U.S. flag

An official website of the United States government, Department of Justice.

Untested Evidence in Sexual Assault Cases

Many jurisdictions across the country are looking at the issue of sexual assault evidence that has not been submitted to a crime lab for testing
Date Published
March 17, 2016

On this page find:

Overview of the Issue

. It is unknown how many unanalyzed sexual assault kits (SAKs) there are nationwide. There are many reasons for this, but one is that tracking and counting SAKs is an antiquated process in many U.S. jurisdictions. And, importantly, the availability of computerized evidence-tracking systems has been a significant issue for many of the nation’s jurisdictions for years.

Also, history matters. Modern DNA forensic analysis was not widely used until the late 1990s. A database against which to compare the DNA of individuals suspected of sexual assault is only as robust as the amount of information in it. The Combined DNA Index System (CODIS) took years to build up the number of profiles in its database to allow law enforcement to routinely submit DNA samples and to enable crime labs to complete the extensive processes to qualify and upload DNA profiles.

In 2011, NIJ also published a special report in response to the recent discoveries of thousands of untested SAKs in police evidence rooms nationwide. The Road Ahead: Unanalyzed Evidence in Sexual Assault Cases (pdf, 29 pages) explores a wide range of issues, including how untested SAKs affect various stakeholders in the nation's criminal justice system: the police and crime laboratories; the courts; victim service agencies; policymakers at the federal, state and local levels; and the victims.

Action-Research Projects in Detroit & Houston to Find Solutions

In 2011, after a competitive selection process, NIJ awarded “action-research” grants to the Houston, Texas, Police Department and the Wayne County (Detroit), Mich., Prosecutor’s Office to form multidisciplinary teams to look at the issue of unsubmitted SAKs in their jurisdictions. On the teams were boots-on-the-ground practitioners who deal with sexual assaults day-in and day-out: police officers, crime lab analysts, prosecutors and victim advocates. And, because NIJ is a research agency — dedicated to finding answers to criminal justice problems through science — it also ensured that social-science researchers with nationwide reputations were on the teams to work hand-in-hand with the practitioners.

The overarching goal in the Detroit and Houston projects was, first, to understand the scope of the issue: How many unsubmitted kits were there, and how and why did the problem develop? Then, the teams were charged with identifying effective, sustainable responses. NIJ’s goal was to not only help Houston and Detroit, but to determine if lessons learned in these two cities might help other jurisdictions.

One of the most interesting — and challenging — features of action research is that it does not capture a single snapshot in time. It takes place in the ever-evolving real world where policies and protocols, technologies, and political realities are changing all the time. In Houston, two relevant changes occurred during the project: (1) the city moved to a state-of-the-art evidence storage facility with sophisticated evidence-tracking capabilities, and (2) the state passed a law requiring that all SAKs (both in storage and current cases) be tested. Therefore, lessons learned in the NIJ-funded project were inextricably intertwined with other changes in that jurisdiction.

Just as in Houston, the Detroit action-research team had to do its work within the larger context of changing realities, including statewide legislation that required testing of all SAKs with the victim’s approval and an infusion of $4 million from the Michigan Attorney General’s Office to test all previously unsubmitted SAKs.

How applicable are the findings in Houston and Detroit — including the protocols they developed for performing a census, DNA testing, and victim notification — to other jurisdictions? Or, as scientists, put it, ‘How generalizable are they?’

Although the projects in Detroit and Houston are only two “case studies” and other jurisdictions are likely to find different results from similar activities, there is no doubt that other jurisdictions may be able to draw upon lessons learned.

Read from the Detroit report Chapter 1: Introduction - The Detroit Sexual Assault Kit Action Research Project (pdf, 551 pages), including an overview of the issue of unsubmitted SAKs, the project’s goals, and the action-research team and how the collaborative process worked. See also Chapter 6: Discussion - Summary of Findings, Implications, and Community Changes (pdf, 551 pages), including the project’s goals, overall lessons learned, and implications and changes in policies and practices.

Additional findings from the Houston action-research project will be released soon.

Lessons Learned: The Value of Multidisciplinary Teams

One of the most important findings to come out of both jurisdictions is the value of forming multidisciplinary teams when addressing the issue of a large number of previously unsubmitted SAKs. Solving sexual assault crimes — bringing rapists to justice and supporting victims — is much more complicated than simply testing the kits: police must perform investigations to develop other potential evidence in the case; prosecutors must file charges; juries must render verdicts; and judges must sentence. The Detroit and Houston team members were able to bring their unique expertise to bear on devising DNA testing, victim-notification, and other protocols.

Among the successes in Houston that the action-research team played a role in:

  • Formation of a CODIS “hit squad” in the Houston Police Department (HPD) to perform investigations when suspects were identified through DNA testing of previously untested SAKs.
  • Hiring a fulltime justice advocate who is now embedded in HPD’s Adult Sex Crimes Unit.
  • Training on the neurobiology of trauma, which has resulted in changes to how police and prosecutors interview victims.
  • Dedication of 11 experienced prosecutors to evaluate cases with CODIS hits.
  • Dedication of a fulltime counselor in the Houston Area Women’s Center.
  • Development of formal protocols to notify victims, based on SAK testing results, including creation of a hotline for victims to obtain information about their case.
  • Creation of a team within the Forensic Science Center’s biology section dedicated to finalizing protocols on kits that were sent to private laboratories for testing and uploading DNA profiles into CODIS.
  • Upgrades to information management systems to ensure that CODIS-hit information was quickly communicated to those who need it.

Among the successes that the Detroit team played a role in:

  • A census of more than 11,000 previously unsubmitted kits.
  • Identifying the “risk factors” that contributed to the development of a large number of unsubmitted kits.
  • DNA testing of nearly 1,600 kits.
  • Development, implementation and evaluation of victim-notification protocols.
  • Increased collaboration among all the stakeholders.
  • Training on the neurobiology of trauma, which has resulted in changes to how police and prosecutors interview victims.

Read from the Detroit report: Chapter 6: Discussion - Summary of Findings, Implications, and Community Changes (pdf, 551 pages), including the project’s goals, overall lessons learned, and implications and changes in policies and practices.

Determining the Number of Untested Kits in a Jurisdiction — Performing a Census

When the NIJ-funded projects began in the spring of 2011, officials in Houston and Detroit did not know how many untested SAKs were stored in their evidence property rooms, but the number was in the thousands.

In Detroit, the first step was to perform a census. Team members manually counted every kit collected from 1980–Nov. 1, 2009, opening each one (approximately 11,300 SAKs) and recording the name of the victim, date of birth, and the date of the assault. The census in Detroit took 15 weeks and 2,365 person-hours.

Detroit found 2,512 SAKs with lab numbers but could not determine how many of these had been tested; 8,707 had never been submitted to the lab.

Read more from the Detroit report: Chapter 2: The Scope of the Problem - How Many Unsubmitted SAKs (pdf, 551 pages), including conducting a census.

In Houston, there were an estimated 16,000 rape kits in police storage. However, the action-research project was occurring in tandem with the HPD’s move to a new evidence-storage facility, so HPD was already performing an audit of all SAKs in their custody. When the audit was completed, HPD determined that 6,663 SAKs had not previously been tested. This included approximately 4,000 kits stored in the property room freezer. Of these 4,000, the researchers randomly selected a sample of 500 to be studied in the NIJ action-research project. Results from that study are expected soon.

Understanding the Number of Unsubmitted Kits

One of the questions NIJ asked the teams in Houston and Detroit to answer was why there were so many unsubmitted SAKs. In Detroit, the researchers examined 20 years of records and interviewed detectives, prosecutors, advocates, elected officials, laboratory personnel and sexual assault nurse examiners. They determined that the problem was rooted in long-standing relationships within and between organizations. The “risk factors” revealed in Detroit include:

  1. Victim-blaming beliefs and behaviors.
    EXAMPLE: Police officers had feelings such as ‘She’s not acting like a real victim,’ or ‘This is a he-said, she-said situation.’
  2. No written policy or protocol for submitting kits to the lab for testing.
  3. Budget cuts that reduced the number of law enforcement personnel assigned to sexual assault cases and crime lab staff, and problems with lab capacity, such as inefficient DNA testing equipment/methods.
  4. High turnover in police leadership, which makes it difficult to identify and remedy front-line practices.
    EXAMPLE: Nine chiefs in a 20-year period; five chiefs in less than three years. 
  5. Strained relationships and lack of training among the necessary partners.

Read more from the Detroit report: Chapter 2: Underlying Reasons Why So Many Unsubmitted SAKs, including the larger historical context and particular risk factors for law enforcement, the prosecutor’s office, the medical system, victims advocacy agencies, and the crime lab.

Results of DNA Testing in Detroit

In Detroit, the SAKs that were tested as part of the action-research project were broken into four testing groups:

  1. Stranger-perpetrated sexual assault (450 cases).
  2. Non-stranger-perpetrated sexual assault (450 cases).
  3. Cases presumed to be beyond the statute of limitations (350 cases).
  4. A comparison of traditional vs. selective degradation DNA-testing methods (350 cases).

Of the 1,595 kits tested in the NIJ-sponsored action-research project in Detroit, nearly half (49 percent; 785 kits) yielded DNA profiles that could be uploaded to CODIS. Out of the 785 profiles uploaded to CODIS, there were 455 “hits.” That means that 28 percent of the SAKs tested in the Detroit action-research project revealed the DNA identification of a potential suspect. Of these, 127 serial assaults were identified. [NOTE: All figures in the Detroit final report are as of Dec. 31, 2013.]

Additional research is underway to understand more about serial assaults in Detroit.

Read more from the Detroit report: Chapter 4: Testing Kits - Developing & Evaluating a SAK Testing Plan (pdf, 551 pages), including DNA-testing outcomes and CODIS hits on the randomly selected sample of 1,600 kits that were tested in the action-research project.

Creating Victim-Centered Notification

The cooperation of the victim can be critical to the success of a sexual assault investigation and prosecution. Determining a victim’s willingness to be part of the criminal justice process can be particularly problematic in older cases where the victim may not have heard anything since agreeing to the rape-kit exam.

Significant research — including that done by Dr. Rebecca Campbell, the lead researcher on the Detroit action-research project — has shown that reactivation of memories of an assault may:

  • Trigger flashbacks and other symptoms.
  • Exacerbate PTSD or other mental health issues.
  • Increase substance use/abuse.
  • Trigger substance abuse relapse.

Download Notifying Sexual Assault Victims After Testing Evidence (pdf, 20 pages), co-published by NIJ and the Office for Victims of Crime. 

Therefore, one of the goals of the projects in Detroit and Houston was to develop victim-notification protocols that were victim-centered and trauma-informed. For more details on victim-notification protocols developed in Detroit:

In Houston, detailed protocols were created for contacting victims depending on the results of SAK testing and whether there was a CODIS hit. A hotline for victims to call was also created.

This report includes step-by-step details regarding victim notification and investigative reviews of case files in Houston. It also includes protocols regarding Houston’s creation of two databases — one containing details of SAK-testing results, one for documenting calls/emails — to ensure that the most up-to-date information is conveyed to victims who call the hotline or send an email.

As part of its work, the Houston action-research team also explored the importance of victim cooperation in a sexual assault investigation. To help address this issue, the Houston Police Department temporarily hired a justice advocate in 2013.

Thereafter, the Houston action-research team conducted focus-group sessions with detectives to explore the impact of the justice advocate position on sexual assault investigations. The response was so positive that when results were presented to HPD leadership, the position was made permanent.

Los Angeles: Judicial Outcomes Study Post-DNA Testing

In 2006, NIJ provided grant support to researchers at California State University to examine the role of DNA testing of untested SAKs in judicial outcomes. The grant was modest — $100,000 — and, therefore, the study had a narrow focus, including time limitations.

The researchers randomly selected 371 SAKs (from approximately 11,000 kits in the property rooms of the Los Angeles Police Department and the Los Angeles Sheriff’s Department) to:

  • Assess the efficacy of DNA testing.
  • Determine the criminal justice outcomes (arrest, charge, conviction) within the first six months after the kits were DNA tested.

The findings with respect to the study's second goal were surprising to many. In the first six months after the 371 SAKs were tested, no new arrests were made, new charges were filed in one case, and there were two convictions. In fact, it is probable that the DNA testing was not responsible for the single filing and the two convictions.

There are a number of important facts to keep in mind when trying to understand these results. First, the study looked at case adjudication in only the first six months after testing. The researchers did not examine whether there have been additional arrests, charges filed or convictions since that time. Second, the sample size was small, and the findings are from only one site. Great caution should be used in trying to extend the findings to other locales. Indeed, the reasons for large numbers of untested SAKs in police property rooms may be very different in other jurisdictions.

Learn more about the project in:

New Orleans: DNA Testing Outcomes

In 2011, NIJ supported a one-year project to perform DNA testing of 1,000 SAKs in New Orleans: 830 SAKs were in the custody of the New Orleans Police Department at the time the project began and 178 SAKs were collected during the project (“current” cases).

There were hits against CODIS that aided police investigations in 13 percent of the cases.

Among the SAKs that had not been tested before the project began, 10 percent yielded a CODIS hit: 9 percent were hits to an individual who had not previously been identified as a suspect in the case, 0.5 percent were forensic hits in which the person who committed the crime was unknown, and 0.5 percent were hits between a person convicted of a crime (either known or named by the victim) and the evidence.

Among the current cases, 31.5 percent yielded a CODIS hit: 21.4 percent were hits to an individual who had not previously been identified as a suspect, 1.7 percent were forensic hits in which the individual was unknown, and 8.4 percent were hits between a person convicted of a crime (either known or named by the victim) and the evidence.

Learn more about the results of the New Orleans project in:

Date Published: March 17, 2016