The purpose of this study was to determine how many SAKs were in Detroit police custody; determine the causes that led 8,717 sexual assault kits (SAKs) to be left untested; develop a plan for testing the SAKs; and to create a victim notification protocol.
This project has four distinct steps:
1) The first step involved conducting a census of all sexual assault kits (SAKs) in police custody in the city of Detroit. The census took 15 weeks to complete and revealed that there were 11,219 SAKs in police custody and 8,717 SAKS had never been submitted for testing.
2) The second step involved investigating why so many SAKs were un-submitted. The research team interviewed current and former employees in all organizations involved (police department, crime lab, prosecutor's office, sexual assault nurse examiners and advocacy groups) as well as examined publicly-available documents and internal organizational records to assess the resources available for serving rape victims and testing SAKs.
3) The third step in the process was to develop a plan and test untested SAKs. A stratified random sample of 1,595 kits were split into four testing groups, each one designed to address specific research questions regarding the utility of SAK testing under different case circumstances:
- Testing Group 1: Stranger-perpetrated sexual assaults
- Testing Group 2: Non-stranger perpetrated sexual assaults
- Testing Group 3: Sexual assault cases that were presumed to be beyond the statute of limitations (SOL).
- Testing Group 4: SAKs were randomly assigned to two different DNA testing methods to examine whether an emerging testing method, selective degradation, could offer faster and less expensive testing options without sacrificing accuracy (relative to traditional DNA testing methods).
A fifth testing group sample, drawn from SAKs outside of the original stratified sample of 1,595 kits, was included to compute sampling weights for the analysis. This sample of 250 SAKs was drawn from the 400 kits tested as part of the earlier 400 Project. While not in the original stratified sample, all 400 of The 400 Project SAKs were included in the 15 week census.
All Testing Groups were compared with respect to their rates of Combined DNA Index System (CODIS) entries (the SAKs contained a DNA eligible profile for CODIS), CODIS hits (a DNA match to a profile in CODIS), and serial sexual assault hits (a DNA match across two or more SAKs). Forty-nine percent of the SAKS tested were CODIS eligible.
4) The fourth step of the project was to develop and evaluate a victim notification protocol. This was achieved by reviewing cases that had CODIS hits, meeting with the victims and notifying them that their SAKs hadn't been tested originally and discussing options for the future, including connecting them with community survivor services. Forty-one cases were selected for notification by the multidisciplinary review team and 31 survivors were able to be found. The contact was made by investigators affiliated with the prosecutor's office and typically occurred at the victim's home. Fifty-seven percent of the victims decided that they wanted to participate in the investigation and prosecution process.
A stratified random sample of 1,595 sexual assault kits (SAKs) were tested out of the 8,717 SAKs in Detroit police custody as of November 1st, 2009. The sample of SAKs were divided into the following testing groups:
- Testing Group One included 450 stranger assailant cases from 2002-2009.
- Testing Group Two consist of 450 non-stranger assailant cases from 2002-2009.
- Testing Group Three consist of 350 cases that occurred before 2002 (or presumed to be after the statute of limitations) and included both stranger and non-stranger cases.
- Testing Group Four consist of 350 cases from 2002-2009 that were then randomly assigned to one of two different DNA testing methods: traditional extraction vs. selective degradation and included stranger and non-stranger cases.
Testing Group Five was drawn from SAKs outside of the original stratified sample of 1,595 kits and was included to compute sampling weights for the analysis. This sample of 250 SAKs was drawn from the 400 kits tested as part of the earlier 400 Project. This group included 250 cases from 2002-2009 for which victim offender relationship was known.
All sexual assault kits (SAKs) in Detroit police property on November 1st, 2009 and current and former employees of all organizations involved with the handling or testing of SAKs in Detroit police custody.
Sexual assault kits (SAKs).
Police forensic testing spreadsheets that determine which SAKs had been submitted to the lab for testing.
Data from the vendor laboratory that tested the SAKs.
Archival records from agencies directly involved in the events surrounding the discovery of the un-submitted sexual assault kits (SAKs).
Data from the Michigan State Police which supervised the DNA testing conducted by the vendor laboratory.
Ethnographic observations of meetings.
Test results from SAKs tested in a separately funded project known as "The 400 Project".
Records from organizations involved with Detroit sexual assault cases: Detroit police records, police advocacy program records, police forensic lab records, prosecution records, state forensic lab records, sexual assault forensic exam program records, and records from community-based advocacy organizations.
Detroit's police database that provided the number of SAK's in police custody.
administrative records data,
This study is comprised of seven SPSS datasets that contain information drawn from the testing of the sexual assault kits (SAKs).
- Dataset one (DSAK_2014-07-06.sav) is the primary SAK-level data file in the Detroit Sexual Assault Kit Action Research Project (DSAK-ARP). It includes 27 variables and 1,595 cases from all four testing groups. Each variable is a description of a particular property of the SAK including victim and assailant demographics (age, gender and race); preexisting victim and perpetrator relationship; whether a weapon or physical force were used; drugs or alcohol involvement; length of time between the offense and the exam; whether the SAK was a Combined DNA Index System (CODIS) entry or CODIS hit; whether the offender was identified in this case or by a hit in another case as well; and what method of DNA testing was used.
- Dataset two (DSAK.g4_2014-07-06.sav) includes 32 variables and 350 cases. This dataset contains information on Testing Group 4 of the DSAK-ARP. The variables include similar information as dataset one, including demographic information about the victims and assailants, factual information about the assailant and the crime, and whether there was a CODIS hit or multiple SAKs hits. Additionally, it contains information on review time, whether DNA testing was continued post-screening, the number of DNA samples in the SAK, and the number of samples that included sperm.
- Dataset three (Lab_Cost_2014-06-03.sav) includes three variables and four cases. Data were provided by the vendor laboratory that tested the SAKs. The dataset contains aggregated data for a subset of Testing Group 4 SAKs from the DSAK-ARP. The three variables indicate DNA testing method, whether a DNA test was given and cost of supplies for DNA testing.
- Dataset four (Lab_Time_2014-06-03.sav) has five variables and 25 cases. Data was provided by the vendor laboratory that tested the SAKs. The cases represent aggregated data for a subset of Testing Group 4 SAKs from the DSAK-ARP. The subsets are defined by the combination of DNA testing method and Lab Analyst ID. The variables describe the vendor ID, method of testing, and time spent testing and reviewing the results.
- Dataset five (MSP_Time_2014-06-03.sav) has four variables and 350 cases. The cases represent data for a unique SAK from Testing Group 4 in the DSAK-ARP. The variables describe the ID, DNA testing method, Michigan State Police ID, and time spent reviewing the DNA results. These data came from the Michigan State Police, which supervised the DNA testing conducted by the vendor laboratory.
- Dataset six (Rape_Kit_Era_Raw_Data.sav) has 22 variables and 10,817 cases. The cases represent all SAKs in police custody from the start of the project to November 1st, 2009. The variables represent information about the SAKs including the year the kit was collected, whether it was submitted to a lab, CODIS status and what sort of additional funding or programs, that could affect SAK testing, were in effect at the time the SAK was collected.
- Dataset seven (400_Project_2014-07-06.sav) contains three variables and 250 cases. These cases are outside of the original stratified sample of 1,595 kits that make up the cases in the other six datasets. These cases are a sub-sample of the 400 Detroit SAKs analyzed as part of The 400 Project. These cases were included in the census conducted as part of The DSAK-ARP. The three variables are an ID, an indicator of whether the kit was expired, and an indicator of whether the offender was a stranger or a person known to the victim. This data set was used to create sampling weights for the DSAK-ARP analyses.
The datasets are accompanied by 17 .R and .Rout syntax files that contain code, scripts and syntax for replicating the results of the original analysis.