Identifying Serial Sexual Offenders Through Cold Cases

By Rachel E. Lovell, Ph.D., Angela Williamson, Ph.D., Thomas J. Dover, Ph.D., Timothy G. Keel, M.S., and Daniel J. Flannery, Ph.D.

An image highlighting the word "Rape" in a police report. Photo provided by The Plain Dealer, Cleveland, Ohio.

Photos courtesy of The Plain Dealer, Cleveland, Ohio.

Nationwide, hundreds of thousands of untested sexual assault kits (SAKs)—also known as rape kits—have languished for decades in evidence storage facilities.1 SAKs are sets of items (e.g., fingernail scrapings, vaginal swabs) collected by medical professionals to preserve evidence from a victim of sexual assault. Investigators then can use this evidence for investigation and prosecution. Many of these kits contain a definitive (and in some cases, the only) piece of evidence linking an offender to a victim—DNA.


Many factors contribute to the large number of untested SAKs. These include poor evidence tracking, outdated and ineffective investigation practices, scarce resources and personnel, misunderstanding of crime lab case acceptance policies, and lack of knowledge among law enforcement personnel about the value of testing the kits.2 Thus, the U.S. Department of Justice, Bureau of Justice Assistance (BJA), developed the National Sexual Assault Kit Initiative (SAKI) to motivate and assist those working to eliminate the backlog.

National Sexual Assault Kit Initiative

To address the country’s large numbers of untested rape kits, BJA developed SAKI to help local and state jurisdictions comprehensively address the backlog. Since 2015, BJA’s SAKI has awarded over $200 million to jurisdictions in need. It will reward an additional $47 million in fiscal year 2020.

As of June 2019, 54 SAKI jurisdictions representing 35 states have inventoried more than 87,325 SAKs. Personnel have sent over 54,200 kits for testing, and new entries already have led to more than 8,537 DNA hits on the national CODIS database. As the thousands of kits already identified through SAKI are tested, DNA evidence may help solve anywhere from 30 to 50 percent of these backlogged cases.

For cases that DNA evidence alone cannot solve, BJA also has formed a partnership with the FBI’s ViCAP, which helps agencies link homicide and sexual assault cases based on case similarities and behavioral evidence. Any case stemming from a SAK that fits ViCAP’s criteria is placed in this national data-sharing system.

Sources: “Sexual Assault Kit Initiative,” accessed November 4, 2019,; and U.S. Department of Justice, Bureau of Justice Assistance, Over 47,000 Kits Tested Through BJA’s Sexual Assault Kit Initiative Since 2015, April 26, 2019, accessed September 27, 2019,

Traditionally, law enforcement sees the utility of testing kits as a way to identify an unknown suspect when the victim wants to prosecute. However, these kits have the potential to provide much more for investigations. They also can connect offenders to previously unsolved crimes, confirm identities of known offenders, exonerate innocent suspects, and populate the federal DNA database, also known as the Combined DNA Index System (CODIS).

Cuyahoga County, Ohio, and Wayne County, Michigan, represent two urban jurisdictions addressing large inventories of untested kits (approximately 22,000 combined) that span decades. Both regularly submitted few kits for forensic testing before the mid-to-late 2000s.3 This makes them ideal case studies for examining sexual offending—they present a more complete picture of the extent of the issue, especially in terms of repeat offenses.

Rachel Lovell

Dr. Lovell is a research assistant professor with the Begun Center for Violence Prevention Research and Education at Case Western Reserve University in Cleveland, Ohio.

Tom Dover

Dr. Dover serves as a crime analyst with the FBI’s Behavioral Analysis Unit-5 and instructs on topics pertaining to behavioral analysis for the FBI National Academy in Quantico, Virginia.

Daniel Flannery

Dr. Flannery, the Dr. Semi J. and Ruth W. Begun Professor, is director of the Begun Center for Violence Prevention Research and Education at Case Western Reserve University in Cleveland, Ohio.

Angela Williamson

Dr. Williamson is the supervisor of National Initiatives with the U.S. Department of Justice’s Bureau of Justice Assistance and serves as the agency’s ViCAP liaison at the FBI.

Timothy Keel

Mr. Keel, a certified behavioral analyst (profiler), is a major case specialist with the FBI’s Behavioral Analysis Unit-4 in Quantico, Virginia.

Financial Benefits of SAK Testing

Researchers assessed the cost-effectiveness of testing previously neglected kits in Cuyahoga County. As of January 1, 2016, personnel had tested 4,347 SAKs as part of the BJA SAKI program. On average, each sexual assault costs a victim more than $200,000 in pain and suffering, earnings lost to injury, medical expenditures, and decreased quality of life. Meanwhile, each SAK costs approximately $950 to test and $1,250 to investigate and provide victim advocacy. This results in a total cost of $2,200 per tested SAK and approximately $9.6 million for all tested SAKs. By testing and following up with an investigation for each SAK, the Cuyahoga County SAK Task Force estimates savings for the community of almost $51,000 per convicted offender, or $48 million total, resulting in a net savings of $38 million. These analyses indicate that cost savings to the community do not come from testing itself but from investigating and prosecuting the offenders, thereby preventing future sexual assaults.

Source: Mendel Singer, Rachel Lovell, and Daniel Flannery, Cost Savings and Cost Effectiveness of the Cuyahoga County Sexual Assault Kit Task Force (Cleveland, OH: Begun Center for Violence Prevention Research and Education, May 2016), accessed October 2, 2019,

Sexual Assault Investigations

Lessons Learned

Beyond just the testing of kits, the findings coming out of SAKI already challenge many assumptions about sexual assault and the offenders who commit it. The authors of this article comprise a unique group of experts—academic researchers, a forensic specialist heading up BJA’s SAKI, and members of the FBI’s Behavioral Analysis Unit. They have collaborated on this project because they bring distinct perspectives from which to contextualize the SAKI findings. As a result, the following five recommendations are suggested for law enforcement agencies investigating sexual assault.


1) Consider that a rapist possibly has done this before, will continue to do this in the future, or both.

Much of what people know about sexual offenders, especially serial sex offenders, comes from convictions and self-reports. This leads to an unrepresentative sampling of sexual perpetrators. In research on self-reported proclivity to commit rape, approximately 35 percent of surveyed male college students indicated some likelihood of committing a rape if given the opportunity to do so without facing consequences.4

It is important to note three things about this statistic. First, it comprises individuals willing to report this proclivity on a survey and most likely undercounts the true number. Second, a significant number of respondents may not actually be willing to offend because they lack the opportunity to do so with impunity. Third, it is possible, and perhaps even likely, that at least some of these individuals did commit rape and either were caught later in life or continued as unknown offenders. Thus, there exist plenty of potentially motivated offenders.

Concerning convicted offenders, prior recidivism research on sexual offenders shows that approximately 5 percent of individuals released from prison for a sexually based offense were rearrested for another such offense within 3 years.5 This is a relatively low rate of recidivism, but, again, these estimates are based on knowledge of repeat offending from official criminal justice records. Additionally, limiting the notion of recidivism to a 3-year window ignores the potential for offenders to revert to offending once initial post-release scrutiny (e.g., probation, parole, mandatory sex offender treatment) ends.

In contrast, DNA evidence collected from these now-tested kits can connect previously unlinked sexual assaults through more objective means and without the prism of self-reporting. Also, this data hinges on evidence gathered at the time the kit was collected, not at conviction, which provides a more representative picture of reported sexual assaults.

Research found that serial sex offending proves much more common than expected—approximately a third to a quarter of subjects identified as part of these SAK initiatives were serial offenders.6 These are low estimates and only represent those cases reported to law enforcement to begin with and where DNA was collected and eventually tested.

Practically, this means several things had to happen:

  • The victim reported the sexual assault
  • A kit was collected
  • The kit included enough DNA from the offender for testing
  • Law enforcement retained the kit
  • Personnel submitted the kit for testing
  • The kit had enough DNA information to meet DNA database eligibility

A breakdown at any of these steps means the DNA did not make it into the system, and the rapes would not be linked.

New Information from Testing

As of 2017, Wayne County identified over 800 serial offenders by testing its backlog of kits. Of those serial offenders, over 170 are unknown, meaning that their DNA links them to multiple cases but they have yet to be identified and most likely are still at large. Additionally, DNA from Wayne County’s kits has hit to offenders or other cases in over 40 different states. This highlights the mobile nature of offenders and the importance of data sharing among law enforcement across the country.

Source: Wayne County Prosecutor, Wayne County Sexual Assault Kit Task Force FAQ April 2017, accessed December 6, 2019,

“Research found that serial sex offending proves much more common than expected—approximately a third to a quarter of subjects identified as part of these SAK initiatives were serial offenders.”

Data coming from these jurisdictions also shows what happens when criminals can continue to offend. One such offender is Nathan Ford. DNA has connected Ford, a former probation officer, to 18 rape kits in northeast Ohio to date and to several other sexual offenses for which there exists no DNA—rapes that span decades.7 These only are the offenses involving Ford that law enforcement knows of. No one could have realized the full extent of his offending had it not been for the now-tested SAKs.

2) Do not assume a consistent modus operandi and victim preference.

While it may be likely that an offender has committed other sexually based offenses, investigators cannot assume the perpetrator kept the same modus operandi (MO) or victim preference across incidents. Standard investigative practice in many law enforcement agencies involves either investigating a sexual assault allegation as an isolated event or using the offender’s MO or victim preference to link other sexual assaults to the offender. However, research calls this practice into question and shows that offenders often vary their MO drastically across assaults.8

Nathan Ford seemingly chose many of his victims at random, with varying MOs and victim preference.9 This made it exceptionally difficult to link his cases together. His victims included high-risk women with substance abuse issues and histories of prostitution, a college student, a schoolteacher, and three teenagers, among others. He dragged some of his victims from the street into secluded areas, assaulted one during a home invasion, and attacked another while she studied on campus.

Serial sex offenders frequently assault both strangers and acquaintances and often exhibit intraserial variations in victim relationship, age, and even gender. For example, one offender identified as part of Cuyahoga County’s SAKI was connected to two sexual assaults. First, he was one of four offenders who sexually assaulted a female juvenile stranger at a party. His second sexual assault occurred 2 months later, and the victim was his 3-year-old son.

Just because an offender and victim have a domestic relationship or know each other does not mean the offender does not also have victims who are strangers. In fact, investigators should expect to find victims who are strangers.10 The best way to link offenses is by collecting and testing SAKs, even in domestic sexual assaults where the offender is a named suspect or has confessed.

Next, officers must prepare to explain to prosecutors, judges, jurors, parole boards, and other related parties that offenders do not necessarily follow substantially similar patterns across offenses. For example, one offender identified as part of the Kansas Bureau of Investigation’s SAKI began committing forcible rapes at the age of 15. His victims often were underage female acquaintances from unstable homes. In less than 3 years, four different victims reported that he had raped them. The victims in the first three incidents were acquaintances of the suspect. The fourth was an intimate partner. A few weeks after raping the fourth victim, he kidnapped and beat her. He strangled the woman to a state of unconsciousness, buried her in a shallow grave for a short time, and then held her captive in his basement for days before allowing her to leave.

3) Do not rely solely on DNA to solve a case.

While SAKs can prove extremely useful to law enforcement for all the reasons detailed, officers cannot assume that DNA is the only investigative tool necessary. Not every kit will contain CODIS-eligible DNA, and not every CODIS-eligible DNA profile will identify a known offender or provide an investigative lead. While advanced methodologies, such as forensic genealogy, certainly can help identify numerous suspects, estimates suggest that DNA will not be the key to solving over 50 percent of cases.11

Behavior cues present at the scene also can help link offenders. The FBI’s Violent Criminal Apprehension Program (ViCAP) proves ideal for helping to connect cases. As such, in early 2017, the BJA’s SAKI team partnered with ViCAP to increase use of the database and sharing of information from SAKI-related cases across jurisdictions and the country.

This may seem to contradict the previous recommendations. In fact, these recommendations complement each other. ViCAP, in addition to DNA testing (where applicable), allows law enforcement to use crime scene evidence more exhaustively. It also proves crucial to remember that most sexual assault cases produce a surviving witness who often can provide key information regarding the suspect, such as physical appearance, speech pattern or accent, MO, and type of vehicle. Investigators can link all these potential identifying features across jurisdictions and states by sharing data in ViCAP.12

An image containing the word DNA provided by The Plain Dealer, Cleveland, Ohio.

4) Do not assume sexual offenders only commit sexual offenses.

Research on the criminal histories of sexual offenders identified via previously untested kits in both Cuyahoga County and Wayne County found that most offenders did not have a prior arrest or conviction for rape in their criminal history. However, almost all of them had at least one arrest for a different serious offense, with violent offenses frequently appearing in their criminal histories.13

Some of these other offenses may have been precursors of, or related to, sexual assaults; however, many of them simply may show that these offenders are criminally versatile. This means that focusing investigative efforts on sexual offenders potentially can reduce not only sex offenses but also other crimes.

For example, one offender identified in Cuyahoga County was linked to two now-tested kits, one from when he was 13 and the other 15. In the decades that followed, he never was arrested for another rape, but he was arrested and convicted for several other crimes, including felony assault, motor vehicle theft, theft, robbery, kidnapping, and aggravated murder.

5) Be patient, understand the limitations of CODIS, and verify that DNA from convicted offenders of interest actually is in the database.

The SAKI already has resulted in the addition of over 16,000 profiles to CODIS. Profiles remain in the database, potentially solving future crimes. The more profiles added to CODIS, the more robust the database becomes. An expanded DNA database enables law enforcement to better identify persons convicted of violent and sexual offenses who also are involved in unsolved crimes and who may reoffend after release.

“The best way to link offenses is by collecting and testing SAKs, even in domestic sexual assaults where the offender is a named suspect or has confessed.”

This is why testing kits associated with sexual assaults committed by nonstrangers, from incidents that occurred outside the statute of limitations, and from victims who do not want to prosecute proves so important. DNA profiles extend beyond the crime in question. Law enforcement may be unable to get justice for a particular victim, but they might for past or future victims.

Sexual offenders may, and often do, have prior arrests or convictions. However, investigators cannot assume that these offenders' DNA already is in CODIS. Typically, eligible convicted offenders from whom DNA lawfully can be collected include those arrested, facing charges, or convicted of murder, sexual abuse, kidnapping, or other qualifying state offenses.14 However, numerous jurisdictions have discovered gaps in the collection of lawfully owed DNA. This refers to a DNA profile from a qualifying offender who should have a sample in CODIS (based on the type and time of the offense in relation to applicable state law) but from whom a sample never has been collected or entered into the database. In some instances, thousands of profiles are missing.15

It is crucial that agencies test unsubmitted SAKs and upload all eligible profiles to CODIS. However, cases will remain unsolved unless evidence profiles can match against convicted offenders’ profiles. The absence of lawfully owed DNA samples in CODIS likely will result in missed opportunities to identify perpetrators of many crimes, including sexual assaults. For example, Michigan demonstrated the importance of lawfully owed DNA collections in 2011 when its department of corrections collected samples from 5,000 prisoners who had slipped through the cracks. Subsequent DNA hits in CODIS were linked to 74 crimes, including 5 murders, 23 rapes, and 3 armed robberies.16

Investigators also need to remain aware of local and state laws impacting their ability to upload convicted offender DNA to CODIS. These include state DNA laws, most of which only allow entry of a suspect’s DNA into CODIS for convictions associated with crimes after 1994 or 1995, when the laws were enacted. As such, many offenders (incarcerated, paroled, or deceased) associated with older cases may not have had their DNA entered into CODIS. Investigators should verify specific individuals of interest with their local or state CODIS administrator.


Testing the evidence contained in SAKs is crucial and, in most states, no longer optional for law enforcement. The benefits explained in these five recommendations make it clear that law enforcement must do all it can with the forensic advancements now available to bring justice to the victims of sexually motivated crimes. However, testing is just the first step—it should be followed by investigations and prosecutions. SAKI grants provide as much funding for prosecutions and investigations as they do for testing kits. Ultimately, both testing evidence and sharing information are key to identifying suspects and reaching successful prosecutions. There is little doubt that doing so will help identify offenders and also prevent those who are likely to reoffend.

“[T]esting is just the first step—it should be followed by investigations and prosecutions.” 

Dr. Lovell can be reached at, Dr. Williamson at, Dr. Dover at, Mr. Keel at, and Dr. Flannery at


1 “What is the Rape Kit Backlog?” End the Backlog, Joyful Heart Foundation, accessed December 6, 2019,
2 Kevin J. Strom et al., The 2007 Survey of Law Enforcement Forensic Evidence Processing (Research Triangle Park, NC: RTI International, October 2009), accessed September 11, 2019,; and Bethany Backes, moderator, “Taking on the Challenge of Unsubmitted Sexual Assault Kits” (webinar, Government Innovators Network, Harvard Kennedy School, October 31, 2014), accessed September 30, 2019,
3 Rebecca Campbell et al., The Detroit Sexual Assault Kit (SAK) Action Research Project (ARP), November 9, 2015, accessed October 2, 2019,; and Misty Luminais, Rachel Lovell, and Daniel Flannery, Perceptions of Why the Sexual Assault Kit Backlog Exists In Cuyahoga County, Ohio and Recommendations For Improving Practice (Cleveland, OH: Begun Center for Violence Prevention Research and Education, August 2017), accessed September 30, 2019,
4 Neil M. Malamuth, “Rape Proclivity Among Males,” Journal of Social Issues 37, no. 4 (Fall 1981): 138-157, accessed September 11, 2019,
5 U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, Recidivism of Sex Offenders Released from Prison In 1994, Patrick A. Langan, Erica L. Smith, and Matthew R. Durose (Washington, DC, 2003), accessed October 2, 2019,
6 A serial sex offender is defined as being connected to more than one untested kit or connected to one or more kits and having a criminal history that includes a sexual offense. Rebecca Campbell, et al., "Connecting the Dots: Identifying Suspected Serial Sexual Offenders Through Forensic DNA Evidence," Psychology of Violence 10, no. 3 (2020): 255-267, accessed April 29, 2020,; and Rachel Lovell et al., “Describing the Process and Quantifying the Outcomes of the Cuyahoga County Sexual Assault Kit Initiative,” Journal of Criminal Justice 57 (July-August 2018): 106-115, accessed September 11, 2019,
7 Amanda Garrett, “New Akron Police Unit Delves into 20 Years of Unsolved Rapes,” Akron Beacon Journal, June 22, 2019, accessed September 11, 2019,
8 Rachel Lovell and Dan Clark, Do Serial Sex Offenders Maintain a Consistent Modus Operandi?: Findings from Previously Unsubmitted Sexual Assault Kits (Cleveland, OH: Begun Center for Violence Prevention Research and Education, September 2017), accessed September 30, 2019,
9 Rachel Dissell, “Nathan Ford, Linked to at Least 15 Cleveland Rapes, Blames ‘Irresistible Impulses,’”, March 28, 2017, accessed September 11, 2019,
10 Rachel E. Lovell et al., “Understanding Intimate Partner Sexual Assaults: Findings from Sexual Assault Kits,” Journal of Aggression, Maltreatment & Trauma 28, no. 1 (2019): 8-24, accessed September 11, 2019,
11 Campbell et al., “The Detroit Sexual Assault Kit”; and Lovell et al., “Describing the Process.”
12 See Federal Bureau of Investigation, ViCAP Homicides and Sexual Assaults, accessed September 11, 2019,
13 Rachel Lovell et al., “Offending Histories and Typologies of Suspected Sexual Offenders Identified via Untested Sexual Assault Kits,” Journal of Criminal Justice and Behavior (forthcoming); and Campbell et al., “The Detroit Sexual Assault Kit.”
14 DNA Analysis Backlog Elimination Act of 2000, Public Law 106-546, 106th Cong., (December 19, 2000), accessed October 2, 2019,
15 Owed DNA Research Project (Cleveland, OH: Begun Center for Violence Prevention Research and Education, Jack, Joseph and Morton Mandel School of Applied Social Sciences, Case Western Reserve University, 2019), accessed September 11, 2019,
16 Michigan Department of Corrections, DNA Profile Collection Completed, accessed October 2, 2019,,4551,7-119--264812--,00.html.