An Examination of Sexual Assault Kit Submission Efficiencies Among a Nationally Representative Sample of Law Enforcement Agencies

Abstract

Drawing on results from a survey of 321 law enforcement agencies, we assess how labor and capital inputs, evidence policies, and other agency characteristics affect the number of sexual assault kits (SAKs) submitted to crime laboratories for testing. We examine to what extent agencies are submitting the maximum number of SAKs possible, given their available resources. Stochastic frontier models are used to analyze the productivity of labor inputs for the submission of SAKs and the extent to which resource inefficiencies contribute to unsubmitted SAKs. Results indicate that agencies are submitting fewer than 60% of SAKs that are possible given their resources. Full-time sworn officers were found to be productive inputs, as the number of SAKs submitted increases by 24% for every 100% increase in the number of officers. Findings also suggest that the accumulation of SAKs is driven partially by technical inefficiencies but more so by a lack of resources.

Keywords

sexual assault kits rape kits sexual assault victimization

Over the past decade, an increasing amount of attention has focused on unsubmitted sexual assault kits (SAKs) that accrue in law enforcement agencies. Many sexual assaults are never reported to the police (Langton, Berzofsky, Krebs, & Smiley-McDonald, 2012). However, when cases are reported, the criminal justice system is expected to aggressively investigate and prosecute reported sexual assaults to deliver justice for victims and also to demonstrate these crimes are treated with highest priority. Nevertheless, in some cases, justice may be delayed, in part because crime laboratories lack the capacity to process forensic evidence in a timely manner. In other cases, justice can be denied because forensic evidence is never submitted to a laboratory for analysis by the agency (Strom & Hickman, 2010).

Instances of justice denied are especially concerning, given the established benefits of testing forensic evidence (R. Campbell, Patterson, Bybee, & Dworkin, 2009; Johnson, Peterson, Sommers, & Baskin, 2012; Nelson, 2010; Peterson, Johnson, Herz, Graziano, & Oehler, 2012; Randol & Sanders, 2015; Ritter, 2012b). Testing SAKs can help identify previously unidentified offenders in unsolved crimes, corroborate identities in known offender assaults within and across cases, identify serial sexual offenders, and exonerate individuals who have been wrongly accused (R. Campbell, Feeney, Fehler-Cabral, Shaw, & Horsford, 2017). Studies have found that 50% to 60% of SAKs contain biological evidence that does not belong to the victim (Ritter, 2011), and the presence of forensic evidence is associated with an increased likelihood that arrests are made and charges are filed (Horney & Spohn, 1996; Peterson, Sommers, Baskin, & Johnson, 2010; but see Spohn & Tellis, 2012).

Law enforcement plays a critical role in either promoting or preventing justice for victims of sexual assault. The law enforcement role includes not only the initial response and interaction with the victim and investigative follow-up opportunities but also the crime scene processing and SAK evidence collection and submission decisions. However, little is known about whether and how the practices, policies, or characteristics of agencies are associated with the handling of SAKs and their movement through the system. Moreover, it is unclear to what extent agencies’ efforts to submit SAKs for testing are hampered by technical inefficiencies (i.e., inefficient uses of available resources), insufficient resources (e.g., too few staff members), or both.

The study of SAK processing efficiency in the criminal justice system is important for several reasons. First, this type of analysis can help identify and contextualize the factors that contribute to the backlog of unsubmitted SAKs that remain in law enforcement custody. Second, and perhaps more importantly, it can identify and describe the conditions that help agencies move kits effectively from collection to submission to the crime laboratory. Ultimately, these results can help put state and local law enforcement leaders and their partners in a position to successfully process SAKs through the system in a timely manner and fully leverage the valuable evidence for investigating and prosecuting these cases. It can also encourage law enforcement to implement the resources, practices, and policies necessary to support and engage victims.

Drawing on results from a nationally representative survey of 321 law enforcement agencies, we assess how labor and capital inputs, evidence policies, and other agency characteristics affect the number of SAKs that agencies submit for testing. We also examine to what extent agencies are submitting the optimal number of SAKs possible to laboratories for testing, given their level of input resources, and draw conclusions regarding the role of resources and technical inefficiencies in SAK processing. We first describe the path by which sexual assault evidence progresses through the criminal justice system, the problem of unsubmitted SAKs, and what factors inhibit SAK submissions before turning to a discussion of our methods and analytical approach.

Literature Review

The Progression of Sexual Assault Evidence Through the Criminal Justice System

Victims of sexual assault may choose to undergo a forensic examination, in which evidence of the crime is collected and potentially used to facilitate the identification or prosecution of perpetrators through DNA analysis. SAKs (i.e., “rape kits”) are often collected as a component of the forensic examination by a medical professional, ideally by a trained sexual assault nurse examiner (SANE). During the medical exam, SANEs document injuries and collect a narrative of the assault from the survivor to guide the collection of forensic evidence (Corum & Carroll, 2014; Valentine et al., 2019). Forensic examinations are intrusive, as the process lasts at a minimum of several hours and involves invasive procedures such as photographic documentation of injuries; hair samples obtained by plucking/pulling; body sample swabs, including oral, vaginal, or anal swabbing; and fingernail scrapings (R. Campbell, Shaw, & Fehler-Cabral, 2015; Scott, 2001; U.S. Department of Justice Office on Violence Against Women, 2013). Victims have described the experience as traumatic, violating, and shaming (R. Campbell, 2005, 2008; R. Campbell & Raja, 2005; Filipas & Ullman, 2001; Starzynski, Ullman, Filipas, & Townsend, 2005; Ullman, 1996). Despite this, many agree to an exam because they assume that evidence from their SAK will be used to aid the capture of their attacker (Martin, 2005; Parnis & Du Mont, 2006; Patterson & Campbell, 2010; Tofte, 2013).

Once a forensic examination has been conducted, an agency takes custody of the SAK and is responsible for submitting it to a forensic laboratory for biological screening and DNA analysis (Butler & Butler, 2010; U.S. Department of Justice Office on Violence Against Women, 2013). Probative DNA profiles that meet the requirements for entry into the Combined DNA Index System (i.e., CODIS, a national repository for DNA samples) may be uploaded to the CODIS database. Agencies can use the investigative leads provided from CODIS to identify or eliminate suspects and to further the investigation. If an agency can identify and build a case against a suspect, the case may be referred to prosecution.

The Problem of Unsubmitted SAKs

Concerningly, recent research and media reports have suggested that, despite the benefits of testing SAKs, there are SAKs that are not submitted to a laboratory for forensic analysis (Lovrich et al., 2004; Reilly, 2015; Ritter, 2011; Sacco & James, 2015; Strom & Hickman, 2010; Telsavaara & Arrigo, 2006). For example, in 2002, it was discovered that nearly 13,000 SAKs collected in Los Angeles County had never been submitted to a crime laboratory (Rubin, 2009; Rubin & Winton, 2009; Tofte, 2010). Likewise, in 2009, the Wayne County (Michigan) Prosecutor’s Office discovered more than 11,000 unsubmitted SAKs in a Detroit Police Department warehouse (Bashford, 2013; R. Campbell, Shaw, & Fehler-Cabral, 2015; Tofte, 2009).

Furthermore, the problem of unsubmitted SAKs is not localized to particular agencies, but rather is common in agencies across the country (Ritter, 2011, 2012a; Strickler, 2009; Strom & Hickman, 2010). Lovrich and colleagues (2004) surveyed approximately 3,400 agencies and found that files from nearly 170,000 unsolved rape cases for 1982-2002 contained untested biological evidence (Pratt, Gaffney, Lovrich, & Johnson, 2006). In a nationally representative survey of state and local agencies, files for 18% of unsolved sexual assault cases for 2003-2007 (an estimated 27,595 cases) were estimated to contain forensic evidence that had not been submitted to a crime laboratory for analysis (Strom & Hickman, 2010). Valentine, Shaw, Lark, and Campbell (2016) study of SAK submission rates by agencies to the Utah state crime laboratory at four sites found that only 38% of SAKs were submitted for testing, whereas Cantwell’s (2002) examination of 17,115 cases of unknown-perpetrator sexual assault in Washington State showed that only 24% of SAKs were submitted. Other studies have shown that fewer than 60% of SAKs were submitted for testing (Johnson et al., 2012; McEwen, 2011; Patterson & Campbell, 2012; Peterson et al., 2012; Shaw & Campbell, 2013).

Although it is uncertain whether the testing of evidence from unsubmitted SAKs would have facilitated identification or prosecution of the perpetrator, decisions to leave some SAKs unprocessed can mean not only that victims of sexual assault do not receive justice but also that perpetrators are free to reoffend (Perkel, 2007; Randol & Sanders, 2015; Sacco & James, 2015). Indeed, studies have found that most undetected rapists are serial rapists responsible for multiple sexual assaults (Lisak & Miller, 2002; see also Zinzow & Thompson, 2015). Thus, the failure to test SAKs can pose a significant threat to public safety. In addition, agencies’ failure to submit SAKs for testing has been described as an unequivocal breach of trust for victims of sexual assault, whose submission to an extensive and often traumatic forensic examination demonstrates their faith in the criminal justice system (Tofte, 2013).

Factors Inhibiting SAK Submissions and Remaining Questions

The attention to unsubmitted SAKs in the past few years has inspired studies that examine why SAKs are not always submitted for forensic testing. One commonly discussed explanation relates to the inability of laboratories to produce timely results when SAKs are submitted for testing (Lovrich et al., 2004; Randol & Sanders, 2015; Strom & Hickman, 2010). Advancements in forensic science collection technology and techniques in the past several years have created a significant demand for forensic testing and, in some cases, have left laboratories unable to process the ever-growing number of requests they receive because of resource or staff shortages (Cantillon, Kopiec, & Clawson, 2009; Durose & U.S. Bureau of Justice Statistics, 2008; Horvath & Meesig, 1996; Lovrich et al., 2004; National Research Council, 2009; Peterson, Hickman, & Bureau of Justice Statistics, 2005; Pratt et al., 2006; Steadman, 2000, 2002). Delays in evidence processing at the laboratory level may discourage agencies from submitting SAKs or encourage practices in which selected SAKs are screened out of submission queues (Lovrich et al., 2004; Peterson et al., 2012; Strom & Hickman, 2010; Tasca, Rodriguez, Spohn, & Koss, 2013).

Unsurprisingly, resource constraints and insufficient staffing in law enforcement agencies have also been cited as important factors inhibiting individual SAK submissions in several studies (R. Campbell et al., 2017; R. Campbell, Shaw, & Fehler-Cabral, 2015; Lovrich et al., 2004; Randol & Sanders, 2015). Luminais, Lovell, and Flannery’s (2017) process evaluation of the Cuyahoga County Sexual Assault Kit Task Force suggests that increased demands for forensic testing in the past few decades, coupled with reductions in law enforcement resources and severe understaffing have created a dysfunctional “conveyer belt” approach to managing sexual assault cases. Because of the continuous flow of cases into the system, agencies must identify ways of keeping the conveyer belt moving and thereby may be incentivized to dismiss or push aside those cases that present any “hiccups.” An effective illustration of this phenomenon in Cuyahoga is offered when the authors explain that “victims who did not respond immediately to phone calls or missed appointments were considered ‘uncooperative’, and often had their cases closed peremptorily by the City of Cleveland’s municipal prosecutor” (see p. 8). The prioritization of specific types of SAKs may be useful or even necessary in agencies with limited resources, but empirical evidence to support such prioritization or that identifies effective modes of prioritization is lacking.

Although Luminais and colleagues’ analysis indicates that a lack of resources (i.e., financial, training, and technological deficiencies, staffing constraints, etc.) is an important reason for why agencies are hindered in their ability to engage in key investigative activities (e.g., they have too few investigators to make contact with every victim or witness) or submit SAKs for testing, it is not the only factor at play. Rather, members of the Task Force perceived a culture of victim-blaming and a lack of a victim-centeredness mentality to also be critical for understanding why some SAKs are not properly investigated, tested, and ultimately prosecuted. For example, members believed that victims’ stories were often discredited because of their behavior at the time of the assault, including being under the influence of drugs or alcohol, having a mental illness, or even walking alone in an unsafe neighborhood (p. 11). Of course, a culture that actively challenges and discredits victims may be deeply interrelated with the problem of insufficient resources. Specifically, a culture of victim-blaming and the discreditation of sexual assault cases may be functional for an overwhelmed system in need of ways to fast-track cases or remove them altogether from the conveyer belt. Although it cannot be assumed that the culture of victim-blaming is ubiquitous across law enforcement agencies, it could perhaps help to explain why the likelihood of some SAKs being submitted for testing is significantly lower in some jurisdictions than others (see Valentine et al., 2019).

Consistent with the idea that a high volume of requests combined with a paucity of resources simultaneously makes it challenging for agencies to fully investigate every sexual case and encourages the screening out of cases to reduce workload, ample research has indicated that not all SAKs are treated equally in the system. For instance, SAKs from victims who admit to drug use before the assault, report a mental or physical impairment, or bathe after the assault are less likely to be submitted for forensic analysis (Burgess, Hanrahan, & Baker, 2005; Patterson & Campbell, 2012; Valentine et al., 2019), as are those in which law enforcement does perceive the victim to be credible (see B. Campbell, Menaker, & King, 2015; R. Campbell et al., 2009; R. Campbell, Shaw, & Fehler-Cabral, 2015; O’Neal, Tellis, & Spohn, 2015; Shaw & Campbell, 2013; Tasca et al., 2013; Valentine et al., 2019). In addition, SAKs deriving from assaults that were drug facilitated or that involved a high number of assaultive actions are more likely to be submitted (Shaw & Campbell, 2013; Valentine et al., 2019). Studies have also found that agencies are less likely to submit SAKs for testing when no charges have been filed against the suspect or when there is no identified suspect, when prosecutors do not request testing, when the suspect is adjudicated, and when investigators do not believe or understand that forensic evidence is useful to the investigation (Lovrich et al., 2004; Luminais et al., 2017; Strom & Hickman, 2010; Tasca et al., 2013). The rationale behind decisions not to test SAKs from cases that do not have an identified suspect is bewildering, given that analysis could offer critical information about the suspect’s identity (Strom & Hickman, 2010). However, as noted by Luminais and colleagues (2017), it is not uncommon for key players in the criminal justice system to lack understanding about the utility of testing all SAKS, or the value of entering all DNA samples into the CODIS as a tool for solving current and future crimes.

Taken together, the aforementioned studies bring to light the pervasive systemic issues that delay or deny justice for survivors of sexual assault and in doing so effectively convey the complexity of both the SAK problem and its solution. Despite the value of this research, these studies have largely focused on why specific types of SAKs are not submitted but have said little about submission behaviors among agencies in aggregate. For instance, it is not well understood whether certain characteristics or practices (e.g., number of staff, budget, and policy structure) of agencies influence higher or lower annual numbers of SAK submissions, nor what effect they have on the number of SAKs submitted in a given year. Although some qualitative research has suggested that resource inputs such as staffing have significant effects on agencies’ abilities to submit SAKs (e.g., Luminais et al., 2017), studies to date have not quantified the precise effect of these various agency inputs. In addition, it is not known to what extent agencies are submitting the maximum number of SAKs possible given the resources they have available to them. Our study elaborates on the productivity of agency inputs and assesses to what extent technical inefficiencies or inadequate resources influence agencies’ overall numbers of unsubmitted SAKs.

Study Overview and Research Questions

This study derives from a larger mixed-methods study on SAK processing funded by the National Institute of Justice (NIJ). Key objectives of that study included understanding agency- and cross-agency-level factors that promote or discourage the submitting and testing of SAKs among a linked sample of law enforcement agencies and the crime laboratories to which they submit biological evidence. Data collection included surveys administered to state, county, and municipal laboratories that conduct biological forensic analysis and to a sample of linked law enforcement agencies. In addition, the project team also conducted site visits in six jurisdictions, during which interviews and focus groups were conducted with (a) a law enforcement agency that submits SAKs to crime laboratories for testing; (b) a crime laboratory that receives and tests those SAKs submitted by the aforementioned law enforcement agency; and (c) a prosecuting attorney that works with the aforementioned agencies to seek convictions for sexual assault offenses. In this study, we focus our analysis on the nationally representative survey data collected from 321 law enforcement agencies. Our primary research questions are as follows:

Research Question 1: How productive are law enforcement personnel in terms of SAK submissions?

Research Question 2: Are there inefficiencies in how law enforcement personnel submit SAKs that contribute to the accumulation of unsubmitted SAKs?

Research Question 3: To what extent are accumulations of unsubmitted SAKs driven by such inefficiencies?

Research Question 4: To what extent are accumulations of unsubmitted SAKs driven by a lack of resources (e.g., not enough staff or administrative budget)?

Inefficiencies in how law enforcement personnel submit SAKs could arise in several ways, including from the specific protocols that law enforcement personnel must follow to submit SAK evidence to crime laboratories and which may explain why some agencies are able to submit more SAKs than others even when they have similar numbers of personnel and administrative budgets.

Sampling and Weighting

Because a core interest of the larger NIJ study was to understand cross-agency dynamics in SAK processing among crime laboratories and a linked sample of their associated law enforcement agencies, our data collection strategy began with a national survey of all state and local crime laboratories that conduct biological forensic analysis. All laboratories (N = 222) in the Bureau of Justice Statistics’ 2009 Census of Publicly Funded Forensic Crime Laboratories were first contacted and 67% of them (n = 147) responded to the survey. Then, a random sample of law enforcement agencies was selected from a full list of all the agencies who submit forensic evidence to these laboratories. This full list of agencies was constructed by matching the jurisdictions reported by each surveyed crime laboratory with the Federal Bureau of Investigation’s (FBI) Uniform Crime Reports to identify all their associated agencies.

Because one law enforcement agency may submit SAKs to multiple laboratories, we began by sorting the list by crime laboratories in order from those associated with the least number of agencies to those associated with the most agencies. To ensure that adequate information could be collected per laboratory, we selected all law enforcement agencies for laboratories associated with four or fewer candidate agencies and then removed those selected agencies from the list so that they would not become candidate agencies for other laboratories. This approach ensured that we would not select the same agencies multiple times if they were associated with multiple crime laboratories. Because these agencies were selected with certainty, their selection probabilities in the survey were equal to 1 and their analysis weights were assigned as 1 (i.e., the inverse of their selection probabilities).

Next, among the laboratories that were associated with more than four candidate agencies, we selected a sample of agencies within each laboratory. Within each laboratory, the candidate agencies were sorted by their population sizes and then a systematic sampling scheme was used to randomly select a subset of m agencies among them (Lehtonen & Pahkinen, 2004). We then removed the selected agencies from the entire merged list, so they would not become candidate agencies for other laboratories. For any laboratory that has less than or equal to 40 candidate agencies, we set m as 4 and selected four agencies per lab. For any laboratory with more than 40 candidate agencies, to prevent large variation among the analysis weights, we divided the number of its candidate agencies by 10, rounded that number up to the next integer and set the integer as m. In this way, the selection probability for each selected agency would not be smaller than 0.10, and its final analysis weight, which is the inverse of the selection probability, would not be larger than 10. The analysis weights are incorporated in all the analyses among the selected agencies in this article.

Key Measures

Our outcome of interest was the number of SAKs submitted by each law enforcement agency to a crime laboratory in the previous year. We captured this with a question on the agency survey and used the log of the response as our dependent variable. To measure productivity of law enforcement personnel, we created a measure that accounts for the number of full- and part-time authorized sworn personnel for each agency (logged). We also control for each agency’s total administrative budget during the fiscal year 2013 (in millions of dollars and logged), the presence or absence of a unit in each agency that specifically investigates crimes of a sexual nature (including but not limited to sexual assault), and the presence or absence of five policies relating to sexual assault evidence: a 100% submission policy for all SAKs collected, a 100% submission policy for all sexual assault evidence collected, a policy describing criteria for submitting SAKs to the crime laboratory, a policy describing mandatory criteria to record decisions justifying why an SAK was not submitted to the crime laboratory, and a policy to prioritize the submission of evidence based on the amount of time elapsed (e.g., new compared with cold cases).

Additional inputs included the method used to obtain SAKs from health care/medical facilities; respondents were asked to select all applicable methods for obtaining kits, including having an officer collect the kit immediately after the sexual assault examination, having an officer wait for notification from the health care/medical facility to collect the SAK, or having health care/medical facilities submit SAKs directly to crime laboratories. Binary variables were created for each method such that 0 = method is not used and 1 = method is used. Respondents were also asked to indicate in an open-ended question what their agencies’ greatest barriers were in submitting SAKs to crime laboratories for forensic analysis (e.g., physical proximity to the laboratory and the inability of the laboratory to produce timely results). Responses were initially coded into thematic categories; however, because of extensive variation in responses and a pattern of statistical nonsignificance across the categories in analytical models, responses were recoded dichotomously to indicate whether the respondent described any barrier to submitting SAKs (= 1) or indicated that there were no barriers to submitting SAKs (= 0). We assumed that respondents who did not provide a response had no barriers to report. Finally, our models control for the total number of SAKs collected by each agency in the previous year. The distribution of this variable was assessed, and agencies were coded as being in the lower, middle, or upper tercile for SAKs collected in the previous year.

Analytical Strategy

We use a stochastic frontier model to measure how productive law enforcement personnel are in terms of submitting SAKs and whether there are any inefficiencies in how personnel submit SAKs that lead to an accumulation of unsubmitted SAKs (Aigner, Lovell, & Schmidt, 1976). This modeling approach is ideal because it can accomplish both of these goals simultaneously and allows post hoc analysis of the modeling results to further understand the extent to which the accumulation of unsubmitted SAKs are driven by inefficiencies versus by an insufficient number of personnel and administrative budget. The model defines technical inefficiency as the extent to which law enforcement agencies submit fewer SAKs than are possible given the personnel, agency budget, and other resources that support the process of submitting SAKs. These kinds of inefficiencies could arise if, for example, law enforcement personnel were obligated to collect a certain quota of SAKs before they were allowed to submit a batch of SAKs to laboratories for DNA testing. This could result in fewer SAKs being submitted than otherwise would be possible.

A generic stochastic frontier specification is presented below:

\begin{array}{l} Number of SAKs submitted = f (Productive Resources) - \\ Technical Inefficiencies + Error, \end{array}

(1)

where f() denotes a generic functional form. In this article, we use a translog specification, which is a flexible specification that can accommodate many nonlinearities in the relationship between productive resources (e.g., number of sworn officers) and the number of SAKs that are submitted. The function “f(Productive Resources)” represents the production frontier—or the maximum number of SAKs that could be submitted—and the “Technical Inefficiencies” parameter measures the extent to which agencies submit fewer SAKs than are possible. The model also includes a standard statistical error term (“Error”).

In this article, we also convert the measure of “Technical Inefficiencies” into an efficiency measure, which is given by:

\begin{array}{l} Efficiency = [f (Productive Resources) - Technical Inefficiencies] / \\ f (Productive Resources) . \end{array}

(2)

Thus, “Efficiency” measures the number of SAKs actually submitted by each law enforcement agency as a proportion of their production frontier. Using this measure, we make a couple additional calculations to inform (a) how many agencies would still have an SAK backlog if it were possible to eliminate all “Technical Inefficiencies,” and (b) among those agencies that would still have an SAK backlog in the absence of “Technical Inefficiencies,” what proportion of the backlog could be eliminated if all “Technical Inefficiencies” were removed. Converting the evidence into these two metrics helps inform the degree to which SAK backlogs are a result of technical inefficiency versus a general lack of resources (e.g., too few staff) at law enforcement agencies to keep up with the number of SAKs being collected in their jurisdictions. Prior to interpreting results from our final models, we examined variance inflation factors to detect any signs of multi-collinearity among the independent variables. These investigations did not show problematic levels of collinearity based on commonly accepted standards for model building (see O’Brien, 2007).

Results

Table 1 displays descriptive statistics for key measures. Although all continuous variables were logged for analysis, we present both logged and unlogged descriptive statistics to facilitate interpretation. As shown, agency respondents in the final sample had an average of 121 sworn personnel and an annual administrative budget of nearly US$18 million. About 63% of agencies reported having a dedicated sex crimes unit within the agency. Nearly one third of agencies (29%) had a 100% SAK submission policy and about one fifth (19%) had a 100% evidence submission policy in sexual assault cases. About one in five agencies (19%) had a policy that required justification decisions be recorded when SAKs were not submitted. Nearly half (49%) of the sample had a policy describing criteria for submitting SAKs to the laboratory, and fewer than 10% had a policy to prioritize the submission of evidence based on time elapsed. Nearly three fourths of agencies (72%) had a practice in place to have their officers retrieve SAKs from medical facilities immediately after sexual assault examinations, whereas 53% of agencies reported that officers waited for notification from the health care/medical facility before picking up SAKs. Of note, only 2% of agencies used a system in which the health care/medical facility submits SAKs directly to the crime laboratory for analysis. A plurality of agencies was in the lowest tercile for number of SAKs collected in the previous year, one third of agencies were in the middle tercile, and one quarter of agencies in the upper tercile. On average, agencies submitted 15 SAKs to crime laboratories in the previous year.

Table 1.

Measures and Descriptive Statistics (n = 321).

Item	M	SD	Range
Number of authorized sworn personnel
Logged	4.04	1.08	0 to 9.20
Unlogged	121.30	375.73	1 to 9,855
Total administrative budget, fiscal year, 2013 (millions US$)
Logged	1.76	1.76	−13.8 to 7.6
Unlogged	17.74	70.59	1.0e–06 to 2,000
Sex crimes unit	0.63	0.48	0 to 1
Policy types
100% SAK submission policy	0.29	0.46	0 to 1
100% evidence submission policy	0.19	0.40	0 to 1
Criteria for submitting SAKs to crime laboratory	0.45	0.50	0 to 1
Mandatory criteria to record decisions justifying why SAKs were not submitted	0.19	0.39	0 to 1
Policy to prioritize the submission of evidence based on time elapsed	0.08	0.27	0 to 1
Method to obtain SAKs from health care/medical facilities
Officer collects kit after sexual assault examination	0.72	0.45	0 to 1
Officer waits for notification from health care/medical facility	0.53	0.50	0 to 1
Health care/medical facility submits SAK directly to laboratory	0.02	0.15	0 to 1
Total number of SAKs collected, 2013
Lowest tercile (reference category)	0.42	0.45	0 to 1
Middle tercile	0.33	0.47	0 to 1
Upper tercile	0.25	0.44	0 to 1
Total number of SAKs submitted to a crime laboratory, 2013
Logged	1.67	1.24	0 to 7.09
Unlogged	15.01	57.10	1 to 1,200
Barriers to submission	0.50	0.50	0 to 1

Note. SAK = sexual assault kit.

Table 2 displays results from stochastic frontier analysis, predicting the number of SAKs submitted by key agency characteristics. Model 1 shows that the number of SAKs submitted by agencies increases by about 25% for every 100% increase in the number of full-time equivalent (FTE) sworn personnel. Unsurprisingly, a higher number of SAKs collected is associated with a larger number of SAKs submitted; agencies in the upper tercile submit around 100 (200%) more SAKs than agencies in the lower tercile, whereas those in the middle tercile submit about 50 (100%) more SAKs than those in the lower tercile. Coefficients for number of FTE sworn personnel and number of SAKs collected are largely the same when method to obtained SAKs and the presence of submission barriers are controlled for in Model 2. Interestingly, agencies that have officers wait for notification before collecting SAKs submit 24% fewer SAKs than agencies that do not use this method. Finally, agencies that reported submission barriers or challenges submit 36% fewer SAKs than those that did not report barriers to submitting SAKs. Results described in Models 1 and 2 remain substantively the same when controlling for the presence of five policy types in Model 3.

Table 2.

Stochastic Frontier Models Predicting the Number of SAKs Submitted by LEA Characteristics and Policies.

Measure	Number of SAKs submitted
Measure	Model 1	Model 2	Model 3
Number of FTE sworn personnel	0.25*** (0.06)	0.24*** (0.06)	0.24*** (0.06)
Administrative budget (millions US$)	0.03 (0.05)	0.03 (0.05)	0.03 (0.05)
LEA has a sexual crimes unit	−0.08 (0.09)	−0.10 (0.09)	−0.08 (0.09)
Number of SAKs collected: middle tercile	1.02*** (0.10)	1.07*** (0.10)	1.08*** (0.10)
Number of SAKs collected: upper tercile	2.04*** (0.13)	2.07*** (0.13)	2.05*** (0.13)
Officer collects SAK immediately after examination		−0.09(0.09)	−0.06(0.09)
Officer waits for notification from medical facility before collecting SAK		−0.24** (0.08)	−0.20* (0.08)
Medical facility directly submits SAKs		−0.13(0.18)	−0.15(0.17)
Any submission barriers		−0.36*** (0.08)	−0.34*** (0.09)
Policy: PD must submit 100% of all kits collected			0.22(0.13)
Policy: PD must submit 100% of all evidence collected			−0.04(0.14)
Policy: Submission criteria			0.18(0.10)
Policy: Must record why a kit was not submitted			−0.15(0.12)
Policy: Submissions are prioritized on the basis of time elapsed			0.07(0.17)
Observations	281	281	281
Average technical efficiency(SD)	0.55(0.19)	0.57(0.18)	0.58(0.18)
z test for negatively skewed residual(p value)	−4.92(.00)	−4.74(.00)	−4.48(.00)

Note. Coefficients are parameter estimates, with standard errors in parentheses. SAK = sexual assault kit; LEA = law enforcement agency; FTE = full-time equivalent; PD = police department.

p < .05. **p < .01. ***p < .001.

Modeling results also suggest that there is a fair amount of inefficiency in how law enforcement personal submit SAKs. For each model, average technical efficiency scores are below 0.6, which indicates that agencies are only submitting around 60% of the maximum number of SAKs that could be submitted (i.e., their productive frontier). Formal statistical significance testing further supports these findings. Supplementary models (not shown) that controlled for agency type (e.g., municipal and county) and jurisdiction size did not show statistically different effects for these variables nor did they alter existing effects in any meaningful way.

Table 3 uses the technical efficiency score for each agency to describe how many more SAKs could be submitted if all agencies operated with perfect efficiency. This helps us to determine the extent to which the accumulation of unsubmitted SAKs is driven by inefficiencies or by agencies being understaffed and lacking enough budgetary resources (or both). Specifically, we estimated the percentage of agencies that could have eliminated their SAK caseload from the previous year under perfect technical efficiency. In addition, among those agencies that could not have eliminated their SAK caseload, we estimated the average percent reduction in their SAK caseload that they could have experienced under perfect technical efficiency. Table 3 shows that, although the model suggests a fair amount of inefficiency exists, this inefficiency only explains a relatively small portion of agencies’ accumulation of unsubmitted SAKs from the previous year. Even under perfect technical efficiency, fewer than 15% of agencies would have been able to eliminate their SAK caseload, defined as processing and submitting all the SAKs that came into their agency from the previous year. Furthermore, among those agencies unable to process all their SAKs, operating at perfect technical efficiency would have eliminated only about 32% of their caseload (on average). Collectively, these findings demonstrate that the inability of agencies to process all the SAKs in their possession is largely driven by a lack of resources rather than by technical inefficiencies in processes used to collect, track, and submit the SAK evidence.

Table 3.

Resource Utilization Inefficiencies and Impact on Size of Backlog.

	Model 1	Model 2	Model 3
Number of unsubmitted SAKs	135.39	—	—
Number of unsubmitted SAKs	(333.44)	—	—
Additional SAKs under perfect technical efficiency	18.14	18.71	17.83
Additional SAKs under perfect technical efficiency	(58.34)	(62.22)	(55.59)
% of LEAs who would have had zero unsubmitted SAKs	14.37	13.87	12.72
Additional SAKs as a percentage of unsubmitted SAKs	32.58%	32.29%	32.21%

Note. Coefficients are mean values/percentages with standard deviations in parentheses. LEA = law enforcement agency; SAK = sexual assault kit.

Discussion

The problem of unsubmitted SAKs in law enforcement agencies’ custody has received unprecedented attention and there has been growing recognition that these types of system breakdowns can result in delays in the justice system and, in some instances, the denial of justice for sexual assault victims (Nelson, 2010; Ritter, 2012b; Strom & Hickman, 2010). Yet, there is limited understanding of what constitutes efficient SAK collection and submission procedures and whether agency organizational policies, staffing, and other characteristics could increase or facilitate the effective collection and submission of SAKs, thereby preventing the accrual of uncategorized and unsubmitted SAKs. In this study, we assessed the productivity of various law enforcement inputs for the submission of SAKs to crime laboratories and examined whether agencies are submitting the maximum number of SAKs given their available resources.

Our analysis revealed several key findings with important implications for agencies and the greater criminal justice and forensic fields. First, results from stochastic frontier analysis showed that full-time sworn agency staff are highly productive inputs, as the number of SAKs submitted to crime laboratories for testing is estimated to increase by about 24% for every 100% increase in the number of sworn officers. Although past studies have shown through qualitative data that law enforcement personnel perceive insufficient staffing or labor cuts to have a negative influence on their ability to submit SAKs or fully investigate sexual assaults (R. Campbell et al., 2017; Luminais et al., 2017), no other known quantitative study to date has been able to demonstrate statistical associations between staffing and agency submission behaviors, nor have studies quantified the effect that adding additional staff can have on investigations. Based on these findings, policy makers at all level of government must recognize that our nation cannot be successful (however defined) without addressing the lack of adequate staffing. Yes, policies and procedures matter, but qualified, well-trained staff are essential not only to making inroads in creating an effective criminal justice system that is responsive to victims of sexual assault but also in establishing steps to sustain and build on the progress that has been made.

Despite the evidence of inefficiency across agencies when it comes to the annual number of SAKs submitted for forensic testing, one of our most important findings is that insufficient resources are the largest contributor to SAK submission backlogs. On average, agencies are submitting fewer than 60% of the SAKs that are technically possible given the level of input resources and estimated productivity of those resources. However, further examination showed that, even operating under perfect technical efficiency, fewer than one sixth of agencies would have been able to process and submit their entire caseload of SAKs submitted in the previous year. The fact that 75% of agencies would still have accumulated unsubmitted SAKs, even if using all of their policies and resources to their maximum potential, suggests that agencies are vastly underresourced for collecting and submitting SAKs. We would argue that this lack of staffing resources also extends to challenges that many agencies face in adequately investigating sexual assault crimes and using promising forensic investigative leads to move cases forward. It is also possible that understaffing and problems with insufficient resources contribute to or perpetuate a culture of victim-blaming, a lack of victim-centeredness, and other attitudes and behaviors aimed at dismissing victims or SAKS to keep the “conveyer belt” moving (Luminais et al., 2017). More research is needed to explore this possibility.

There are also relevant findings regarding the policies in place in agencies for retrieving kits from hospitals, prioritizing kits for submission, and recording information about why kits were not submitted. Overall, about three out of 10 agencies reported having a 100% SAK submissions policy. For agencies that did not have “submit all” policies, fewer than half (45%) had written policies detailing under what circumstances SAKs should and should not be submitted for testing. Among all agencies, only about one in five agencies (19%) had a policy that required their police officers to record the justification explaining why SAKs were not submitted. These findings are noteworthy, as the lack of clear policy and standards guiding submission decisions can lead to inconsistent practices. Agencies that do not have submit-all policies should at the minimum have clear and detailed guidance for when SAKs should be submitted. Furthermore, agencies should have a mandatory policy for providing written and standardized justification for why SAKs were not submitted, which will help to provide some measure of quality assurance and protect against scrutiny about submission decisions down the road. Failure to have such a policy can exacerbate problems resulting from unsubmitted SAKs being discovered at some later time. Such policies would have benefited agencies in Los Angeles, Detroit, Houston, and other cities that had to dedicate tremendous resources and staff time to respond to public scrutiny about unsubmitted SAKs.

Finally, the study found that agencies that wait for notification from the health care/medical facility before collecting the SAK submit 24% fewer SAKs than agencies that do not use this method. This finding supports the need for agencies to implement policies that require routinely and proactively retrieving SAKs from health care/medical facilities in addition to policies that help push agencies to develop specific plans (through their multidisciplinary partnerships) for coordinating SAK evidence retrieval and chain of custody with their SANE counterparts. We cannot claim that the negative effect of agencies’ waiting on facility notification is causal, in that employing this collection method causes agencies to leave more SAKs unsubmitted than other types of agencies. However, this notification approach possibly suggests a latent characteristic of agencies, reflecting a more reactive attitude toward SAK processing. Although more research is needed to fully understand why this approach is negatively associated with the number of SAKs submitted or how this situation could be improved, research has suggested that the presence of community- or hospital-based victim advocates can help to bridge the gap between law enforcement agencies and hospitals to ensure that SAKS are picked up and submitted in a timely manner (Luminais et al., 2017).

Other limitations to the study should also be considered when interpreting the results. For instance, efficiency measures in the analysis are static and represent only a snapshot in time. Furthermore, “efficiency” in processing SAKs only covers the procedural components of SAK processing; it considers neither criminal justice outcomes pertaining to case closure or prosecution nor factors related to the victims and their recovery. Future longitudinal studies that explore the impact of resources on key outcomes over time (e.g., submissions to crime laboratories, testing of the kits, case closures, prosecutions, and victim recovery) could potentially provide more nuanced information about the value of various resources, technical inefficiencies, and the need for additional resources in different types of agencies. Likewise, it may be considered a limitation that our questionnaire only asked respondents to report the number of SAKs that had been submitted for testing, which may not accurately reflect the complex reality of SAK submissions. In practice, the submission of SAKs can sometimes transpire across multiple transactions, for instance, when DNA is not found on the items that the agency originally submitted for testing and which in turn might encourage them to submit additional items to the laboratory. Although this level of detail would have been difficult to capture in a nationally representative survey, future research should more carefully capture the back-and-forth nature of SAK submissions and identify whether this feature influences agency decisions to submit the SAK in the first place or whether these transactions are impacted by departmental policies or other factors.

Another important limitation is that the data collected on numbers of SAKs received and submitted to crime laboratories were self-reported by agency contacts rather than based on official audits or administrative records. It is possible that respondents inflated the number of SAKs that were submitted in the previous year to portray their agency in a more positive light, especially given the mounting criticism of agencies that do not submit SAKs for testing by victims’ advocates, the media, and political figures. In fact, the average percentage of SAKs submitted to crime laboratories in our sample (excluding those with a 100% submission policy) was 67%, which exceeds estimates provided in several other studies (see Cantwell, 2002; Johnson et al., 2012; McEwen, 2011; Patterson & Campbell, 2012; Peterson et al., 2012; Shaw & Campbell, 2013; Valentine et al., 2016) and thus may bolster the notion that agencies may have exaggerated their commitment to submitting SAKs for testing. Future studies should attempt to replicate our findings using official records of SAK submissions when available.

One final limitation of this study is that although this is one of the first to examine the connections between departmental policies and the submission of SAKs, our operationalization of such policies likely oversimplified the role they play in law enforcement agencies and their impact on the number of SAKs that are submitted for testing. Specifically, the policy variables were binary in nature, measuring only the presence or absence of these policies and therefore do not accurately capture the content of these policies, the amount of detail they provide, or the extent to which they are implemented with fidelity. However, because the law enforcement survey asked respondents to indicate whether a given policy was “formal” or “informal” in nature, we were able to run a series of models that tested the effects of formal, informal, or any (formal or informal) policy on SAK submissions. Ultimately, separating out formal from informal policies did not show meaningful differences from models that tested the effects of policy when coded as present if it existed either formally or informally. Therefore, for parsimony, our final models presented the effects of policies, formal or informal, on SAK submissions. We believe that our analysis provides a critical jumping-off point for future research, which should work to unpack the nuances associated with policy implementation and how various factors (e.g., completeness, level of detail, and mode of implementation) can have an impact on SAK submissions.

Conclusion

The law enforcement community and the larger criminal justice system have received a considerable amount of scrutiny for unsubmitted SAKs. Fortunately, there are efforts in place to address the problem of unsubmitted SAKs, including an influx of federal and state funding to not only inventory and test previously unsubmitted SAKs but also to build capacity through training and policy development in agencies. As an example, since 2015, the U.S. Department of Justice–funded Sexual Assault Kit Initiative (SAKI) has supported state and local agencies in strengthening their multidisciplinary response to sexual assault while also inventorying over 70,000 previously unsubmitted SAKs across 54 SAKI state and local jurisdictions (see www.sakitta.org). Although some of these SAKs are still in the testing process, to date, SAKI participating jurisdictions have reported nearly 8,000 CODIS hits from these kits. Although these are extremely positive results and indicate the potential for reform, we must also acknowledge the challenges that agencies —along with their partners in forensics, prosecution, victim advocacy, and medical care—face in the struggle to fully respond to sexual assault. Undoubtedly, these challenges are complex and will require thoughtful and holistic solutions. It is imperative that researchers continue to build the knowledge base and produce actionable research that identifies and informs effective strategies and practices for ensuring sexual assault evidence is properly collected, investigated, and submitted for forensic testing.

Footnotes

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: This study was funded by the National Institute of Justice, award number 2013-NE-BX-0006.

ORCID iD

Joshua A. Hendrix

Author Biographies

Joshua A. Hendrix is a research scientist at RTI International in the Center for Community Safety and Crime Prevention. He specializes in school safety, school emergency preparedness, victimization, and behavioral health. His recent work has been published in the Journal of School Violence, Surveillance & Society, Policing and Society, Organization & Environment, Police Quarterly, and the Journal of Experimental Criminology.

Kevin J. Strom is the Director of the Center for Policing Research and Investigative Science at RTI International and a nationally recognized expert in policing, forensics, and homeland security. Dr. Strom is the project director for the Sexual Assault Kit Initiative Training and Technical Assistance Program, an effort funded by the Bureau of Justice Assistance to support the testing of unsubmitted sexual assault kits.

William J. Parish is a research economist in RTI International’s Behavioral Health Services, Policy, and Economic Research Program with expertise in applied econometrics and health economics. His current work includes a mix of primary and behavioral healthcare program evaluation and he has recently published in the International Review of Psychiatry, Psychiatric Services, Addiction, and the American Journal of Preventive Medicine.

Patricia A. Melton is nationally recognized for her work in sexual assault reform and forensic DNA analysis. As a senior research forensic scientist at RTI International, she develops and implements training, technical assistance and educational strategies which promote the derivation of polices focusing on sexual assault evidence to solve violent crimes as well as participate in national initiatives addressing sexual assault.

Amanda Royal Young is a public health analyst in RTI International’s Investigative Science Program. Her research interests include sexual assault, human trafficking, policing and investigative sciences, and domestic and intimate partner violence.

References

Aigner

D. J.

Lovell

C. A. K.

Schmidt

(1976). Formulation and estimation of stochastic frontier production function models. Santa Monica, CA: RAND Corporation.

Bashford

(Producer). (2013, March). How New York City tackled its backlog. New York City’s sexual assault kit backlog project: Lessons learned. Retrieved from http://victimsofcrime.org/docs/DNA%20Trainings/brashford-slides.pdf?sfvrsn=0

Burgess

A. W.

Hanrahan

N. P.

Baker

(2005). Forensic markers in elder female sexual abuse cases. Clinics in Geriatric Medicine, 21, 399-412. doi:10.1016/j.cger.2004.10.005

Butler

J. M.

Butler

J. M.

(2010). Fundamentals of forensic DNA typing. Amsterdam, The Netherlands: Academic Press.

Campbell

Menaker

King

(2015). The determination of victim credibility by adult and juvenile sexual assault investigators. Journal of Criminal Justice, 43, 29-39

Campbell

(2005). What really happened? A validation study of rape survivors’ help-seeking experiences with the legal and medical systems. Violence and Victims, 20, 55-68.

Campbell

(2008). The psychological impact of rape victims. American Psychologist, 63, 702-717. doi:10.1037/0003-066X.63.8.702

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, 363-376. doi:10.1177/1524838015622436

Campbell

Patterson

Bybee

Dworkin

E. R.

(2009). Predicting sexual assault prosecution outcomes: The role of medical forensic evidence collected by sexual assault nurse examiners. Criminal Justice and Behavior, 36, 712-727. doi:10.1177/0093854809335054

10.

Campbell

Raja

(2005). The sexual assault and secondary victimization of female veterans: Help-seeking experiences with military and civilian social systems. Psychology of Women Quarterly, 29, 97-106. doi:10.1111/j.1471-6402.2005.00171.x

11.

Campbell

Shaw

Fehler-Cabral

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

12.

Cantillon

Kopiec

Clawson

(2009). Evaluation of the impact of the forensic casework DNA backlog reduction program. Fairfax, VA: ICF International.

13.

Cantwell

(2002). A report on the potential for apprehending rapists in Washington State using DNA analysis. Washington, DC: Office of Senator Maria Cantwell.

14.

Corum

Carroll

(2014). Forensic analysts’ perspectives: Sexual assault kits under the microscope. Journal of Forensic Nursing, 10, 50-57. doi:10.1097/JFN.0000000000000022

15.

Durose

M. R.

, & U.S. Bureau of Justice Statistics. (2008). Census of publicly funded forensic crime laboratories, 2005. Washington, DC: U.S. Department of Justice, Office of Justice Programs.

16.

Filipas

H. H.

Ullman

S. E.

(2001). Social reactions to sexual assault victims from various support sources. Violence and Victims, 16, 673-692.

17.

Horney

Spohn

(1996). The influence of blame and believability factors on the processing of simple versus aggravated rape cases. Criminology, 34, 135-162. doi:10.1111/j.1745-9125.1996.tb01200.x

18.

Horvath

Meesig

(1996). The criminal investigation process and the role of forensic evidence: A review of empirical findings. Journal of Forensic Sciences, 41, 963-969. doi:10.1520/JFS14032J

19.

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, 193-222. doi:10.1177/1077801212440157

20.

Langton

Berzofsky

Krebs

C. P.

Smiley-McDonald

(2012). Victimizations not reported to the police, 2006-2010. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

21.

Lehtonen

Pahkinen

(2004). Practical methods for design and analysis of complex surveys (2nd ed.). West Sussex, UK: John Wiley.

22.

Lisak

Miller

P. M.

(2002). Repeat rape and multiple offending among undetected rapists. Violence and Victims, 17, 73-84.

23.

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). Washington, DC: U.S. Department of Justice.

24.

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. Retrieved from https://case.edu/socialwork/begun/sites/case.edu.begun/files/2018-09/backlog-brief_8_14_17_Final.pdf

25.

Martin

P. Y.

(2005). Rape work: Victims, gender, and emotions in organization and community context. New York, NY: Routledge.

26.

McEwen

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

27.

National Research Council. (2009). Strengthening forensic science in the United States: A path forward. Washington, DC: The National Academies Press.

28.

Nelson

M. S.

(2010). Making sense of DNA backlogs, 2010—Myths vs. reality (NIJ Special Report). Washington, DC: National Institute of Justice.

29.

O’Brien

R. M.

(2007). A caution regarding rules of thumb for variance inflation factors. Quality & Quantity, 41(5), 673-690.

30.

O’Neal

E. N.

Tellis

Spohn

(2015). Prosecuting intimate partner sexual assault: Legal and extra-legal factors that influence charging decisions. Violence Against Women, 21, 1237-1258. doi:10.1177/1077801215591630

31.

Parnis

Du Mont

(2006). Symbolic power and the institutional response to rape: Uncovering the cultural dynamics of a forensic technology*. Canadian Review of Sociology/Revue Canadienne de Sociologie, 43, 73-93. doi:10.1111/j.1755-618X.2006.tb00855.x

32.

Patterson

Campbell

(2010). Why rape survivors participate in the criminal justice system. Journal of Community Psychology, 38, 191-205. doi:10.1002/jcop.20359

33.

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, 2259-2275. doi:10.1177/0886260511432155

34.

Perkel

(2007, May 21). DNA labs facing huge backlog: Canadians and Americans failing to meet target processing times for thousands of forensic DNA samples. The Scientist. Retrieved from https://www.the-scientist.com/daily-news/dna-labs-facing-huge-backlog-46448

35.

Peterson

J. L.

Hickman

M. J.

, & U.S. Bureau of Justice Statistics. (2005). Census of publicly funded forensic crime laboratories, 2002. Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.

36.

Peterson

J. L.

Johnson

Herz

Graziano

Oehler

(2012). Sexual assault kit backlog study. Washington, DC: National Institute of Justice.

37.

Peterson

J. L.

Sommers

Baskin

Johnson

(2010). The role and impact of forensic evidence in the criminal justice process. Washington, DC: National Institute of Justice.

38.

Pratt

T. C.

Gaffney

M. J.

Lovrich

N. P.

Johnson

C. L.

(2006). This Isn’t CSI: Estimating the national backlog of forensic DNA cases and the barriers associated with case processing. Criminal Justice Policy Review, 17, 32-47. doi:10.1177/0887403405278815

39.

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, 1-13.

40.

Reilly

(2015, July 16). Tens of thousands of rape kits go untested across USA. USA Today. Retrieved from https://www.usatoday.com/story/news/2015/07/16/untested-rape-kits-evidence-across-usa/29902199/

41.

Ritter

(2011). The road ahead: Unanalyzed evidence in sexual assault cases. Washington, DC: U.S. Department of Justice, Office of Justice Programs, National Institute of Justice.

42.

Ritter

(2012a). Solving sexual assault: Finding answers through research. NIJ Journal, 270, 4-17.

43.

Ritter

(2012b). Untested evidence in sexual assault cases: Using research to guide policy and practice. Sexual Assault Report, 16, 33-48.

44.

Rubin

(2009, May 31). DNA going untested in cases throughout L.A. County. Los Angeles Times. Retrieved from http://articles.latimes.com/2009/mar/31/local/me-dna31

45.

Rubin

Winton

(2009, October 29). L.A. announces plan to reduce backlog of unexamined DNA evidence from violent crimes. Los Angeles Times. Retrieved from https://www.latimes.com/local/la-me-dna29-2008oct29-story.html

46.

Sacco

L. N.

James

(2015). Backlog of sexual assault evidence: In brief. Washington, DC: Congressional Research Service.

47.

Scott

R. K.

(2001). A comparison of rape evidence collected by hospitals with that collected by sexual assault nurse examiners in Macomb County, Michigan (Doctoral dissertation). Michigan State University, East Lansing.

48.

Shaw

Campbell

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

49.

Spohn

Tellis

(2012). The criminal justice system’s response to sexual violence. Violence Against Women, 18, 169-192. doi:10.1177/1077801212440020

50.

Starzynski

L. L.

Ullman

S. E.

Filipas

H. H.

Townsend

S. M.

(2005). Correlates of women’s sexual assault disclosure to informal and formal support sources. Violence and Victims, 20, 417-432.

51.

Steadman

(2000). Survey of DNA crime laboratories, 1998 (Bureau of Justice Statistics Special Report). Washington, DC: U.S. Department of Justice, Office of Justice Programs.

52.

Steadman

(2002). Survey of DNA crime laboratories, 2001 (Bureau of Justice Statistics Bulletin). Washington, DC: U.S. Department of Justice, Office of Justice Programs.

53.

Strickler

(Writer). (2009, November 10). Thousands of rape kits wait to be tested. CBS News. Retrieved from https://www.cbsnews.com/news/thousands-of-rape-kits-wait-to-be-tested/

54.

Strom

Hickman

(2010). Unanalyzed evidence in law-enforcement agencies. Criminology & Public Policy, 9, 381-404. doi:10.1111/j.1745-9133.2010.00635.x

55.

Tasca

Rodriguez

Spohn

Koss

M. P.

(2013). Police decision making in sexual assault cases: Predictors of suspect identification and arrest. Journal of Interpersonal Violence, 28, 1157-1177. doi:10.1177/0886260512468233

56.

Telsavaara

T. V.

Arrigo

B. A.

(2006). DNA evidence in rape cases and the Debbie Smith Act: Forensic practice and criminal justice implications. International Journal of Offender Therapy and Comparative Criminology, 50, 487-505. doi:10.1177/0306624X05285929

57.

Tofte

(2009). Testing justice: The rape kit backlog in Los Angeles City and County. New York, NY: Human Rights Watch.

58.

Tofte

(2010). “I used to think the law would protect me”: Illinois’s failure to test rape kits. New York, NY: Human Rights Watch.

59.

Tofte

(2013, August). Guest blog Sarah Tofte: Police often abandon rape kits and investigations. Retrieved from http://www.endthebacklog.org/blog/guest-blog-sarah-tofte-police-often-abandoned-rape-kits-and-investigations

60.

Ullman

S. E.

(1996). Do social reactions to sexual assault victims vary by support provider? Violence and Victims, 11, 143-157.

61.

U.S. Department of Justice Office on Violence Against Women. (2013). A national protocol for sexual assault medical forensic examinations: Adults/adolescents (2nd ed.). Washington, DC: U.S. Department of Justice.

62.

Valentine

J. L.

Sekula

L. K.

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, 3547-3573. doi:10.1177/0886260516681881

63.

Valentine

J. L.

Shaw

Lark

Campbell

(2016). Now we know: Assessing sexual assault criminal justice case processing in an urban community using the Sexual Assault Nurse Practitioner Evaluation Toolkit. Journal of Forensic Nursing, 12, 133-140. doi:10.1097/JFN.0000000000000115

64.

Zinzow

H. M.

Thompson

(2015). A longitudinal study of risk factors for repeated sexual coercion and assault in U.S. College men. Archives of Sexual Behavior, 44, 213-222. doi:10.1007/s10508-013-0243-5