Behavioral nudges reduce failure to appear for court

In New York City alone, for example, we calculate that approximately 40% of defendants (or 100,000 people) missed their court date for low-level offenses in 2015.

Other felony and misdemeanor defendants are asked to post monetary bail, which acts as collateral to incentivize appearance in court.

Although defendants are given all of the relevant information they need (e.g., when and where to appear for court, what the consequences are for missing court), they might be insufficiently aware of this information.

This could happen for various reasons—the information might not be salient enough, or defendants might simply forget it as their court date approaches.

That is, many failures to appear may occur not because defendants are intentionally showing contempt of court, but rather because existing policies do not allow enough room for error.

It might in fact be more cost-effective, while also more humane, to make court information salient for defendants.

In the first study, we redesign court summons forms to simplify how information is presented.

In the second study, we augment the redesigned form by sending text messages highlighting critical court information for defendants.

These interventions reduce failures to appear on average by 13% and 21%, respectively, suggesting that a meaningful proportion of defendants who fail to appear are not intentionally skipping court, but are effectively unaware of court.

In a series of lab experiments, we find evidence in support of this hypothesis, as the redesigned summons form improved participants’ identification and recall of court information.

Instead, they are given a summons form and are required to appear in court 60-90 days later, with some flexibility the week before the scheduled court date.

If defendants voluntarily show up to court at a later date, the warrant will often be vacated.

Historically, failure to appear rates are around 40% for summonses that require a court appearance.

The data contains defendant gender, date of birth, and address, information about the violation, and court outcomes (see supplementary materials for more details).

At the time of our first study, the only way defendants were notified of their court dates was on the summons ticket they received at the time of the original offense.

The entire policy to inform defendants about their court date and deter them from skipping court depended on this form.

However, the form’s design prioritized information about the original offense, rather than information about the defendant’s court appearance.

The defendant’s court date was written at the bottom of the form, below details about the officer issuing the summons form.

Given that this information was so easy to overlook, many defendants might have been insufficiently aware of when they were expected to appear in court and what the consequences were for missing court.

We worked with the Mayor’s Office of Criminal Justice, New York City Police Department, and the New York State Office of Court Administration to redesign the summons ticket to make relevant information more salient.

The new form prominently features the appearance date and court location at the top of the ticket, where people are more likely to see it.

And it clearly states in bold typeface, on the front of the form, that missing the assigned court date will lead to a warrant (see Fig. 1 for old and new forms).

If this form reduces failures to appear, then that suggests many defendants might have missed court simply because they were unaware of important information.

We evaluated the effectiveness of this intervention using a regression discontinuity design, comparing whether failure to appear rates were lower for defendants who happened to be among the first to receive a new form given by a particular police officer, versus one of the last to receive the old form given by that same officer.

Indeed, we find no differences in prior summonses or failures to appear, types of offenses, or their predicted failure to appear likelihood based on observables; and just a small difference in gender (see figs. S2 and S3 and table S2).

Since the only notable difference before and after the switch date is the kind of form issued, any difference in failure to appear rates for defendants on either side of the switch date can plausibly be attributed to the redesigned form (see supplementary materials for details and robustness checks regarding our identification strategy).

Figure 2 presents failure to appear rates for defendants issued forms just before and just after officers’ switch dates.

Failure to appear rates are lower just after the introduction of new summons forms.

This is because most officers switched to the new form between May and July, when failure to appear rates are highest.

We observe this seasonality of failure to appear rates in other years, but the drop in failure to appear rates following the switch date appears to be the specific result of introducing the redesigned forms (see supplementary materials discussion of robustness checks and fig. S5).

In this study, we tested whether failures to appear could be further reduced by texting defendants court information (date, location) and information about the consequences of missing court.

Defendants who received the new summons form could provide their cell phone number to the citing officer, though it was not mandatory.

In the other groups, summons recipients received three messages: seven days before, three days before, and one day before their scheduled court date.

In the “consequences” group, defendants received messages that described their court date and location, and also told them a warrant would be opened and they might be arrested if they missed their court date.

In the “plan-making” group, defendants received messages that described their court date and location, and also prompted them to make a plan to attend court, including marking their calendars, setting an alarm, and looking up directions (but there was no mention of consequences).

The “plan-making” messages, which did not mention the consequences of failure to appear, also significantly reduced failures to appear by 6 percentage points (15.8%, P < 0.001).

These results build on a prior, smaller scale study examining the effectiveness of postcard reminders for defendants in a context where baseline failure to appear rates were significantly lower than the current context (13).

The fact that these reminders are effective suggests that a significant proportion of defendants missed court because they lacked the most basic information about their scheduled appearance.

But even the simple consequences message, which just contained information about their court appearance, reduced failures to appear.

It is possible that our interventions were effective for other reasons besides just making defendants more aware of court information.

Participants were then shown a summons form and asked to identify three pieces of information on the form: the defendant’s court date/time, the defendant’s court location, and the defendant’s alleged offense.

The new form simply moves court information to the top, but leaves unchanged the position of information about the alleged offense.

This was also true for identifying the court location (New form: MLog(RT) = 4.59, SD = 0.40 vs.

Clearly, people more easily identify information at the top of the form, and moving court information there makes it more accessible.

Lab experiment 2 extends these results by testing whether the new form actually improves recall of court information.

Most importantly, they were asked to recall the penalty for failure to appear, the court date, and the court location.

First, we find that participants who saw the new forms were more likely to correctly recall their court date (new form: 38%, old form: 19%, P < 0.001) and court location (new form: 46%, old form: 26%, P < 0.001).

Moreover, we find that participants who saw the new forms were more likely to correctly recall that the penalty for failure to appear was a warrant (new form: 52%, old form: 41%, P = 0.003, see table S7).

This suggests that defendants who received the old form could have been unaware of court information simply because it was not communicated effectively.

The main difference appears to be that the new form made it easier to find information about court and the consequences of missing court (see table S8).

Relative to most other actions, participants rated failures to appear for court as less likely to be due to forgetting (Mcourt = 3.86, SD = 2.06; Mother actions = 4.24, SD = 1.45; paired t test, t(300) = 3.79, P < 0.001) and more likely to be intentional (Mcourt = 5.17, SD = 1.75; Mother actions = 4.82, SD = 1.29; paired t test, t(300) = 3.92, P < 0.001).

Participants (N = 304, MTurk sample) read background information on summonses and failure to appear rates in New York City.

In the “intentional” condition, after reading the background information, participants wrote down one reason why someone might purposely skip their court appearance, and then they made their policy choice.

(B) Laypeople’s (left) and criminal justice experts’ (right) judgments of whether the redesigned summons form would improve recall of court information and the rate at which defendants showed up for court (from lab experiments 4 and 5).

In the second part of the study, participants were shown pictures of both the old and new form (they were not labeled as such), and they indicated whether they thought recipients of the old or new forms would be more likely to remember their court information and to show up for their court appearance.

A clear majority of experts thought recipients of the new form would be more likely to remember their court date (86%) and court location (68%) and to show up to court (69%).

5, experts thought the new form would be more effective than did laypeople, who showed no clear preference for the new form (court date: 49%; court location: 50%; show up to court: 47%; all Ps < 0.001).

The form redesign reduced failures to appear by 8 percentage points (15%) for defendants living in the bottom quintile of neighborhood wealth, compared to 5.7 percentage points (13%) for defendants living in other neighborhoods.

These differences in summonses issued are accompanied by failure to appear rates that are higher for defendants living in poorer neighborhoods (poorest quintile among summonses recipients: 53% vs. wealthiest quintile: 37%), and neighborhoods with the highest proportion of Black and Hispanic residents (highest Black and Hispanic population quintile: 53% vs. lowest Black and Hispanic population quintile: 34%), compounding the negative consequences of summonses on poor and minority citizens.

By anticipating how human error can lead to failures to appear, our interventions have clear benefits for both defendants and the court system.

From our data, we cannot estimate how often warrants for summons failures to appear lead to arrests in New York City.

Of course, the proportion of defendants who are arrested for these warrants is necessarily higher than the proportion of residents, but if even 1% of defendants in our sample were arrested for failure to appear, then our interventions would have saved approximately $140,000 from August 2016 to September 2019.

Yet failure to appear rates for these offenses did not significantly increase when the threat of warrants was removed (33).

The fact that these material consequences have no impact on failures to appear suggests many defendants are not engaged in a careful calculus of whether to skip court.

First, the old summons form’s heading read: “Complaint/Information.” On the new form, we changed the heading to read: “Criminal Court Appearance Ticket,” to emphasize that the recipient was required to appear in court.

Second, the old summons form listed the court date at the bottom.

On the new form, we moved this to the top, and we made it easier for officers to clearly indicate the court location.

Fourth, the old summons form only noted on the back of the form that arrest warrants are issued for failures to appear.

The background information on summonses included the types of offenses for which summonses are issued and the requirement to appear in court 60 to 90 days later.

On the screen containing the summons form, there was a text box that reminded participants of the three pieces of information participants were searching for.

Next, they responded to multiple choice questions in which they were asked to recall the court date and court location listed on the form.

There are also stiff penalties for failing to appear for court, less so for other failures.

All participants read the same background information about summonses in New York City as in lab experiment 1, with additional text explaining that arrest warrants are issued for missing court (and defendants are warned about this), along with the base rate of failures to appear?

After the first two parts of the experiment (described above), participants then completed several demographics questions, mostly related to their profession: What state/territory they worked in, their current role, the number of years they have been in that role, the kind of court they work in (if applicable), and how many defendants they observe failing to appear for court (if applicable).