Perpetrator Involvement in Domestic Homicide Reviews in England and Wales

Data Collection

Participants included stakeholders who had participated in a DHR, including independent chairs, review panelists—like DA coordinators (DACs) (who oversee local partnership responses), specialists from DA services, and other agency representatives—and family members and advocates supporting families. Participants were initially recruited for an anonymous online survey using a purposeful and snowballing strategy. Before completing the survey, respondents received information, and informed consent was obtained. A total of 117 survey responses were received. At the end of the survey, respondents were invited to an interview. If respondents agreed, they were asked to provide contact details and then approached, offered further information, and asked to complete a consent form. Forty stakeholders were subsequently interviewed by phone/video conference. Semi-structured interviews explored experiences and perceptions of DHRs. Each interview was audio-recorded, and a transcript produced (if desired, participants could see and comment on this, with “Interviewee Transcript Review” appropriate in sensitive research; see Rowlands, 2021). Pseudonyms protect participant identities, and their role is reported at first reference. See Table 1.

OPEN IN VIEWER

Table 1.

Participant Roles.

Involvement	n	%
Independent Chair	9	31.0
Family Member	3	10.3
DA Specialist	6	20.7
DAC	4	13.8
Other Agency Review Panelist	4	13.8
Family Advocate	3	10.3
Total	29	100.00

Data Analysis

Reflexive thematic analysis was undertaken (Braun & Clarke, 2021). In the doctoral study, one of the themes generated was “practices of review.” Within this theme, the sub-theme “perpetrator involvement” was coded across 29 interviews. These data were extracted, re-analyzed, and re-coded through a close, interpretative reading. Coding drew on my practice experience.

Data Collection

Assisted by my insider status, I approached the Home Office for information on DHR reports submitted to the QA panel in 2018 and approved for publication. While this information was incomplete, it enabled a targeted search of community safety/other partnership websites. Searches were conducted at the end of 2019 and repeated in 2020, allowing up to 2 years between submission and publication. Exclusions included DHRs I had chaired or where only executive summaries were available. A total of 60 DHR reports were located online: 43 IPHs, 11 AFHs, and six other types of DA-related deaths.

Data Analysis

Document analysis was undertaken (Bowen, 2009). Within the coding schedule/manual, one category was “family, informal network and perpetrator involvement.” Data about the nature of involvement was generated using pre-defined codes based on my practice experience. Coding was usually explicit, although sometimes interpretation was necessary (Rowlands & Bracewell, 2022). Analysis was in Excel with descriptive statistics through SPSS.

Agency Information

The doctoral study examined DHR practices, including information gathering from agencies. This was no different for perpetrators and could be important if they had “loads of. . . agency contact” (Caroline, a DA specialist). Examples included contact with CJ and health agencies and also inter-agency work. This reflects the requirement in the statutory guidance for information sharing, including from agencies working with perpetrators (Home Office, 2016, p. 11). Yet, there could be barriers to accessing this information, either generally (because of agency confidence or willingness to participate) or specifically (especially for DA-related suicides, where agencies might feel unable to share information without a CJ outcome, that is, there was not a convicted perpetrator). Given this paper’s focus on perpetrator involvement directly or by proxy, these issues are not further explored.

Direct Perpetrator Involvement

Of the different ways to involve perpetrators, the most discussed was direct involvement, with participants sharing experiences or perceptions of challenges and risks. Illustrating the varying degrees of concern or support for this practice, Lily (a family advocate) explained that involvement was necessary because “it’s part of the guidance.” More purposefully, Margaret identified how, as an independent chair, she tried to “get as much as I can about the perpetrator because. . . [it’s] half the equation.” The responsibility for facilitating perpetrator involvement was identified as mainly falling on independent chairs. Usually, this was facilitated by the police or prison service following conviction, with several independent chairs describing visiting a perpetrator in prison or seeking involvement by letter. While no detailed accounts of seeking or gaining access were provided, Emma (an independent chair) implied that involvement did not always happen, distinguishing between DHRs “where I have talked to the perpetrator.” Supporting the suggestion that perpetrator involvement was not routine, Mia (a DA specialist) referred to participating in a DHR that was “the first one. . . [where] the perpetrator also contributed.” In DA-related deaths by suicide, several participants explained that involvement was often not sought, reflecting concerns about (a) the absence of a conviction, as noted above, or (b) fears of or potential risk to a victim’s family, including harassment by the (alleged) perpetrator or impact on the care of surviving children.

Involvement of Perpetrator’s Family

Other participants discussed involvement by proxy, mainly by the perpetrator’s family and potentially other testimonial network members. Examples included a perpetrator’s sister, who provided information about a perpetrator’s childhood, and members of the community who were “outside of the standard agency contact” (Caroline). However, two issues were of note. First, engagement with a perpetrator’s testimonial networks was not always considered. Bobby (a family advocate) explained, “it’s not consistently that we do interview them, or even contact them to give them an opportunity to contribute.” In Bobby’s view, this inconsistency was because these testimonial networks were seen as compromised because their family member “was. . . a perpetrator of DA.” Second, if involved, this could have—as for a victim’s family, which is discussed below—an adverse impact. Thus, Hudson (an independent chair) shared an example where:

The perpetrator’s father was a really nice guy, and he was, you know, very upset about it all, and he didn’t see it coming and blames himself and all of that.

Case Background and Circumstances

Leilani (a DAC) summarized this sub-theme, describing the involvement of the perpetrator as about “gaining a better understanding. . . [of] their background.” Background information might be about a perpetrator’s own history and/or relationship(s) (including abusive behavior). Meanwhile, circumstances might include understanding a perpetrator’s help-seeking, including access to and/or contact with agencies, not least “what. . . [agencies] did and didn’t do and [a perpetrator’s] understanding of what was out there to help support them” (Hudson). Several participants emphasized that, without perpetrator involvement, there might be less focus on the individual responsible for the abuse, including their pathway to that behavior, leaving learning less “rounded” (Hazel, a review panelist). Finally, perpetrator involvement might help build a picture of what “led up to it [the death event]” (Harper, a DAC), including the perpetrator’s actions (e.g., a homicide) or the potential impact of their violence and abuse (e.g., a death by suicide).

Case Learning

Perpetrator involvement could enable the learning of lessons because, as Marie (a family advocate) noted, they might “disclose something significant that may have helped prevent the homicide from happening,” which, along with information about case circumstances and background, could help answer the question “why have they done what they did?” (Hazel). Thus, perpetrator involvement could contribute to an improved understanding of the long-term background and/or the events preceding a DA-related death. For example, Owen (a review panelist) described a case which uncovered learning about information sharing between the police and mental health services. More thematically, Hudson explained how learning included “broadening the understanding of what the landscape is out there to support victims and perpetrators” (emphasis added).

Among participants, perpetrator involvement was identified as important for learning vis-a-vis CJ processes. First, if the perpetrator pleaded guilty, there would not be a full criminal trial, so “you never heard the background” (Harper). Consequently, perpetrator involvement could ensure that this background was explored, leading to learning and helping answer a victim’s family’s questions. Second, as DHRs are about learning, while CJ is about culpability, perpetrator involvement could enable an examination of complexities in DA-related deaths. An example several participants provided was where the homicide victim was the perpetrator of DA (and so the homicide offender was the substantive victim). Illustratively, Grace (an independent chair) explained how, in a DHR where a victim of DA had been convicted of manslaughter, by meeting with her, it was possible to “get an understanding of what the relationship had been like” and so contextualize the killing.

Positive Outcomes for Family

As a final sub-theme, although there could be an adverse impact on a victim’s family (discussed below), some participants identified how perpetrator involvement might have positive outcomes and, for some, be welcomed. Several participants highlighted what they felt was an assumption that family might be against perpetrators’ involvement. Yet, although most family members interviewed—like Isabella and Avery (and also some professional participants)—were critical of, or resistant to, perpetrator involvement, this was not always the case. Lily, a family advocate, explained that some families were “keener for that than most chair’s think they would be” and that “by and large they are like, ‘yeh, go and ask him’.” The reason for this was twofold. Perpetrator involvement could be about getting answers “because they [the victim’s family] want to know why, they want to hear from the horse’s mouth as it were, what triggered this or what happened.” Alternatively, a family might be clear about where responsibility lay. Amelia (an independent chair) described one family’s determination that “the focus should be on him and what he did, not their family member”; this then shaped how they engaged in the DHR process.

Whose Story Is Told

Participants identified how perpetrator involvement could lead to questions about whose story was being told and how (i.e., a DHR’s focus). However, it is essential to note that while these concerns are related to perpetrator involvement, they also bring wider issues with the quality of a DHR’s conduct into scope. Nonetheless, first, regardless of their involvement, participants suggested a DHR could focus on a perpetrator’s story. Thus, Amelia described how a DHR might end up “being his story. Rather than hers [the victims]” because “he’s the one who has had agency contact.” This could be particularly challenging if a victim’s family were not involved. In her example, Amelia was concerned that, as an independent chair, she had not had “enough . . . to be able to articulate any of her story.”

Second, if a perpetrator was involved, participants questioned their intent and/or veracity. As Joshua (an independent chair) noted, in some cases, a perpetrator “denies wholeheartedly any abuse.” More fundamentally, Grace asked:

If the perpetrator has lied consistently through a criminal process, why are they suddenly going tell us the truth? And how will we know if they were telling us the truth?

Even if a perpetrator acknowledged their actions, some participants felt that the information provided might be suspect. For example, Victoria (a DAC) suggested a perpetrator might be “minimizing and making excuses. . . [thereby] making him look better than he was. . . and diminishing the experience of the victim.” There was no explicit discussion of perpetrator testimonial networks in this respect; clearly, a similar tension about the veracity of their information could exist. Yet, conversely, a perpetrator’s testimonial networks have different subjective experiences and do not necessarily participate in these tactics: as noted above, Hudson described a perpetrator’s father’s involvement positively.

Participants were concerned not only about perpetrator involvement in its own right but also about what this might mean for a victim who might, consequently, fall from focus. This could include questioning perpetrator involvement per se, given they were responsible for a victim’s death. Illustratively, while Claire’s concerns about the treatment of her loved one reflected wider issues with the DHR process, they also speak to a perpetrator over-focus:

I think one of the things for me was she just like another case. There was nothing personalised that she was actually a human being. And I think a lot of it was perpetrator based and not really about my daughter (emphasis added).

Role of Professional Stakeholders

Many participants recognized that perpetrator involvement could be challenging for professional stakeholders, as for families. Significantly, there was an emphasis on the responsibility of the independent chair to manage this process, including usually facilitating perpetrator involvement and ensuring “that only relevant information is used” (Marie). By relevance, Marie meant critically engaging with information gathered from or about the perpetrator. However, in some cases, participants identified how the independent chair lacked the skills, knowledge, or experience to do this, forcing a review panel intervention. Thus, after a perpetrator interview in prison, Victoria described how “I remember us [the review panel] having to say to the chair. . . ‘you need to couch that’. . . rather than put it boldly in the report.” Here, Victoria identified how an independent chair could uncritically take a perpetrator’s view.

Conversely, review panelists might present a challenge to perpetrator involvement. Bobby noted that sometimes professionals might, in effect, be “siding with a perpetrator” (by which she meant accepting their account of events and victim blaming). Alternatively, review panelists might be reluctant about perpetrator involvement. Thus, Emily (an independent chair) described how “members of the panel were very resistant to him having any input at all, because they felt well, ‘where’s her voice?’” This reluctance was such that it could arise even when the offender was the substantive victim. Emma described one case where:

I went and visited her in prison. But the Panel were really not interested in what her experience was. . . And she was tricky for them, you know? She was volatile, she was intimidating, she was a lot of things that we are used to perpetrators being. But she was also a victim. A serial victim. And again, that’s hard for Panels to get their heads around.

Negative Outcomes for Family

While there could be benefits for a victim’s family from a perpetrator’s involvement, it could nonetheless be “difficult and upsetting for the [victim’s] family” (Marie). Avery explained that this was because even if the perpetrator’s involvement “will help in the long run. . . it won’t help . . . right now, this day.” Moreover, while families might have different views about perpetrator involvement, as already discussed, Avery also highlighted how impact could vary too. On the one hand, a family might feel that “why [should] that person would get the chance to say something when they have just been convicted of causing the death.” Conversely, “if they decide that they don’t want to say anything [i.e., to be involved], that can be really hurtful again.” Given the potential for such affective harm, Bobby highlighted that a decision might be made not to contact a perpetrator. However, as already noted, perpetrator involvement was also sometimes not pursued because of fears of potential risk to a victim’s family, including physical or other harm. Illustratively, several participants discussed reviews into deaths by suicide where the (alleged) perpetrator was living near a victim’s family and/or had a role in terms of care of the children. In these examples, it was reported that families felt they could not participate, although it was unclear whether the perpetrator had been involved or was aware of the DHR. However, the possibility that a DHR could precipitate tension was clear: in another example, an (alleged) perpetrator tried to resist the commissioning of a DHR. While understanding these decisions, Bobby highlighted that a perpetrator’s absence might limit learning.

Accountability

The first aspect of finding balance was accountability. As Harper highlighted, it is the responsibility of the CJ process to determine culpability, and thus, “actually, having a. . . [DHR]. . . isn’t going to make that difference” in terms of “sanctions.” Yet, for several participants, a CJ outcome provided assurance by establishing the perpetrator’s status and thereby framing their involvement because, as Margaret explained, “the blame. . . remains with the perpetrator” (although, as already noted, this was not usually the case in reviews into deaths by suicide). This assurance shaped perpetrators’ involvement, directly or by proxy. For Iris (an independent chair), perpetrator involvement consequently meant “not colluding, but at the same time being sufficiently sympathetic that they feel able to talk about their experiences with services.” In another example, Amelia reflected on a DHR she had chaired where most agency contact had been with a perpetrator, so she had been—as discussed previously—concerned that it had “ended up being his story.” To address this and ensure accountability, Amelia described “some of the speculation that happened. . . and my musings with the minimal information that we had” to try and re-focus the review, including talking to a former partner whose information about the perpetrator’s abusive behavior “took the spotlight off him a bit.”

Interpreting Information

As a result of concerns about veracity or intent and any potential impact on a victim’s family, participants emphasized that interpreting information gleaned from or about perpetrators was necessary to deliver accountability. Interpretation required contextualization, in particular, an understanding of the dynamics of DA. For example, Hazel emphasized the importance of a perpetrator’s background (including their childhood) to understand their pathway to abuse but, in doing so, ensuring accountability by interpreting this without “making any excuses for what they’d done, but just to see where it all fitted in.” However, this might also mean recognizing that some information could not be used or, if used, was done so in a way that illuminated abusive tactics. In one example, as an independent chair, Hudson met a perpetrator in prison: he described coming away with an understanding of how everything he had said had been to “justify” what they had done. This might mean “putting some boundaries around what information is put in or not” (Bobby).

However, if interpretation was not managed successfully, it could affect both findings and, critically, a victim’s family and their sense of whether a perpetrator was being held accountable. Family member Isabella felt it had been “insensitive” of an independent chair to use information from the perpetrator, who she felt had been seeking to “justify herself, trying to protect herself.” In contrast, another family member—Claire—felt the independent chair had achieved this balance because they “took what he needed to take from us about him, but it was a lot about [my loved one], her feelings and thoughts.”

Two final points are of note. First, interpretation is not solely necessary for information gleaned via perpetrator involvement. Interpretation is sometimes also required regarding what a victim’s family wanted to say about a perpetrator and whether this could be included in any findings. (Although this point has universal application: all information, including from agencies, needs to be interrogated). Illustrating this tension, Bobby gave the specific example of a victim’s family who wanted information included about a perpetrator that “had absolutely no foundation of evidence or. . . you know, we didn’t know if was true.” Second, and echoing earlier discussions, interpretation is necessary to resolve case complexities, including where the offender was the victim of DA.

In summary, participants described how, in different ways, perpetrator involvement could play a part in, indeed be of benefit to, a DHR. Yet, at the same time, perpetrator involvement could be complex because of perceived or actual challenges and risks, not least because of concerns about a perpetrator’s possible motives or truthfulness. However, participant accounts also suggested that perpetrator involvement and/or that of their families is uncommon. To explore the frequency and nature of perpetrator involvement, I turn next to the DHR reports.

Direct Perpetrator Involvement

Attempts were made to engage perpetrators in just over half of the DHR reports (n = 31, 51.7%). However, perpetrators were little involved, with participation only recorded in 5 DHRs (8.3%). In these DHR reports, there was relatively little information on the nature of this involvement, although all five perpetrators appear to have been interviewed once in prison by the independent chair. In two of the DHR reports, specific procedural issues were reported, with one report indicating that consent was sought for access to medical records and agreement of pseudonym(s). In three of the DHR reports, how the perpetrator’s information was used was also reported, including statements about the necessity of verifying any claims or reporting a claim but being clear that the review panel did not accept it.

Otherwise, perpetrators declined to be involved (n = 13, 21.7%), did not respond (n = 11, 18.3%), or were sent the DHR report but did not then respond (n = 2, 3.3%). In 15 DHRs (25%), no approach was made, reflecting concerns about possible risks from or quality of any contribution (n = 8, 13.3%) or because the perpetrator had died by suicide (n = 7, 11.7%). No information was recorded about perpetrator involvement in 14 DHR reports (23.3%). See Table 2.

OPEN IN VIEWER

Table 2.

Direct Perpetrator Involvement.

Involvement	n	%
Approached—Declined	13	21.7
Approached—Involved	5	8.3
Approached—No response	11	18.3
Approached—Other a	2	3.3
Decided not to approach	8	13.3
Not able to approach—Suicide	7	11.7
Missing	14	23.3
Total	60	100.00

a

Shared report, but no response received.

Involvement of Perpetrator’s Family

Attempts were made to engage with the perpetrators’ families in over half of the DHR reports (n = 35, 58.3%), with over a third then involved (n = 17, 28.3%, with a further 6 DHRs (10.0%) including no details). Across these DHR reports, family involvement was facilitated by the independent chair, with meetings taking place either in person or sometimes via phone. There were varying levels of detail about the nature of this involvement, including any follow-up (such as sharing a draft report). Notably, family makeup varied, including parents, children, former partners, or other connected people.

Otherwise, families declined to be involved (n = 8, 13.3%) or did not respond (n = 4, 6.7%). In one case, a decision was made not to attempt to engage the family (n = 1, 1.7%). No information about family involvement was recorded in 24 DHR reports (40.0%). See Table 3.

OPEN IN VIEWER

Table 3.

Involvement of Perpetrator Family—All.

Involvement	n	%
Approached—Declined	8	13.3
Approached—Involved	17	28.3
Approached—Involvement missing	6	10.0
Approached—No response	4	6.7
Not able to approach—Other	1	1.7
Missing	24	40.0
Total	60	100

However, there were differences between case types. In IPHs, no information was recorded for just over half of the families (n = 23, 53.5%). Otherwise, just under half of families were approached (n = 19, 44.2%), with just under a third then involved (n = 10, 23.3%. In 1 DHR (2.3%), no details of involvement were recorded). Otherwise, families declined to be involved (n = 5, 11.6%) or did not respond (n = 3, 7.0%). In one case, a decision was made not to attempt to engage with the family (n = 1, 1.7%). See Table 4.

OPEN IN VIEWER

Table 4.

Involvement of Perpetrator Family—IPH.

Involvement	n	%
Approached—Declined	5	11.6
Approached—Involved	10	23.3
Approached—Involvement missing	1	2.3
Approached—No response	3	7.0
Not able to approach—Other	1	2.3
Missing	23	53.5
Total	43	100

In AFHs, bar one family for whom no information was recorded (9.1%), all families had been approached (n = 10, 90.9%). Almost two-thirds were involved (n = 6, 54.5%, with one family approached but no details recorded, 9.1%), and just under a third declined (n = 3, 27.3%). See Table 5.

OPEN IN VIEWER

Table 5.

Involvement of Perpetrator Family—AFH.

Involvement	n	%
Approached—Declined	3	27.3
Approached—Involved	6	54.5
Approached—Involvement missing	1	9.1
Missing	1	9.1
Total	11	100.0

For other case types, no information was recorded for most families (n = 4, 66.7%), with two families being approached and then being involved or not responding (n = 1, 16.7%). See Table 6.

OPEN IN VIEWER

Table 6.

Involvement of Perpetrator Family—Other.

Involvement	n	%
Approached—Involved	1	16.7
Approached—No response	1	16.7
Missing	4	66.7
Total	6	100.0

Using Perpetrators as a Source of Information

While a DHR is victim-centric, any information contributing to learning should be considered. Thus, as noted earlier, the subjective experience of the perpetrator is potentially useful. This possibility is evident from questions in the statutory guidance that help determine the scope of a DHR. Covering issues from timeframe through to focus (like immigration status or contact with multi-agency risk management processes), these are framed regarding both the victim and/or the perpetrator (Home Office, 2016, pp. 13–15).

Yet, a perpetrator has committed—or, in a DA-related suicide, is suspected of having a contributory role toward—fatal DA. Given their actions, drawing on Goffman (1963), we can understand a perpetrator’s identity as spoiled and so marked out as discredited, thereby affecting how a perpetrator and/or their family are used as source(s) of information. Thus, stigma may affect if and how perpetrators respond to an invitation to be involved in a DHR. Indeed, as already noted, research has highlighted difficulties in obtaining information from perpetrators, including achieving access and/or a reluctance to discuss their actions (Dobash & Dobash, 2015; Evans et al., 2023). This difficulty was evident in the DHR reports: perpetrators did not participate (n = 26 or 43.3%) more than take part (n = 5, 8.3%). These difficulties may also be exacerbated by timing. DHRs can be a lengthy process (Haines-Delmont et al., 2022). Based on my experience as an independent chair, in most cases, contact attempts occur after the criminal trial and within a year of the victim’s death. However, bar the indication that independent chairs often work through the police or prison service—a route I have used—the data do not clarify how approaches are made or the issues that need to be managed. Yet, considerations include the best approach. Based on the data reported here and my own experience, perpetrator interviews may be time-limited and so focused on the relationship and death event. Yet, gathering the fullest information possible may require more extensive involvement. For example, when Adams (2007) interviewed men who killed their wives, he did so over several sittings. Meanwhile, a life story approach may elicit more learning about a perpetrator’s trajectory (Di Marco & Evans, 2021). Such considerations are not merely an instrumental concern with data collection. The timing of perpetrator involvement may mean there is limited time for their rehabilitation and recovery, including coming to terms with their actions (Adshead et al., 2015), upon which their willingness to be involved may depend. They are also a matter of ethics given, however uncomfortable it may be, the potential for a perpetrator’s own trauma and its implications (Mohamed, 2015).

Furthermore, this spoiled identity is relational. As Dilmon and Timor (2014, p. 1126) describe it, a perpetrator may be a subject from whom we “recoil from and perhaps even hate.” Thus, other actors are also engaged with a perpetrator’s subjectivity. Illustratively, for some participants, the prospect of perpetrator involvement was unjust because their (in)direct actions meant a victim could not. Thus, Emily highlighted how review panelists might resist perpetrator involvement, feeling, “well, where’s [the victim’s voice]?” Such was the potential strength of this feeling that review panelists might even be reluctant to engage with a DA victim who had killed their abuser. Further complicating usefulness is a victim’s family’s feelings: while some families may want involvement, others may not.

This relationality also encompasses the perpetrator’s family, who were involved in 23 DHRs (38.3%). Yet, in contrast, estimates of victims’ families’ involvement suggest about half take part (Rowlands & Cook, 2022). Thus, while potentially able to provide information, these data support Bobby’s suggestion that perpetrators’ families are not consistently approached, perhaps because of their relationship. Drawing again from Goffman, a perpetrator’s family might be perceived as potentially discreditable because they are associated with someone responsible for serious violence. Goffman (1963, p. 30) describes this as “courtesy stigma,” with Condry (2009) offering an account of such vicarious experiences for the families of men convicted of serious violence. This may explain why only a third of DHR reports involved perpetrator families. I recognize this possibility from my practice: I have rarely engaged with a perpetrator’s family. While often a practical decision (e.g., access issues), upon reflection, this is also perhaps because a perpetrator’s family is not prioritized. The exception is in AFH cases, where, in my experience, engagement occurs because of the shared victim-perpetrator relationship. The findings from the DHR reports appear to support this, given almost all families in AFH cases were approached (a higher rate than reported elsewhere; see Rowlands & Cook, 2022). Thus, conceivably, courtesy stigma is, at least partly, neutralized by a familial relationship. Finally, if a perpetrator’s family is invited, they may participate. What is not clear from the data is what barriers may prevent involvement, with these potentially influenced by a family’s experience of courtesy stigma and whether the DHR process—including the information and support available—overcomes this and enables participation. As with perpetrators, further consideration is required around the best approach in terms of when and how such engagement is conducted, as well as ethical implications.

Whether Perpetrator Involvement Is Useful

Compounding the issue of using perpetrators as a source of information was, for participants, whether the information obtained was useful. Many participants accepted that perpetrator involvement could be useful, providing information about a perpetrator’s history, help-seeking, the relationship and/or the death event. However, participants also felt there might be or had experienced challenges and risks. Effectively, a perpetrator has a dual aspect. First, a DHR’s concern with a perpetrator is victim-orientated (e.g., describing and analyzing their behavior and its impact on a victim). However, a DHR’s concern may also turn toward the perpetrator (e.g., considering if and how agencies managed their risk or responded to need). This duality is evident in the data. Moreover, this duality makes explicable the benefit of understanding case background and circumstances—be that about a perpetrator’s abusive behavior or their experience of help-seeking and agency interventions—but also the challenges and risks of overly focusing on their story at a victim’s expense. Specifically, in using perpetrator information, there is a concern about “colluding with the perpetrators’ world view and description of events” (Enander et al., 2021, p. 64). Generally, this concern arises because perpetrator talk can minimize abuse and obscure responsibility (Hearn, 1998). More specifically, perpetrator accounts may reflect how they manage their identity (Evans et al., 2023), ranging from denial to mediated acknowledgement, to degrees of responsibility taking, to acceptance and change (Dobash & Dobash, 2015, pp. 85–91). Moreover, even if a perpetrator does discuss their violence and abuse, a distinction must be drawn between what they disclose, including if and how this minimizes, denies or blames others for their actions—itself, part of the endeavor to manage their spoiled identity—and what might have been a victim’s account and/or an objective assessment of the same (Di Marco & Evans, 2021). In the small number of perpetrator interviews I have conducted, I have often encountered such identity management, with perpetrators often engaging in victim blame and denial, obfuscation, or being unable—whether intentionally or because of disassociation—to recall or provide information. Even so, these narrative distortions can be useful, given a perpetrator might “unwittingly” reveal their attitudes and beliefs (Adams, 2007, p. 9).

Although the data do not speak to this, a similar question needs to be asked about a perpetrator’s family, including their recognition, understanding of, and response to a perpetrator’s violence and abuse. This includes the complexity of a family member’s feelings toward a loved one who is an offender (Condry, 2009).

Taken together, while documenting perpetrator accounts is useful and can inform prevention (Dobash & Dobash, 2015; Evans et al., 2023; Hearn, 1998), if used uncritically within a DHR, the use of this information may affect findings and risk the perpetuation of a forensic narrative that rationalizes a perpetrator’s behavior (Monckton Smith et al., 2014), either by weighting any account to a victim’s detriment (Bracewell et al., 2022) or enabling perpetrator’s to shape and/or distort it (Dangar et al., 2023). As a result, in seeking perpetrator involvement, a DHR may risk both testimonial and hermeneutical injustice (Cook et al., 2023). In my practice, I have sought to manage this concern, as did participants, by finding a balance between victim family involvement and the management of perpetrator information and whose story was told. Importantly, this is not solely an epistemological question: any (re)produced forensic narratives risk adversely affecting victims’ families (Mullane, 2017). However, it is also important to recognize this is not only about perpetrator involvement, as some of the complexity identified in this paper relate as much to the general functioning and conduct of DHRs as to the specific practice of involving perpetrators. Thus, in use, perpetrator information must be critically analyzed and triangulated with other information.

Unused and Unusable

Thus far, I have argued that the prospect of using a perpetrator’s information and whether this is useful is complex because of the tensions between potential benefits, challenges, and risks. Indeed, the involvement of perpetrators directly or by proxy is uncommon and these sources of information are mainly unused or seen as unusable. The essential tension has been articulated by Mullane (2017), who points out that perpetrator involvement is premised on it helping “achieve the twin purposes of preventing DA and homicide,” and yet, “they rarely express remorse and often struggle to accept responsibility” (p. 275). Perpetrator family involvement can also be limited, although perhaps structured by the nature of the relationship (i.e., AFH or IPH).

In this light, as described in the introduction, the limits of the statutory guidance are notable. Indeed, most of the complexity identified here is not addressed. Why is this the case? One possibility is that the difficulty of this practice is a barrier, exacerbated generally by the limits to the framework, leadership, and oversight for DHRs to date (Haines-Delmont et al., 2022; Rowlands, 2023).

However, we might consider perpetrator involvement in terms of how, as a source of information, perpetrators and/or their families are (un)used and how this comes about in terms of the choices made and who makes them. Thus, using perpetrator information in DHRs is not simply an operational matter about whether they or their proxies might be useful. Instead, decisions around use shape what information is or is not available in a DHR and any findings, generating the potential for discursive and symbolic presence or absence. Drawing on Mary Douglas’s work about purity and danger, we might posit that the danger posed by perpetrator involvement means there is a reluctance to engage with them or their proxies (albeit perhaps this is reduced if there is a shared family relationship), with this both a product of and, indeed, producing a boundary. Thus, a perpetrator is dangerous both because of their past (i.e., because of their violence and abuse to the victim ³ ) and in the present (given their spoiled identity, the actual or perceived veracity of any information they provide and potential impact on a DHR in terms of distortion of learning and or toward stakeholders, e.g., a victim’s family). So, not only is perpetrator involvement dangerous instrumentally, but given the “symbolic load” that can be carried by that which is dangerous, the use of perpetrator information also brings with it the risk of a “contagion” (Douglas, 1966, p. 4). Consequently, the lacuna concerning this practice may reflect a desire for separation that arises from a perpetrator’s violence and abuse by which they both challenge the social order and are a danger to it. Thus, the possibilities of perpetrator involvement—and so the use and usefulness of their information, directly or by proxy—are bounded because it is “out of place. . . and so. . . regarded as objectionable and vigorously brushed away” (Douglas, 1966, p. 197)