Career Dimensions of Stalking Victimization and Perpetration

Abstract

Prior literature concerning stalking, particularly in the field of criminology, finds wide variation in fundamental trends regarding stalking victimization and perpetration. There seems to be little consensus regarding when and how stalking is manifested. Furthermore, prior research to date has not addressed the etiology of stalking‐related behaviors by applying principles from criminal career research, including participation, frequency, onset, and duration. The present study builds upon prior research by addressing trends in age of onset for stalking victimization and perpetration, the duration of stalking‐related behaviors, and the relationship between those behaviors and other types of crime over the life course using primary data from a sample of young adults. Findings indicate that stalking victimization and perpetration share important career attribute similarities, and that self‐reported history of intimate partner violence and sexual assault are strongly associated with stalking outcomes.

Keywords:

• stalking
• victimization
• perpetration
• criminal career
• life course

Introduction

Stalking has only recently been recognized as a criminal offense (Bjerregaard, 2000). Celebrity victims of stalking initially captured media attention, which quickly lead to an unprecedented outbreak of legal action against the perpetration of unwanted pursuit behavior (Perez, 1993). Given that the USA has only considered stalking a crime since 1990, ¹ research investigating the issues surrounding stalking victimization and perpetration is still developing.

Definitions of stalking vary among states and researchers. While many states have adopted similar legislation as California, defining stalking as the “willful, malicious, and repeated following or harassing” of another person, many discrepancies still exist among state stalking laws. For example, state stalking legislation varies regarding the inclusion of threats, cyber‐stalking (i.e., stalking via electronic communication such as emailing, text messaging, etc.), and the stalking of victims’ friends or family members (Leiter, 2005). Stalking definitions adopted by social scientists have also varied considerably. Some researchers have chosen to implement legal definitions determined by an individual state (Meloy & Gothard, 1995) whereas others have chosen to exclude a definitive definition of the construct by allowing research participants to individually interpret the definition of stalking (Mustaine & Tewksbury, 1999). Further, some scholarly definitions of stalking provide participants with specific examples of stalking behaviors (Tjaden & Thoennes, 1998) while others offer a broad description of stalking without delineating specific conduct (Wright, Burgess, Burgess, Laszlo, McCreary, & Douglas, 1996). Overall, the majority of research relies upon a definition of stalking that stresses repeated negative behavior in which the perpetrator pursues the victim on two or more occasions and in which the victim perceives the behavior as undesirable, frightening and/or harassing in nature.

While research has started to address stalking victimization and perpetration, much remains unknown about the most basic descriptive aspects of stalking. Specifically, research has not yet explored stalking victimization and perpetration using a life course perspective. The objective of the present study is twofold: (1) to introduce a basic quantitative description of criminal careers in stalking, providing comparisons from extant stalking research that has not explicitly addressed these issues in a consistent fashion or with robust data and (2) to relate those findings to a meaningful framework for careers in stalking perpetration and victimization in such a way as to guide future research on stalking that implements a life course perspective.

Prior Literature

Life Course/Career Criminal Research

A life course perspective of crime aims to examine trajectories of involvement with crime over the course of the lifetime of individual offenders and victims (Sampson & Laub, 1993). More specifically, a life course examination of crime considers the potentially linked experiences with offending and/or victimization over time rather than treating such episodes as separate events, and as such a focus on the longitudinal pattern of offending and victimization, including description and etiology, is at the cornerstone of this perspective (Blumstein, Cohen, Roth, & Visher, 1986; Piquero, Farrington, & Blumstein, 2003). Criminal career research is comprised of four focal components and examines patterns of offending with respect to (1) participation, (2) frequency, (3) seriousness, and (4) length. In light of this perspective, several patterns have been detected among criminal offenders (see reviews in Blumstein et al., 1986; Farrington, 2003; Piquero et al., 2003). For example, demographic characteristics (specifically, sex, race, and age) are related to participation but not necessarily to frequency; offenders tend to begin engaging in crime during youth (ages 8–14), peak in adolescence (ages 15–19), and desist from crime in early adulthood (ages 20–29); and while offenders often repeat certain crimes (especially as they age), they do not often specialize in one type of crime.

Whereas criminal career research has examined the nature and extent of criminal offending over time, research investigating victimization over the life course is much less developed. In adopting the criminal career framework, Farrell, Tseloni, Wiersema, and Pease (2001) discuss the importance of studying victimization in terms of victim careers or “career victims.” While prior research has employed a life course perspective to examine many types of offending and a handful of types of victimization (Wittebrood & Nieuwbeerta, 2000), it has not yet examined stalking victimization or perpetration over the life course. Although little is known about stalking from a life course perspective, a small number of researchers have provided insightful information about age of stalking onset, duration of stalking, frequency/intensity of stalking, desistance from stalking, and the relationship between stalking and other crimes. We next turn to a brief overview of this line of research.

Participation in Offending and Victimization

Prevalence rates for stalking vary, perhaps due in part to variations in stalking definitions among researchers. National rates for stalking victimization are limited to one influential study, conducted by Tjaden and Thoennes (1998), as part of the National Violence Against Women Survey (NVAWS). Tjaden and Thoennes (1998) found the lifetime prevalence of stalking victimization among the sample of 8,000 women to be between 8% and 12% and between 2% and 4% among the sample of 8000 men. ² Researchers investigating the prevalence of stalking among college students report substantially higher rates of victimization, typically between 6% and 27% (Logan, Leukefeld, & Walker, 2000; McCreedy & Dennis, 1996). Average rates of stalking victimization range from 13% to 30% for female college students (Fisher, Cullen, & Turner, 2002; Fremouw, Westrup, & Pennypacker, 1997) and from 11% to 19% for male college students (Bjerregaard, 2000; Haugaard & Seri, 2001). While a growing body of literature has examined stalking victimization among adult and college student samples, the prior literature sampling children and adolescents is sparse. To date, stalking victimization prevalence rates among those who are pre‐college age are virtually un‐documented.

Previous research on stalking perpetration among the general public, college students, and adolescents is more limited than the existing research on stalking victimization. Of the studies that have investigated the prevalence of stalking perpetration, most have found substantially lower rates of perpetration in comparison with reported rates of victimization. Large‐scale studies examining stalking perpetration within the general public are also more limited compared with stalking victimization research. Studies measuring stalking perpetration among samples of non‐college participants have examined samples of convicted stalkers (Logan, Nigoff, Walker, & Jordon, 2002; Meloy, 1996; Mullen, Pathè, Purcell, & Stuart, 1999; Schwartz‐Watts & Morgan, 1998), batterers and stalkers (Burgess, Baker, Greening, Hartman, Burgess, Douglas, & Halloran, 1997; Burgess, Harner, Baker, Hartman, & Lole, 2001), and case studies of stalkers (Kurt, 1995). Some studies collected data on stalking perpetration in the general population from samples of stalking victims (Brewster, 2000; Emerson, Ferris, & Gardner, 1998). Stalking perpetration rates reported by college students range from 1% (Fremouw et al., 1997) to 8% (Haugaard & Seri, 2003).

Frequency of Offending and Victimization

Prior research on the frequency of stalking is limited to an influential study by Fisher et al. (2002). Among a large sample of college women, these authors determined that one‐third of the stalking victims reported experiencing stalking behaviors less than once per week whereas two‐thirds of the victims experienced stalking weekly, daily, or more than daily. Although the research on the frequency of stalking episodes is sparse, Blumstein et al. argue that frequency of offending is the most essential criminal career component. Therefore, it is clear that research adopting a life course perspective of stalking would highly benefit from examining the frequency of offending.

Further, very little is known about repeat stalking victims and perpetrators. Examining continuation/desistance from stalking victimization and perpetration is critical to gaining an understanding of the nature of stalking. Evidence suggests that some crime victims are repeatedly victimized (Doerner & Lab, 2005) and, similarly, that a small number of crime perpetrators have lengthy criminal careers (Blumstein et al., 1986). Consequently, it is imperative to determine whether stalking victims and perpetrators experience victimization once or repeatedly in order to ascertain the nature of stalking and to inform public policy decisions and efforts regarding stalkers.

Seriousness of Offending and Victimization

To date, research examining the seriousness of stalking has focused on the impact(s) of victimization. Victims of stalking frequently experience a variety of physical and psychological effects. For example, psychological effects of stalking victimization often include feelings of paranoia, fear, anxiety, and post‐traumatic stress symptoms, while physical symptoms may consist of appetite disturbances, headaches, asthma attacks, persistent nausea, and chronic sleep disturbance (Hall, 1998; Pathé & Mullen, 1997). Notably, these symptoms overlap greatly with post‐traumatic stress disorder (PTSD) indicating the severity and persistence of negative effects that some stalking victims endure. Given that victims of stalking often experience a number of traumatic aftereffects from exposure to repeated harassment, and in light of the findings that stalking is not uncommon, stalking is an important issue for social concern. In light of the unique nature of stalking, it is important to recognize that stalking victims may encounter entirely different experiences. For example, stalking often consists of a variety of behaviors that vary substantially in frequency and duration. Therefore, it is important to examine the seriousness of stalking in order to gain a better understanding of the etiology of stalking, which may be particularly useful for public health and policy responses. To date, research assessing the seriousness of stalking has not quantitatively calculated a scale or formula designed to assess the seriousness of stalking, an approach that we believe offers an important perspective into the nature of stalking.

Length of Offending and Victimization

The extant research on stalking victimization indicates that victims tend to be younger (under 29 years old) rather than older (Haugaard & Seri, 2003; Tjaden & Thoennes, 1998). The age of stalking perpetrators appears to be more varied, with some research determining that young adults are more often stalking perpetrators (Bjerregaard, 2000; Haugaard & Seri, 2003) and other research concluding that persons middle‐aged and older are more likely to perpetrate stalking (Harmon, Rosner, & Owens, 1995; Meloy & Gothard, 1995; Mullen & Pathè, 1994). Given the limited research attention given to age of onset in prior studies of stalking perpetration and victimization, and in light of the mixed findings for stalking perpetration, it is clear that an enhanced understanding of age of onset is needed to determine the types of individuals most at risk for experiencing and committing stalking.

The duration of stalking episodes has also received little attention among prior research. Among samples of adults in violent relationships, victims experienced stalking, on average, between two (Mullen & Pathè, 1994) and five years (Meloy, 1992) prior to experiencing violence from the same partner. Female college student victims report stalking episodes lasting on average between 12 and 21 weeks whereas male college student victims report being stalked an average of 26 weeks (Bjerregaard, 2000; Fisher et al., 2002).

Relationship between Stalking and Other Crimes

It is important to understand that victims and perpetrators of stalking may also experience or perpetrate other types of crimes during their lives. Prior research has begun to examine the extent to which stalking victims also experience other types of crimes, and findings indicate that stalking victimization is associated with physical assault victimization and sexual assault victimization (Slashinski, Coker, & Davis, 2003; Spitzberg & Rhea, 1999; Tjaden & Thoennes, 1998). However, prior research examining the extent to which stalking perpetrators also perpetrate other types of crimes is notably absent.

Present Study

Using the backdrop of the life course perspective, the current research explores the extent to which young adults have experienced stalking perpetration and victimization. The current study offers several unique contributions to the existing literature on stalking and criminal careers (of offenders and victims). First, this is the first study to employ a life course perspective to determine age of onset, duration of stalking perpetration and victimization, repeat perpetration and victimization, and desistance from perpetration and victimization. Second, the current study is among the first to examine the relationship between stalking perpetration and victimization and experiencing other types of crimes over the life course. Third, the current research uses one of the largest samples of young adults to examine stalking behavior. While the majority of prior stalking research has used smaller and highly select samples (typically ranging from 200–300 participants), it is important to use a large sample in order to make meaningful conclusions about findings given that stalking perpetration and victimization are relatively rare events.

Method

Participants and Procedure

An extensive web‐based survey was designed for this study. The survey consisted of modified versions of previously validated scales regarding stalking victimization and perpetration, intimate partner violence victimization and perpetration, sexual assault victimization and perpetration, and demographic questions. Data for this analysis were collected in April and May, 2007. The population of interest was all college students at a major southeastern university. The entire student body, except those under age 18 who were legally minors, formed the sampling frame. A simple random sample of 10,000 individuals from the population of 48,237 students was e‐mailed an invitation to anonymously participate in the survey. From this initial sample of 10,000, there were 103 e‐mail addresses found to be invalid, and the response rate was 19.4%, ³ resulting in a final sample of 1,921 individuals, of which 61% was female (N = 1,171) and 39% male (N = 750). Ages of respondents at the time of the survey ranged from 18 to 72 years old, with 75% of the sample age 25 or younger. Seventy‐eight percent of the sample was White, 6% Black, 8% Asian, and 8% ‘other’ races. Ninety percent of the sample was non‐Hispanic and 96% of the sample identified their sexual orientation as heterosexual. Table 1 presents descriptive statistics.

Table 1 Descriptive statistics

	Minimum	Maximum	Mean	SD
Stalking victimization (0 = No, 1 = Yes)	0	1	0.27	0.444
Stalking perpetration (0 = No, 1 = Yes)	0	1	0.06	0.233
Sexual assault victimization (0 = No, 1 = Yes)	0	1	0.32	0.467
Sexual assault perpetration (0 = No, 1 = Yes)	0	1	0.03	0.177
Intimate partner violence victimization (0 = No, 1 = Yes)	0	1	0.18	0.383
Intimate partner violence perpetration (0 = No, 1 = Yes)	0	1	0.06	0.239
Sex (1 = Male, 2 = Female)	1	2	1.61	0.488
Age	18	72	24.03	7.448
Race (1 = White, 2 = Non‐White)	1	2	1.22	0.412
Ethnicity (1 = Non‐Hispanic, 2 = Hispanic)	1	2	1.10	0.301
Sexual orientation (1 = Heterosexual, 2 = Other)	1	2	1.06	0.311

Measures

Survey participants were queried regarding their experiences in current or past episodes of stalking in which they were the victim, as well as episodes in which they were the offender. For purposes of the survey, an “episode” of stalking was defined as events or behaviors that (1) were harassing, frightening, intrusive, unwanted, and threatening; (2) involved the same individual offender/victim; (3) occurred more than once; and (4) included one or more measures from the modified stalking behavior scale presented in the NVAWS (Tjaden & Thoennes, 1998). The original scale was modified to include cyber‐stalking behaviors such as unwanted electronic messaging. Accordingly, each participant was asked to recount frequency of stalking‐related behaviors on an 11‐item scale (0 = behavior never occurred, 1 = behavior occurred once, 2 = behavior occurred more than once). For stalking victimization, respondents were asked if anyone had ever: (1) followed, watched, or spied on you, (2) stood outside your home, school, or workplace, (3) showed up at places uninvited, (4) vandalized your property or destroyed something you loved, (5) sent you unsolicited letters, written correspondence, or unwanted emails, (6) made unwanted phone calls to you, (7) left unwanted messages for you, (8) left unwanted items for you to find, (9) tried to communicate with you in other ways against your will, (10) sent unwanted messages electronically, and (11) posted unwanted messages/pictures to internet websites. Similar to the measure of stalking perpetration, respondents were asked if they had initiated any of the above acts. Summed scores of two or more of the eleven stalking criteria met the definition for stalking (meaning that at least two different behaviors occurred or that one behavior occurred at least twice). Importantly, a careful distinction must be made between a stalking behavior and a stalking episode. Stalking behaviors consist of individual items such as the modified NVAWS items used in the present study (e.g., following, telephone calling, leaving unwanted messages, etc.). Consistent with prior research, experiencing a single stalking behavior once does not constitute as stalking. However, experiencing two or more of the stalking behavior(s) constitutes as a stalking episode (provided all other definitional criteria mentioned above in this section are met). Stalking victimization and perpetration were separate measures based on the same set of eleven stalking items. Prevalence and frequency measures were derived from responses to these scales of stalking victimization and perpetration. These measures were recoded into the dependent variables for our logistic regression models for both victimization and perpetration (0 = did not meet stalking definition, 1 = met stalking definition). Further, respondents were asked to indicate the duration of the stalking episode. Finally, respondents were asked to indicate the age at which each stalking incident, either victimization or perpetration, began. This measure serves as an indicator of age of onset, and thus, the starting point for tracking career trajectories. Through the use of skip patterns in the web‐based survey, respondents indicating that they had experienced or perpetrated more than one stalking episode were guided to subsequent web pages that allowed them to report identical information about other stalking episodes (including types of behaviors, age of onset, duration, etc.).

In addition to the stalking victimization and perpetration items, respondents were asked questions regarding their experiences with two other types of violent interpersonal crimes: sexual assault and intimate partner violence (IPV). The sexual assault items were derived from a modified version of Fisher, Cullen, and Turner’s (2000) measures and included ten types of behavior ranging from sexual coercion to rape. The IPV survey items were derived from a modified version of the Revised Conflict Tactics Scale (Straus, Hamby, Boney‐McCoy, & Sugarman, 1996) and consisted of three types of behavior that included physical, psychological, and sexual abuse in the context of a romantic relationship. In both cases, respondents were asked about victimization and perpetration for sexual assault and IPV in parallel tracks (e.g., “Have you ever been forced to have sexual intercourse?” and “Have you ever forced anyone to have sexual intercourse?”), and values for sexual assault and IPV victimization and perpetration were dichotomized based on the presence or absence of any of the sub‐types. ⁴

Analytic Plan

Many dimensions of career trajectories for victims and offenders have the potential to inform broader understanding of the etiology of stalking behaviors. For example, earlier age of onset for perpetration may be linked to a greater number of victims or to increased risk of violent outcomes (see Moffitt, 1993). Frequency and duration, among other factors, may be similarly informative. Although the nature and scope of this study prevents an exhaustive examination of all of these career dimensions, a basic analysis of career trajectories in stalking begins with identification and description of these attributes. To do this, descriptive statistics are used to examine the prevalence, frequency, seriousness, onset, and duration of stalking perpetration and victimization. Further, we investigate the involvement in stalking victimization and perpetration over the life course with relation to histories of victimization or perpetration in sexual assault and intimate partner violence. Because the variables in the model are dichotomous, logistic regression is employed to examine the relationship between stalking (victimization and perpetration separately) and experiences with other types of crimes, including intimate partner violence (IPV) and sexual assault.

Results

Participation in Offending and Victimization

Almost 27% of the respondents (N = 517) acknowledged that they have been victims of stalking during their lifetime. Of the stalking victims, 23.4% were male and 76.2% were female. Seventy‐nine percent of the stalking victims were White, 6.4% Black, 5.6% Asian, and 8.1% specified their race as ‘other’. Approximately 88% of the stalking victims were non‐Hispanic. Over 95% of the stalking victims considered themselves heterosexual. One hundred and eleven respondents (5.8%) admitted that they have perpetrated stalking during their lifetime. Of the stalking perpetrators, 34.2% were male and 64.9% were female. Over 80% of stalking perpetrators were White, 8.1% Black, 3.6% Asian, and 7.2% specified their race as ‘other’. Eighty‐nine percent of the stalking perpetrators were non‐Hispanic and 91.9% were heterosexual.

Eighteen percent of the sample reported IPV victimization (N = 261) whereas 25% of the victims were male and 75% were female. Overall, 6% (N = 88) of respondents reported that they perpetrated IPV whereas 40% of the perpetrators were male and 60% of the perpetrators were female. Four hundred and ninety‐nine respondents indicated they had been victims of sexual assault (32.2%), of which 14% were male and 86% were female. About 3% of the sample admitted perpetrating sexual assault (N = 49), whereas 63% of sexual assault perpetrators were male and 37% were female.

The nature of the relationships between victims and offenders was also assessed. For stalking victims, the majority of respondents, 46% (N = 276), reported that their stalker was a friend, acquaintance, or stranger, whereas 30% (N = 179) of the respondents reported that the perpetrator was a current or ex‐intimate partner, spouse, or someone they had dated. Results were slightly different for stalking perpetrators: 56% (N = 153) reported that they had stalked ex‐intimates, spouses, or someone they had dated, while 33% (N = 51) reported that they stalked a friend, acquaintance, or stranger. The remainder of the victim and offender relationships were represented by a relatively small proportion of classmates, co‐workers, family members, and cases where the relationship was unknown.

Frequency of Offending and Victimization

Frequency is conceptualized as the number of behaviors occurring within a given time frame. For the purposes of this study, it is logical to define the time frame of interest as the bounds for the stalking “episode”—specifically, the time from the first stalking‐related behaviors involving a single perpetrator or victim to the last stalking‐related behaviors involving that same perpetrator or victim. Within each episode, frequency may vary considerably, with stalking behaviors occurring only twice (by definition) in an isolated manner, or conceivably with stalking behaviors occurring in a constant and uninterrupted manner. ⁵ It is reasonable to conceptualize stalking episodes as highly variable in terms of frequency as victims and offenders will also vary substantially in terms of risk or opportunity exposure, vulnerability or motivation, and a host of other factors.

For individuals who reported being stalking victims, the frequency of stalking behaviors ranged between two and more than twenty events (Mean = 6.92, SD = 4.47). Among victims, females reported only slightly more frequent victimization (Mean = 6.98, SD = 4.36) compared to males (Mean = 6.74, SD = 4.86). Conversely, stalking perpetrators reported fewer stalking‐related behaviors (Mean = 4.55, SD = 3.38). Female perpetrators committed fewer stalking‐related behaviors (Mean = 3.99, SD = 2.39) than did male stalkers (Mean = 5.49, SD = 4.53). Results from t‐tests indicate that there are no statistically significant differences in frequency of stalking behaviors between men and women who are victimized; likewise, there are no statistically significant differences in the frequency of stalking behaviors between men and women who perpetrate stalking (see Table 2).

Table 2 Comparison of stalking frequency by gender

	Male	Female	Mean difference
Stalking victimization	Mean = 1.08	Mean = 1.12	−0.039
	SD = 0.305	SD = 0.350	t = −1.192p = 0.235
Stalking perpetration	Mean = 1.13	Mean = 1.08	0.048
	SD = 0.475	SD = 0.325	t = 0.628p = 0.531

Very few of the individuals in the sample reported more than a single episode of stalking, whether in the context of victimization or perpetration. However, for individuals who reported a second victimization experience, the frequency of stalking‐related behaviors for the second episode increased relative to the mean (Mean = 8.48, SD = 5.39). For those individuals who reported a third perpetration episode, the frequency of stalking‐related behaviors decreased in the third episode relative to the mean (Mean = 4.38, SD = 2.33).

Seriousness of Offending and Victimization

Seriousness is difficult to quantify because of the vast range of experiences reported relative to stalking. In self‐report studies, seriousness is a subjective value that is filtered through individual perception, and serves as a function of coping mechanisms, resilience, support structures, and a host of other personal‐psychological attributes. Instead, we attempt to quantify seriousness quantitatively, using data gathered on the frequency, duration, and relative severity of stalking behaviors. This is not to suggest that our approach is the only or necessarily best approach to measuring seriousness; however, we intend it to be an attempt at standardization within our sample so that the range of victims’ and perpetrators’ experiences can be more fully described herein and elsewhere.

The value for episodic stalking seriousness was calculated with the following expression:

where i = individual and j = episode, a_i represents the weighted frequency of stalking behaviors deemed “more severe”, ⁶ b_i represents the weighted frequency of stalking behaviors deemed “less severe”, ⁷ and c_j represents the episode duration. The first set of terms is weighted more heavily than the second set due to our assessment of these behaviors as being more disruptive, frightening, and harassing compared to other potential outcomes. ⁸ Finally, we take the natural log of the product in order to reduce dispersion to manageable levels. A handful of cases resulting in negative scores on the seriousness scale were dropped from the analysis.

Results for the analysis of stalking seriousness are depicted in Figures 1 (victimization) and 2 (perpetration). Values for the first reported stalking episode range from 0 to a maximum of 8.78 for victimization and 9.47 for perpetration. Overall means diverge slightly, with victims showing higher seriousness (Mean = 3.14, SD = 1.76) compared to perpetrators (Mean = 2.14, SD = 1.76). Female victims show only slightly higher episodic seriousness than their male victim counterparts (Means = 3.15 vs. 3.09, SD = 1.73 and 1.85, respectively). Male perpetrators have slightly higher episodic seriousness than female perpetrators (Means = 2.28 vs. 2.06, SD = 2.23 and 1.49, respectively). Results from t‐tests indicate that there are no statistically significant differences in stalking seriousness scores between men and women who are victimized; likewise, there are no statistically significant differences in stalking seriousness scores between men and women who perpetrate stalking (see Table 3).

Figure 1 Distribution of stalking seriousness scores for the first victimization episode.

Figure 2 Distribution of stalking seriousness scores for first perpetration episode.

Table 3 Comparison of stalking seriousness by gender

	Male	Female	Mean difference
Stalking victimization	Mean = 3.41	Mean = 3.46	−0.052
	SD = 1.745	SD = 1.546	t = −0.296, p = 0.768
Stalking perpetration	Mean = 3.16	Mean = 2.50	0.656
	SD = 2.272	SD = 1.364	t = 1.396, p = 0.172

Length of Offending and Victimization

For the purposes of this study, length of offending is operationalized using self‐reported data for age on onset and duration, which is further sub‐classified as (1) the overall length of time between the first and last reported episodes and (2) the length of time from the beginning to the end of stalking behaviors within a given episode. Both onset and duration provide critical information regarding the nature of stalking victimization and perpetration. Onset may be defined as the age at which the first episode of stalking victimization or perpetration occurs. Figures 3 and 4 show a histogram depicting age of onset for the first reported stalking incident for victimization and perpetration, respectively. The distributions are interesting in several respects. First, the earliest onset for stalking victimization is reported at age 10, the earliest onset for perpetration is reported at age 12, and a sizable portion of the sub‐sample of stalking victims and offenders report onset before age 18 (22.4% and 27.9% for victims and offenders, respectively). Second, both stalking victimization and perpetration onset appear to peak at the beginning of the college years. The mean age of onset for both victims and offenders in this sample is approximately 20 (Mean = 20.57 and 19.86, SD = 5.68 and 4.45 for victims and offenders, respectively). Third, onset seems to center around the transition to young adulthood for both victims and offenders. Approximately 41.97% of victims and 54.1% of offenders report their first stalking experience between ages 18 and 22. Fourth, although less prevalent, stalking onset appears to occur well after the “traditional” college years in a sizable portion of cases. The data indicate that 17.4% of victims and 9.9% of offenders in this sample experience onset after age 25. However, although there is variation in age of stalking onset, it is important to note that common social and environmental factors (e.g., exposure to new people, increased dating behaviors, etc.) may be common to college students across age groups. Finally, there are some observed gender differences in onset. Females experience onset for victimization earlier (Mean = 20.03, SD = 4.85) compared to males (Mean = 22.45, SD = 7.66), while males experience onset for perpetration only slightly earlier (Mean = 19.69, SD = 3.72) compared to females (Mean = 19.96, SD = 4.80). Results from t‐tests indicate that there is a statistically significant difference in onset age between men and women who are victimized, with males experiencing stalking victimization onset at a mean age of 22.45 compared to 20.03 for female victims. However, there are no statistically significant differences in onset age between men and women who perpetrate stalking (see Table 4).

Figure 3 Age of onset for the first stalking victimization episode.

Figure 4 Age of onset for the first stalking perpetration episode.

Table 4 Comparison of stalking age of onset by gender

	Male	Female	Mean difference
Stalking victimization	Mean = 22.45	Mean = 20.03	2.423
	SD = 7.658	SD = 4.847	t = −3.184∗, p = 0.002
Stalking perpetration	Mean = 19.69	Mean = 19.96	−0.270
	SD = 3.732	SD = 4.804	t = −0.289, p = 0.773
∗p < 0.05.

Career duration may be defined as the length of time between the first occurrence and the last (or most recent) occurrence of stalking. Investigation of criminal careers has drawn special attention to career duration in light of the substantial policy implications for identifying and selectively incapacitating individuals who pose the highest risk for lengthy careers in crime (Blumstein et al., 1982; Piquero et al., 2003). In the case of stalking offenders, career duration may encompass an individual’s likelihood of pursuing the same victim over a long period of time, or pursuing different victims in a serial or overlapping fashion. Although the survey included questions about the nature of the relationships between victims and offenders for each episode, it is impossible to conclude that subsequent episodes involved the same victims and/or offenders because identifying information was not provided. Thus, when considering career duration for stalking offenders, both the time from the first reported behavior to the last reported behavior and the length of the episode must be taken into account.

Results from the analysis of career duration for stalking offenders and victims are intriguing. First, very few victims and even fewer offenders reported more than one episode of stalking. Only 10.57% of all stalking victims in the sample report experiencing a subsequent episode, and only 0.78% report a third episode. Findings are similar for offenders, with 7.2% of all offenders reporting involvement in a subsequent episode and only 2.7% reporting a third perpetration episode. The data support the conclusion that the vast majority of stalking episodes are singular rather than serial in nature. However, when stalking victimization or perpetration does reoccur, it seems to be clustered in time during the college/early adulthood transition. The mean age of onset for reported stalking victimization is very close for multiple episodes (20.65, 21.86, and 23.14 for the first through third victimizations, respectively), as is the mean age of onset for reported stalking perpetrations (19.88, 21.58, and 20.75 for the first through third perpetrations, respectively). Thus, in terms of the length of time between stalking episodes, career duration appears to be comparatively short.

Episodic duration exhibits a different pattern. Although relatively few participants report multiple victimizations or multiple perpetrations, the episodic duration apparently increases with subsequent episodes. For victimization, the mean episodic duration increases from 6.82 months for the first episode to 7.09 months for the second episode, then again to 8.21 months for the third episode. Perpetration episode duration shows different trends, beginning with a mean of 6.36 months for the first episode, nearly doubling to 11.42 months for the second episode, and declining slightly to 10.17 months for the third episode. The inconsistency in these means may be due to relatively few outliers of multiple‐episode stalking, as the data indicates that second and third episodes are comparatively rare events. These trends are depicted in Figures 5 and 6.

Figure 5 Duration for the first stalking victimization episode.

Figure 6 Duration for the first stalking perpetration episode.

Relationship between Stalking and Other Crimes

The final objective is to identify other compositional or behavioral factors that are significantly associated with stalking victimization and perpetration outcomes. Because stalking is often characterized as a delusional, obsessional, or otherwise dysfunctional relationship‐centric behavior that may be precipitated by rejection (Zona, Sharma, & Lane, 1993), there is some theoretical expectation for an association between stalking and other potentially related or patterned outcomes, such as intimate partner violence and sexual assault (especially in cases where the perpetrator is a current or former intimate partner). Extant research shows that the majority of offenders engage in similar types of crimes (i.e., interpersonal crimes). Therefore, separate models featuring stalking victimization and perpetration as binary outcomes are examined to determine if relationships exist among intimate partner violence victimization and perpetration as well as sexual assault victimization and perpetration. ⁹ Control variables for the models include age, race, gender, ethnicity, and sexual orientation. Of particular interest is the role played by gender. It is often assumed that females disproportionately experience risk for stalking victimization and that males disproportionately engage in stalking perpetration (Harmon et al., 1995; Meloy, 1996; Mullen & Pathè, 1994).

Results from logistic regression analysis show that when stalking victimization is the principal outcome, a significant association is detected for intimate partner violence victimization (B = 0.793) as well as for sexual assault victimization (B = 1.369), indicating that stalking victims are also highly likely to be victims of other interpersonal crimes (see Table 5). Likewise, gender is strongly and significantly associated with stalking victimization (B = 0.636), indicating that stalking victimization is more prevalent for females than males. None of the remaining control variables attain significance, including intimate partner violence perpetration and sexual assault perpetration.

Table 5 Relationship between stalking victimization and other crimes

	B	SE	Wald	Exp(B)
Sexual assault victimization	1.369∗∗	0.141	94.125	3.933
Sexual assault perpetration	0.147	0.340	0.188	1.159
IPV victimization	0.793∗∗	0.174	20.756	2.211
IPV perpetration	0.352	0.275	1.632	1.421
Sex	0.636∗∗	0.161	15.666	1.888
Age	0.006	0.009	0.465	1.006
Race	−0.097	0.174	0.312	0.908
Ethnicity	0.133	0.222	0.363	1.143
Sexual orientation	0.121	0.206	0.348	1.129
Constant	−3.040∗∗	0.519	34.301	0.048
∗∗p < 0.001.

When stalking perpetration is considered as the outcome, the results are somewhat different. Consistent with the findings from the stalking victimization model, a reported history of intimate partner violence perpetration (B = 1.261) and sexual assault perpetration (B = 1.264) are both significantly associated with stalking perpetration. However, a reported history of intimate partner violence victimization (B = 0.653) and sexual assault victimization (B = 1.221) are also significantly associated with stalking perpetration while controlling for age, race, gender, and ethnicity. The only other variable approaching significance in the model was sexual orientation (B = 0.487) (see Table 6). These findings suggest that individuals who experience certain types of interpersonal violence as a victim may also be at risk for interpersonal violence perpetration, although these findings do not explicitly support a causal relationship between these factors.

Table 6 Relationship between stalking perpetration and other crimes

	B	SE	Wald	Exp(B)
Sexual assault victimization	1.221∗∗	0.258	22.481	3.391
Sexual assault perpetration	1.264∗∗	0.384	10.820	3.540
IPV victimization	0.653∗∗	0.270	5.860	1.922
IPV perpetration	1.261∗∗	0.332	14.383	3.529
Sex	−0.138	0.276	0.248	0.871
Age	−0.009	0.016	0.284	0.991
Race	0.123	0.288	0.183	1.131
Ethnicity	−0.219	0.392	0.312	0.803
Sexual orientation	0.487	0.257	3.577	1.627
Constant	−3.569∗∗	0.851	17.598	0.028
∗∗p < 0.001.

When considering which factors relate to the seriousness (rather than simply the prevalence) of stalking, one of the key issues is that the objective seriousness is somewhat dependent upon the respondent’s opportunity to be a victim or offender. Thus, we also examine the relationships between these factors and stalking victimization seriousness in a supplemental OLS regression model. ¹⁰ Results indicate that sexual assault victimization (B = 0.626), intimate partner violence victimization (B = 0.646), and respondent age (B = 0.032) are positively and significantly associated with stalking victimization seriousness. Also, stalking victimization age of onset is negatively and significantly associated with stalking victimization seriousness (B = −0.068). The remaining factors were not significant. These results indicate that those who are also victims of sexual assault and intimate partner violence are more likely to score higher on the stalking victimization seriousness measure, as are respondents who were older and those who reported an earlier age of stalking victimization onset (see Table 7).

Table 7 Relationship between stalking victimization seriousness and other crimes

	B	SE	Beta	t
Sexual assault victimization	0.626∗∗	0.177	0.178	3.531
Sexual assault perpetration	0.206	0.375	0.027	0.570
IPV victimization	0.646∗∗	0.195	0.171	−1.082
IPV perpetration	−0.319	0.296	−0.055	3.322
Sex	−0.335	0.216	−0.078	−1.534
Age	0.032∗	0.015	0.146	2.055
Race	0.042	0.217	0.010	−0.016
Ethnicity	0.276	0.276	0.050	1.019
Sexual orientation	0.291	0.236	0.059	1.217
Age of stalking victimization onset	−0.068∗∗	0.023	−0.209	−2.970
Constant	3.130∗∗	0.691	—	4.532
∗p < 0.05; ∗∗p < 0.001.

Discussion

The objective of this study was to introduce new evidence regarding stalking offending and victimization from a large sample of young adults using a life course perspective. Prior studies generally do not concur about even the most basic attributes of the stalking experience, for example, how common it is, when it begins, how it develops over time, and when it ends. The descriptive information presented in this paper moves the field forward and provides a starting point for theoretical development and further research of career dimensions in stalking victimization and perpetration. Next, we briefly discuss the study’s main findings organized around the various criminal career dimensions.

Participation

The results indicate that almost 27% of respondents report being victims of stalking during their lifetime. The prevalence rate for victimization is substantially higher than the rates reported by a national sample (between 8% and 12% for women and 2% and 4% for men; Tjaden & Thoennes, 1998) and on the highest end of the spectrum among college samples. Recall that between 13–30% of female college students (Fisher et al., 2002; Fremouw et al., 1997) and 11–19% of male college students report stalking victimization (Bjerregaard, 2000; Haugaard & Seri, 2001). The prevalence rate for stalking perpetration indicates that almost 6% of the sample reported perpetrating stalking during their lifetime. Similar to stalking victimization, the current study found a prevalence rate for stalking perpetration toward the uppermost range in comparison with rates from prior research. Stalking perpetration rates previously reported by college students ranges from 1% (Fremouw et al., 1997) to 8% (Haugaard & Seri, 2003).

The gender differences for stalking victimization indicate that many men (almost one‐quarter of the sample) are victims of stalking, but the majority of victims are women. This finding supports prior research that suggests women are more often victims of stalking than men (Harmon et al., 1995; Meloy, 1996; Mullen & Pathè, 1994). However, the substantial number of male victims indicates that stalking cannot be considered a risk for females only. Gender differences in stalking perpetration are perhaps more surprising. Almost 36% of the stalking perpetrators are male whereas 64% are female. This finding indicates that males are less likely to be stalking perpetrators than females, an interesting finding in light of prior research that identifies perpetrators as male (Harmon et al., 1995; Meloy, 1996; Mullen & Pathè, 1994).

A full explication of the gender differences in stalking patterns (e.g., male stalker vs. female victim, and female stalker vs. male victim, etc.) is beyond the scope of this study. However, the results with respect to prevalence by gender are intriguing in that they suggest some same‐sex stalking behavior despite the majority of respondents in the sample reporting that they are heterosexual. Several possible explanations account for these gender differences in prevalence, all of which should be more fully specified and explored in future research. First, it is plausible that females are simply more likely to report involvement in stalking‐related victimization and perpetration behaviors than are males; thus, the differences could be due to male under‐reporting. Second, the range of possible relationships between victim and offender appears to be considerably more diverse than only ex‐intimates. Results indicate that most stalking victims in our sample believed that their stalker was a friend, acquaintance, or stranger rather than a current or former intimate partner or spouse. In cases where heterosexual female offenders are stalking other females, the relationship could extend to new partners of ex‐intimates or to other females who are viewed as competing for the same male. Third, it is possible in some cases that stalking victims did not know who was responsible for stalking them (e.g., the stalker was a stranger to the victim and they never met face‐to‐face). Additional explanations may exist, and we anticipate greater attention to gender differences in stalking victimization and perpetration in future work.

Frequency

Like the more general criminal career research, results indicate almost no difference in frequency of stalking behaviors between genders for victims, suggesting that males and females may share similar victimization experiences. Also, results for perpetration show that males engage in more frequent stalking‐related behaviors, potentially offering insight into the chances for violent escalation, to the degree that violence is a function of opportunity rather than premeditation.

Seriousness

This analysis diverged somewhat in terms of attempting to quantify seriousness based on victims’ and perpetrators’ self‐reported frequency and duration of stalking‐related experiences. The analysis was unconcerned with self‐reported assessment of seriousness, which introduces vast subjectivity based on individual personality, vulnerability, resilience, and a host of other psychological factors. Regarding findings for seriousness, there appears to be few major gender differences for mean scores within victimization and perpetration. This suggests that the experiences of males and females for victimization and perpetration are reasonably similar, at least within this sample. This diverges somewhat from prior studies on stalking that suggest that females are disproportionately victimized in cases of stalking and that males are disproportionately severe in perpetration. Our data do not support that conclusion in the aggregate. It may, therefore, be useful to account for the “classic” stalking pattern in terms of differences in the personal‐psychological makeup (e.g., personality traits, resilience, support networks, etc.) of individuals involved in stalking episodes, rather than as a function of gender‐based sociological constructs.

Results from OLS regression indicate that sexual assault victimization, intimate partner violence victimization, respondent age, and age of stalking victimization onset are significant predictors of stalking victimization seriousness. These results are consistent with expectations given the hypothesized association between intimate partner or other relationship stressors and eventual stalking. At the same time, it is important to note the potential issue of causal ordering. Our assertion is not that intimate partner violence or sexual assault necessarily causes stalking or even precedes it in time, but simply that these types of violence are associated. It is possible, for example, that stalking and intimate partner violence share common psychosocial mechanisms and may also be similar in terms of the likelihood of self‐report. Additionally, this evidence lends support to the notion that an earlier age of onset is associated with more serious stalking. More specifically, at earlier ages of onset, victims report experiencing more frequent and more severe stalking behaviors for longer durations compared to those with a later onset age.

Length of Offending and Victimization

Turning to age of onset, the data show convergence for victims and offenders, with both experiencing a mean age of onset around age 19, although the data indicate that victimization occurs significantly earlier for females than for males. In the context of stalking offenders, the mean age of onset seems slightly higher than age–crime curves for other types of crime, depending on the nature and source of the data being employed. It is noteworthy that the mean age of onset seems higher still in comparison to other self‐report data. However, a substantial proportion of stalking occurred prior to age 18, suggesting that stalking is occurring with regularity in pre‐adolescence, a hypothesis that is, to our knowledge, thus far untested in a systematic fashion using robust samples and appropriate multivariate statistical techniques. Finally, the observed trend in onset after age 25 may be due to the trend in divorce following the first marriage. Pearson’s correlations indicate a moderate and statistically significant positive association between self‐reported divorce and age of stalking onset for both victimization (r = 0.315; p < 0.01) and perpetration (r = 0.373; p < 0.01). One reason for the trend in divorce is the increasing presence of non‐traditional students on college campuses, many of whom return to school later and possess substantially more life experience, including failed marriages. This period represents a substantial and potentially dangerous transition in the life course, particularly when one or both parties maintain contact. As Emerson et al. (1998, p. 312) surmise: “Trouble arises when neither party changes their irreconcilable definitions of the relationship, yet some sort of contact continues. This is the breeding ground for stalking.”

Results for duration indicate that stalking seems to be an isolated occurrence for victims as well as perpetrators. Most individuals involved in stalking reported only a single episode, and most also reported that the episode was limited to approximately one month. The circumstantial evidence supports a view of stalking as a situational act that tends to occur in isolation from other stalking episodes rather than an activity that occurs with relative frequency and long duration over the life course. Identifying patterns of stalking episodes for victims or offenders appears rather difficult, perhaps because these types of behaviors occur infrequently and with a low base rate overall.

Although we make certain inferences about desistance from stalking in addressing duration of stalking careers, desistance is not explicitly discussed with respect to system outcomes such as arrest or civil litigation. This is due in large part to the apparently episodic nature of stalking victimization and perpetration rather than persistent involvement over multiple episodes according to self‐report in this sample. While general criminal careers may range in duration, with many estimates put average career duration between five and twelve years depending on the sample, data source, and length of follow‐up period (see Piquero et al., 2003), it appears that exposure to multiple stalking episodes is a phenomenon with an exceedingly low base rate. A more complete accounting of desistance from stalking will require new attention to this phenomenon, though data collection may prove challenging.

Limitations

The present study is a step forward in furthering knowledge about the etiology of stalking careers over the life course, but some limitations should be noted. First, although this study features an adequate sample size, the overall response rate was relatively low. Also, as with most surveys of any kind, we recognize that there may be some degree of self‐selection among the participants, but as mentioned earlier, it is important to note that the respondents’ gender and racial composition closely reflect that of the entire university population. Second, the current study asked respondents to retrospectively report details of incidents which may have resulted in a failure for respondents to recall certain events, especially with respect to specific dates. While this is certainly a limitation of the data, some research suggests that respondents have reliable recall for salient life events (such as those examined in this study, including stalking, intimate partner violence, sexual assault, etc.) even if the specific details of those events are more problematic (Henry, Moffitt, Caspi, Langley, and Silva, 1994). Third, due to its cross‐sectional nature, the present study could not definitively address the causal order between stalking, intimate partner violence, and sexual assault victimization and perpetration. From the small literature on stalking, it seems reasonable to presume that the relationship between intimate partner violence, sexual assault, and stalking may function in any number of ways, including co‐occurring in time. Further study will be required to fully explore the role of each of these experiences as risk factors for future victimization or perpetration. Fourth, it is also conceivable that stalking may beget stalking. For example, experience as a stalking victim may be associated with the same individual later perpetrating stalking. This overlap is virtually unaddressed in the prior literature on stalking etiology, and most data sources are unequipped to disentangle the potential relationship due to tight focus on the experiences of victims or offenders, but not both. The more general offending/victimization research indicates a broad overlap between offenders and victims (Gottfredson, 1981; Lauritsen, Sampson, & Laub, 1991). Finally, our attempts to quantify seriousness are not the only or necessarily the best of the many options; however, we contend that techniques for objectively comparing stalking experiences in a large dataset must begin somewhere. For example, our derivation of the seriousness equation may be further refined by additional insight from future empirical work on stalking in order to more accurately quantify the degree to which various aspects of the victim–offender relationship influence both perceived and objective seriousness. We invite critique of this method and anticipate further development of investigation into stalking seriousness as the field advances.

With these limitations in hand, the findings from this study suggest several future directions for both research and policy. First, the data support the conclusion that stalking victimization and perpetration needs to be addressed in younger samples. Extant literature on this topic is extremely limited, and policy has only recently taken note of the phenomenon. It will be critical to assess risk of victimization by peers as well as adult offenders. Stalking may not only be a safety concern for younger victims targeted by older perpetrators, but some juvenile stalking offenders may fit a well‐known developmental typology in which more serious life‐course persistent offenders begin earlier and engage in violent crime (Moffitt, 1993). Second, new attention should be devoted to modeling career trajectories for stalking perpetrators and victims. Data from the present study indicate that most perpetrators (and victims) are involved in single episodes, but the nature of repeated offenses should be investigated further. For example, do “career” stalkers maintain a single victim, or is there overlap in selecting victims? Do multiple stalking offenders transition from victim to victim immediately, or is there a time gap between episodes? To what extent are these trends related to age of onset for stalking, and do younger stalkers or their victims experience markedly different (e.g., more serious, longer duration) stalking episodes? Research on stalking etiology has only begun to address these issues, and the limited generalizability of case study data should be augmented with larger, quantitative datasets wherever possible. Additional directions in research may benefit from a full exploration of the relationship between sexual assault, intimate partner violence, and stalking, particularly with regard to establishing a causal mechanism that would inform risk assessment. Principle here will be the further unpacking of the theoretical mechanisms underlying stalking perpetration, victimization, and their overlap. Agnew’s (1992) general strain theory, with its focus on negative life events as sources of strain, is likely to provide a promising framework. Collectively, attention to the aforementioned and related issues will provide much needed descriptive and etiological information on the nature of stalking and the stalking process. It is one of the associated number of crime types that has received scant attention among criminologists, and given its physical and emotional toll, it appears to be the one that requires sustained attention in the years ahead.

Notes

1. California was the first state to recognize stalking as a crime. By 1995, all 50 states and the District of Columbia followed suit and implemented anti‐stalking legislation.

2. The NVAWS rates of stalking victimization are dependent upon the definition of stalking used. The more stringent definition (requiring the victim to feel a great amount of fear) is associated with a lower prevalence rate for both men (2%) and women (8%) whereas the broader definition of stalking (requiring the victim to feel somewhat fearful) yields higher victimization rates for both males (4%) and females (12%).

3. Research regarding web‐based response rates indicates substantially lower rates than for other traditional methods of survey distribution (Couper, 2000). Couper (2000, p. 484) explains that some web‐based surveys obtain less than a 10% response rate for “single invitation surveys.” Despite the relatively low response rate of the current study, the composition of survey respondents closely resembles the overall university population. For example, the entire university population is comprised of 53% females and 65% of the university population is White, 8% Black, 7% Asian, and the remaining 20% represent other races.

4. Survey items are available upon request from the authors.

5. Given the definition of stalking, behaviors must occur at least twice. A frightening, threatening, intrusive and harassing behavior occurring only once cannot be considered stalking.

6. The “more severe” contact behaviors included being “followed, watched, or spied on you,” “stood outside your home, school, or workplace,” “showed up at places uninvited,” and “vandalized your property or destroyed something you loved.”

7. The “less severe” behaviors included distal behaviors such as “Sent you unsolicited letters, written correspondence, or unwanted emails,” “Made unwanted phone calls to you,” “Left unwanted messages for you,” “Left unwanted items for you to find,” “Tried to communicate with you in other ways against your will,” “Sent unwanted messages electronically,” and ‘Posted unwanted messages/pictures to internet websites.”

8. Despite repeated efforts to locate a “model” weight derivation from extant research, we were unable to find a comparable technique that incorporated all of the key behavioral frequencies, episodic duration, and categorical distinctions necessary for our calculation. Thus, the “more severe” behaviors were weighted three times heavier than the “less severe” behaviors in our equation. No attempt is made to distinguish different weights for individual stalking behaviors. As noted later, we consider this to be an exploratory method of quantifying stalking seriousness and encourage refinement in subsequent research.

9. Standard diagnostics (bivariate correlations and variance inflation factors) indicated no collinearity problems among any of the variables used in the multivariate analyses.

10. Following the suggestion of an anonymous reviewer, we also estimated a second supplemental regression model with stalking perpetration seriousness as the dependent variable. In this model, none of the factors were significantly associated with stalking perpetration seriousness. These results indicate that, contrary to findings from the larger stalking literature, “traditional” factors associated with stalking perpetration (including gender) do not appear to have a statistically significant relationship to more serious stalking perpetration in this sample.