Moving Forward While Looking Back: Understanding the Influence of a “Test All” SAK Mandate on Sexual Assault Case Attrition

Abstract

The discovery of thousands of untested sexual assault kits (SAKs) in police jurisdictions across the United States has prompted federal, state, and local agencies to enact policies regarding SAK testing. One legislative response has been to implement “test all” SAK policies. Such policies are expected to remove discretion from the kit submission process and provide crucial forensic evidence needed to guide sexual assault investigations. This study presents the impact of a “test all” policy in a single jurisdiction and investigates whether the increased number of DNA matches expected to come from more testing after implementation increased the likelihood of arrest. Results indicate that the impact of a DNA match on arrest increased after the mandate. The study concludes with a discussion of implications for sexual assault investigations and recommendations for future research.

Keywords

rape police decision-making police discretion

The discovery of thousands of untested sexual assault kits (SAKs) in police jurisdictions across the United States has prompted federal, state, and local agencies to make considerable efforts to address the issue in recent years. Government sponsored projects in Los Angeles, Houston, Detroit, New Orleans, and Cuyahoga County, Ohio shed light on how SAK testing practices and forensic results from previously untested kits resulted in missed opportunities to identify suspects in sexual assault incidents reported to the police (R. Campbell et al., 2017; Lovell et al., 2018; Pinchevsky, 2018; Waltke et al., 2018). Resources for testing previously untested SAK’s and resolving backlogs have been made available from special programs such as the Bureau of Justice Assistance’s Sexual Assault Kit Initiative (SAKI). Legislators have also responded by introducing state bills to count, test, and track rape kits and to give rights to survivors (End the Backlog, 2020). Included among these legislative remedies are “test all” mandates that require all future SAK’s collected be submitted to crime laboratories for testing (Davis et al., 2020; Strom & Hickman, 2016).

Efforts to document and address untested SAK inventories are ongoing, and recent estimates reveal that there are between 200,000 and 400,000 kits in storage across the United States (R. Campbell et al., 2017; Pinchevsky, 2018; Wang & Wein, 2018). The volume of untested kits varies significantly across jurisdictions with testing rates ranging from as low as 18% (Strom & Hickman, 2010) to as high as 89% (McEwen, 2011). Investigations into factors that influence SAK testing practices in jurisdictions have centered on the roles that organizational conditions (Hendrix et al., 2020; Luminais et al., 2017) and case-level characteristics play in the police decision to submit SAKs for testing (B. A. Campbell, Menaker, & King, 2015; Patterson & Campbell, 2012; R. Campbell, Shaw, & Fehler-Cabral, 2015; Valentine et al., 2019). This body of work has generated important insights into the discretion used by police in testing decisions and how the investigative and scientific complexities that drive those decisions may run counter to the interests of victims.

As a legislative response, the intent of “test all” SAK policies is to limit law enforcement discretion in making the decision to test. Such policies are expected to remove bias from the kit submission process, may assure victims that their complaint is being taken seriously, and may provide crucial forensic evidence needed to guide sexual assault investigations. Testing can aid investigations by identifying suspects unknown to the victim, confirming the identity of suspects known to the victim, identifying serial offenders, and exonerating those who are wrongfully accused (R. Campbell, Shaw, & Fehler-Cabral, 2015; Davis et al., 2021; Lovell et al., 2017, 2018, 2020).

Evaluations of these policies examined changes in sexual assault arrest measures before and after “test all” policies were implemented and did not find a significant effect (Davis et al., 2020; Mourtgos et al., 2021). While informative, these evaluations were limited in that they were not able to directly consider how forensic results from testing SAKs following the mandate impacted arrest outcomes at the incident-level. This study presents the results of an evaluation of one jurisdiction’s “test all” mandate on incident-level outcomes of sexual assault reports made to the police. The study begins with a review on SAK testing policies and practices followed by a description of the “test all” mandate to be evaluated. Next the evaluation plan and analytic approach are described followed by a discussion of the findings. The study concludes with recommendations for future research.

A Review of SAK Testing Practices

After decades of research documenting sexual assault case attrition at the arrest and prosecutorial stages, the discovery of thousands of untested SAKs provide yet another example of how victims of sexual assault are often treated in the criminal justice system (Corrigan, 2013). Following the report of a sexual assault, victims are typically advised to visit a hospital to undergo a sexual assault forensic medical exam (SAFME) and to complete an SAK that contains physical evidence of the assault gathered by health care professionals (Valentine et al., 2019). It is reasonable for victims to expect that following the collection of the SAK, all forensic evidence will promptly be tested and used in the investigation of the assault. Once completed, it is typically the responsibility of law enforcement to submit the SAK to a forensic laboratory for DNA testing (Department of Justice, 2013). Research and media reports indicate that many jurisdictions do not routinely submit SAKs for testing (Ayesh, 2020; R. Campbell et al., 2016; Hopkins, 2020; Strom & Hickman, 2010). Critics have argued that failure to test SAKs means that justice is delayed at best, and, in some cases, justice is denied altogether for victims of sexual assault (Pinchevsky, 2018; Strom & Hickman, 2010).

Over the past decade, scholars have investigated why SAKs are not routinely submitted for forensic testing. Some studies have described how organizational context influences the capacity to test and consider factors such as departmental resources and policies that drive SAK test submissions (Hendrix et al., 2020). Many crime laboratories have limited capacity to test due to underfunded and understaffed facilities and struggle to process large numbers of requests in a timely fashion (R. Campbell et al., 2017; Lovrich et al., 2004; Peterson et al., 2012; Strom & Hickman, 2010). Insufficient staffing resources in law enforcement agencies and lack of authorization to accessing the FBI Combined DNA Index System (CODIS) system have also been identified as barriers to SAK submissions (R. Campbell, Shaw, & Fehler-Cabral, 2015; Davis & Wells, 2019; Hendrix et al., 2020; Lovell et al., 2018; Menaker et al., 2017; Randol & Sanders, 2015).

The working relationship between law enforcement and other stakeholders involved in responding to sexual assault has also been empirically linked to testing rates. For example, higher testing rates have been reported in jurisdictions connected to strong sexual assault nurse examiners (SANEs) and sexual assault response team (SART) programs (Patterson & Campbell, 2012; Shaw & Campbell, 2013). Other studies describe how submission practices can be shaped through discretion and negotiation between police and laboratory personnel and that power imbalances between them may shift in favor of laboratories emerging as the determiners of acceptable testing standards (R. Campbell & Fehler-Cabral, 2022; Strom & Hickman, 2016). Given the importance of victim cooperation to an investigation and prosecution (Kaiser et al., 2017), the police may base a decision to test on whether the victim is willing to cooperate in pressing charges and may wait to test SAK’s until they are confident the victim wants to proceed (B. A. Campbell, Menaker, & King, 2015; Ritter, 2011). Under these circumstances, decisions about testing become further removed from more immediate investigatory considerations.

Probative Value and the Decision to Test

Formal and informal policies that limit testing capacity may discourage agencies from routinely submitting SAKs for testing and motivate practices in which SAKs are screened for submission (Lovrich et al., 2004; Peterson et al., 2012; Strom & Hickman, 2010). Discussions about if and how cases should be tested often center on whether investigators believe that results will yield DNA with “probative value” to the investigation and prosecution (Peterson et al., 2012; Ritter, 2011). SAK testing results are considered probative to the extent that they may bolster probable cause for an arrest. In the sexual assault literature, SAKs have been described as having probative potential when there are indicators of offense seriousness and/or when DNA test results will identify an unknown suspect, confirm the identity of a known suspect, enhance victim credibility, or corroborate the victim’s story (Alderden et al., 2021; Davis et al., 2021; Valentine et al., 2019). In contrast, probative potential is reduced when police personnel perceive that a test will unlikely yield forensic material such as time-to-report, the victim bathed before the kit was collected, the suspect claims consent or was known to the victim, there was already corroborating evidence to support an arrest, or the suspect had already been adjudicated (Johnson et al., 2012; Peterson et al., 2012; Ritter, 2011; Valentine et al., 2019).

In jurisdictions with large numbers of unsubmitted SAKs, a primary reason for not sending SAKs for forensic testing is that investigators do not perceive the results as contributing probative value for case processing (Peterson et al., 2012; Wells et al., 2016). In a survey of the Houston Police Department, Fallik and Wells (2014) found that detectives perceived laboratory results as adding probative value to their investigation in only 3% of cases. Facing the uncertainty that a test result will yield a suspect identification match, along with lengthy waiting times for results in many jurisdictions, law enforcement officers may perceive SAK testing as adding little to investigations (R. Campbell et al., 2016).

In discounting the potential probative value of SAK testing as a general practice, police may be submitting SAKs from cases in which probable cause for an arrest already exists (Fallik & Wells, 2014) and may explain why Schroeder and Elink-Schuurman-Laura (2017) report that arrests are often made before DNA testing is complete. Relatedly, there is some evidence to suggest that SAKs do not get submitted for analysis due to action or sometimes inaction by the prosecutor’s office (Wallenborn, 2022). Prosecutor’s offices may have limited resources such as sufficient personnel to track and manage the large numbers of SAKs that move through crime laboratories (Patterson & Campbell, 2012). Adequate staffing alone, however, is not a guarantee for more extensive testing if prosecutors are not committed to the charging and prosecution of sexual assault cases. By not supporting SAK testing or bringing forth charges in sexual assault cases, prosecutors are both indirectly and directly influencing the submission of SAKs.

Facing the logistical challenges associated with clearing up large untested SAK inventories, along with plans to increase testing moving forward, some have suggested that SAK testing may be more efficient and beneficial to investigations if SAKs were prioritized for submission based on certain case characteristics. For example, there has been much discussion about prioritizing testing based on the relationship between victim and offender (B. A. Campbell et al., 2021; Ritter, 2011; Strom & Hickman, 2016; Wells, 2016). The argument in favor of prioritization on these grounds is that the probative value of forensic results is higher in cases in which the victim and offender are strangers because in those incidents of sexual assault, DNA testing and a possible CODIS match may offer the best chance to identify a suspect and solve the case (Johnson et al., 2012; Peterson et al., 2012). Others, however, have challenged the idea that prioritizing is a viable approach to testing based on evidence that DNA CODIS suspect matches are equally likely to come from cases involving strangers as those in which the victim and offender knew each other (R. Campbell et al., 2019; Lovell et al., 2018).

Legal and Extra-Legal Factors

Studies have also examined the influence of case characteristics on the police decision to test SAKs (Patterson & Campbell, 2012; Shaw & Campbell, 2013; Valentine et al., 2019). Guiding the approach is the vast body of literature on sexual assault case attrition that examines the impact of case characteristics on arrest and prosecution outcomes. Literature has shown that arrest outcomes can be influenced by legal and extra-legal case factors (B. A. Campbell, Menaker, & King, 2015; Lapsey et al., 2022; O’Neal et al., 2015; Pattavina et al., 2021; Spohn & Tellis, 2012). Legal factors indicate evidence of a crime as defined by statute and strength of evidence (Spohn & Tellis, 2012). Extra-legal factors are legally irrelevant factors that describe the victim, suspect, or other circumstances associated with an incident (O’Neal et al., 2015); they capture concerns with victim character, credibility, and behavior that are consistent with rape myth adherence and often serve to contest victims’ claims to lack of consent (see review by Lapsey et al., 2022). The influence of extra-legal factors on case attrition occurs when key justice system actors—police and prosecutors—use them to rationalize ending complaints early in the processing stage (B. A. Campbell, Menaker, & King, 2015; Pattavina et al., 2021).

Valentine et al. (2019) has advocated for a similar approach to predicting the decision to test an SAK. Legal and extra-legal case factors may signal the perceived probative potential of a DNA match from an SAK test in shoring up probable cause for an arrest. Consistent with this approach are studies that have investigated the role case-level factors play in the test submission decision. In a multi jurisdiction examination of SAK test submissions Patterson and Campbell (2012) reported that legal factors influence the decision to submit. They found that SAKs collected from victims with reported physical injuries were more likely to be submitted for testing, while the SAKs from those who bathe or shower after the assault were less likely to be submitted. Valentine et al. (2019) reported SAKs with suspected drug facilitated assaults were more likely to be tested. Shaw and Campbell (2013) found that the number of assaultive acts against the victim increased the likelihood that the SAK would be tested. Extra-legal factors have also been found to influence the decision to test. SAKs from cases in which the victim is white and older (Shaw & Campbell, 2013), is known by the suspect, used drugs prior to the assault, or reported a mental or physical impairment were less likely to be tested (Valentine et al., 2019).

This body of research demonstrates that the SAK testing decision should be considered part of the sexual assault case attrition story. More importantly, because extra-legal factors appear to influence submission decisions, there exists the possibility that the probative value of a test (and potential DNA match) might add to an investigation and prosecution may be underestimated in certain cases. Evidence of this concern is reported in the regular updates provided by the SAKI Initiative. To date, efforts to test previously untested SAKs have resulted in 23,426 follow-up investigations, 2,423 cases charged, and 1,554 convictions (SAKI, 2022).

Test All Mandates

The purpose of “test all” mandates is to limit discretion regarding which SAKs should be tested and, by extension, the use of a probative value standard in making the testing decision. Indeed, the current literature on SAK testing practices lends strong support for the adoption of “test all” practices. This is further echoed by The National Institute of Justice’s endorsement of testing all rape kits as a best practice. Thirty states and Washington, DC have mandated the timely submission and testing of newly collected SAKs (End the Backlog, 2020). Wang and Wein (2018) note that there are two main ways that jurisdictions can achieve the goal of “test all” approaches. First, there is the “forklift approach” where jurisdictions test all SAKs in the backlog without prioritization. Second, jurisdictions can test all SAKs in the backlog with a method of prioritization—deciding that some cases should be processed before others.

Comprehensive testing has other potential outcomes that are worthy of note. For example, “test all” policies can result in a more extensive population of the Combined DNA Index System (CODIS). DNA analysis requires the comparison of two samples (R. Campbell et al., 2017) so more comparison samples can lead to more CODIS matches and greater likelihood of obtaining forensic matches. These CODIS matches link to offenders identified in the database as well as forensic matches in which evidence matches another crime scene—thus connecting to both known and unknown offenders. Comprehensive testing and submission of DNA profiles is also important for the identification of serial offenders. Evidence suggests that most offenders commit more than one sexual offense (Lisak & Miller, 2002) and analysis of SAKs has underscored these findings (R. Campbell et al., 2019; Lovell et al., 2017). Accordingly, even if the statute of limitations has passed, there remains value in testing previously untested kits (R. Campbell et al., 2019).

Mandatory testing policies can also serve to build trust with victims and their families (Mourtgos et al., 2021) by signaling that these crimes are a priority and solving them is important to the community. Finally, “test all” policies can exonerate individuals who have been wrongly accused of sexual assault (R. Campbell et al., 2017). While there are few false rape reports to police (Lonsway & Archambault, 2012), it is nonetheless crucial to the administration of justice to identify those who are unjustly accused. At the same time, however, “test all” policies also put a strain on existing resources (Davis & Wells, 2019) by adding to the workload of crime laboratories to conduct the forensic analysis but also on the detectives and prosecutors who track the cases throughout the already lengthy testing process (Wallenborn, 2022).

Recent evaluations examined the effect of “test all” mandates on arrests for sexual assault complaints reported to law enforcement. Davis et al. (2020) examined arrest rates for the state of Texas before and after the implementation of a “test all” mandate. Mourtgos et al. (2021) examined the effect of a “test all” mandate on arrest rates in a single midwestern jurisdiction. The focus on arrest rates as the outcome in these studies is important. If arrests for sexual assault increase in the aftermath of a “test all” policy, then the policy would be considered a meaningful improvement to the system’s response to sexual assault because more offenders will be held accountable and more justice will be brought to victims (Mourtgos et al., 2021). Both the Davis et al. (2020) and Mourtgos et al. (2021) studies reported a downward, but insignificant trend in arrest rates after the mandate implementation. The use of aggregate arrest outcome measures is limited, however, in that they may mask important incident level mechanisms by which the mandate may translate into arrest outcomes. This study addresses this limitation by examining the impact of a “test all” mandate on arrest outcomes at the incident level in one large midwestern city.

Data and Methods

The police jurisdiction where the study takes place employed more than 1,000 sworn officers and had a special unit where detectives were assigned to investigate sexual assault complaints. The jurisdiction had a population of around 650,000. Approximately 80% of the population identified as White and approximately 10% as African American. Thirty percent of the population identify as Hispanic or Latino. There was an active SANE program serving victims from the jurisdiction.

In the early 2010s, the jurisdiction passed a mandate requiring that all previously untested SAK’s be submitted to the accredited crime laboratory(s) for testing as soon as practically possible and that newly collected SAKs be submitted within 3 weeks of collection for testing and comparison to DNA databases. The law also required the establishment of procedures for victims to consent to submit a collected SAK kit for testing and law enforcement investigation. While the mandate provided some funds to the forensic laboratory for DNA testing, no funding was provided to support efforts by the police department to submit and track new and previously untested SAKs or to follow-up on any investigative leads that may come from previously untested kits.

The evaluation of the mandate began after receiving permission to extend a study on sexual assault case attrition that was taking place at the site at the time the mandate was implemented. That study involved collecting case-level data from all police reports of sexual assaults for the first three of the 5 years immediately prior to the mandate. The data collection period was extended to include sexual assault reports where SAKs were collected for 2 years after the mandate was implemented. Data collection was completed by the end of 2016.

Case-level details included in this study were gathered by trained researchers from 881 police reports of incidents involving sexual assault, sexual assault with an object, and forcible sodomy involving females 13 and over and where a rape kit was collected from the victim.¹ Details were gathered on victim, suspect, and witness characteristics, as well as incident location and circumstances, investigative decisions, and case outcomes.

Also collected for each incident was whether an SAK was submitted for testing and if there was a DNA match to an identified suspect or person unknown to the victim. A DNA match occurred when a DNA proﬁle from the SAK matched to a suspect’s DNA sample in CODIS or to a forensic sample that may have been collected from a suspect during the investigation. SAKs with no CODIS match describes cases in which there was no DNA match to a suspect that could be named or that the test process did not yield viable DNA material to upload to CODIS. Although there was more capacity to test SAKs after the mandate and police expressed concern about the additional workload, no significant changes were observed regarding case processing practices of reported sexual assaults post-mandate implementation. Five hundred seventy-four (574) incidents occurred over a 3-year period prior to the mandate and 307 occurred in the 2 years following the mandate. Fifty-six percent of the premandate kits were sent for testing, and 94% were sent postmandate.²

Prior to the legislation, the decision to test a kit within this jurisdiction was made on an individual case basis, and according to key personnel from the site, this decision was largely guided by the confidence of the investigator in the potential probative value of the results. In removing police discretion that “test all” practices prescribe, it is reasonable to assume that more tests will lead to more CODIS matches.³ But a full understanding of the impact of testing mandates must also consider the influence of test results on case outcomes. The analysis that follows examines the ways in which the shift to the “test all” SAK policy influences the relationship between a DNA match and arrest outcomes. Beginning with a focus on the premandate cases, logistic regression was used to examine the predictive influence of legal and extra-legal case characteristics on the decision by the police to submit SAKs for testing prior to the mandate. The results build on prior work by providing additional insight into the factors that are considered to be probative in making the decision to test and allow for a more in-depth assessment of the impact of the mandate on arrest outcomes. The following research questions guide the evaluation plan:

What case-level factors explain the police decision to submit SAKs under a probative value standard used before the mandate?

Does the relationship between a suspect DNA match and an arrest change from pre- to post-“test all” mandate when other case-level factors are taken into account?

Does the impact of a suspect DNA match on arrest change from pre- to post-mandate for cases where police have varying levels of confidence in the probative value of an SAK test leading to an arrest?

To address the first question, we determined the influence of case-level characteristics on whether an SAK was submitted for testing before the mandate. Guided by prior studies on sexual assault attrition (see Alderden & Ullman, 2012; Morabito et al., 2019; Spohn & Tellis, 2012) and recent studies that examined predictors of kit testing described earlier (Shaw & Campbell, 2013; Valentine et al., 2019), we considered comparable legal factors such as measures of offense seriousness and the presence of other evidence, as well as extra-legal factors that capture victim behavior and credibility as described in the police reports.

Legal factors included in the analysis were whether the suspect used a weapon (0 = no, 1= yes) and whether the victim sustained any physical injury according to victim and/or police records (0 = no, 1 = yes). Also included are the number of witnesses and if there was other nonbiological evidence (i.e., text messages, photos, video footage) (0 = no, 1 = yes) indicated by the official records. Extra-legal factors were operationalized as: if the victim was engaged in so-called “risky behaviors” (0 = no, 1 = yes),⁴ if there was a delay in reporting of more than one day (0 = no, 1 = yes), if there were reported concerns by the police about the victim’s credibility as indicated by the case notes (i.e., inconsistencies in stories) (0 = no, 1 = yes), if there was more than one suspect (0 = no, 1 = yes) and if the victim was cooperative during the investigation (0 = no, 1 = yes). Other predictors included the relationship between the victim and suspect (intimate partner, acquaintance, or stranger), victim age ranging from 13 to 87, and victim race (0 = BIPOC,⁵ 1 = white).

For question two, logistic regression was used to estimate change in the likelihood that a suspect DNA match ended in an arrest from pre- to post-mandate controlling for the same legal and extralegal case-level factors. Finally, for the third question, we used a marginal effect comparison to examine whether the probability that cases in which SAKs are considered to have relatively little probative value were more likely to end in arrest following the mandate.

Results

Table 1 provides the descriptive statistics for the full sample, the premandate sample, and the postmandate sample. Significant differences existed between premandate and postmandate periods in some of the variable distributions. There were significantly fewer incidents with nonbiological evidence postmandate (46% vs. 24%), fewer between acquaintances (57% vs. 50%), fewer with more than one suspect (20% vs. 11%), more with white victims (55% vs. 65%), smaller number of witnesses (3 vs. 2) and slightly older victims on average (27 vs. 30). Fifty-six percent of SAKs from the premandate period were tested, and 94% from the postmandate period. Arrest rates were similar premandate and postmandate at 26% and 24%, respectively.

Table 1.

Descriptive Statistics for Premandate and Postmandate Periods.

Indicator	% YesFull sample (N = 881)	% YesPre-mandate (n = 574)	% YesPost-mandate (n = 307)	Pre–post Difference X²
Presence of nonbiological evidence	39	46	24	43.01**
Suspect used a weapon	8	8	9	0.34
Police or hospital reported injury	67	69	64	2.14
Delayed reporting one day	14	13	17	3.19
More than one suspect	16	20	11	12.61**
Number of witnesses (mean, t-value)	3	3	2	8.42**
Victim cooperated	65	64	68	1.56
Victim engaged in “risky” behavior	69	70	68	0.41
Police concerns of victim credibility	22	24	19	2.68
Victim race (white)	58	55	65	8.5**
Intimate partner	12	11	14	1.04
Acquaintance	54	57	50	3.90*
Stranger	34	32	37	0.17
Victim age (mean, t-value)	27	27	30	–3.72**
Kit submitted	70	56	94	99.18**
Arrest	25	26	24	0.5
Submissions	—	(n=255)	(n=207)	—
DNA hit	—	28	37	4.53*

p < .05, **p < .01.

Bivariate associations between each variable and whether a kit was submitted for testing on the premandate cases are provided in Table 2, Column A. Variables with significant bivariate associations were considered for a multivariate logistic regression model and results are presented in Table 3. Several case-level factors emerged as significant predictors of an SAK test submission. SAKs were more likely to be submitted for testing if the suspect used a weapon (OR = 3.21, p < .01), if there were more witnesses (OR = 1.17, p < .01), and if the victim was noted as cooperative (OR = 3.16, p < .01) or older (OR = 1.02, p < .05). SAKs were significantly less likely to be submitted if there were credibility issues with the victim (OR = .55, p < .01) and the suspect was an intimate partner (OR = .29, p < .01) or an acquaintance (OR = .46, p < .01) as compared to a stranger.⁶

Table 2.

Bivariate Statistics for SAK Test Submissions and Arrest Outcomes.

	Column A.			Column B.
	% SAKs Submitted			% SAKs Submitted with
	Premandate			Arrest Premandate and Postmandate
	(n = 574)			(N = 462)
Indicator	Yes	No	X ²	Yes	No	X ²
Presence of non-biological evidence	60	53	2.98	46	23	26.90**
Suspect used a weapon	77	54	8.81**	34	33	0.03
Police or hospital reported injury	56	56	0.01	36	26	5.16*
Delayed reporting one day	47	57	2.53	21	35	4.27*
More than one suspect	55	59	0.48	19	37	8.99**
Number of witnesses (mean, t-value)	4	3	–4.58**	4	3	–5.50**
Victim cooperated	65	40	34.63**	41	12	34.00**
Victim engaged in “risky” behavior	59	50	3.88*	32	34	0.12
Police concerns of victim credibility	46	59	6.66**	17	37	11.90**
Victim race (white)	55	57	0.23	29	39	5.28*
Intimate partner	46	57	2.83	59	30	15.00**
Acquaintance	52	61	5.10*	37	29	3.92*
Stranger	67	51	12.57**	21	41	18.50**
Victim age (mean, t-value)	28	26	–1.86*	27	29	1.58
DNA hit	—	—	—	60	21	68.90*
Mandate period	—	—	—	29	36	3.2

p < .05, **p < .01.

Table 3.

Logistic Regression Results Predicting SAK Test Submission Premandate (N = 574).

Indicator	OR	B	SE (B)	95% CI (B)
Suspect used a weapon	3.21**	1.17	0.41	[0.363, 1.972]
Number of witnesses	1.17**	0.15	0.03	[0.087, 0.221]
Victim cooperated	3.16**	1.15	0.20	[0.761, 1.538]
Police concerns of victim credibility	0.55*	–0.60	0.22	[–1.039, –0.160]
Victim engaged in “risky” behavior	1.32	0.28	0.20	[–0.120, 0.674]
Intimate partner	0.29**	–1.23	0.32	[–1.861, –0.590]
Acquaintance	0.46**	–0.77	0.21	[–1.189, –0.357]
Victim age	1.02*	0.02	0.01	[0.003, 0.036]
Constant	0.35**	–1.05	0.37	[–1.773, –0.328]
Wald X² = 77.13**
Pseudo R² = .13

p < .05, **p < .01.

These results are consistent with prior studies reporting that legal and extra-legal factors predict whether an SAK will be submitted for testing. The significant impact of legal factors suggests that the decision to submit an SAK is more likely when there is some indicator of offense severity such as use of a weapon and the number of witnesses. Even with legal factors considered, extra-legal factors including concerns with victim credibility, when the suspect is known to the victim (i.e., intimate partner, acquaintance), and uncooperative victims reduce the likelihood of an SAK submission. One scenario offered in discussions with law enforcement at the site and consistent with these findings is that kit submissions may be less likely to provide probative value for incidents where the victim and suspect know each other, such that the case becomes a so-called, “he said, she said” scenario. A DNA match to the suspect can confirm that sexual acts had occurred but would probably fail to add additional value to the investigation. This may explain the finding that SAKs from cases involving an intimate partner or acquaintances are less likely to be submitted than those involving strangers.⁷

The next task was to determine if an increase in DNA matches on suspects in the postmandate period translated to a greater likelihood of arrest. A DNA suspect match before the mandate occurred in about 28% of 246 cases for which testing results were documented and 37% of 207 cases in the postmandate period. This difference was statistically significant (X² = 4.53, p < .05). However, even if the “test all” mandate resulted in more DNA matches, it does not necessarily signify that the forensic evidence will increase the likelihood of arrest for those cases. Empirical investigation into this question provides a de facto test of the probative value standard. If the likelihood of arrest on cases with a DNA match does increase postmandate, then the mandate would be effective, and conversely, the probative value standard did not result in selecting for testing SAKs from cases for which DNA matches would bolster probable cause for an arrest. No difference suggests that forensic evidence from more DNA matches under the “test all” mandate is no more or less effective in leading to arrests than using a probative value standard.

To investigate this outcome, a logistic regression model was estimated to predict arrest outcomes on the 462 cases for which SAK test results were reported.⁸ We considered the same legal and extra-legal factors as those for the kit submission analysis, and descriptive measures are presented in Table 2, Column B. Variables with significant bivariate associations with arrest were considered for the multivariate model. Also included in the model was a binary variable indicating whether the test resulted in a DNA match (0 = no, 1 = yes), a binary variable for pre–post mandate (0 = premandate, 1 = postmandate) as noted in police records, and an interaction term between the mandate period and whether there was a DNA match (0 = no, 1 = yes) as independent variables. A positive interaction term coefficient would mean that the likelihood of arrest for a match increased moving from the premandate to postmandate period. Conversely, a negative coefficient would mean that the likelihood of arrest involving a match decreased postmandate.

One additional independent variable for this model was created to consider the police perception of the probative value of a DNA match in leading to an arrest. The score was created using the logit coefficients (B) for the intercept, and each variable in the model used to predict SAK submission (see Table 3). Holding values of the independent variables (X) as observed for each incident, an SAK testing probability for each is generated using the formula 1/(1 + e^–z), where Z is the linear combination Z = B₀ + B₁X₁ + B₂X₂ + . . . B₈X₈.

A higher probability of testing score reflects greater confidence ascribed by police in the potential of test results to support an arrest, thus it serves as an indicator of the predicted probative value of testing (PPT). The PPT score ranges from .08 to .97 and the average PPT score is similar between the premandate (.55) and postmandate cases (.53). Including this measure as a control variable in the model accounts for police perception of the value of a possible DNA match on the likelihood of arrest. To reduce collinearity with other independent variables, it was recoded into a binary measure where cases at or below the 50th percentile are coded as 0 to indicate a relatively weak PPT and those above the 50th percentile are coded as 1 indicating a moderate to strong PPT.

The results of the logistic regression model are reported in Table 4. With the mandate period by DNA match interaction included in the model, the interpretation of each individual variable that comprises the interaction represents the effect of that variable on arrest when the other variable is held constant at zero. For the pre-mandate period (when mandate = 0), the effect for a DNA match on arrest is positive and significant (OR = 3.89, p < .01). The interaction term coefficient indicates an increase in the likelihood of arrest when there is a DNA match moving to the postmandate period; it is positive and significant, meaning that the mandate did increase the likelihood of arrest in cases with a DNA match (OR = 2.87, p < .05). Other variables that significantly predict an arrest include the number of witnesses (OR = 1.11, p < .05), the presence of nonbiological evidence (OR = 2.96, p < .01), victim cooperation (OR = 4.16, p < .01), and if the suspect is an intimate partner (OR = 4.87, p < .01) or acquaintance (OR = 2.17, p < .01). Arrest is less likely if there are concerns about victim credibility (OR = .39, p < .05), there is more than one suspect (OR = .28, p < .01), and cases involving white victims (OR = .55, p < .05). The PPT score was not significant suggesting that the police perceptions of the value of a test have little influence on case outcomes. In sum, these results show that the “test all” mandate resulted in an increase in the likelihood of arrest in cases in which there was a DNA match on a suspect when other case factors are held constant.

Table 4.

Logistic Regression Results Predicting Arrest for Cases With Submitted SAKs (N = 462).

Indicator	OR	B	SE (B)	95% CI (B)
Mandate period	0.49	–0.71	0.38	[–1.45, 0.03]
DNA match	3.89**	1.36	0.36	[0.66, 2.06]
Mandate period × DNA match	2.87*	1.05	0.53	[0.01, 2.10]
PPT (1 = above 50th percentile)	1.34	0.29	0.38	[–0.46, 1.04]
Police or hospital reported Injury	1.06	0.06	0.29	[–0.52, 0.63]
Number of witnesses	1.11*	0.10	0.05	[0.01, 0.19]
Presence nonbiological evidence	2.96**	1.08	0.29	[0.52, 1.65]
Police concerns of victim credibility	0.39*	–0.94	0.39	[–1.70, 0.17]
Victim cooperated	4.16**	1.42	0.39	[0.67, 2.18]
Delayed reporting one day	0.47	–0.76	0.42	[–1.57, 0.06]
Intimate partner	4.87**	1.58	0.56	[0.49, 2.67]
Acquaintance	2.17**	0.78	0.28	[0.22, 1.33]
Victim was white	0.55*	–0.59	0.26	[–1.11, –0.07]
More than one suspect	0.28**	–1.27	0.44	[–2.14, –0.41]
Constant	0.04**	–3.12	0.50	[–4.09, –2.14]
Wald X² = 115**
Pseudo R² = .3190

Note. PPT = probative value of testing.

p < .05, **p < .01.

Using the PPT score, we can further explore whether SAKs from cases that have a low score also have a low likelihood of ending in an arrest if there is a DNA match. Such a finding could serve as evidence to validate the use of a probative or discretionary testing policy. To consider this possibility, the predicted probabilities of arrest and the marginal effects of a DNA match on arrest were compared across the two PPT levels (see Table 5). If cases have a weak PPT (at or below the 50th percentile), the average probability of arrest in cases with a DNA match is slightly larger in the postmandate period for cases (.38 vs .44) but is not statistically significant. The marginal effect comparison of a DNA match on arrest for the weak PPT group is higher in the postmandate period than the premandate but is not statistically significant. If cases have a moderate to strong PPT, the average probability of arrest in cases with a DNA match is slightly larger in the postmandate period for cases (.56 vs .62) but again not statistically significant. The marginal effect comparison of a DNA match on arrest from the premandate to postmandate period (.24 vs .41) is larger and statistically significant. These results show that the move toward the “test all” policy is more likely to benefit cases that have a stronger (PPT) score.

Table 5.

Predicted Probabilities and Marginal Effects of DNA Match on Arrest.

	Premandate probability of arrest	Postmandate probability of arrest	Equivalency test (X²)
Weak PPT (≤50th percentile)
No DNA match	0.18	0.11	3.67*
DNA match	0.38	0.44	0.68
Marginal effect of a match	0.19	0.32	2.48
Moderate to strong PPT (>50th percentile)
No DNA match	0.32	0.21	3.97*
DNA match	0.56	0.62	0.71
Marginal effect of a match	0.24	0.41	3.79*

Note. Values of covariates in the model as observed. PPT = probative value of testing.

p ≤ .05, **p < .01.

Interestingly, the average probability of arrest for cases in which there was no DNA match was significantly smaller in the postmandate period for both PPT levels. This finding might suggest that DNA tests are exonerating suspects when there is no DNA match to a suspect, which could explain the decrease in arrests with no DNA match premandate to postmandate, or that police are beginning to rely more on suspect DNA matches in the decision to arrest postmandate, or the consideration of both. Based on this study’s findings, moving from a probative value standard in deciding which cases to test to a “test all” policy yielded an increase in both the percentage of cases with a DNA match as well as the likelihood that cases with a match will significantly contribute to an arrest. However, cases with a weaker PPT under a probative standard, in comparison to those with a moderate to strong PPT score, did not benefit as much from the “test all” policy in terms of the impact of a DNA match on arrest. Overall, the findings do suggest that the “test all” strategy is a more successful testing practice for sexual assault investigations than the probative standard because this strategy results in DNA matches that uniquely contribute to the arrest decision.

Discussion

In this police jurisdiction where slightly more than half of the SAKs were being tested, the move toward a “test all” policy appears to have increased the number of DNA matches and the likelihood that a DNA match would result in an arrest. This study provides support for “test all” policies that generate forensic results from SAKs to aid in the identification of suspects and increase the likelihood of offenders being held accountable. As research on “test all” policies advances, it is important for each study to be considered in the relevant time, organizational, and cultural contexts, especially given that many departments and forensic labs continue to grapple with the scientific, resource, and logistical challenges that accompany ramping up testing capacity (R. Campbell & Fehler-Cabral, 2018). Furthermore, the composition of this population may limit the generalizability to other communities. Nonetheless, this study contributes to the literature on SAK testing policies and is consistent with studies from other jurisdictions in reporting that legal and extra-legal factors drive submission decisions under a probative value standard. Relatedly, this study also highlights the complexities of evaluating the implementation of new testing policies and the challenges they present for researchers. One limitation of this study is that some key measures such as test submission dates and other specific details on the nature of a “no DNA match” (i.e., untestable sample vs. no match) test outcome were not available for this analysis.

While additional resources were provided to the jurisdiction to test SAKs after the mandate, none were allocated to support the added detective work. More matches would be expected if processing capacity and efficiency increased postmandate. A reduction in processing time and a quicker turnaround may translate to an increased likelihood of arrest when there is a match, which may keep the investigation active and the victim cooperative—key factors in moving cases through the system (Ahrens et al., 2020; Lovell et al., 2021; Valentine et al., 2019). The lack of additional resources for investigative work may have resulted in incomplete record keeping on the testing results, considering that about 20% of cases for which SAKs were tested did not have results documented in the police files. This degree of missingness on testing outcomes is a limitation of this study. As explained earlier, it could be that those cases were resolved before results were available, but complete and current record keeping is essential to understand the results of SAK testing and how they influence arrest.

The science surrounding DNA continues to improve and funding continues to be made available from federal agencies to support greater testing capacity. As CODIS becomes increasingly populated with DNA profiles, the chances of offender or forensic DNA matches will increase, which may also increase the likelihood of arrest. Exactly when a growing inventory of profiles will reach its “tipping point” to the extent that more identified suspects will consistently lead to more arrests is difficult to say, but police departments and researchers will hopefully continue to engage in this area of research and document the process.

The results of this study also generated more questions about how DNA matches are used. Both Mourtgos et al. (2021) and Davis et al. (2020) report that there was a decrease (yet not statistically significant) in arrests after the implementation of “test all” mandates. While they were not able to consider the impact of testing directly, the results of this study reveal that the probability of arrest for cases without a DNA match significantly decreased after the mandate. The reasons for this finding remain unclear. One explanation is that more testing is exonerating wrongfully accused suspects. A more likely reason may be that “test all” mandates are facilitating a reliance on the presence of forensic evidence in making the case for arrest as suggested by Quinlan (2020). Both possibilities may have concerning implications for victims. Yu and colleagues (2022) describe how forensic results that are inconsistent with the victim’s account of what happened may be used to undermine victim credibility, especially when SAKs are processed by health and law enforcement professionals without specialized training. Placing the findings of this study in context of the evaluative work of Davis et al. (2020) and Mourtgos et al. (2021) highlights the need to examine the unintended ways in which system actors adapt “test all” policies moving forward. In addition, more research is needed to determine if our findings are jurisdiction-specific or have applicability beyond this one site. This would allow us to explore the effects of local context on testing outcomes.

Conclusion

Mandatory testing of SAKs has become standard policy in some jurisdictions across the country. Prior impact evaluations of these policies using aggregate arrest measures reveal no statistical effect of the policy. This study investigated the impact of one “test all” mandate on case-level outcomes in a single U.S. jurisdiction. The results show that the mandate increased the number of DNA hits and the likelihood of arrest after the mandate. Although these findings suggest that the policy is having an intended effect by holding more offenders accountable, some noteworthy nuances were also observed. It was found that not all cases are likely to benefit from an increase in the probability of arrest after the mandate. Cases likely to benefit most from a DNA hit on arrest postmandate were the same cases that would have also had a stronger chance of being submitted if a probative value standard had been applied. While these results generally provide promising support for “test all” mandates, more research is needed to understand the processes associated with mandate implementation, and how resources are allocated to handle testing, tracking, and follow-up.

Footnotes

Data Availability Statement

Data used for this study will be made available upon request to the authors.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this study was provided by National Institute of Justice award 2012-IJ-CX-0052.

ORCID iDs

April Pattavina

Sabrina S. Rapisarda

Notes

Author Biographies

April Pattavina, Ph.D., is a Professor in the School of Criminology and Justice Studies at the University of Massachusetts Lowell. Her research focuses on gender-based violence, correctional reform, and program evaluation.

Melissa S. Morabito, Ph.D., is a Professor in the School of Criminology and Justice Studies at the University of Massachusetts Lowell. Her research focuses on the police response to vulnerable populations.

Sabrina S. Rapisarda, M.A., M.Ed., is a Doctoral Candidate and Instructor of Record in the School of Criminology & Justice Studies at the University of Massachusetts Lowell. Her research is situated where public health intersects with the criminal-legal system; she studies in various substantive areas including intimate partner and sexual violence, substance use and related harms, treatment courts, and corrections.

Linda M. Williams, Ph.D., is a Senior Research Scientist and Director of the Justice and Gender Based Violence Research initiative at Wellesley Centers for Women, Wellesley College, and professor emerita at the University of Massachusetts Lowell. The focus of her current work is on the justice system response to sexual violence, commercial sexual exploitation, intimate partner violence and child maltreatment. Her work has appeared in journals such as Violence Against Women; Social Policy and Society; Child Maltreatment; Journal of Consulting and Clinical Psychology; Journal of Aggression, Conflict and Peace Research; Journal of Contemporary Criminal Justice; Victims and Offenders; Journal of Crime and Justice.

References

Ahrens

Dahlgren

Howard

(2020). Rape kit notification: Recommendations and barriers to reconnecting with survivors. Journal of Trauma & Dissociation, 21(4), 419–436. https://doi.org/10.1080/15299732.2020.1770911

Alderden

M. A.

Cross

T. P.

Vlajnic

Siller

(2021). Prosecutors’ perspectives on biological evidence and injury evidence in sexual assault cases. Journal of Interpersonal Violence, 36(7–8), 3880–3902. https://doi.org/10.1177/0886260518778259

Alderden

M. A.

Ullman

S. E.

(2012). Creating a more complete and current picture: Examining police and prosecutor decision-making when processing sexual assault cases. Violence Against Women, 18(5), 525–551. https://doi.org/10.1177/1077801212453867

Ayesh

(2020, May 13). Untested rape kits leave thousands of serial rapists at large. Axios. https://www.axios.com/systemic-neglect-rape-kits-thousands-untested-dc4443be-54c4-484f-9020-4ad07969d40d.html

Campbell

B. A.

Menaker

T. A.

King

W. R.

(2015). The determination of victim credibility and juvenile sexual assault investigations. Journal of Criminal Justice, 43(1), 29–39. https://doi.org/10.1016/j.jcrimjus.2014.12.001

Campbell

B. A.

Wells

King

W. R.

(2021). What happens when sexual assault kits go untested? A focal concerns analysis of suspect identification and police pre-arrest decisions. Journal of Criminal Justice, 76, 101850. https://doi.org/10.1016/j.jcrimjus.2021.101850

Campbell

Feeney

Fehler-Cabral

Shaw

Horsford

(2017). The national problem of untested sexual assault kits (SAKs): Scope, causes, and future directions for research, policy, and practice. Trauma, Violence & Abuse, 18(4), 363–376. https://doi.org/10.1177/1524838015622436

Campbell

Fehler-Cabral

(2018). Why police “couldn’t or wouldn’t” submit sexual assault kits for forensic DNA testing: A focal concerns theory analysis of untested rape kits. Law & Society Review, 52(1), 73–105. https://doi.org/10.1111/lasr.12310

Campbell

Fehler-Cabral

(2022). “Just bring us the real ones”: The role of forensic crime laboratories in guarding the gateway to justice for sexual assault victims. Journal of Interpersonal Violence, 37(7–8), NP3675–NP3702. https://doi.org/10.1177/0886260520951303

10.

Campbell

Pierce

S. J.

Sharma

D. B.

Feeney

Fehler-Cabral

(2016). Should rape kit testing be prioritized by victim–offender relationship? Empirical comparison of forensic testing outcomes for stranger and nonstranger sexual assaults. Criminology & Public Policy, 15(2), 555–583. https://doi.org/10.1111/1745-9133.12205

11.

Campbell

Pierce

S. J.

Sharma

D. B.

Feeney

Fehler-Cabral

(2019). Developing empirically informed policies for sexual assault kit DNA testing: Is it too late to test kits beyond the statute of limitations? Criminal Justice Policy Review, 30(1), 3–27. https://doi.org/10.1177/0887403416638507

12.

Campbell

Shaw

Fehler-Cabral

(2015). Shelving justice: The discovery of thousands of untested rape kits in Detroit. City & Community, 14(2), 151–166. https://doi.org/10.1111/cico.12108

13.

Corrigan

(2013). The new trial by ordeal: Rape kits, police practices, and the unintended effects of policy innovation. Law & Social Inquiry, 38(4), 920–949. https://doi.org/10.1111/lsi.12002

14.

Davis

R. C.

Auchter

Wells

Camp

Howley

(2020). The effects of legislation mandating DNA testing in sexual assault cases: Results in Texas. Violence Against Women, 25(5), 417–437. https://doi.org/10.1177/1077801219838330

15.

Davis

R. C.

Jurek

Wells

Shadwick

(2021). Investigative outcomes of CODIS matches in previously untested sexual assault kits. Criminal Justice Policy Review, 32(8), 841–864. https://doi.org/10.1177/0887403421990723

16.

Davis

R. C.

Wells

(2019). DNA testing in sexual assault cases: When do the benefits outweigh the costs? Forensic Science International, 299, 44–48. https://doi.org/10.1016/j.forsciint.2019.03.031

17.

Department of Justice. (2013). A national protocol for sexual assault medical forensic examinations: Adults & adolescents (2nd ed.). Office on Violence Against Women.

18.

End the Backlog. (2020). We have a clear and ambitious goal: End the rape kit backlog. https://www.endthebacklog.org/

19.

Fallik

Wells

(2014). Testing previously unsubmitted sexual assault kits: What are the investigative results? Criminal Justice Policy Review, 26(6), 598–619. https://doi.org/10.1177/0887403414528001

20.

Hendrix

J. A.

Strom

K. J.

Parish

W. J.

Melton

P. A.

Young

A. R.

(2020). An examination of sexual assault kit submission efficiencies among a nationally representative sample of law enforcement agencies. Criminal Justice Policy Review, 31(7), 1095–1115. https://doi.org/10.1177/0887403419884730

21.

Hopkins

(2020, December 30). After 3 years and $1.5 million testing rape kits, Alaska made one new arrest. Pro Publica. https://www.propublica.org/article/after-3-years-and-1-5-million-testing-rape-kits-alaska-made-one-new-arrest

22.

Johnson

Peterson

Sommers

Baskin

(2012). Use of forensic science in investigating crimes of sexual violence: Contrasting its theoretical potential with empirical realities. Violence Against Women, 18(2), 193–222. https://doi.org/10.1177/1077801212440157

23.

Kaiser

K. A.

O’Neal

E. N.

Spohn

(2017). “Victim refuses to cooperate”: A focal concerns analysis of victim cooperation in sexual assault cases. Victims & Offenders, 12(2), 297–322. http://doi.org/10.1080/15564886.2015.1078864

24.

Lapsey

D. S.

Campbell

B. A.

Pumlee

B. T.

(2022). Focal concerns and police decision making in sexual assault cases: A Systematic review and meta-analysis. Trauma, Violence, and Abuse, 23, 1220–1234.

25.

Lisak

Miller

P. M.

(2002). Repeat rape and multiple offending among undetected rapists. Violence & Victims, 17(1), 73–84. https://doi.org/10.1891/vivi.17.1.73.33638

26.

Lonsway

K. A.

Archambault

(2012). The “justice gap” for sexual assault cases: Future directions for research and reform. Violence Against Women, 18(2), 145–168. https://doi.org/10.1177/1077801212440017

27.

Lovell

Huang

Overman

Flannery

Klingenstein

(2020). Offending histories and typologies of suspected sexual offenders identified via untested sexual assault kits. Criminal Justice & Behavior, 47(4), 470–486. https://doi.org/10.1177/0093854819896385

28.

Lovell

Luminais

Flannery

D. J.

Bell

Kyker

(2018). Describing the process and quantifying the outcomes of the Cuyahoga County sexual assault kit initiative. Journal of Criminal Justice, 57, 106–115. https://doi.org/10.1016/j.jcrimjus.2018.05.012

29.

Lovell

Luminais

Flannery

D. J.

Overman

Huang

Walker

Clark

D. R.

(2017). Offending patterns for serial sex offenders identified via the DNA testing of previously unsubmitted sexual assault kits. Journal of Criminal Justice, 52, 68–78. https://doi.org/10.1016/j.jcrimjus.2017.08.002

30.

Lovell

Overman

Huang

Flannery

D. J.

(2021). The bureaucratic burden of identifying your rapist and remaining “cooperative”: What the Sexual Assault Kit Initiative tells us about sexual assault case attrition and outcomes. American Journal of Criminal Justice, 46, 528–553. https://doi.org/10.1007/s12103-020-09573-x

31.

Lovrich

N. P.

Pratt

T. C.

Gaffney

M. J.

Johnson

C. L.

Asplen

C. H.

Hurst

L. H.

Schellberg

T. M.

(2004). National forensic DNA study report (Final report) [Report No. 203970]. U.S. Department of Justice.

32.

Luminais

Lovell

Flannery

(2017). Perceptions of why the sexual assault kit backlog exists in Cuyahoga County, Ohio and recommendations for improving practice. Begun Center for Violence Prevention Research and Education, Case Western Reserve University.

33.

Menaker

T. A.

Campbell

B. A.

Wells

(2017). The use of forensic evidence in sexual assault investigations: Perceptions of sex crimes investigators. Violence Against Women, 23(4), 399–425. https://doi.org/10.1177/1077801216641519

34.

McEwen

(2011). The role and impact of forensic evidence in the criminal justice system, final report. U.S. Department of Justice.

35.

Morabito

M. S.

Williams

L. M.

Pattavina

(2019). Decision making in sexual assault cases: Replication research on sexual violence case attrition in the US [Report No. 252680]. U.S. Department of Justice. https://www.ojp.gov/library/publications/decision-making-sexual-assault-cases-replication-research-sexual-violence-case

36.

Mourtgos

S. M.

Adams

I. T.

Nix

Richards

(2021). Mandatory sexual assault kit testing policies and arrest trends: A natural experiment. Justice Evaluation Journal, 4(1), 145–162. https://doi.org/10.1080/24751979.2021.1881410

37.

O’Neal

E. N.

Tellis

Spohn

(2015). Prosecuting intimate partner sexual assault: Legal and extralegal factors that influence charging decisions. Violence Against Women, 21(10), 1237–1258. https://doi.org/10.1177/1077801215591630

38.

Pattavina

Morabito

M. S.

Williams

L. M.

(2021). Pathways to sexual assault case attrition: Culture, context, and case clearance. Victims & Offenders, 16(8), 1061–1076. https://doi.org/10.1080/15564886.2021.1970661

39.

Patterson

Campbell

(2012). The problem of untested sexual assault kits: Why are some kits never submitted to a crime laboratory? Journal of Interpersonal Violence, 27(11), 2259–2275. https://doi.org/10.1177/0886260511432155

40.

Peterson

Johnson

Herz

Graziano

Oehler

(2012). Sexual assault kit backlog study [Report No. 238500]. U.S. Department of Justice.

41.

Pinchevsky

G. M.

(2018). Criminal justice considerations for unsubmitted and untested sexual assault kits: A review of the literature and suggestions for moving forward. Criminal Justice Policy Review, 29(9), 925–945. https://doi.org/10.1177/0887403416662899

42.

Quinlan

(2020). Visions of public safety, justice, and healing: The making of the rape kit backlog in the United States. Social & Legal Studies, 29(2), 225–245. https://doi.org/10.1177/0964663919829848

43.

Randol

B. M.

Sanders

C. M.

(2015). Examining the barriers to sexual assault evidence processing in Washington State: What’s the hold up. Criminology, Criminal Justice, Law, & Society, 16(2), 1–13.

44.

Ritter

(2011). The road ahead: Unanalyzed evidence in sexual assault cases. The National Institute of Justice. https://www.ojp.gov/pdffiles1/nij/233279.pdf

45.

Schroeder

D. A.

Elink-Schuurman-Laura

(2017). The impact of forensic evidence on arrest and prosecution [Report No. 250721]. https://www.ojp.gov/pdffiles1/nij/grants/250721.pdf

46.

Sexual Assault Kit Initiative. (2022). Performance metrics & research. https://www.sakitta.org/metrics/

47.

Shaw

Campbell

(2013). Predicting sexual assault kit submission among adolescent rape cases treated in forensic nurse examiner programs. Journal of Interpersonal Violence, 28(18), 3400–3417. https://doi.org/10.1177/0886260513504496

48.

Spohn

Tellis

(2012). Policing and prosecuting sexual assault in Los Angeles city and county: Collaborative study in partnership with the Los Angeles Police Department, the Los Angeles County Sheriff’s Department, and the Los Angeles County District Attorney’s Office [Report No. 237582]. National Institute of Justice.

49.

Strom

K. J.

Hickman

M. J.

(2010). Unanalyzed evidence in law-enforcement agencies: A national examination of forensic processing in police departments. Criminology & Public Policy, 9(2), 381–404. https://doi.org/10.1111/j.1745-9133.2010.00635.x

50.

Strom

K. J.

Hickman

M. J.

(2016). Untested sexual assault kits. Criminology & Public Policy, 15(2), 593–601. https://doi.org/10.1111/1745-9133.12213

51.

Valentine

J. L.

Sekula

Cook

L. J.

Campbell

Colbert

Weedn

V. W.

(2019). Justice denied: Low submission rates of sexual assault kits and the predicting variables. Journal of Interpersonal Violence, 34(17), 3547–3573. https://doi.org/10.1177/0886260516681881

52.

Wallenborn

J. K.

(2022). Why sexual assault kits were not tested: A systematic review. Journal of Criminology, Criminal Justice, Law & Society, 23(3), 22–44.

53.

Waltke

LaPorte

Weiss

Schwarting

Nguyen

Scott

(2018). Sexual assault cases: Exploring the importance of non-DNA forensic evidence [Report No. 250704]. National Institute of Justice. https://www.ojp.gov/pdffiles1/nij/250704.pdf

54.

Wang

Wein

L. M.

(2018). Analyzing approaches to the backlog of untested sexual assault kits in the USA. Journal of Forensic Sciences, 63(4), 1110–1121. https://doi.org/10.1111/1556-4029.13739

55.

Wells

(2016). Some considerations when making decisions about prioritizing sexual assault kits for forensic testing. Criminology & Public Policy, 15(2), 585–592. https://doi.org/10.1111/1745-9133.12206

56.

Wells

Campbell

Franklin

(2016). Unsubmitted sexual assault kits in Houston, TX: Case characteristics, forensic testing results, and the investigation of CODIS hits. U.S. Department of Justice. https://www.ojp.gov/pdffiles1/nij/grants/249812.pdf

57.

Walsh

Zweig

J. M.

(2022). The link between the SAMFE and police perceptions of victim credibility. Feminist Criminology, 17(1), 26–49. https://doi.org/10.1177/15570851211012468