Blog

Recently, we became aware of at least two blog posts (see here and here) lifting passages from, and selectively highlighting a result of, a draft paper the five of us authored. These posts provide a distorted picture of what our paper said and did.  The paper is not intended as an analysis of the Public Safety Assessment-Decision Making Framework (“PSA-DMF”) System risk assessment, nor is the data the paper analyzes anywhere close to final.  Rather, the paper proposes new statistical methodology.  It uses, for illustration purposes, partial data made available to us in the middle of a still-ongoing field experiment of the PSA-DMF System.  The study that produced the partial, interim data used to illustrate our statistical methods is still ongoing and our illustration uses less than 20% (in a rough sense) of the information the study will ultimately produce.  An interim report on the study, which has been public for some time now, makes all of this clear.  Moreover, there were many results from the paper’s illustrative application, only one of which the blog posts highlight.

We are scientists.  We want our work, both our new methodologies and our applied findings, to inform public debates and decision making.  But it is not helpful to mistake illustrative applications based on partial interim data in a paper intended to propose new statistical methods for final study results, nor to highlight selectively only certain results of an overall analysis.  We hope this post will clarify mistaken impressions.

Here are some details.

Our paper proposes new statistical methodology for evaluating risk assessment instruments and uses, as its applied example, interim data from a still-ongoing randomized control trial (“RCT”) study.  The RCT study evaluates the use of a predisposition risk assessment instrument called the Public Safety Assessment (“PSA”) and the accompanying, jurisdiction-specific, Decision Making Framework (“DMF”) in Dane County, WI.  Arnold Ventures supported the development of the PSA-DMF System.

The two blog posts purport to make much of one of the several results stemming from the illustration of our new methodology via the application of our methods to interim data from the Dane County study.  The results of our illustration are many and varied.  Here is a sampling of some these varied results as applied to the interim data from this one RCT:

• The availability of the PSA-DMF System had no statistically significant effect on the prevalence of predisposition new criminal activity (“NCA”) in any of three conceptually defined classes of individuals appearing at a first appearance hearing.
• The availability of the PSA-DMF System had no statistically significant effect on the prevalence of predisposition new violent criminal activity (“NVCA”) in any of three conceptually defined classes of individuals appearing at a first appearance hearing.
• The availability of the PSA-DMF System had no statistically significant effect on the prevalence of predisposition failure to appear (“FTA”) in any of three conceptually defined classes of individuals appearing at a first appearance hearing.
• The availability of the PSA-DMF System had no statistically significant effect on the measure of the racial fairness of the Dane County judges (actually, “Commissioners” in Dane) that two authors of our paper proposed in separate work, called “principal fairness.
• The availability of the PSA-DMF System had a statistically significant effect on the principal fairness measure with respect to gender comparisons, in that it increased the strictness of Commissioner decisions for men while decreasing the corresponding strictness for women, thus widening somewhat an already-existing gender difference in those decisions.

The two blog posts seize upon the last of these results (ignoring almost all of the others) to articulate an attack on the PSA-DMF System and upon the use of algorithms or risk classification instruments in criminal justice more generally.

Advocates for a particular position often simplify, distort, and selectively quote.  In a democracy such as ours, committed to free expression of ideas, it is not a mistake for them to do so.  The mistake is for anyone else to pay attention.  Good decision making in a democracy requires readers to distinguish between what is worthy of attention and belief and what is not.

One indication of whether a report on research deserves attention is whether the report’s authors contacted the relevant researchers before publishing, to request comment or clarification about the nature of the research.  To our knowledge, the authors of neither blog post did here.  Had they done so, we would have been happy to bring certain facts to their attention.  First, the paper they quote is about statistical methodology, as a cursory review of it reveals.  It is quite nerdy.  The “results” in the paper are intended to illustrate how the statistical methodology works on a dataset.  They are not intended to form the basis of conclusions relevant for policy.  That is why, for example, we made no attempt in the paper to adjust for the fact that we conducted multiple tests on the same data.  Including an illustrative application in a paper proposing new statistical techniques is traditional in the field, and again, the results of an illustration are not intended to form the basis of policy making.  Second, as a lengthy report and a more accessible FAQ sheet (both publicly available on the website of the Access to Justice Lab, where two of us work) make clear, the data used for the paper were an incomplete subset of about 20% of the data the Dane County RCT will eventually produce.  Moreover, the Dane County RCT is one of five field RCT studies the A2J Lab has underway (a sixth RCT field operation will produce a technical report only due to IRB restrictions).  In a rough, hand-wavy sense, we are talking about one-fifth of one sixth of the studies underway in this area.  Third, as the report and the FAQ sheet both discuss, based on the data produced from the Dane County study at this time, the availability of the PSA-DMF System had no statistically significant effect on the number of predisposition days that individuals appearing at a first appearance hearing spent incarcerated, suggesting (although not proving) that any disparity in the strictness or permissiveness of Commissioner decisions (which is what the blog posts focus on) did not translate into a difference in predisposition jail time.

Stepping back, it is not accurate to say that any statistically significant difference between two groups immediately means actionable discrimination.  Even taking the illustrative gender result upon which the two blog posts focus as the final word (it isn’t) from the only field RCT (it isn’t) on the only risk assessment instrument used in criminal law (not even close), one should be cautious.  Gender disparities have long been present in criminal justice, with women generally being arrested with less frequency than men and receiving more lenient treatment from the court system.  Saying that gender disparities exist, or even increase, because of some intervention requires careful thought about whether we want to do anything about those disparities, and if so, what we might want to do.  In this case, suppose the presence of the PSA-DMF System increased gender disparities by increasing the leniency of the criminal justice system’s treatment of women vis-à-vis similarly situated men (as appears to be partially the case from our illustrative result).  One way the criminal justice system could solve that “problem” would be by treating women more like men, i.e., treating women more strictly/harshly.  Is that what we want?

Good science is slow, sometimes maddeningly so.  Credible research takes time.  Useful inference requires careful attention to context, and policy decisions require the weighing of alternatives.  It is not helpful to any of these processes for those with a political ax to grind to present results designed to illustrate the application of new statistical methods, results based on a subset of preliminary data from one of several ongoing studies, as though those results were the last word on anything, including that one study.

One final thought:  The Access to Justice Lab, where two of us work and which conducted the Dane County field operation, is supported by Arnold Ventures, and the Dane County study itself is supported by Arnold Ventures.  That said, all five of us are agnostic at this stage about whether any criminal justice risk assessment instrument, including the PSA-DMF System, is a good or a bad thing.  In our view, credible evidence one way or the other does not yet exist.  What little that does exist suggests to us that the sturm und drang about risk assessments, from proponents and opponents, may be overblown.  Moreover, what one thinks about whether it is a good or bad idea to use risk assessment instruments of any kind should turn in huge part about what one would do if risk assessments are not used.  The most common alternative to the use of risk assessments, in criminal justice at least, is unguided, or loosely guided, or less guided human decision making.  And those opposed to the use of risk assessments in criminal justice apparently prefer these unguided human decisions.  The United States has had decades of experience with unguided human decision making in its criminal justice systems.  How has that gone?

James Greiner
Ryan Halen
Kosuke Imai
Zhichao Jiang
Sooahn Shin

* Note:  authors are listed in alphabetical order by last name

We’re closing out 2020 with a bang: We have three new studies in the field. Our amazing partners have worked with us to prepare and launch these projects in the face of the unprecedented challenges posed by COVID-19. Focusing on different topics and in different geographies, these new studies have the potential to improve access to justice across the U.S.

---

Transformative Justice Program (Williamson County, TX)

The Problem

In Texas, young adults aged 17 – 24 are overrepresented in the adult criminal justice system, accounting for 29% of the state’s arrests while only making up 11% of the population. Emerging adults also have the highest short-term recidivism of any age group due to underlying factors that are not addressed by the current criminal justice system. These factors include, among others, mental health struggles, substance abuse, and co-occurring disorders.

The Program

The program is Williamson County’s first felony diversion program. It diverts emerging adults charged with a low-level felony offense from the traditional criminal justice system. The program is a court-based system that combines release into the community with developmentally appropriate, intensive, community-based services. County staff work with participants to create individualized plans to address health, housing, educational, and other needs.

Eligible participants can be part of the program for up to 18 months, and they will be connected with community social services providers to help them meet their individualized program goals. Participants who successfully meet the goals identified in their individual plan will graduate from the program and will be eligible for expungement of the record of the arrest and charges.

The Study

In partnership with Williamson County, the Public Policy Research Institute at Texas A&M University, and the University of Texas Health Science Center at Houston (UTHealth) School of Public Health, the A2J Lab has launched a randomized evaluation of the diversion program.

The research team will conduct an RCT to identify the effects of the program on several outcomes including recidivism rates and quality of health outcomes. The study will also supplement criminal justice data with quarterly surveys and other data sources.

What We’ll Learn

This study will provide important evidence about the impact of community-based services. The population has a high level of need. Because they are so young, any benefit they experience could improve their lives and their communities for decades to come. With this research, we will learn the extent of this potential improvement.

---

Online Dispute Resolution (Carroll County, IA and Miami-Dade County, FL)

The Problem

Miami-Dade will likely issue tens of thousands of traffic compliance tickets in 2020. These are minor infractions that often simply require a motorist to show proof of license, registration, or insurance to resolve the case. The complicating factor is that the motorist needs to get to the courthouse on the right day and time and wait until it is their turn to give this proof. Similarly, the three different law enforcement agencies that police the Carroll County roadways issued thousands of traffic tickets last year. All of these tickets require an in-person appearance if there’s a discrepancy to resolve.

Attendance can be difficult, and the consequences for not being in the right place at the right time can be expensive. While the process of having these tickets resolved is often inexpensive, leaving the issue unattended can result in increased fines and fees. The courts’ solution to this?  Put it online.

The Program

With the installation of Online Dispute Resolution (ODR), motorists can now avoid the hassle of getting to the courthouse–which comes with all of the typical inconveniences of parking, reliable transportation, taking time off of work, finding childcare–and upload all of the compliance documents or negotiate discrepancies online. Prosecutors and court staff can then review what is submitted and handle the ticket without ever causing the motorist to set foot in the courtroom.

The Study

The A2J Lab is running RCTs in both counties. The research team will measure outcomes such as perceptions of the justice system, time to disposition, settlement success, and failure to appear.

What We’ll Learn

This change is happening at a critical time for state courts in public perception. As courts are moving practices online to cope with the COVID-19 pandemic, they have very little information about how those changes impact people’s perceptions of court experiences and other outcomes. This study will collect data on how using ODR impacts users’ overall perception of the justice system. This data, along with information about how cases resolve, will create evidence about how ODR improves the experiences of people in Miami-Dade and Carroll counties.

---

Plain Language Court Forms (DuPage County, IL)

The Problem

Reducing technical jargon and “legal-ese” is of major interest to access for justice advocates. For pro se litigants (people without lawyers), forms can be so complex that they require large sets of self-help materials to understand. Changing the forms themselves to be easy-to-read documents is a more efficient way to assist people without attorneys file the paperwork they need.

The Program

The Illinois Supreme Court Commission on Access to Justice created a Standardized Forms Committee to explore the creation of pro se forms that implement standardization and plain language practices. The new forms attempt to provide better guidance and support to pro se litigants by integrating plain language and simplification.

The study will include individuals who attempt to download a pro se divorce form for filing in DuPage County, Illinois. Any attempt to download a relevant form through the various websites, such as the DuPage County website (www.dupageco.org), the computers in the self-help center at the courthouse, or local legal aid organization websites will redirect to a study-based webpage. This webpage will verify eligibility for the study, provide information about the study, and subsequently provide either the new standardized pro se form or the previously created DuPage-specific pro se divorce form. These forms will be provided at random.

The Study

Through metadata and time-stamp information attached to the downloaded forms, as well as information documented in the court’s case management system, the A2J Lab will track a number of outcomes. These outcomes will include success of filing the form, time to filing, time to disposition, number of procedural errors, and whether the study participants use the downloaded form or end up reaching out to legal aid for assistance. The A2J Lab expects to conduct the field randomization for approximately six months and will track each case through disposition.

What We’ll Learn

This evaluation will give concrete evidence about what if any impact access to plain language forms has. Increasing the body of evidence about its efficacy will help courts decide how to deploy their resources.

---

Preliminary data for all of these studies will be available in the next year.

Today, the A2J Lab presented the findings from the interim report of the randomized evaluation of the Public Safety Assessment–Decision Making Framework System (PSA-DMF), a pretrial risk assessment tool and related decision-making framework, in Dane County, Wisconsin. The full interim report is now available on the A2J Lab website.

The presentation at the Dane County Criminal Justice Council was the first result of a years-long study, which is not yet complete. In the late spring of 2017, Dane County started providing the PSA-DMF System information to judicial officers deciding how much and what kind of bail and supervision to assign to individuals who have been arrested; such decisions affect whether the individual will be released or remain in jail until trial. Working with the A2J Lab and Arnold Ventures, Dane County’s randomized field experiment began a month later.

The PSA-DMF System is one tool in the toolbox that the judicial officer can draw upon in the exercise of their professional discretion. Scientists supported by Arnold Ventures produced the PSA by reviewing past data on criminal history, demographics, new crimes, rates at which participants failed to appear at their court hearing, and other potential risk factors. The PSA scores are applied to the Decision Making Framework, which brings together the information from the PSA with a community’s local policies and values, its laws, and its resources to provide a recommendation regarding pretrial release and, if release is obtained, a supervision level. Judicial officers may use the PSA-DMF System report when deciding whether to release an individual before trial, and this decision rests always with the judicial officer.

The experiment used something like a coin flip to divide cases into two groups. In the treated group, the judicial officer who was deciding how much and what kind of bail and pretrial supervision to assign received a paper printout with the PSA-DMF System information on it. In the control group, the judicial officer did not receive the paper printout. In other words, the control group received standard practice, and the treated group received the new system in which the judicial officer got the PSA-DMF System printout.

The report released this week analyzes data from one year of follow-up for cases that were included in the study between the start of the study and the middle of 2018. Dane County randomized cases from the middle of 2017 until the end of 2019. In this interim report, the A2J Lab analyzed data the County provided to compare the treated and control groups on (among other things) measurements of racial fairness, number of days incarcerated (if any), rates at which participants failed to appear at their court hearing, new criminal activity, and new violent criminal activity. Criminal justice officials in Dane County worked to provide information to the A2J Lab to facilitate the A2J Lab’s independent evaluation in a spirit of learning and a desire to improve.

Quantitatively, the A2J Lab does not yet have enough cases, or enough of a follow-up period on those cases, to make firm conclusions about whether it is better or worse to make the PSA-DMF System printout available to the judicial officer before assignment of bail (if any) and release conditions. The A2J Lab’s studies, like all randomized control trials, require a number of participants sufficient to detect policy-relevant differences between treatment groups (here, the PSA-DMF System report versus business as usual). The number of participants must be sufficient to analyze for statistical significance. Because of the two-year follow-up period, the County will provide the A2J Lab full information on all arrestees in the study sometime in early 2022. If the data are provided then, the A2J Lab’s final report will be available in the summer of 2022.

The limited data available thus far, not enough to draw firm conclusions, suggested several findings:

• There is some evidence that providing the PSA-DMF System printout to the judicial officer caused a change in the officer’s decisions.
• Generally, when the printout indicated that an individual presented lower risk, the judicial officer was less likely to require cash bail or, if cash bail was required, the amount was lower than in comparable control group cases. The opposite was generally true in the treated group when the printout indicated that the individual presented higher risk, as compared to the control group. This change was statistically significant but mild, and we cannot yet tell whether the change is policy-relevant.
• Treated group cases varied less in bail types and amounts than did control group cases. This change was strong and statistically significant.

As of the time of the interim report, there was no statistically significant difference between treated versus control group cases with respect to:

• various measures of the racial fairness of the judicial officer’s decisions;
• the number of days (if any) of pretrial incarceration;
• the frequency with which arrestees failed to appear at court dates; or
• the frequency with which arrestees were arrested for new crimes, including new violent crimes, during the pretrial period.

These finding might change, however, when the A2J Lab finishes analysis of the final dataset.

While it is too early to draw conclusions about whether the PSA-DMF System is positive, negative, or neutral for Dane County, this data provides a first look at the type of report that the final dataset will support.

Texas A&M University’s Public Policy Research Institute (PPRI) and the Access to Justice Lab at Harvard Law School (A2J Lab) are launching a randomized study of counsel at first appearance (CAFA) in Texas—the first study of its kind.

At first appearance, magistrates decide whether a person will be released while awaiting trial or held in jail. Yet only 4 of 254 Texas counties provide lawyers at this important stage. This week, the research team launches the evaluation in Hays County. Previously, counsel was typically assigned after first appearance. Going forward, the Texas Indigent Defense Commission (TIDC) will provide the funding for counsel to be present at magistration some of the time to evaluate the effect of provision of counsel.  The study will randomize days on which counsel is present for a first appearance. Eligible study participants will participate in brief interviews with study attorneys before being represented.

This study will use a randomized control trial (RCT) to test whether CAFA:

• Lowers money bond amounts
• Increases the likelihood of non-monetary bonds (personal bonds)
• Increases pretrial release
• Affects the likelihood of defendants returning to court
• Reduces jail costs and overall system costs

The Hays County COVID-19 response means that these appearances are now occurring remotely. This adjustment will allow researchers to observe a change in outcomes when appearance is remote versus when appearance is in-person upon resumption of normal operations. This observational analysis may pave the way to understanding the value of online courts.

Hays County Precinct 3 Commissioner Lon Shell said, “Hays County is excited to participate in this study. We believe it will provide vital information that can be used to continue improvements to our justice system.”

George Naufal, associate research scientist at the PPRI at Texas A&M University, says, “Understanding the impact of appearance at this stage will provide a huge benefit to the people of Texas. As more jurisdictions consider adding counsel at this stage, providing them with information about the effectiveness will help them decide if it’s a prudent use of their resources.”.

Jim Greiner, Faculty Director of the Access to Justice Lab at Harvard Law School, says, “The decisions made at the first court appearance have a huge impact. Those decisions mean the difference between spending days, weeks, months, or, in some cases, years in jail before a trial even occurs. With this study, we’ll get concrete evidence about whether providing counsel at those appearances makes a meaningful difference. If it does, this program could become a model for others.”

The study is supported by Arnold Ventures and the Texas Indigent Defense Commission.

Are you interested in learning about managing RCTs in the legal profession?

Work with us—we’re hiring!

We have an opening for a new Research Specialist to support the operations of some of our current research studies. If you’re interested in hands-on experience of quantitative research in law, we’d love to hear from you.

All inquires must be submitted through Harvard Law School’s online application portal.

We’re currently accepting applications. If you’re interested, please apply as soon as possible.