Theorising victim decision making in the police response to domestic abuse

The paper provides a conceptual argument for the importance of understanding victim decision making in responding to cases of domestic abuse. It applies core elements of decision-making theory to illustrate how victims may undergo different forms of thinking, could be affected by bias, consider gain and loss differently, and suffer from decision inertia. Suggestions are provided on how this perspective could be used in practice to deliver a victim empowerment approach in policing. However, the paper also expresses the importance of the theory’s limitations, especially around external validity, and suggests that research across various disciplines is critical to determine whether the perspective could be placed onto an empirical footing.


Introduction
Police forces are often considered to have a unique role in handling domestic abuse, since they must respond to frontline calls and act as gatekeepers to the criminal justice system (CJS) (Corcoran and Allen, 2005;Tasca et al., 2012). As the main respondents to abuse, it is paramount to develop best practice into how the police can ensure effective engagement with victims throughout the investigation. Currently, victim withdrawal (also commonly labelled non-cooperation, recantation, retraction, and disengagement) presents an ongoing challenge for police and prosecutors in various jurisdictions (Sleath and Smith, 2017). Successful police investigations that result in a prosecution rarely proceed without victim cooperation (Dawson and Dinovitzer, 2001;Ellison, 2002). In cases where a charge was successful, national statistics from England and Wales illustrate how victim retraction and non-attendance made up 51.6% of all unsuccessful prosecutions in the year ending 2018, with similar trends occurring each year prior to this (Elkin, 2018). Consequently, legal practitioners have explicitly stated that victim non-cooperation is one of the biggest influences in a decision to drop charges .
In jurisdictions where the CJS is the main response to abuse, victims are considered a vital investigative resource (Wilson and Segrave, 2011). This can lead to difficulty for police officers when a victim does not want to cooperate, as it causes conflict between using the aggrieved (the complainant or victim) as a source of information and responding to their needs as a victim (Barrett and Hamilton-Giachritsis, 2013). This conflict is further deepened when victims state they do not want to pursue a criminal sanction, leading to an almost paradoxical issue (Kasturirangan, 2008). That is, how do the police, as a criminal justice agent, foster cooperation with victims who do not want a criminal justice response? To address the issue, Mills (1998) argues that more sensitivity is needed in distinguishing between cases where victims may be empowered to make their own decision to withdraw, in comparison to cases where the victim may have been coerced into dropping the charges.
Despite positive changes to the attitudes of police officers attending domestic abuse incidents (Myhill, 2019), as well as the greater understanding and empathy from officers towards the victims' situations (Jakobsen, 2019), a large percentage of cases still subsequently results in withdrawal (Dawson and Dinovitzer, 2001;Robinson and Cook, 2006;Sleath and Smith, 2017). This behaviour can manifest when the victim provides a formal retraction statement; changes, minimises, or justifies the incident; provides a statement that supports the suspect; or denies that the incident occurred altogether. To provide more structure in understanding this behaviour, Sleath and Smith (2017) established two distinct types of withdrawal from a police investigation. The first relates to formal retraction, which involves a written statement explaining that the victim is withdrawing from the process. The second category relates to informal disengagement, which involves the victim ceasing contact with the police without formal retraction. Whilst those who informally disengage often do not provide reasons for withdrawal, those who compile formal retraction statements tend to explain their decision (Sleath and Smith, 2017). Such information is invaluable to victimology researchers, as it provides insight into why some victims prefer no action despite suffering injuries, pains and losses from their victimisation (Wilson and Segrave, 2011).
Yet research examining the victims' reasons for retraction has been slow to develop, likely due to previous theories, such as Battered Women Syndrome, carrying significant limitations and considering victims as defective decision makers (Hamel, 2013). More recent research, however, illustrates a range of external issues faced by victims when they consider their engagement with a police investigation (Birdsall, 2018;Kingsnorth and Macintosh, 2004;Sleath and Smith, 2017). Although the body of knowledge is growing and police officers are increasingly aware of the difficulties faced by victims, there is still an underlying frustration with the volume of victims who either withdraw their complaint, or refuse to provide a statement altogether (Hester, 2006;Horwitz et al., 2011). This frustration may also differ from one officer to another, due to individual differences in each officer's awareness, training and use of discretion in cases of domestic abuse (Myhill and Johnson, 2016;Saxton et al., 2020). This paper, therefore, argues that for the police to provide a victim-centred approach, more attention needs to be placed on understanding the decision making of victims. Subsequently, it aims to apply decision-making theory as an underlying template for considering and potentially recording the difficulties faced by victims when they are required to engage with the police. Through the lens of victim empowerment (Hoyle and Sanders, 2000), officers could then work with victims to understand why some may initially withdraw, as well as change their minds as the investigation is ongoing. The approach may also allow officers to identify potential cooperation or withdrawal at the initial stages of the police investigation, which could be systematically recorded, contribute towards risk assessments, and be used as a vital piece of information in managing the case. Yet, such training must work in partnership with Independent Domestic Violence Advocates and other domestic abuse agencies who aim to tackle the variety of practical issues faced by victims as they enter into a lengthy investigative and prosecutorial system.
In outlining the approach, the paper will first establish core concepts of decision-making theory, before then applying this to victim decision making in cases of domestic abuse. The paper compiles this argument with the recognition that the perspective suffers from a lack of external validity within the realms of sociology and psychology, and therefore strongly suggests that future empirical research should be conducted to determine whether the perspective can be evidenced to work in practice.

Domestic abuse victims as rational decision makers
Multiple studies have examined how rational choice theory relates to perpetrator decision making during offending (Cornish and Clarke, 2014), such as victim selection (Walsh, 2017) as well as pre-, during and post-crime behaviour choices (Beauregard and Leclerc, 2007). Other research has focused on how these theoretical frameworks (i.e. uncertainty avoidance, causal attribution and focal concerns) influence the judicial decisions made in specialised domestic abuse courts (Pinchevsky, 2017). However, few studies have focused on victims of crime, and fewer still on the decision-making process of domestic abuse victims (Taylor-Dunn, 2016). Instead, a focus on the victim tends to examine external factors that influence their cooperation with the investigation of a crime, rather than the internal mechanisms impacting their decision making.
Some research does however begin to examine relevant issues to this aim. For instance, Kingsnorth and Macintosh (2004) used a rational choice perspective to explore the factors influencing victim cooperation in intimate partner violence cases. They concluded that intimate partner violence victims are 'engaged in a complex decision-making process in which they seek to weigh the costs and benefits of involving criminal justice system officials in their lives' (Kingsnorth and Macintosh, 2004: 322). However, their study does not specify the decision processes or rules underlying a victim's assessment of costs and benefits (Burgess-Proctor, 2012). Instead, Kaiser et al. (2017) suggest that the focal concerns perspective provides a broader and more nuanced way of exploring victim decisions to cooperate in these types of cases. Through an adaption of the perspective to sexual assault victim decision making (Kaiser et al., 2017), their study found that victim focal concerns relate to (a) seriousness of the offence/dangerousness of the offender, (b) costs of cooperation, and (c) likelihood of conviction. The second focal concern of victims -costs of cooperating with police and prosecutors -relates to not only tangible costs, such as the time required of the victim as the case moves forward, but also intangible costs such as reputational damage resulting from cross-examination in court. This could be used to connect domestic abuse victim cooperation decisions to cognitive and decisional theory relating to loss aversion and the increased power of loss-framed decisions. For instance, Wydell and Zerk (2017) explored helpseeking behaviours in older victims of domestic abuse and found that their interconnectedness with family, social embeddedness in the community, and meanings of the home influenced helpseeking. In this sense, potential for loss of those things (i.e. loss of access to grandchildren, threat of social isolation, etc.) through pursuing prosecution may influence the decision to disengage with the criminal justice process or even dissuade reporting in the first place. Therefore, officers could potentially harness knowledge of the core principles of victim decision making to assess and capture the huge variation in victim circumstances and how these may affect victim cooperation with the investigation.

Key concepts in decision making and their application to domestic abuse victimology
This paper is devoted to establishing some of the most important concepts in explaining decision making. Borrowing rationale from economics and psychology, this paper explores the decisions, or lack thereof, some victims may make based upon their circumstances. The paper first outlines the dual system theory of cognition, explaining how victim decision making may be processed differently during the immediate abuse through to the closure of their case. Then the rationalemotional model (Anderson, 2003) and prospect theory (Kahneman and Tversky, 1979) are used to aid explanations of victim decision making, demonstrating how rational choices are made based upon the victim's circumstances.

Dual processing theory
Based on decades of research conducted with his collaborator Tversky, Kahneman (2011) suggests there is a general dichotomy between two modes of thought and decision-making processes: System 1 thinking, which is fast, automatic and instinctive and System 2 thinking, which is slow, effortful and conscious. This dichotomy could also be summarised as the difference between intuitive and analytical thinking. System 1 decisions are led by emotion and appear to be subconscious, whereas coming to System 2 decisions requires conscious effort.
System 1 decisions tend to reflect experience: once a response to a recognised pattern of stimuli is learned well through reinforced practice, it becomes automatic and therefore feels easy, intuitive and can be quick. This experience allows for pattern recognition within a situation and helps the decision maker draw upon previously learned associations to swiftly arrive at a choice with limited information and without the cognitive demands of careful and systematic comparison of alternatives. This holds the benefit of making decisions quickly and is said to be generally adaptive and useful (Croskerry and Norman, 2008;Eva and Norman, 2005). As initial calls to the police during a domestic abuse incident have previously been linked to victims fearing for their lives and sustaining further injury (Saxton et al., 2020), System 1 (an instinctive survival response) may take control as the victim interprets the situation and acts to preserve their life and physical integrity. The result of this may often be a call to the police for assistance in handling the immediate abusive situation, without thought or desire for wider criminal justice implications (Hirschel and Hutchinson, 2003).
However, after the initial police response, the victim then faces a gateway into the CJS, where there is a lengthy process of investigation, charging and prosecution (and potentially sentencing) of the suspect. As this process can last months (Hester, 2006), the victim is likely to enter into System 2 thinking where they analyse and rationally assess the realistic options before them. System 2 decisions follow a goal-orientated and systematic process, whereby alternative options are evaluated in order to identify the optimal choice. This fits with rational choice theory, a theoretical framework often used in economics, which suggests that individuals use objective calculations (i.e. System 2 thinking) to make rational decisions to achieve outcomes in line with their own personal objectives and self-interest (Buskens, 2015). Whilst this approach is more likely to generate 'better' decision making compared with potentially biased intuition, this thinking is time intensive, requires more cognitive resources and is reliant on complete situational information in order to systematically compare options (Gilbert, 2002;Kahneman, 2011).
Whilst the theory itself is common in economics, rational choice theory has been widely criticised in other social sciences (i.e. sociology and psychology) for its lack of relevance to real-life decision making (lacking external validity) and lack of empirical testing (reliability) (Green and Shapiro, 1994). Instead, it is argued that humans have bounded rationality in that they are limited in rational decision-making ability by multiple factors including access to information, cognitive limitations and time (Simon, 1987). Such limitations are echoed in the victim choice approach to domestic abuse (Hoyle and Sanders, 2000: 17). Subsequently, many Western jurisdictions argue the use of pro-arrest approaches, which aim to act in the victim's best interests regardless of the views or choices they express (Hoyle and Sanders, 2000: 18). However, whilst bounded rationality may be a limitation to a pure victim choice approach, the limitations to human decision making could be accounted for and enacted upon through a victim empowerment approach (Hoyle and Sanders, 2000: 19). Therefore, in an application of theory to practice, awareness would be needed as to how victims often do not have the time, information or ability to apply laws of logic to assess options. Instead, in many cases, some victims may use laws of probability and rely on heuristics (Klein G, 2008), or make quick decisions due to bias (Aschuler, 2007).

Heuristics and cognitive biases
Heuristics serve as general and automatic cognitive shortcuts for simplifying complex tasks (Schaeffer, 1989). They are particularly helpful in time pressured and uncertain settings where analytical strategies cannot be used. Tversky and Kahneman (1974) demonstrated that heuristics underlie a wide range of intuitive judgements. These strategies allow decision makers to typically come to an acceptable decision quickly, without generating excess cognitive load. However, sometimes these heuristics can lead to systematic deviations from logic, probability or rational choice. The resulting errors are called cognitive biases and many different types have been documented (Tversky and Kahneman, 1982). These have been shown to affect people's choices in situations like valuing a house, deciding the outcome of a legal case, or making an investment decision.
For instance, omission bias is the tendency to favour an act of omission (inaction) over one of commission (Anderson, 2003;Spranca et al., 1991). It can occur due to a number of processes, including psychological inertia (discussed later), the perception of transaction costs, and a tendency to judge harmful actions as worse than equally harmful inactions (van den Heuvel et al., 2012). This manifestation of cognitive bias in victim decision making could be illustrated by the differences in withdrawal types highlighted by Sleath and Smith (2017). Within their study, they found that an equal proportion of victims withdrew from the investigation through a formal retraction statement (n ¼ 113, 21.6%) and through informal disengagement (n ¼ 140, 26.7%). It is the latter type of engagement that could relate to the issues highlighted above, with some victims potentially disengaging from the police investigation due to omission bias or decision inertia. This would be in comparison to the victims who formally retracted through signing a statement, as this more likely demonstrated an active decision to withdraw from the CJS.

Rational choice and prospect theory
For victims who make active decisions concerning the investigation and prosecution of a suspect, prospect theory suggests that decision making is comprised of two phases: framing and valuation (Tversky and Kahneman, 1992). The theory is largely concerned with perceived gain and loss, which can apply to many situations including the risks involved in cooperating with the police. This decision making may also include a wide range of other people, such as the victim's interactions with the suspect, wider justice system and support services, each of which form a potential strategy in dealing with abuse. In each strategy, a victim would likely consider the available options and the potential gains and losses of each, resulting in the decision of an option to take forward.
Whilst heavily steeped in economics, Kahneman (2011) explains how prospect theory was born from limitations to utility theory, a concept first established by Daniel Bernoulli (Bernoulli, 1954). Utility theory explains how we place value on prospects based upon their utility value, as opposed to their actual monetary value. For example, a 10% rise in pay may have the same utility to everybody, however a raise of specifically £100 may not (Kahneman, 2011). Following the logic, Bernoulli (1954) concluded that when individuals make decisions, or gamble, it is the utility value that is assessed against the proportion of risk/uncertainty, as opposed to monetary value. This explanation helped underpin an understanding of a further concept called loss aversion, where individuals are more likely to settle for a lower amount of monetary value if the proportion of certainty is higher, because the certainty contributes towards greater utility.
In decision theory, loss aversion refers to the human tendency to prefer avoiding losses to acquiring equivalent gains: it feels better to not lose £10 than it feels to find £10 (Kahneman and Tversky, 1979). The key mechanism behind loss aversion is that people react differently to positive and negative changes. More specifically, losses are twice as powerful compared with equivalent gains (Tversky and Kahneman, 1992). This idea is one of the foundations of prospect theory, which describes how people choose between different prospective options and how they estimate the perceived likelihood of these different options. For instance, even though the likelihood of a costly event may be small, we would rather agree to a smaller and sure loss, such as in the form of making a monthly or annual insurance payment, than risk a large expense (Outreville, 1998). As a result, people weight potential costs and failures more heavily than potential benefits and rewards. Therefore, people may be unwilling to make decisions that represent loss (i.e. prosecute partner/ terminate a relationship), even though the decision itself may be the best option. Furthermore, people prefer a probable larger loss to a sure loss if these are the only available prospects (Jou et al., 1996).
Furthermore, the framing of gain and loss could be largely based on a victim's personal circumstances at the time of the abuse incident. Using the rationale from Kahneman (2011), if two people have different personal circumstances as their initial reference point, a proposition may have very different meanings for both individuals. This is a limitation to utility theory, in that it does not account for the change to an individual's circumstances, which Kahneman (2011) termed as Bernoulli's error. This paper argues that a similar error applies to many Western CJSs, because there is an expectation that a prosecution of the abuser carries the same utility to each victim of domestic abuse. However, this is unlikely since it does not take into account the victim's personalised circumstances when they consider their options (Cerulli et al., 2015;Hoyle, 1998), and may well be the reason for high levels of victim withdrawal (Dawson and Dinovitzer, 2001;McLeod, 1983;Robinson and Cook, 2006;Sleath and Smith, 2017).
Applying the theory to policing, if the investigation and prosecution of the suspect is framed as a loss in any way to the victim, then a victim may be more likely to choose risky prospects in order to minimise the more certain risk of loss. For example, Victim A lives separately to the suspect, whereas Victim B shares a home that is reliant on the income of the suspect. Both aim to become domestic abuse free and both are offered the option of either, Certainty to prosecute the suspect to stop the abuse [OR] equal chance to enter back into the relationship and the abuse restarts, or enter back into the relationship and the abuse stops.
Theoretically, both victims should be more likely to choose the certainty to prosecute the suspect in order to achieve the aim of becoming abuse free. However, this assumption is similar to Bernoulli's error as it does not account for changes in victim circumstance. As Victim A frames their prospects, they consider that a prosecution will result in the suspect being sent to prison and the abuse will stop, meaning this is a more certain gain than the alternative. However, Victim B has the certainty of losing their home if the suspect is sent to prison, meaning that it would be more rational for this victim to take the riskier option of returning to the relationship in order to minimise the certainty of loss. Whilst the example is overly simplistic, it does explain the issues faced by numerous victims of abuse and encapsulates the plethora of reasons why a victim may withdraw from the CJS.
Taking into account the valuation stage, these prospects may further change again depending on the severity of the abuse suffered. For example, if Victim B has suffered an extremely violent incident resulting in severe injury, they may be more likely to value the certain loss of their home as a smaller loss in comparison to a future abuse episode. This would align with the earlier example of insurance payments, where individuals would then logically choose the smaller certain loss to avoid the much larger probable loss (Outreville, 1998). Moreover, this valuation is likely to be subjective to each individual depending upon their experience with violence and/or previous victimisation. The complexity involved in victims' circumstances illustrates how victims often have to weigh up risks and choose between physical and psychological safety, against their financial and practical security (Carey and Soloman, 2014). Furthermore, this framing and valuation of their prospects also occurs after a particularly emotional and stressful incident, perhaps resulting in some victims avoiding decision making altogether.

Decision inertia
When faced with a decision situation that requires a choice to be made between competing negative options, decision makers can struggle to make that choice. This phenomenon is called decision inertia. In the context of critical incidents (political, security, military, law enforcement), Power and Alison (2017) found that rather than disengaging and avoiding difficult choices, decision makers are acutely aware of the negative consequences that might arise if they fail to decide (i.e., the incident would escalate). This can lead to intense deliberation over possible choices and their consequences and, ultimately, can result in a failure to take any action in time (or at all). This could be related to the personally critical decisions faced by victims of domestic abuse, and especially those who suffer repeat incidents. This is because when faced with a decision that involves competing options of potential negative outcomes, victims may exhibit decision inertia in the form of repetition of previous choices regardless of the outcome (Jung et al., 2019).
Early decision-making research proposed two hypotheses as possible explanations of inertia in repeated decision making: (a) expectancy consideration (i.e. the tendency for decision makers under risk to favour one option and attribute bad outcomes of decisions as bad luck resulting in repetition of negative outcome-generating judgements) and (b) consistency-seeking after commitment (i.e., commitment to a decision results in resistance to change in order to reduce feelings of cognitive dissonance) (Alison et al., 2015;Geller and Pitz, 1968;Grabitz, 1971). However, more recent research suggests that decision inertia is more complicated and driven by multiple cognitive processes, such as action orientation and decision autonomy, rather than a simple preference for consistency or indecisiveness (Jung et al., 2019). Other factors found to cause or increase decision inertia in critical situations are non-time-bounded choices and a lack of strategic direction (Alison et al., 2015), factors that could potentially be harnessed by police officers within the initial stages of an investigation to encourage active victim participation.

Use of victim decision-making theory in police practice
The aim of the current paper was to establish how decision-making theory could be applied to victims of domestic abuse. The rationale for the theoretical examination is that, by taking account of victim decisions, officers may have the ability to proactively address issues with victim cooperation. The main suggestion of this paper is that officers could use concepts of victim decision making as a means of enacting victim empowerment within police investigations of domestic abuse (Hoyle and Sanders, 2000). This would aim to address the criticisms to both the victim choice and pro-arrest approach, by providing a response to abuse that aims to understand the value of the CJS to each victim before then enacting a response to their needs (Hoyle and Sanders, 2000: 16-19).
As such, the theoretical template could be used to encourage officer-victim dialogue into the difficulties faced by the victim when they are required to engage with the police. This template would focus on ensuring empathy, patience and an understanding of why victims may not want to press charges, or change their minds as they reframe their prospects throughout the CJS. This perspective could be built into a training package to work alongside existing risk assessments that aim to identify coercive control and reabuse, with an overall aim of identifying the range of victim needs at the initial scene. Furthermore, the use of the theoretical template may also allow officers to make early determinations of victim cooperation or withdrawal, which could be recorded at the initial stages of the police investigation. This assessment would be vital in building a picture of risk to the victim, whereby the identification of non-cooperative victims could trigger a response of enhanced communication to explore the reasoning behind their withdrawal. From this point, partnership working would be critical to tackle the identified practical issues faced by these victims in order to remove the prospect of loss and introduce the prospect of gain.

Encouraging gain and avoiding loss
In theory, victims will choose more certain gains over risky gains (Kahneman, 2011). In instances where the victim considers prosecuting the suspect for retributive justice (which is considered the gain), then effort must be placed into ensuring certainty of the prosecution. This aligns with the current pro-prosecutorial system present in many Western justice systems, where effort is focused on mandatory arrest and no-drop prosecutions against a suspect (Han, 2003). In these cases, victims should be made to feel as secure as possible in the prospect of a prosecution, as they become risk averse to avoid any negative feelings involved in loss (Kahneman and Tversky, 1979).
A practical difficulty, however, would occur when a victim wants the violence to stop, 'but not at all costs' (Hoyle and Sanders, 2000: 21) or without the need for retribution (Hare, 2006). Throughout the CJS these victims will likely reframe their prospects, which may involve circumstances where the victim has already managed to become abuse free before the prosecution of the suspect (i.e., cessation of stalking/harassment, or the relationship ends). When reframing, a victim may be presented with a decision to either settle with the current cessation of abuse, or continue with the prosecution of the suspect, which may or may not result in retaliatory abuse. In this instance, victims may not consider the prosecution to be worth the effort (Hoyle, 1998), or place more value on the suspect leaving them alone (Hare, 2006) and therefore be more likely to choose to withdraw from the CJS.
Yet, this outcome is often considered a failure within Western justice systems, as there is an expectation that the suspect should be punished for their behaviour because 'criminalisation serves a symbolic purpose [of] indicating the moral unacceptability of domestic violence' (Hoyle and Sanders, 2000: 14). Yet the punishment and incarceration of the suspect may often be considered a tangible loss to victims of abuse, especially if the victim only desired support with the immediate abuse incident (Hirschel and Hutchinson, 2003). Subsequently, these victims may be more likely to become risk seeking to avoid this certain loss if they were to cooperate with the investigation and prosecution. This means that if a riskier approach was to arise that also resulted in the abuse ceasing, then the victim may be likely to choose this gamble because of loss aversion. Therefore, logically speaking, if a suspect promised to never abuse the victim again and wanted to rekindle the relationship, then this prospect is a realistic option for the victim since it avoids the certain loss of the relationship or something that the suspect may provide. Furthermore, for the period of time where the victim does not experience further abuse (or perceived further abuse), their decision has ultimately provided an outcome that has resulted in less loss than if they were to cooperate with the police investigation and prosecution, reinforcing their confidence in that decision to withdraw.
The above dynamics may well be one of the reasons why police officers often express frustration and powerlessness in the face of wider criminal justice mechanisms when responding to domestic abuse (Horwitz et al., 2011). As frontline responders, it is likely that some officers feel constrained by the criminal justice options available and how they may not allow for the level of discretion needed for them to avoid the large proportion of cases that result in victim withdrawal (Dawson and Dinovitzer, 2001;McLeod, 1983;Robinson and Cook, 2006;Sleath and Smith, 2017). This has been identified and termed by Stark (2016: 347) as the 'negative feedback loop', where officer demand is largely made up of continued calls from the same population of victims who have not received useful CJS outcomes. A potential remedy would be to create a disposal to cases of abuse that evidences victim empowerment. The potential disposal, or range of disposals, would need much further research and consideration. For example, they may only be considered available for low risk cases, whereby officers and/or support professionals have liaised with a victim to provide an outcome that best suits their needs. This approach may also develop trust in law enforcement, which would be especially critical if a victim is to suffer future incidents of abuse. This approach could allow officers greater discretion when handling cases, especially as some officers believe that the police response to all cases is too aggressive (Myhill, 2019). Yet this approach would need very strict training, monitoring and auditing to ensure that the disposals are genuinely being used for victim empowerment as opposed to achieving a quick disposal of the officers' workload.

Understanding victim inertia
Further to the concept of gain and loss, the use of a theoretical template of victim decision making by police officers could also involve officers identifying when victims are making active decisions as opposed to being passive entities in the police intervention. For example, if an incident is brought to the attention of the police via a third-party report, some officers may already take extra care in such cases as they recognise the victim has not made an active decision to involve the police in the abuse incident (Birdsall, 2018). Subsequently, these cases may have a higher likelihood of victim withdrawal through informal disengagement (Sleath and Smith, 2017) as the victim avoids or delays making decisions about further police intervention.
Taking into account the more recent research that suggests decision inertia is driven by multiple cognitive processes, such as action orientation and decision autonomy (Jung et al., 2019), the police could also harness certain techniques to encourage active victim participation in the decision-making process of the investigation. This could involve forming decisions at certain criminal justice checkpoints (McLeod, 1983), where a strategy is formed in partnership with the victim (Alison et al., 2015). It is important to note that this strategy could involve the enacting of a victimless prosecution, whereby the victim fully supports the prosecution of the suspect, but just does not want to be practically involved in the process (Ellison, 2002). Furthermore, officers and criminal justice agents could also set avoid goals if the victim's situation is particularly complex, and perhaps allow flexible goal setting with the victim in the initial stages of the police investigation (Power and Alison, 2017).

Limitations to the theory
The foundations of the decisional theories previously discussed differ greatly from each other. Some are based on experimental results whilst others are rooted in naturalistic approaches. Furthermore, the contexts in which these decisional theories have been applied and/or tested not only differ from each other, but differ from the context that this current paper is targeting. On the whole, the majority of the decision theories discussed target practitioner-level decision makers (i.e. police, firefighters, medical professionals, etc.). This paper theorises the application of these perspectives to a decision maker, which has typically been absent in decision research -the victims of domestic abuse. Therefore, caution must be taken when considering this application.
It has been shown previously that new innovations that seek to reduce crime, improve police practice or increase victim and/or public satisfaction can have both intended and unintended consequences (i.e. increased crime and reoffending, displacement of offending or decreased victim satisfaction) (Braga, 2006;Kirby, 2013;Lum et al., 2017). For instance, whilst there is a strong body of literature that supports that hot-spot policing can significantly reduce crime and increase the effectiveness of policing (Braga et al., 2014;Skogan and Frydl, 2004), there is a growing concern that this approach has unintended consequences in terms of increasing police abuse of power and decreasing public assessment of police legitimacy (Kochel, 2011;Kochel and Weisburd, 2017;Rosenbaum, 2006). In order to avoid these unintended consequences, the development of new policies and practices -or the remodelling of existing ones -must be evidence-based and evaluated using high quality research (Byrne and Marx, 2011;Weisburd et al., 2010). This paper does not seek to make conclusions regarding the underlying decision-making process of victims, nor does it seek to make recommendations for changes to police practice in response to domestic abuse incidents based on this. Instead, this paper is a conceptual call for research into victim decision making. We believe that victim decision making research needs to be placed on an empirical footing in order to evidence whether it can improve the CJS response to victims and/ or should be utilised to change police practice.

Conclusion
Applying the logic of prospect theory to victim decision making, in order to increase victim cooperation with the police and CJS there needs to be clear and certain gains for each victim. Officer dialogue and actions should be geared towards making prosecutions a valid and useful option for victims, as opposed to repeating Bernoulli's error of presuming that prosecutions carry the same utility to each victim without taking cognisance of their personal circumstances (Cerulli et al., 2015;Kahneman, 2011). This is especially important considering the growing body of research illustrating how arrests and prosecutions of the suspect in every case may not always be in the best interests of the victim (Hoyle and Sanders, 2000).
The introduction of new disposals to cases that evidence victim empowerment could alleviate the negative feedback loop (Stark, 2016) and officer frustrations of otherwise having to apply a pro-prosecutorial approach to all victims of abuse (Horwitz et al., 2011). The introduction of a victim empowerment disposal could also be examined directly to determine whether it results in greater levels of victim satisfaction and an increased likelihood the victim would use the police again to handle the abuse, in comparison to a group of victims that either had their cases dropped, or faced a prosecution without their support. Training for officers to spot when victims are making active decisions as opposed to being passive entities in the police intervention could allow for more action to prevent decision inertia. This would be as important as encouraging a system of gains, since Sleath and Smith (2017) found similar levels of victim withdrawal due to informal disengagement alongside the formal retraction process.
However, this perspective would need to be rigorously tested and placed on an empirical footing. This is especially due to the arguments around how the approach does not have external validity within the fields of psychology and sociology. If found to be scientifically viable in understanding victim engagement and decision making, then the theory should be applied through the lens of a victim empowerment approach. Doing so could overcome the difficulties of both a victim choice perspective and a pro-arrest perspective to ultimately formulate CJS responses that address the needs of all victims (Hoyle and Sanders, 2000).