because behavioral approaches to criminal justice reform have been largely
overlooked, we believe that there are many “easy wins” to be had. Of course, eective nudges are
not substitutes for substantial policy change, but they could be an eective complement and can be
more readily implemented and scaled than broad policy changes. A concerted eort toward low-cost,
incremental benefits could add up to make a significant dierence both for the criminal justice system
and for people’s lives.
This research by ideas42 and the University of Chicago Crime Lab, in collaboration with MOCJ, NYPD,
and OCA, is a promising step toward incorporating behavioral science in criminal justice. We are
eager to continue eorts to better understand how novel, low-cost strategies could be used by NYC
and other jurisdictions to make progress on persistent policy challenges.
11
Another early example is the work of Bornstein et al (2013)." http://ppc.unl.edu/wp-content/uploads/2012/08/Bornstein-et-al-Re-
ducing-courts-failure-to-appear-rate-by-written-reminders-Psychology-Public-Policy-and-the-Law-2013.pdf
12
https://www1.nyc.gov/assets/dca/MobileServicesStudy/Research-Brief.pdf
USING BEHAVIORAL SCIENCE TO IMPROVE CRIMINAL JUSTICE OUTCOMES 20Next Steps
ideas42 uses the power of behavioral science to design
scalable solutions to some of society’s most difficult problems.
To find out more, visit us at ideas42.org or follow us @ideas42
The Crime Lab partners with policymakers and practitioners to help
cities identify, design, and test the policies and programs with the
greatest potential to reduce crime and improve human lives at scale.
To learn more visit us at urbanlabs.uchicago.edu/labs/crime
