A critical analysis of stalking theory and implications for research and practice

Abstract

This article comprehensively reviews and critiques theories providing an aetiological account of stalking. We evaluate applications of preexisting psychological theories to stalking (attachment theory, evolutionary theory, social learning theory, information processing models of aggression, coercive control theory, and behavioural theory) as well as the only novel theory of stalking to date: Relational goal pursuit theory. Our aim was to identify which are supported by research, identify gaps in theoretical scope and explanatory depth and examine how current theories might inform clinical practice. This evaluation suggests that theories of stalking are underdeveloped relative to other areas of forensic clinical psychology and the theoretical literature is relatively stagnant. Consequently, there is limited research into clinically meaningful constructs that can guide the assessment, formulation and treatment of this client group. We identify similarities across existing theories, discussing implications for future research and clinical practice with people who stalk.

1. INTRODUCTION

Stalking is broadly defined as a pattern of repeated, unwanted intrusive behaviours imposed by one person upon another, which evokes fear or significant distress in the target(s) (McEwan, et al., 2020; Purcell, et al., 2004; Spitzberg & Cupach, 2000). Knowledge of people who stalk, their victims and their behaviour has grown substantially over the past three decades, but the stalking research literature remains largely atheoretical. Most studies have focussed on describing the phenomenon, but there has been less research into the psychological and social mechanisms that produce and maintain this common behaviour (with some notable exceptions; Brownhalls et al., 2019; Johnson & Thompson, 2016; Spitzberg & Cupach, 2014). In contrast to other forensically‐relevant behaviours such as interpersonal violence or harmful sexual behaviour, there are relatively few explanatory theories of stalking, and none have been subject to comprehensive and sustained evaluative research (Spitzberg & Cupach, 2007, 2014). This is a significant gap in the research literature as, without an appropriately tested theory, treatment of stalking behaviour is reliant on expert opinion and guesswork (Purcell & McEwan, 2018).

From a scientist–practitioner perspective, treatments should reflect explanatory theory that has been subject to evaluation and to outcomes research using appropriately designed studies. The stalking literature has not achieved even the first part of these aims. There is a clear need to address the fundamental question of what leads some people to stalk while the vast majority of others do not. Only with an understanding of the psychological, psychopathological, situational and social factors that contribute to stalking can we hope to develop effective intervention strategies that can be subject to evaluative research. This article takes a first step in this direction by providing a critical analysis of existing stalking theories according to how well they (a) explain what is already known about stalking, (b) are consistent other accepted theories of human behaviour, and (c) provide adequate guidance for future research and clinical practice (Gannon & Pina, 2010; Ward et al., 2006). We aim to identify areas of theoretical consistency that may present opportunities for theory knitting (Kalmar & Sternberg, 1988) and draw attention to where research could focus to advance the stalking literature ways that can inform practice innovation.

1.1. Types of psychological theories

Effective assessment and treatment of any problem behaviour must be underpinned by a solid understanding of relevant aetiological theory. Theoretical frameworks enable clinicians to understand an individual's presentation through the lens of what is known about the psychological variables that contribute to the behaviour (Ward et al., 2006). This, in turn, creates a base from which to build broad treatment approaches for the problem behaviour and to inform treatment plans for specific individuals (Gannon, Collie, et al., 2008). At present, aetiological theories of stalking remain underdeveloped relative to theories of other problem behaviours such as aggression, fire setting, and harmful sexual behaviour (Anderson & Bushman, 2002; Finkel, 2014; Gannon et al., 2011; Ward et al., 2006).

Ward et al. (2006) describe four different approaches to understanding and explaining human behaviour in a forensic context: typologies, single‐factor theories, multi‐factor theories, and micro‐theories. Typologies are often developed to sort and classify a diverse group in a way that can usefully inform assessment, treatment, and management (Gannon & Pina, 2010). Given the complexity and heterogeneity of stalking, typologies of the behaviour have flourished (e.g. Spitzberg, 2007, lists 24 published typologies of stalking, and no doubt more have been proposed in the years since). The most useful typologies combine the nature of the prior relationship between victim and stalker, the presence and nature of mental illness, and the apparent motivation or function of the stalking to group stalking cases into types that have meaningful implications for subsequent action (McEwan & Davis, 2020). Such typologies provide a valuable function in applied settings, helping to make sense of the mass of offence information and suggesting hypotheses about motivation or the function of behaviour (McEwan & Davis, 2020). However, typologies are inherently limited in their ability to guide more nuanced intervention as they have limited explanatory function and many are purely descriptive. Therefore, although they can be a useful first step in helping to guide hypotheses when assessing an individual, typologies must give way to theory to progress research and practice in a more sophisticated way.

Where typologies are primarily descriptive, single‐factor theories rely on a single underlying characteristic or mechanism (e.g. social learning, evolutionary theory) to explain the relevant phenomenon, in this case, stalking (Gannon & Pina, 2010). Single‐factor theories often rely on an existing theoretical basis (e.g., attachment theory) and then apply and adapt this theory to a unique context (e.g., stalking behaviour). In contrast, multi‐factor theories draw together different constructs or theories to explain how, when combined, the problem behaviour results (Gannon & Pina, 2010; Ward et al., 2006). Multi‐factor theories potentially offer great utility in explaining complex behaviour as they can acknowledge different components that may contribute to the phenomenon. For example, a multi‐factor theory may encompass preexisting psychological vulnerabilities (i.e. insecure attachment and stalking‐supportive cognition), social contexts (i.e. a relationship breakdown) and exacerbating factors (i.e., depressed mood and substance use). Both single‐ and multi‐factor theories have been proposed to explain stalking.

In the subsequent sections of this article, each published psychological theory of stalking is appraised using epistemological criteria originally developed by Hooker (1987) and elaborated by Newton‐Smith (2002) and Ward et al. (2006) to provide a method of consistently and rigorously assessing competing scientific theories. This approach to theory appraisal rests on a scientific realist perspective. Theoretical claims that are well developed and predictively successful are taken to represent real knowledge about the world and its workings (Ward et al., 2006). These appraisal criteria allow us to evaluate not only the empirical adequacy of a theory (how well it explains and predicts) but also whether the theorised processes and entities have sufficient explanatory depth, are consistent with other accepted theories and produce new paths for elaboration and improvement.

Versions of the criteria used in this article have previously been applied by Ward et al. (2006) to evaluate theories of sexual offending, by Gannon, Rose, and Ward (2008) to evaluate rape theories and by Gannon and Pina (2010) to evaluate theories of deliberate fire setting. The five appraisal criteria used in this article are as follows:

1. Empirical adequacy: is the theory consistent with existing research findings regarding the phenomenon?

2. External consistency: is the theory coherent with other background theories that are currently accepted?

3. Unifying power: does the theory draw together previously isolated research findings or theories?

4. Fertility: does the theory provide ideas for clinical practice or research?

5. Explanatory scope and depth: does the theory explain the full scope of the observed phenomenon and propose detailed ways of explaining it?

(Gannon & Pina, 2010).

2. SINGLE‐FACTOR THEORIES OF STALKING

2.1. Attachment theory adaptations

Explaining stalking as an expression of an insecure attachment style was first proposed by Reid Meloy in the late 1980s, and he has elaborated on this explanation in intervening years (Meloy, 1989, 1998, 2007). Attachment theory was originally described by John Bowlby and Mary Ainsworth through the middle decades of the 20^th century, explaining people's interactional dispositions as adults through the prism of their attachment experiences with their primary carer as an infant (Bretherton, 1992). Attachment theory has undergone multiple revisions and elaborations (Bartholomew & Horowitz, 1991; Hazan & Shaver, 1987; Mikulincer & Shaver, 2007), but the key proposition remains, that is, those who report insecure attachment struggle to trust others and may experience problems in adult relationships. In contrast, those who can form secure attachments are more easily able to trust others, to have satisfying interpersonal relationships and are less likely to encounter relationship difficulties (Bartholomew & Allison, 2006). There is clear evidence of external consistency in adapting attachment theory to explain stalking, given the wealth of research devoted to the topic and the evidence of it being a useful way of understanding adult interpersonal styles and relationship behaviours.

Meloy (1992, 1996) suggested that insecure attachment could lead to stalking via its contribution to the development of borderline and narcissistic personality characteristics, which have been commonly observed among those whose stalking attracts criminal justice attention (McEwan & Strand, 2013; Nijdam‐Jones et al., 2018). These personality patterns share the common feature that maintaining a positive self‐concept is reliant on the positive attention of others (i.e., their sense of self‐worth is externally contingent; Senkans et al., 2020). Relational threats lead to feelings of vulnerability and intense negative arousal to which insecurely attached individuals respond with maladaptive coping strategies (Dykas & Cassidy, 2011; Mikulincer & Shaver, 2019). This can lead to misguided and damaging attempts to maintain the prized relationship and/or to punish the individual who is perceived as a source of threat, leading Meloy (2007) to describe stalking as “a behavioural expression of attachment pathology” (p. 2).

Meloy's (1992, 1996) theorising about the role of insecure attachment in stalking had clear fertility, laying the groundwork for numerous subsequent studies investigating stalking and attachment (see Meloy, 2007). His original work also influenced Davis et al. (2012), who drew more broadly on the modern adult attachment literature to suggest that insecure attachment could underpin stalking in the same way that it has been hypothesised to produce intimate partner violence. In this model, the anxiously attached person's need for reassurance and acceptance makes them prone to coercively controlling behaviours when threatened by the loss of a partner (discussed further below). Davis and colleagues explicitly link anxious attachment to failures in self‐regulation of emotional distress in relational situations, providing this application of adult attachment theory to stalking with greater unifying power and external consistency, as it also draws on the wider empirical literature regarding self‐regulation.

There is a sizeable body of evidence demonstrating that many, though not all, people who stalk report insecure attachment. Investigations in clinical and forensic populations have shown insecure attachment to be more commonly self‐reported by stalkers than by offender and community controls (MacKenzie et al., 2008; Tonin, 2004). MacKenzie et al. (2008) specifically identified significantly higher rates of preoccupied attachment (as hypothesised by Meloy, 2007), although Tonin (2004) did not, and a recent Italian study suggested that most of their small offender sample endorsed a dismissing attachment style (Civilotti et al., 2020). Disrupted relationships with early caregivers have been identified in both clinical/forensic and college‐based samples and interpreted as evidence of disturbed attachment (Kienlen et al., 1997; Langhinrichsen‐Rohling & Rohling, 2000). Investigations of college students engaging in unwanted pursuit have also found higher rates of self‐reported insecure attachment, specifically anxious attachment (Dutton et al., 2006; Dye & Davis, 2003; Patton et al., 2010). There is also evidence from college students indicating that the relationship between anxious attachment and stalking is moderated by emotional distress, particularly the experience of anger and jealousy (Davis et al., 2000; Dye & Davis, 2003), which is highly consistent with Civilotti and colleagues' qualitative findings in a forensic sample. Evidence for this theory can also be found in adult samples recruited in community settings. De Smet et al. (2012) demonstrated that insecure attachment, anxiety and negative affect was associated with unwanted contact with a former partner in a sample of 3924 divorced adults recruited in Belgium. A subsequent study by the same authors (De Smet et al., 2013) showed that anxious attachment was a robust predictor of unwanted pursuit behaviour in male adults following divorce, interacting with dyadic and situational factors in multivariate modelling. In an Australian sample of 637 self‐reported ex‐intimate stalkers, Johnson and Thompson (2016) found that increased stalking persistence was associated with insecure attachment. This body of research clearly demonstrates that attachment‐based explanations for stalking behaviour have some empirical adequacy.

The explanatory scope and depth of attachment‐based accounts of stalking might be improved by considering the vast and rich literature on adult attachment, rather than the specific focus on romantic attachment that has been taken to date. Adult attachment theory posits that maladaptive expectations about the self, others and the self in relation to others affect all aspects of social functioning (Mikulincer & Shaver, 2019). This theory would seem to have relevance to stalking behaviour outside of romantic pursuit contexts, but this has not been examined at all in stalking research. Modern adult attachment theories, such as that outlined by Mikulincer and Shaver (2007), also offer more detailed accounts of attachment‐related behaviour with greater explanatory depth and potential for operationalisation in research than that those that have currently been applied to stalking. For example, Mikulincer and Shaver overtly adopt some of the language and techniques of modern social cognitive psychology in explaining the activation and functioning of their theorised attachment system. Drawing on the wealth of adult attachment research could also assist in overcoming some of the other shortfalls of how attachment theory has been adapted to explain stalking. With the exception of De Smet et al.'s (2013) work, current applications of attachment theory to stalking attribute the behaviour almost entirely to the psychological characteristics of the person stalking, neglecting most of the situational, dyadic and cultural factors that undoubtedly also contribute to whether and when stalking occurs (Mullen et al., 2001).

While application of attachment theory to stalking has considerable explanatory potential, there are also potential gaps or flaws in this approach. Questions remain about whether an adult's general attachment style is subject to accessibility effects, leading them to respond in ways that reflect their current relationship experiences rather than an overall ‘style’ of attachment (Holmes, 2000). This poses problems for assessing the attachment styles of stalkers who are often experiencing relationship problems at the time they come to attention or are primed to focus on relationship problems when asked about stalking. Related to this is the broader issue of how to assess adult attachment style. Measures differ in their emphasis on ‘state’ versus ‘trait’ elements of attachment, and even in whether attachment is viewed as a specific biological and psychological system (cf. Mikulincer & Shaver, 2007) or a cognitive style characterised by particular beliefs and attitudes that are reflected in behaviour (cf. Bartholomew & Horowitz, 1991). Already the stalking and attachment literature is divided between these different approaches, using different measures and constructs to link insecure attachment to stalking behaviour. A further challenge is that some theorists believe that attachment style can differ in different relationships (Baldwin & Sinclair, 1996; Ravitz et al., 2010) while those taking a systems approach suggest it is more consistent across relationships (Mikulincer & Shaver, 2019). If the latter approach is taken (which the stalking literature has to date), there is a significant challenge using insecure attachment as a singular explanation for stalking, as on it cannot address why someone would stalk in one relationship context but not in another, or why they would stop stalking if the insecure attachment remains. Moreover, if insecure attachment is the sole feature used to explain and predict stalking behaviour, stalking would have to exist wherever and whenever disordered attachment exists, which is clearly not the case (MacKenzie et al., 2008; Tonin, 2004). This critique is not to suggest that attachment theories cannot be usefully applied to stalking behaviour; indeed modern attachment theories would provide a fertile framework for hypotheses about the role of self and relationship cognition, and associated affect, in stalking behaviour. However, for this to occur, a more nuanced and detailed application of modern adult attachment theory to stalking behaviour is required that also takes into account other contributing factors.

2.2. Social learning theory

Fox et al. (2011) described the application of Social Learning Theory (SLT) to stalking behaviour. SLT explains crime as a function of social factors that influence what a person learns (Bandura, 1971). Fox and her colleagues proposed that individuals who stalk would know others who had engaged in similar behaviour (differential association), imitate the behaviour of others (modelling), hold attitudes favourable to stalking, and balance the risks and rewards associated with stalking in a pro‐stalking manner (differential reinforcement). This application clearly has strong external consistency, drawing on not only well‐established tenets of wider social learning theory but also being consistent with findings from the psychology of crime literature showing that antisocial peers and antisocial attitudes are two prominent factors linked to broader offending behaviour (Bonta & Andrews, 2016). The unifying power of the theory is also good, with social learning theory positioned by Fox and colleagues as complementary to attachment‐oriented explanations of stalking, with a greater focus on proximal social factors that may contribute to stalking in particular contexts.

Social learning theory's focus on proximal factors provides this model with explanatory depth: concepts such as differential peer association and attitudes supportive of stalking can clearly be operationalised and tested. In addition to proposing the theory, Fox et al. (2011) undertook the only test of this proposal to date, using a large sample of American college students. They found that the hypothesised social learning factors were present and associated with self‐reported stalking perpetration, meaning that this is one of a handful of stalking theories with some direct evidence of empirical adequacy. Other studies demonstrating the prevalence of antisocial personality characteristics among individuals who stalk (McEwan & Strand, 2013; Nijdam‐Jones et al., 2018; Ogloff et al., 2021) provide indirect evidence consistent with the idea that people who stalk might endorse attitudes supportive of the behaviour. The empirical scope of social learning theory is adequate, with the emphasis on stalking as a crime of power and control potentially applying to non‐intimate contexts and stalking by people of different genders, as acknowledged by Fox and colleagues. While individual characteristics such as psychopathology or self‐regulation deficits were not considered by Fox and colleagues, such characteristics could be quite easily integrated in future investigations to broaden the empirical adequacy of SLT when applied to stalking. SLT is clearly a fertile theory, both in creating opportunities for new research and also suggesting directions for clinical interventions focussing on peer relationships and beliefs and attitudes. Unfortunately, in the decade since Fox and colleagues' study was published, such research has not been pursued.

2.3. Evolutionary theory

Evolutionary theory suggests that human behaviour adapts in response to evolutionary pressures that make particular behaviour either necessary or desirable in order for the individual to survive and thrive (Buss, 2005). Duntley and Buss' (2012) application of evolutionary theory to stalking hypothesises that it may be one of a number of strategies shaped by evolutionary processes to help solve problems of mating and within‐gender competition. They propose that some individuals (particularly men) developed stalking strategies to overcome adaptive problems like attracting, retaining or regaining a mate or fending off competitors. Duntley and Buss argue that the physical and psychological costs of stalking on victims would result in adaptations to defend against stalking. Over generations, successful stalkers would be defined by their adaptation of new strategies to overcome the victim's defences and so successfully mate. This theory views stalking as a functional adaptation that, while often ineffective, produces enough of a reproductive advantage to be selected over generations and evolutionarily shape a “psychology of stalking” (p. 312).

Duntley and Buss (2012) cite findings from the broader stalking literature related to observed gender disparities and apparent stalker motivations to support their theory (i.e. indirect evidence of empirical adequacy). This theory has rarely been used to explain others' findings, though De Smet et al. (2013) suggested that evolutionary pressures may explain gender differences observed in their study, which suggested that males were less likely than females to pursue former partners when alternative romantic options were available. Evolutionary theory also has some unifying power, explicitly building upon and integrating ideas from extant stalking taxonomies and models and having potential to integrate with the more proximal explanatory hypotheses proposed in social learning theory. As evidenced by the array of research hypotheses outlined in Duntley and Buss' article, it clearly has fertility. The empirical scope of the evolutionary theory is also good; it accounts for gender differences in stalking behaviour by proposing that, while males and females face similar mating pressures and within‐gender competition, the function and effectiveness of stalking would be greater for males over time due to their physical size and ability to use sexual coercion. They acknowledge different functions or motivations for stalking, including the presence of non‐relationship‐based stalking as a way of solving social adaptation problems. The proffered explanation is limited but shows awareness of the range of stalking phenomena that have been observed outside of relationship pursuit contexts.

Like any evolutionary model of behaviour, Duntley and Buss (2012) minimise the role of social learning and cultural influences on development. Additionally, the theory focuses almost exclusively on a heteronormative dynamic with men as the perpetrators. This causes problems for the theory's empirical adequacy, as it fails to acknowledge research demonstrating that women perpetrate approximately 15% of stalking (Dreßing et al., 2020) and a similar proportion of cases involve victims and stalkers of the same sex (Strand & McEwan, 2011). The large body of research supporting the notion that various antisocial behaviours have a learnt element is not discussed, limiting the external consistency of a wholly evolutionary model. In some ways, the theory has substantial explanatory depth, proposing a series of specific distal causal mechanisms that produce stalking in a population, although its explanatory depth when applied to a specific individual is weaker. While Duntley and Buss identify individual stalking adaptations as ‘cognitive biases’, ‘cognitive adaptations’ and ‘false beliefs’, they do not elaborate in any detail on what form these might take or the processes by which they might produce stalking compared to other behaviours. As with the ‘attitudes supportive of stalking’ proposed in social learning theory, it is possible that this aspect of evolutionary theory could be expanded, increasing both its depth and fertility.

3. MULTI‐FACTOR THEORIES OF STALKING

Multi‐factor theories may have greater potential than single‐factor theories to explain a complex phenomenon such as stalking. Some attempts at multi‐factor theories of stalking rely heavily on the adaptation of theories from other contexts (i.e. violent offending), whereas other contemporary proposals have unified previously isolated explanatory factors (i.e. deficits in self‐regulation and increased rumination) to develop unique theoretical approaches.

3.1. Behavioural theory

The tenets of behavioural theory were applied to stalking by Westrup and Fremouw (Westrup, 1998, 1998) in their description of the use of functional analysis to understand the precipitants and maintenance of stalking behaviour. They proposed that stalking behaviours, like any other behaviour viewed from a functional standpoint, repeat and escalate due to the interaction of antecedents (preceding circumstances and events) and consequences (reinforcing events that, if rewarding, lead to the behaviour being repeated when antecedents are again present). Functional analysis has been applied to other types of problem behaviour such as fire setting (Jackson et al., 1987) and inpatient aggression (Daffern et al., 2007; Daffern & Howells, 2009), and functional analysis is based on more than half a century of behavioural theory and research that is integral to the wider clinical and social psychology literature (i.e. providing the approach with strong external consistency).

Westrup and Fremouw's (1998) approach has good explanatory scope and depth, proposing that individual acts of stalking behaviour have particular functions—to either escape or avoid something or to gain something. In a specific situation, engaging in stalking may make the stalker feel less lonely or uncomfortable (negatively reinforcing the behaviour), or it may provide feelings of power or control or gain attention (positively reinforcing the behaviour). They suggest that stalking behaviours may be multiply maintained by having both a negatively and positively reinforcing function. They also note that quite different behaviours (e.g. a telephone call and property damage) might have quite a similar function—to get the victim's attention and express anger. The empirical adequacy of the theory has not yet been directly tested and the focus of most of the stalking literature on broad description of the phenomenon does not translate easily to this very detailed application of behavioural theory. Westrup and Fremouw's approach clearly has excellent clinical fertility as it provides a means of identifying individual factors that can be modified to alter or stop the behaviour. Through research applying functional analysis to stalking, the approach's theoretical fertility could be improved by identifying patterns in antecedents and consequences that can inform further theory development.

Westrup and Fremouw (1998) have little to say regarding the types of antecedents and reinforcers that might be present in different cases, or the interactive mechanisms that produce stalking in some individuals and contexts but not others. Nonetheless, this approach has a strong capacity for unifying other disparate theories if used as part of a wider multi‐dimensional approach. It is quite possible that behavioural theory could be used to combine attachment‐based explanations for stalking with other theories that consider social or psychological determinants (e.g. social learning theory and evolutionary theory). This approach explicitly directs the user to examine both proximal and distal factors related to the stalking behaviour and could be applied to ex‐intimate and other relationship contexts in a common manner, although potentially incorporating different types of antecedents as different pathways into the behaviour. The detailed nature of functional analysis, and the ease with which it can be integrated with other theories, means that it could provide a framework for unifying existing theories that has considerable clinical utility.

3.2. Integrative contextual developmental (ICD) model of male stalking

The ICD model is not a theory per se, but a framework that combines existing theories of stalking and research findings into a unified whole. White et al. (2000) proposed a hierarchical model that attempts to explain the interaction of stalking‐relevant phenomena occurring at sociocultural, interpersonal, dyadic, situational and intrapersonal levels. The model is underpinned by a feminist social learning viewpoint and is premised on the assumption that male dominance colours the cultural and social norms that children learn from their family, peer and intimate networks, and is reflected in relationships and in individual personality and behaviour.

The approach and individual components of the ICD model have several strengths. It is truly integrative, linking specific single‐ or multi‐factor theories to different levels in the hierarchy, providing the model with strong unifying power. At the sociocultural level, the authors identify traditional gender roles, cultural scripts for sexual relations and dating and normative beliefs about sex and violence as potentially relevant. A large body of sociological and psychological research supports the existence of these constructs in English‐speaking industrialised societies, providing external consistency (Berkowitz et al., 2022). The IDC model suggests that these cultural norms are expressed in an individual's behaviour only if the social networks to which the individual is exposed teach and reinforce cultural norms supportive of male dominance (consistent with social learning theory). White and colleagues also cite early development of insecure attachment due to disturbed family environments as further evidence of how social networks may impact future behaviour, and Fox et al.'s (2011) later work showing differential peer association among those who stalk is consistent with this hypothesis.

At the dyadic level, the authors turn to Spitzberg and Cupach's (2000, 2014) Relational Goal Pursuit (RGP) theory to explain why stalking by men often occurs during attempts to form or at the dissolution of intimate relationships (described further below). They also make reference to the wider interpersonal relationship literature, which has provided evidence that the perpetrator and target's behaviours and emotional responses are interdependent in situations involving relationship rejection and pursuit (a contention subsequently supported by De Smet and colleagues' 2012 and 2013 findings).

At the situational level, White and colleagues hypothesise that specific situational factors are relevant to when and how particular stalking behaviours might occur, although they do not elaborate on the nature of such factors. There is little investigation of situational factors in the stalking literature, although the importance of context is a central premise of the RGP theory and functional analytic approaches, and is implicit in attachment theory. Finally, at the intrapersonal level, the authors turn to the psychological and psychiatric literature to incorporate evidence of attitudinal, motivational and characterological features of the perpetrator, primarily highlighting Meloy's (1998) work on insecure attachment, discussing the potential role of attitudes supportive of violence and making reference to the range of evidence linking stalking with particular types of psychopathology.

The empirical adequacy of the ICD model is variable, reflecting variability in the level of support for the theories that White and colleagues are knitting together in the model. The model's strengths are its unifying power and elements of its empirical scope. The definition used by the authors restricts stalking to the context of an ‘undesired relationship’ but otherwise takes a broad view, applying across both stalking and other, lesser forms of unwanted relationship pursuit behaviour. This is a particular strength because the authors assume a continuum of unwanted intrusive behaviour and develop a theory that can account for variations along the continuum. It also gives equal attention to a wide range of sociocultural and individual factors and explicates the links between the two. While the authors acknowledge the role of proximal factors such as intoxication and emotional arousal, the model places greater focus on distal factors, making explanatory depth and clinical fertility relative weaknesses. This could be overcome by proposing psychological mechanisms that might link the various levels together within an individual to produce stalking behaviour. The major criticism of the ICD model is its narrow scope, focussing exclusively on relationship pursuit and gender. While this is certainly necessary given the gendered nature of stalking and the frequency with which it is perpetrated by ex‐partners, the limited scope restricts the applicability of the model to stalking that emerges in other contexts.

White et al. (2000) make a clear argument for the need to further develop a multi‐factor theory that can accurately and adequately describe the characteristics, situations, and processes that contribute to stalking. They position the ICD model of male stalking as a base from which to generate and test further hypotheses and theories (implying strong research fertility), and the model has been cited in some subsequent studies of stalking behaviour, often those examining gender‐related characteristics (e.g., Logan & Walker, 2009; Yanowitz, 2006). The relatively limited impact of the ICD model relative to some other stalking theories may be because it so overtly draws on extant theories (which are cited in favour) rather than proposing novel hypotheses of its own.

3.3. Coercive control theory

The concept of coercive control developed in the intimate partner abuse literature over the latter half of the 20^th century but has attracted considerable research and policy attention in the years since Stark's (2007) reconceptualization. Davis et al. (2012) drew on Stark's work and the earlier work of Dutton and Goodman (2005) to suggest that highly coercive and controlling behaviours observed in intact relationships characterised by intimate partner violence “parallel many of the behaviours included in legal definitions of stalking and harassment” (p. 330). From this starting point, the authors went on to theorise that stalking behaviour might be usefully understood as a form of coercive control occuring outside of an intimate relationship.

Coercive control theory as presented by Stark (2007) is a sociological rather than psychological theory and so is not reviewed further in this article, given our aims. Dutton and Goodman (2005) provided a detailed psychological model of how coercive control dynamics might develop in an intimate relationship, which is of more relevance. They define coercive control as “a dynamic process linking a demand with a credible threatened negative consequence for non‐compliance” (Dutton & Goodman, 2005, p. 746). Focussing mainly on the perspective and experience of the target, they suggest that successful coercive control requires five components: surveillance of the target's behaviour; demands with a credible threat of harm if not met; delivery of threatened consequences; control of the target's ecology; and arranging the target's social environment so that they are isolated from normal supports (Davis et al., 2012; Dutton & Goodman, 2005). Davis and colleagues compared a small number of studies of stalking with these components and concluded that, when examined this way, stalking and coercive control overlap considerably and so may be understood in theoretically similar ways. The primary difference identified by Davis and colleagues is that coercive control behaviours had traditionally been conceptualised as occurring in an intact relationship, whereas stalking occurs pre‐ or post‐relationship. ¹

Coercive control theory is derived from feminist perspectives on the gendered social contexts in which intimate partner violence and stalking occur (providing a measure of external consistency). Reflecting this, the construct of coercive control rests on sociological analyses of the expression of power in relationships. Because power is distributed unequally between those of different genders, coercive control is understood as a highly gendered phenomenon perpetrated by men toward women, though less frequent occurrences in other relationship contexts are recognised (Dutton & Goodman, 2005; Stark, 2012). Davis et al. (2012) elaborate on this, proposing that stalking may be an exaggerated expression of gendered courtship scripts in which a man should pursue his love interest despite her protestations. They recognise that these scripts may equally apply to women but, similar to broader coercive control theory (Dutton & Goodman, 2005; Stark, 2007), argue that patriarchal social structures mean that women's attempts to persistently and aggressively seek a relationship are not as successful and are not viewed as threatening or frightening.

There is some research evidence to support the idea that stalking is an extension of coercive control when it begins following the end of an abusive relationship, providing some empirical adequacy for this theory. Several studies have shown that stalking is more likely if there was controlling behaviour or intimidation during the prior relationship (Cloonan‐Thomas et al., 2022; Davis et al., 2000; Dye & Davis, 2003; Katz & Rich, 2015; Ornstein & Rickne, 2013; Roberts, 2005). However, these findings also highlight the most significant challenge to the explanatory scope and depth when applying coercive control theory to stalking. It is unclear how Dutton and Goodman's (2005) necessary components for coercive control can be met in cases where there is no previous relationship between stalker and victim, only a very tenuous relationship existed, or where the prior intimate relationship was not already characterised by coercion and control.

Dutton and Goodman's (2005) description of coercive control in intimate partner violence relies heavily on the creation of an environment in which coercion can occur because the contingent threat is deemed credible by the victim. Whether this environment would be present when the stalker is pursuing the beginning of a relationship with the victim is questionable. It may be that over time, the stalking itself would create an ecology in which the threat is deemed to be credible by the victim, but that would not necessarily account for victims' initial reactions to stalkers' behaviour. There is little evidence that those who stalk in an attempt to begin a relationship frequently use violence (McEwan, Mullen & MacKenzie, 2009; McEwan, Mullen, MacKenzie & Ogloff, 2009), so how these stalkers ‘prove’ any threats to be credible is unclear. A related issue is the ability of coercive control theory to explain stalking episodes in which the perpetrator is persistently ‘romantic’ rather than threatening. Having someone persistently express their love when it is unwanted is undoubtedly disconcerting and can be frightening for the victim; however, whether such a scenario would equate to a ‘credible threat of a negative consequence if the demand is not met’ is unclear.

The other significant limitation of this theory is its failure to explain why and how stalking emerges from different contexts that do not fit within the ‘male perpetrator–female victim’ dyad or the relationship pursuit paradigm. Criticising the empirical adequacy of this theory on the basis that it does not address stalking outside of a relationship context is perhaps unfair, as Davis et al. (2012) very specifically link it to the context of relationship pursuit. However, even within this domain, the model does not address the role of various types of psychopathology that have been shown to be present among those who stalk potential or former partners (Albrecht et al., 2022) and would undoubtedly influence perpetrators' use of coercive control tactics, nor does it address other proximal factors relevant to stalking. There may be considerable fertility and unifying power in combining elements of coercive control theory focussed on gendered cultural scripts and social structures with more proximally relevant single‐factor theories such as attachment, social learning or relational goal pursuit theory. Davis and colleagues identify this as a goal in theory integration and suggest that also applying learnings from self‐regulation theory could offer additional benefits to such an integrated theory.

3.4. Relational goal pursuit theory

Relational goal pursuit theory (RGP) was proposed by Spitzberg and Cupach (2000, 2014, Cupach & Spitzberg, 2004) to suggest that stalking (and the related concept of obsessive relational intrusion) develop from everyday relationship striving. The theory draws on broader literature about relationship formation and dissolution and goal pursuit (Spitzberg & Cupach, 2014). In RGP theory, success in intimate relationships is conceptualised as a goal—a desired end state that the individual wants to attain. Drawing on the vast research literature on goal‐driven behaviour, Spitzberg and Cupach (2014) posited that goals are pursued when they are perceived as highly desirable, attainable and not substitutable. Those who engage in abnormal behaviour in pursuit of a relationship are hypothesised to elevate the personal importance of relational goals by linking them to superordinate goals concerning happiness and self‐worth. Achievement of these higher‐order goals becomes contingent on the successful pursuit of the relational goal, exaggerating the latter's urgency and importance and making the person more inclined to persist with relationship pursuit, even in the face of apparent failure. Spitzberg and Cupach (2014) also posit the presence of specific cognitive processes that make the relational goals appear attainable and so worth persistently pursuing. These include biased appraisals of the likely benefits of success and costs of failure, high self‐efficacy (belief that they can enact goal pursuit behaviour) and positive outcome expectancies—the belief that their behaviour will result in goal attainment.

Providing considerable explanatory depth, RGP theory hypothesises about specific cognitive and affective mechanisms that maintain stalking behaviour. Goal frustration is thought to increase rumination, which maintains the person's focus on the unattained goal but also leads to intense negative emotional states (described as ‘emotional flooding’). Emotions such as anger, shame, guilt, jealousy, fear and sadness become overwhelming and motivate one to act to reduce emotional arousal, including unwanted intrusive behaviour and further rumination. The person comes to believe that achieving the relational goal is the only way to reduce their distress and so persists in their efforts. Over time, they will also engage in cognitive rationalisation that legitimises their pursuit behaviour, allowing it to continue. These are proposed to include idealising the relationship, misconstruing the target's responses, justifying objectively inappropriate behaviour and minimising its consequences to self and others.

Hypotheses arising from RGP theory have been tested in several studies investigating self‐reported unwanted pursuit behaviour or stalking in college student and general adult samples. Demonstrating the empirical adequacy of much of RGP theory, these studies have provided support for the hypothesised relationships between stalking and rumination (Cupach et al., 2011; De Smet et al., 2015; Johnson & Thompson, 2016; Senkans, 2016; Spitzberg et al., 2014); goal linking (Cupach et al., 2011); high self‐efficacy (Brownhalls et al., 2019; Cupach et al., 2011; Spitzberg & Cupach, 2014; Spitzberg et al., 2014); intense negative affect (Brownhalls et al., 2019; Johnson & Thompson, 2016; Spitzberg et al., 2014); and rationalisation (Brownhalls et al., 2019). In the most recent test of RGP theory, Brownhalls and colleagues concluded that affective flooding and rationalisation were the most important components in the multivariate modelling of stalking behaviour in a sample of 379 American adults.

RGP theory provides perhaps the most robust theoretical explanation of stalking behaviour to date. It displays good external consistency with existing theoretical explanations of human behaviour in relationships (Spitzberg & Cupach, 2000). The concepts of relationship goals and self‐worth also clearly share similarities with concepts central to stalking explanations based on insecure attachment, suggesting that findings from this approach could be unified under RGP theory with some additional work. RGP theory's more detailed account of stalking‐supportive cognition gives greater explanatory depth than other authors who have suggested a role for cognitive variables (e.g. Duntley & Buss, 2012; Fox et al., 2011; White et al., 2000) while conceptualising stalking as a goal‐driven behaviour makes the concept of self‐regulation relevant, offering linkages with Davis et al.'s (2012) application of adult attachment theory and stalking via self‐regulation deficits.

RGP theory also has strong fertility and clinical utility, proposing quite specific psychological constructs that can be operationalised in both research and practice. The explanatory scope of RGP theory is limited to relationship pursuit, though the proposed cognitive and affective mechanisms could be relatively easily expanded into other contexts for stalking behaviour by hypothesising other kinds of personal goals (unrelated to romantic relationships) and could be linked to superordinate personal goals, providing the necessary affective arousal and motivation to persist. Perhaps the greatest weakness of RGP theory in its current form is that it does not explain why stalking behaviour stops, although the sizeable literature concerning goal disengagement (e.g. Barlow et al., 2020) suggests that this gap might be relatively easily filled in a way that maintains the internal coherence of the theory.

3.5. Integrated Model of Information Processing

The most recent application of an existing multi‐factorial theory to stalking is by Birch et al. (2018). They consider how Huesmann's (1998) Integrated Model of Information Processing, originally developed to explain aggression, might be relevant to understanding and therefore treating stalking behaviour. Birch and colleagues conceptualise stalking as involving direct, indirect, explicit and implicit components of aggression throughout a stalking episode. Drawing on aggression theory (and providing some external consistency), they suggest that the cognitive and affective processes related to aggression could be equally related to stalking behaviour, particularly highlighting the role of behavioural scripts, normative beliefs and hostile attributions, and the schemas that tie these knowledge structures together. All are well‐supported and investigated constructs in the aggression literature (Gilbert et al., 2013). Birch and colleagues suggest ways in which these knowledge structures and associated affective arousal might influence the use of different kinds of stalking behaviours (all of which are conceptualised as forms of aggression). Birch et al. (2018) directly apply Huesmann's Integrated Model of Information Processing to suggest a cyclical set of cognitive and affective processes that underpin stalking behaviour, which is then maintained over time through the effects of positive and negative reinforcement.

Birch et al.'s (2018) application of information processing theories of aggression to stalking is novel but highly consistent with many earlier psychological explanations for stalking (giving it unifying power). Reflecting the role of behavioural theory in social information processing models, there are similarities with Westrup and Fremouw's (1998) suggestion that particular events trigger cognitive activation and affective arousal that leads to stalking and that the consequences of the stalking behaviour reinforce its use over time. Information processing models also draw on social learning theory, meaning there is overlap with Fox et al. (2011) application of this theory to stalking. More distantly, there is consistency between Birch and colleagues' approach and that of Davis et al. (2012), who suggested a central role for self‐regulation (or dysregulation) in stalking. While not a central component of Birch and colleagues' approach, they do suggest that effective regulation of affect arising from activation of knowledge structures may be a key part of preventing impulsive acts of stalking. Broadly, the information processing approach is also consistent with the tenets of RGP theory, with its focus on goal‐directed behaviour and self‐regulation, while the concept of behavioural script rehearsal has considerable overlap with rumination and fantasy (though the constructs also have some important differences; Gilbert & Daffern, 2017; Hosie et al., 2022).

The major strength of Birch et al.'s (2018) application of an information processing model to stalking is the explanatory depth and fertility of such an approach. They suggest how this approach could be used in clinical practice and hypothesise about specific psychological constructs and processes that could be tested in future research. There is no research to date examining these kinds of knowledge structures in stalking, but Birch and colleagues' work presents a framework within which such research could be pursued. Perhaps the most significant gap in Birch and colleagues' theorising is the lack of attention to the content of the knowledge structures (schemas and scripts) that might be particularly relevant to stalking. There seems to be clear space to integrate information about relational schemas and scripts, and self‐ and other schemas that have been discussed when applying attachment theory to stalking and are implicit in RGP theory, and the cultural norms and cultural scripts that were suggested by White et al. (2000). Hypothesising about the content of knowledge structures that are so central social information processing models would add considerable explanatory depth to this work. Further depth could be added by theorising about the differences between stalking and other forms of overt aggression, which are not targeted and persistent, and what elements of information processing may contribute to prolonged and targeted stalking behaviour rather than one‐off acts of aggression.

4. SUMMARY OF THEORETICAL CRITIQUE

The essential nature of a theory is that it places an activity within a wider framework which imparts a meaning that is not apparent in the event itself. This allows understanding, critical appraisal, prediction and the possibility of fundamental change in the phenomena being explained. The first task when theorising a behaviour such as stalking must be to identify the scope of the phenomenon that needs to be explained and then, to develop a theory with sufficient explanatory depth, which is consistent with substantiated theories of similar behaviour, and that accounts for existing research findings. Finally, the theory should offer avenues for future research, and in the case of psychological theories of forensically relevant behaviour, it should be able to inform clinical practice.

The strengths and weaknesses of the theories and models appraised in this article are summarised in Table 1. It should be noted that while many achieve a positive rating against various criteria, not all meet each criterion to an equivalent degree, and the table does not provide a summary of the critiques of each theory provided above. Even where they demonstrate adequate explanatory scope and depth, there were also important limitations in this area for most of the theories and models. Limitations tended to be that theories only address stalking in a particular context (e.g. pre‐ or post‐romantic relationship) or by a particular group of people (e.g. men or people who are insecurely attached). However, as noted throughout the body of the article, several extant theories could be relatively easily adapted to take in a broader scope of stalking behaviour by a wider range of people. This would have significant benefits for research and practice as constructs could be operationalised and tested in larger samples and be applied with a wider variety of clients. Empirical adequacy was also found wanting for many of the theories discussed here, despite the fact that fertility was a clear area of strength for most theories and models appraised. Opportunities for useful research is not the problem—it is the absence of such research that needs to be overcome with more targeted and theory‐driven research, and by conducting research into areas of conceptual similarity across theories. Almost all theories proposed to date have good external consistency. Most have the potential to unify different research findings from the stalking literature and to be integrated with other theoretical explanations for stalking to provide a more complete account of the behaviour.

TABLE 1.

Summary table of performance of evaluated theories against predefined epistemological criteria

	Empirical adequacy	External consistency	Unifying power	Fertility	Explanatory scope and depth
Attachment theory adaptations	+	+	+	+	+
Social learning theory	+	+	+	+	+
Evolutionary theory	/	/	/	+	/
Behaviourism	‐	+	+	+	+
Integrative contextual development model of male stalking	/	+	+	+	‐
Coercive control theory	/	+	+	+	/
Relational goal pursuit theory	+	+	+	+	+
Integrated model of information processing	/	/	+	+	+

Note: + indicates that the criteria was adequately met, / indicates that this is present but either indirect or incomplete, ‐ denotes that the criteria has not been met.

4.1. Areas of theoretical consistency

There is some consistency in aspects of how the different theories conceptualise stalking and the kinds of social and psychological variables that they hypothesise to be relevant to the onset and maintenance of stalking behaviour. These areas of convergence between existing theories suggest constructs or processes that might be particularly fruitful for future research and for further theory development. This approach, known as theory knitting, was originally proposed by Kalmar and Sternberg (1988), who suggested that theory development can be aided by drawing together the strongest elements of existing theory with the researcher's original ideas, which is a more fruitful approach to theory development than a traditional, adversarial model (Kalmar & Sternberg, 1988; Ward & Siegert, 2002). Thus, the following paragraphs highlight where existing theories of stalking appear to converge or have similarities, suggesting that these areas might be worth focussing on in future research designs and incorporating into further theory development using a theory knitting approach.

Stalking is conceptualised as a purposeful, goal‐directed behaviour in almost all existing theories (Davis et al., 2012; Duntley & Buss, 2012; Fox et al., 2011; Meloy, 1989; Spitzberg & Cupach, 2000; Westrup & Fremouw, 1998). In some, this is overt (e.g. RGP theory, evolutionary theory, behavioural theory, and coercive control theory), while in others, the goal‐directed nature of the behaviour is implied (e.g. in the adaptations of attachment theory, the goal is to avoid negative affective arousal and self‐concept arising from relationship conflict). From this, it might be concluded that stalking is universally theorised as a maladaptive attempt (or series of attempts) to achieve a specific interpersonal goal or set of goals.

Stalking is understood in all extant theories as a product of the interaction of the individual (and their cognitive and affective processes) with their environment. The context in which the behaviour begins is crucial in almost all theoretical perspectives and is closely linked to the kinds of goals thought to drive stalking. Different authors all suggest that the kinds of situations that lead to stalking must be appraised as relevant to the sense of self of the person who goes on to stalk, leading to affective arousal and providing motivation to engage in stalking behaviour. Again, this is more explicit in some theories than others. RGP theory specifically identifies situations with potential for relationship success or failure as being linked to the individual's sense of self‐worth and happiness (Spitzberg & Cupach, 2014). Approaches based on insecure attachment propose essentially the same argument, though using a different set of language and literature to make the link (Davis et al., 2012; Meloy, 1998). Westrup & Fremouw (1998) do not suggest a particular kind of situation as a universal precipitant but emphasise situational characteristics and the individual's appraisal of them as an important antecedent to stalking behaviour. This link is perhaps least obvious in coercive control theory, though analysis of Dutton and Goodman's (2005) original model shows that this is likely because the model is conceptualised largely from the perspective of the person being abused, with little attention to the psychological experiences of the abuser.

Theories addressing the psychology of the stalker differ in their explanations of why some situations are more relevant to an individual's self‐worth or self‐concept than others. They all suggest some sort of predisposing psychological vulnerability is necessary, though with different levels of explanatory depth. For example, attachment theorists suggest that those with anxious attachment are vulnerable to experiencing threatened self‐worth when faced with relationship dissolution because they rely on others to maintain a positive sense of self (Davis et al., 2012; Meloy, 1998). Relational goal pursuit theory describes the vulnerability of linking relational success to self‐worth, though it does not suggest why some people might do this while others do not (Spitzberg & Cupach, 2014). This approach has considerable overlap with concepts relevant to insecure attachment but goes beyond this to include those who wish to commence a relationship and tie their self‐worth to success in that realm. Birch et al. (2018) highlight the predisposing vulnerability of schemas that bias people toward expecting mistreatment and hostility from others. While the original proponents of coercive control theory did not hypothesise about psychological vulnerabilities of the abuser, subsequent research using this framework suggests that shame and negative self‐worth is associated with increased need for control, very consistent with attachment theory (Kaplenko et al., 2018). How such psychological vulnerabilities might intersect with gender‐related cognition to create a particular vulnerability is an area in need of considerably greater research (cf. Senkans et al., 2020).

These different kinds of predisposing psychological factors might be usefully understood as creating vulnerability to stalking in different kinds of interpersonal situations. This maps on rather well to some stalking typologies that incorporate both situation and motivation to explain when stalking is more likely to occur (e.g. Mullen et al., 2009). It is also consistent with the idea of different ‘pathways’ into stalking that are a feature of some other theories of forensically relevant behaviour (Gannon et al., 2011; Ward & Siegert, 2002). Unfortunately, with the exception of those associated with insecure attachment, the kinds of self‐, other‐, and relational schemas that might create vulnerability to stalking in particular situations are poorly understood and have not been subject to any research to date.

Almost all theories critiqued suggest that for situational appraisals to lead to stalking, they must result in intense affective arousal (which is perhaps more likely when situations are appraised as highly relevant to the individual's sense of self). Intense negative affect is a central component of attachment theories, RGP theory, information processing models, and behaviour theory (Birch et al., 2018; Davis et al., 2012; Meloy, 1998; Spitzberg & Cupach, 2014). It is consistently hypothesised across multiple theories that one of the functions of stalking behaviour is to change the situation and so reduce negative affective arousal. This hypothesis suggests that stalking behaviour has a self‐regulatory function. Only Davis et al. (2012) and Birch et al. (2018) make this link explicit, both hypothesising that problems with self‐regulation may be a key deficit that contributes to stalking behaviour.

Interestingly, the role of positive affective arousal in stalking has not been theorised to date, despite typologies suggesting that positive emotions are relevant (e.g., Mullen et al.'s [2009] incompetent suitors and intimacy seeking types, who stalk out of love, lust, and hope). A self‐regulation framework may provide a coherent way of theorising the role of positive emotions in stalking, given that such emotions have been integrated into self‐regulation theories more broadly (Carver, 2003; Carver & Scheier, 1990). Despite the hypothesised self‐regulatory function of stalking behaviour, there is no research examining self‐regulation (and the related constructs of self‐control and impulsivity) or how people who stalk respond to and cope with affective arousal or intense emotional states. This is a clear topic for future research.

Moving away from affective factors and back to cognition, there are some similarities across theories that suggest a role for stalking‐supportive cognition. This is most clearly elaborated in RGP theory (Spitzberg & Cupach, 2014) and by Birch et al. (2018), but is mentioned by Fox et al. (2011) and Duntley and Buss (2012), and is implicit in both Westrup and Fremouw's (1998) application of functional analysis to stalking and adaptations of attachment theory (Davis et al., 2012; Meloy, 1998). Outside of RGP theory, what ‘stalking‐supportive cognition’ actually entails is poorly operationalised and so difficult to test. Spitzberg and Cupach (2014) specifically hypothesise the presence of a belief that the stalking victim secretly wants the attention, and so minimises any evidence of alternatives and reframes their own behaviour accordingly. This account has received some empirical support, though it has only been tested in a single study (Brownhalls et al., 2019). Birch and colleagues and Fox and colleagues both suggest a role for broader aggression‐supportive and antisocial beliefs, but the presence of such cognition among people who stalk has never been explicitly investigated outside of diagnoses of antisocial personality disorder. The role of cognition in precipitating and maintaining stalking behaviour is sorely understudied even though it would be a clear target for any psychological interventions designed to stop stalking. This is a further area in need of future research.

The final construct highlighted as potentially relevant during this review and critique was the role of rumination in stalking. While only explicitly appearing in RGP theory and mentioned by Birch et al. (2018), rumination has been shown to relate to stalking in multiple studies, suggesting that it should be more closely examined in future research and theory development. Spitzberg and Cupach (2000, 2014) discuss rumination as a normal response to goal frustration, but rumination can also be conceptualised as a maladaptive emotion regulation strategy (Skinner et al., 2003), offering a second avenue to incorporate this construct into explanatory theories of stalking. Notably, attachment anxiety has been associated with increased emotion dysregulation and rumination in the broader literature (Mikulincer & Shaver, 2019), suggesting further opportunity for unifying existing research findings when theorising stalking in future.

5. IMPLICATIONS AND CONCLUSIONS

This review and critique of existing stalking theory highlights the potential to develop a comprehensive multi‐factorial theory of stalking that builds on and unifies existing approaches while overcoming some of their problems with empirical scope and depth. By considering these theories together, and examining the many areas of clear similarity, this paper suggests multiple avenues for theoretically‐informed future stalking research. Such research is essential if we are to progress the stalking literature beyond description of the phenomenon toward evidence‐based explanations that can inform treatment and prevention approaches.

5.1. Implications for research

Our analysis of existing theories suggests a range of individual and situational factors that should be the focus of future research to inform both theory development and practice. There is a clear need for research into the kinds of predisposing knowledge structures (schemas, and their component parts such as behaviour scripts and normative beliefs) that relate to stalking in different contexts. This obviously includes the self‐, other‐ and relational schemas observed in anxious attachment that already have some empirical support but should go beyond this to consider other combinations of schemas that may be important to stalking in other contexts, such as those concerning gender roles and expectations (see Senkans, 2016; Senkans et al., 2020). There is a need to examine the potential role of ‘stalking‐supportive cognition’ in maintaining stalking behaviour, including the role of rationalisation as proposed in RGP theory. It is possible that this kind of distorted thinking might actually prolong stalking behaviour if understood as part of an offence process model, or they could be simply post‐hoc neutralisation reported by people who have previously engaged in stalking, rather than being criminogenic in and of themselves (cf. Maruna & Mann, 2006). Research into ‘stalking‐supportive’ cognition should also examine whether and how generally antisocial cognition and aggression‐supportive cognition are related to the behaviour, given the large literature on these constructs that has already been translated into clinical practice.

A self‐regulatory function for stalking is proposed in many extant theories, making this an area in need of urgent investigation. This includes research testing Davis et al.'s (2012) suggestion that deficits in self‐ or emotion‐regulation contribute to stalking. Such research might also usefully examine the potential role of self‐control and different types of impulsivity in stalking behaviour, given their demonstrated importance in aggression more broadly (Denson et al., 2012; Derefinko et al., 2011). However, it is also entirely possible that stalking is not a product of self‐regulatory deficit, but results from the use of active self‐regulation or coping strategies that the person perceives as appropriate but are maladaptive because they involve harm to others and the individual using them (e.g. maladaptive efforts at negotiation or help‐seeking; opposition; and rumination; Skinner et al., 2003). Developing a good understanding of the self‐regulatory function of stalking behaviour is an area with considerable research and practice potential, but self‐regulatory mechanisms may play a different role depending on the context in which the stalking emerges and continues, the nature of the stalking behaviours used and the individual's perceived need to inhibit stalking behaviour (cf. Finkel, 2014).

Addressing research questions in each of these areas will require innovative research designs and methodologies. At present the stalking research literature is bifurcated into studies involving forensic/clinical samples and those involving college students (or to a lesser extent, adult community members) (see Spitzberg & Cupach, 2014). There is no way to know whether findings in one group translate to the other, meaning that comparing stalking behaviour in these groups, and testing the presence and role of different psychological constructs across these groups is essential. It may not be possible to generalise from community studies to treatment of people whose stalking attracts criminal justice attention, but at present, this is not at all clear. Recent advances in measurement of self‐reported stalking may assist with this kind of research design (McEwan et al., 2020). Research using mixed methods and qualitative research designs would also clearly be useful to address some research questions. This might be particularly important to develop offence‐process models of stalking behaviour and to understand how different cognitive and affective content and processes interact (e.g., Stairmand et al., 2021; Wheatley et al., 2020). Given the apparent importance of intense affective arousal in stalking behaviour, finding ways to measure both ‘hot’ and ‘cold’ cognition that contributes to stalking will be challenging, but likely essential to understand how social information processing and self‐ and emotion regulation interact to lead to someone choosing stalking as a response to affective arousal (Senkans et al., 2020). Finally, the holy grail of stalking research would be longitudinal studies that were able to track stalking behaviour over time and examine how different kinds of interventions might be more or less effective in reducing it. As part of longitudinal research, the use of a daily diary research design would allow the recording and subsequent analysis of proximal cognitions and corresponding behaviours, which are more likely to be specific to the stalking context than general attitudes and behaviours (Gunthert & Wenze, 2012).

This leads to one of the clear gaps in theoretical explanations of stalking at present, which is the lack of attention to why stalking stops. None of the theories reviewed here consider desistance from stalking, even though the vast majority of stalking episodes end, some quite quickly and some after many months or years (McEwan, Mullen & MacKenzie, 2009; McEwan, Mullen, MacKenzie & Ogloff, 2009). Developing a coherent and testable account of why stalking stops is an obvious area for future theorising and enquiry.

A second clear gap in existing stalking theory is the lack of attention to the role of psychopathology in stalking behaviour. Research indicates that over half of people whose stalking attracts criminal justice attention have a diagnosable mental disorder (Nijdam‐Jones et al., 2018), and in a proportion of cases, the symptoms of that disorder are implicated in their stalking behaviour (Albrecht et al., 2022). Psychopathology has played a major role in typologies of stalking but is virtually unmentioned in stalking theory. Only attachment‐based explanations address psychopathology in any real sense (and even then only considering personality disorder). Integrating the full scope of relevant psychopathology into future stalking theory in a coherent way is necessary to ensure that interventions designed to stop stalking are appropriately targeted (whether that involves treating mental disorder or not). It is possible that symptoms of psychopathology could be integrated through the inclusion of the self‐ and emotion‐regulation constructs previously discussed, given that such constructs are trans‐diagnostically associated with many forms of mental disorder (Santens et al., 2020). Regardless of how symptoms of psychopathology are conceptualised, it is essential that they are integrated into explanatory theories of stalking given their clear centrality to the behaviour in many cases (Mullen et al., 2009).

5.2. Implications for practice with people who stalk

The evidence base regarding psychological interventions for people who stalk is dismal and lags far behind equivalent treatment literature for violence, harmful sexual behaviour, intimate partner abuse, and even fire setting. At present, there is a single published treatment trial involving people who stalk (Rosenfeld et al., 2019) and clinicians must rely on ‘best practice’ guidelines to inform their thinking (e.g. Purcell & McEwan, 2018). The lack of research into treatment has multiple causes, not least that stalking is often invisible in criminal justice systems because people who stalk are typically not charged with stalking (Brady & Nobles, 2017). The wide range of offences committed by people who stalk may see them channelled into other offending programs (e.g. for violence or intimate partner violence), but whether such programs address needs related to stalking is unclear (Purcell & McEwan, 2018). However, it is likely that another important reason for the absence of evidence‐based treatment for stalkers is the lack of both theory and research than can inform the content of such interventions. As discussed above, there are multiple avenues to improve this research base and to further develop theory that can guide both research and practice in future.

That said, the potential for future research is not particularly useful for clinicians who are required to assess and treat people who stalk now. The theoretical review and critique suggests several areas that may be useful to attend to when assessing and formulating an individual's stalking behaviour, given that they appear across multiple theories and have some level of empirical and/or theoretical support. Specifically, based on our review, we recommend that clinicians consider:

1. The function of the stalking for the individual (assessed through functional analysis of individual stalking behaviours focussing on cognitive, affective and situational factors);

2. Self‐, other‐ and relational schemas whose activation may be relevant to the onset and maintenance of affective arousal related to the stalking behaviour;

3. The individual's experience of affective arousal and contexts in which the person experiences intense affective arousal;

4. The person's strategies for regulating and coping with stress and intense affect (including their experience of rumination generally or on specific topics relevant to the stacking situation).

5. Cognition associated with the individual's stalking behaviour or other problematic interpersonal behaviour. This may include the role of stalking‐related cognition as an antecedent and reinforcing consequence in a functional analysis, but might also consider the role of broader aggression‐supportive cognition or antisocial beliefs.

This approach to assessment and formulation is theoretically informed and consistent with existing research findings. It would appear to provide the most appropriate approach at present, given the current state of the research literature.

6. CONCLUSION

We have described and critiqued the small number of single‐ and multi‐factor theories that have been used to try to explain stalking to date. While single‐factor theories add important “pieces of the puzzle” when attempting to explain stalking, in isolation, they seem insufficient to explain the complexity of features that appear to interact to produce stalking behaviour. Multi‐factor models seem to offer a more nuanced and detailed account of stalking behaviour but are still subject to limitations in their scope and depth. There is clearly space to further develop a more complete multi‐factor explanatory theory of stalking that builds on and integrates existing work and fill gaps in those accounts. There is also space to develop ‘micro‐theories’ or offence process models that can improve clinicians' ability to provide evidence‐based relapse to prevent work for people who stalk (e.g. Barnoux et al., 2015; Gannon, Collie, et al., 2008; Stairmand et al., 2021).

Our critique of existing explanatory theories of stalking highlights the urgent need to develop a better understanding of the psychological, social and cultural factors that cause stalking and so should be targets for intervention and prevention. At present, theory in this field is stagnant and consequently, most research continues to describe rather than explain. This makes it of limited use to clinicians trying to formulate or develop treatment plans for the wide variety of people who present with stalking behaviour. We urge both researchers and clinicians working in this field to consider how they can contribute to meaningful, theoretically informed research and practice with people who stalk. We hope that this article provides some useful guidance about where to direct these efforts to have the greatest impact in the short to medium term.

Parkhill, A. J. , Nixon, M. , & McEwan, T. E. (2022). A critical analysis of stalking theory and implications for research and practice. Behavioral Sciences & the Law, 40(5), 562–583. 10.1002/bsl.2598