American Journal of Criminal Justice

, Volume 38, Issue 3, pp 369–391

Stalking Strain, Concurrent Negative Emotions, and Legitimate Coping Strategies: A Preliminary Test of Gendered Strain Theory

Authors

    • College of Arts and SciencesUniversity of South Florida-Sarasota/Manatee
  • Raymond Paternoster
    • Department of Criminology and Criminal JusticeUniversity of Maryland
Article

DOI: 10.1007/s12103-012-9179-x

Cite this article as:
Ngo, F.T. & Paternoster, R. Am J Crim Just (2013) 38: 369. doi:10.1007/s12103-012-9179-x
  • 353 Views

Abstract

Using data from the Supplemental Victimization Survey of the NCVS and relying on theoretical direction provided by Broidy and Agnew’s gendered strain theory, we examine gender differences in the concurrent emotional responses to a type of strain that has not been examined by GST researchers: stalking. In particular, we assess whether males and females experience similar levels of concurrent negative emotions and whether concurrent negative emotions are similarly associated with legitimate coping resources for males and females. We found the co-occurrence of emotions is more typical among females than males and the impact of concurrent emotions on the strain/non-crime relationship appear to affect females more than males. One notable finding that emerged from our results is that the co-occurrence of emotions can have both proscriptive and precipitating effects on legitimate outcomes. The implications of our findings for theorists and researchers are also discussed.

Keywords

StalkingGeneral strain theoryReactions to victimizationSupplemental Victimization SurveyNational Crime Victimization Survey

Introduction

The general strain theory (GST) of crime and delinquency, crafted by Robert Agnew (1992), is distinguishable from other criminological theories by its explicit recognition of emotions as salient predictors of crime and delinquency. While other criminological perspectives such as social control, social learning, and rational choice ignore the role of emotions in their explanation of criminal and delinquent behaviors, GST argues that emotions are crucial to understanding criminal and delinquent activities. According to GST, strain or stressful events trigger negative emotions such as anger and depression, and these emotional states provide the motivation for corrective action to ease the emotional distress, with crime being one type of coping mechanisms.1 GST further specifies that “only some strained individuals turn to delinquency” (Agnew, 1992, p. 66) even among those who are experiencing negative emotions because there are available legitimate coping resources that they could turn to solve the problem of strain.

Given the emphasis on emotions inherent in GST, scholars and researchers have explored this aspect of the theory in some detail. To date, a vast stream of research has assessed the role of negative emotions in the strain-crime relationship (e.g., Agnew, 1985; Aseltine, Gore, & Gordon, 2000; Bao, Haas, & Pi, 2004; Brezina, 1996; Broidy, 2001; Jennings, Piquero, Gover, & Perez, 2009; Manasee & Ganem, 2010; Mazerolle & Maahs, 2000; Mazerolle & Piquero, 1997; Moon, Morash, McCluskey, & Hwang, 2009; Ostrowsky & Messner, 2005; Piquero & Sealock, 2000). Recent research on GST has also started to examine the co-occurrence of negative emotions in producing deviant behavior (De Coster & Zito, 2010; Ganem, 2010) as well as the gendered nature of emotional responses to strain (Broidy, 2001; Hay, 2003; De Coster & Zito, 2010; Jang, 2007; Jang & Johnson, 2003; Kaufman, 2009).

Findings generated from these studies raise several issues for strain researchers. First, there is evidence that situational-based measures of negative emotions, rather than trait-based measures, are more appropriate as mediating variables in the strain-crime relationship (Manasee & Ganem, 2009; Mazerolle, Piquero, & Capowich, 2003; Moon et al., 2009). Second, although some emotions do predict certain crimes better than others, a number of emotions appear to co-occur in influencing criminal activities (Ganem, 2010). Third, the robust gender difference in crime and delinquency is plausibly due to the gendered expressions of emotional responses to strain (De Coster & Zito, 2010) rather than in the gendered experiences of emotions as theorized by GST (Broidy & Agnew, 1997).

It is noteworthy that prior tests of GST have mainly focused on the link between strain and illegitimate coping mechanisms. This attention is understandable since GST was developed to account for crime and delinquency. However, GST postulates that only some strained individuals turn to illegitimate coping strategies, and that a strained individual is generally likely to adopt illegitimate coping strategies when legitimate coping resources are either unavailable or ineffective. Further, there is evidence that both legitimate and illegitimate coping strategies can be simultaneously adopted in response to distinct strain-caused negative emotions (see Broidy, 2001). Accordingly, exploring the nature of the link among GST’s central variables and legitimate outcomes will not only make a valuable contribution to the growing scholarship on GST but also help advance the status of GST as a general theory.

In this paper, we focus on a type of strain that has not to date been examined within the GST framework—the experience of being stalked. We also highlight the other side of Agnew’s theory—that persons may respond to strain and negative emotions with conventional and legitimate coping mechanisms—and examining gender differences in the concurrent emotional responses to strain. In particular, relying on Broidy and Agnew’s gender/GST perspective and employing six measures of negative emotions, 13 measures of concurrent or co-occurring negative emotions, and five measures of legitimate coping responses, we examine whether: 1) males and females experience similar levels of concurrent negative emotions and 2) concurrent negative emotions are similarly associated with legitimate coping resources for males and females.

Our paper proceeds as follows. First, we begin with a summary of Broidy and Agnew’s gender/GST perspective. We also present the empirical evidence on gender and GST, concurrent negative emotions in the strain-crime relationship, and the gendering of emotional responses to strain. Next, we delineate the nature and extent of stalking and describe our hypotheses and methods. Finally, we discuss our findings and their implications.

Broidy and Agnew’s Gender/GST Perspective

Since its introduction in 1992, Agnew has continued to revise and extend GST.2 One particularly notable extension to GST was Broidy and Agnew’s (1997) application of GST to account for the robust gender difference in crime and delinquency. According to Broidy and Agnew (1997), gender differences in crime and delinquency are due in part to differences in the type of strain experienced by males and females. Specifically, since males are more concerned with material success and extrinsic achievements while females appear to be more concerned with the establishment and maintenance of close relationships and with the meaning/purpose in life, males are hypothesized to be more likely than females to experience financial strain and problems with peers while females are more likely than males to experience network-related stressors (i.e., stressors involving family and friends), gender-based discrimination, and restrictions on their behavior. They argued that the specific types of strain experienced by males (i.e., financial strain and interpersonal conflict) relative to the types of strain experienced by females (i.e., family and friends problems and gender discrimination) are more conducive to crime and delinquency.

Broidy and Agnew (1997) also claimed that males and females differ in their emotional responses to strain, particularly with regard to the most criminogenic emotion—anger. They acknowledge that females are just as likely to respond to stress with anger as are males; however, the anger women experience is qualitatively different from that experienced by men in that the anger experienced by men tends to be characterized by moral outrage or “righteous indignation” that is directed outward (Katz, 1980) while the anger experienced by women is typically accompanied by emotions such as fear, anxiety, guilt, and shame (directed inward): “[t]he moral righteousness of the angry male may propel him into serious violent and property crime, whereas the depression and serious misgivings of the angry female may lead her into more self-destructive forms of deviance” (Broidy & Agnew, 1997, p. 281).

Finally, given the evidence that the relationship between strain/anger and crime is conditioned by factors such as coping resources, coping skills, social support, constraints to delinquent coping, and one's disposition toward delinquency (see Agnew, 2006a), Broidy and Agnew (1997) theorize that gender differences in crime and delinquency are also due to gender differences in these moderating variables. That is, males have higher rates of crime and delinquency than females because they have more opportunities to engage in certain types of crime, have lower types of social control and lower levels of emotional support, are limited in their legitimate coping resources, and are more likely to associate with deviant others.

Prior Research

Gender and GST

A number of researchers have sought to test Broidy and Agnew’s gender/GST propositions using data from national data sets (Agnew & Brezina, 1997; Hoffman & Cerbone, 1999; Hoffman & Su, 1997; Jang, 2007; Jang & Johnson, 2005; Kaufman, 2009; Mazerolle, 1998), as well as from samples of college students (Broidy, 2001; Sharp, Terling-Watt, Atkins, Gilliam, & Sanders, 2001; Sharp, Brewster, & Redhawk Love, 2005), conventional adolescents (Hay, 2003), and detained juveniles (Piquero & Sealock, 2004). A review of the literature on gender and GST reveal mixed results, with some studies reporting no significant gender differences while other studies provide evidence of gendering in the strain-crime process. However, the null findings pertaining to gender differences in the strain-crime association may be due to the cumulative measures of strain employed in prior research. For example, drawing data from the National Youth Survey and employing a composite scale as well as specific measures of delinquency, Mazerolle (1998) reported that while the effects of various strain measures on a composite scale of delinquency did not differ across gender, in crime-specific analyses, negative life events and negative relations with adults significantly predicted violent delinquency among males.

Empirical studies on gender and GST have also revealed some evidence of gender differences in the types of strain experienced. For instance, utilizing a sample of adolescents attending a public school and examining five measures of family strain (unfair discipline, coming from a non-intact family, physical punishment, parental rejection, psychological control), Hay (2003) found physical punishment, the family strain most predictive of delinquency, was experienced at significantly higher levels by boys. Similarly, employing data from Add Health and examining two measures of strain (violent victimization and suicide attempts by family/friends), Kaufman (2009) found males were more likely to be a victim of violence than females but females were more likely than males to have a family member or friend who attempted/completed a suicide over the prior year.

Previous studies have also examined the emotional reactions and behavioral outcomes to strain experienced by males and females. The results generated from these studies reveal that males and females differ in their emotional response to strain in that while males and females are similarly likely to respond to strain with anger, females are more likely to experience non-angry negative emotions than males (Broidy, 2001; Hay, 2003; Jang & Johnson, 2005; Kaufman, 2009; Sharp et al., 2005; Sigfusdottir, Farkas, & Silver, 2004; but see Piquero & Sealock, 2004). The results also indicate that non-angry emotions are negatively associated with crime (Broidy, 2001; Hay, 2003; Hay, 2003; Jang & Johnson, 2005; Sharp et al., 2005; Sigfusdottir et al., 2004; but see Piquero & Sealock, 2004), as well as the fact that females are more likely than males to adopt legitimate mechanisms to cope with strains (Broidy, 2001; Jang, 2007; Jang & Johnson, 2005).

Finally, several studies have also expanded the application of gender and GST by considering deviant outcomes that recognize the gendered nature of deviant choices (Steffensmeier & Allan, 1996). For example, utilizing a sample of college females, Sharp et al. (2001) examined the effect of negative emotions on purging. They found that anger was associated with purging behavior only at high levels of depression. In a related study, Sharp et al. (2005) examined the impact of childhood neglect, abuse and parental hostility on anger, non-angry emotions, criminal behavior, and eating disorders using a sample of male and female college students. Sharp and colleagues uncovered that while anger and non-angry emotions were related to criminal behavior for males and females, non-angry emotions were associated with disordered eating only among females (see also, Kaufman, 2009).

The Co-Occurrence of Negative Emotions

The literature on emotional research suggests that people are predisposed to experience clusters of emotions. Specifically, the evidence indicates that the experience of positive emotions is often bundled or co-occur (e.g., a person who is prone to feel happy is also prone to feel proud and optimistic), and the same is true for negative emotions (Watson, Clark, & Tellegen, 1988). To date, only a handful of research projects have examined the concurrence of emotions in influencing criminal and deviant behavior using the GST framework (see De Coster & Zito, 2010; Ganem, 2010). In a study involving a sample of college students and employing a vignette design, Ganem (2010) examined whether different types of strain cause different types of negative emotions and whether different types of negative emotions cause different types of crime. The author constructed three vignettes to elicit feelings of anger, frustration, and fear. For example, in the anger vignette, study participants were presented with a scenario in which the main actor was pushed by another actor who had begun to flirt with the main actor’s girlfriend. Participants were asked to report how angry they would feel if they were the main character in the vignette. Each vignette also concluded with a criminal act and participants rated the extent to which they would engage in the same criminal act as the vignette main character.

Ganem (2010) discovered that while some emotions predict certain crimes more than others (e.g., anger predicted intentions to hit someone and fear predicted intentions to cut class), emotions do co-occur in influencing criminal behavior (e.g., frustration and fear, in addition to anger, also predicted intentions to hit someone). Ganem (2010) also found that concurrent emotions can have both proscriptive and precipitating effects on criminal involvement. In particular, she found that anger combined with frustration inhibited intentions to hit someone while fear combined with frustration encouraged intentions to cut class.

In a different study, De Coster and Zito (2010) drew from the core proposition posited by Broidy and Agnew (1997) that the concurrence of anger and depression, which is more typical among females than males, is crucial to understanding the gender gap in delinquency. Using a sample of middle school students, they examined the join impact of anger and depression on delinquency and found that the effect of the interaction between anger and depression on delinquency was significant only for males. De Coster and Zito also uncovered that depression exacerbated the effect of anger on delinquency among male students. They contended that the explanation for the gender gap in delinquency/crime may lie in the gendered expression of emotions as proposed by the literature on the sociology of emotions (Simon & Nath, 2004; Stets & Turner, 2006; Shields, 2002) rather than the gendered experience of emotions as postulated by gendered strain theory.

The Gendering of Emotional Experiences and Expressions

Relying on a sample of middle school students, De Coster and Zito (2010) examined whether 1) female anger is more likely than male anger to be accompanied by depression, 2) depression and anger interact in producing delinquency, and 3) depression mitigates the effect of anger on delinquency. They found support for the first and second hypotheses in that anger in females was more likely to be accompanied by depression than anger in males and that both anger and depression mediated the relationship between stress and delinquency among males and females. As for the third hypothesis, they found that depression alters the effect of anger on delinquency among males but not among females but that anger exacerbated instead of mitigated the relationship between stress and delinquency among males.

From their findings, De Coster and Zito (2010) concluded that males and females express their emotions in accordance with definitions of masculinity and femininity (Shields, 2002; Simon & Nath, 2004; Stets & Turner, 2006). That is, males tend to express both anger and depression outwardly and females tend to internalize or talk about these emotions without acting outwardly. De Coster and Zito also maintained that the qualitative distinction in the experience of anger across gender does not explain the gender gap in delinquency and “Instead, gender differences in the expression of the concomitant emotions of anger and depression prove to be relevant because males are more likely than females to express these dual emotions with delinquency” (p. 236).

Extending the work undertaken by Ganem (2010) and De Coster and Zito (2010), we seek to contribute to the gender/GST scholarship by examining gender differences in the concurrent emotional responses to one type of strain that has not been investigated by GST researchers—stalking. We also examine gender differences in conventional and legitimate coping to strain. Before we present our hypotheses and describe our data and methods, we present a brief review on the nature and extent of stalking.

Stalking as Strain

In the U.S., stalking was not considered a crime until 1990 when California became the first state to enact an anti-stalking law (National Institute of Justice, 1996). The passage of the California anti-stalking statute was an instantaneous response to the 1989 brutal murder of Rebecca Schaeffer, an actress who was shot to death by an obsessed fan who stalked her for almost 2 years, and the murders of five women in Orange County who had been stalked by former boyfriends or spouses (Fisher, Cullen, & Turner, 2002). While the legal definition of stalking varies from state to state, the Model Antistalking Code for States3 defines stalking as " a course of conduct directed at a specific person that involves repeated visual or physical proximity, nonconsensual communication, or verbal, written or implied threats, or a combination thereof, that would cause a reasonable person to feel fear” (Tjaden & Theonnes, 1998). Today, all 50 states and the District of Columbia have legislation that addresses the problem of stalking (National Institute of Justice, 1996).

Due to diverse stalking definitions employed by researchers, the reported prevalence of stalking varies from study to study. For instance, results from a nationally representative telephone survey of 8,000 women and 8,000 men indicate that the lifetime prevalence of stalking victimization among the sample women is between 8 % and 12 % and between 2 % and 4 % among the sample men (Tjaden & Theonnes, 1998). On the other hand, the reported prevalence of stalking from studies involving college students is substantially higher, ranging from 6 % to 27 % (Logan, Leukefeld, & Walker, 2000; McCreedy & Dennis, 1996). Specifically, average rates of stalking victimization among female college students vary from 13 % to 30 % (Fisher et al., 2002; Fremouw, Westrup, & Pennypacker, 1997) and from 11 % to 19 % among male college students (Bjerregaard, 2000; Haugaard & Seri, 2001).

The levels of distress reported by stalking victims are quite substantial and the psychological impact is also great. For example, results from a nationally representative study reveal that over 40 % of stalking victims (both male and female) expressed that they were very concerned about their safety, over a quarter reported that they sought counseling during the process, and many resort to extreme actions such as getting a gun (17 %), moving out of town (11 %), and changing addresses (11 %) to cope with their victimization (Tjaden & Theonnes, 1998). There is also evidence that the risk for stalking victimization appears to be highest among individuals who were divorced or separated, and women exhibit a substantially higher risk of stalking victimization relative to men. The risk of stalking seems to diminish with age and with regard to race, non-whites have higher risk of experiencing stalking relative to other racial and ethnic groups. Additionally, approximately 1 in 8 victims with a job reported that they lost time from work because of a fear for their safety or to pursue activities such as obtaining a restraining order or testifying in court, and about 3 in 10 victims had to accrue out-of-pocket for things such as attorney fees, damage to property, child care costs, or moving expenses (Baum, Catalano, Rand, & Rose, 2009).

Relative to research on stalking victimization, empirical studies on stalking perpetration are very limited. To date, the most comprehensive study on stalking perpetration was conducted by Nobles, Fox, Piquero and Piquero (2009) that utilized the life course framework and the criminal career paradigm to examine the onset, duration, and desistance from stalking perpetration and victimization. Drawing on data from a sample of college students attending a major southeastern university, Nobles and colleagues reported several notable findings. First, they found the lifetime prevalence rate for stalking victimization among sample respondents was 27 % and the lifetime prevalence rate for stalking perpetration was approximately 6 %. It is noteworthy that in comparison with rates from prior research, their results fall in the uppermost range (the average rate of stalking victimization for female college students is between 13 % and 30 %, Fisher et al., 2002; Fremouw et al., 1997; the average rate of stalking victimization for male college students is between 11 % and 19 %, Bjerregaard, 2000; Haugaard & Seri, 2001;the average rate of stalking perpetration among college students is between 1 % and 8 %, Fremouw et al., 1997; Haugaard & Seri, 2001). Second, similar to prior research, Nobles et al. (2009) found that the majority of stalking victims were women but unexpectedly, they also discovered that many of the stalking perpetrators (64 %) were women.

Pertaining to the frequency of stalking victimization and perpetration, Nobles et al. (2009) uncovered that while males engaged in more frequent stalking-related behaviors relative to females, there were no gender differences with regard to stalking victimization experiences. They also found the mean age of onset for both stalking victimization and perpetration was 19 years (albeit females appeared to experience stalking victimization earlier relative to males), stalking appeared to be an isolated occurrence for victims as well as perpetrators (with the episode generally lasted for about 1 month), and with regard to victimization and perpetration seriousness, the experiences of males and females were comparable.

The Current Study

In this paper, we focus on two issues: 1) whether males and females experience similar levels of concurrent negative emotions and 2) whether concurrent negative emotions are similarly associated with legitimate coping resources for males and females. Our specific interest is in the relationships among two measures of stalking strain, six measures of situational-based negative emotions, 13 measures of situational-based concurrent negative emotions, and five measures of conventional behavioral responses to strain. Our comprehensive inclusion of diverse emotions was motivated by the fact that a majority of prior tests of GST have included a limited range of negative emotional states and employed trait-based measures of negative emotions. The employment of trait-based or dispositional emotions may not be appropriate in tests of GST because they do not capture the emotional experiences at particular moments in time (Ganem, 2010; Mazerolle et al., 2003).

Research Hypotheses

The first issue that we examine pertains to whether males and females experience similar levels of concurrent negative emotions. According to extant evidence, males and females are similarly likely to respond to strain with anger but females are more likely than males to experience non-angry negative emotions (see Broidy, 2001; Hay, 2003; Jang & Johnson, 2003; Kaufman, 2009; Sharp et al., 2001; Sigfusdottir et al., 2004). Further, there is evidence that the concurrence of anger and depression is more typical among females than males (De Coster & Zito, 2010). Accordingly, our Hypothesis 1 predicts that females experience higher levels of concurrent negative emotions than males.

The second issue that we examine is whether concurrent negative emotions are similarly associated with legitimate coping resources for males and females. Given the evidence that males and females are similarly likely to respond to strain with anger (see Broidy, 2001) and that concurrent negative emotions involving anger (e.g., anger and depression) are related to deviant outcomes (see De Coster & Zito, 2010; Ganem, 2010), our Hypothesis 2 predicts that concurrent negative emotions that involve anger (e.g., anger and fear, anger and depression) are not related to legitimate coping strategies for both males and females. On the other hand, given the evidence that non-angry emotions tend to be related to legitimate coping outcomes (see Broidy, 2001), Hypothesis 3 predicts that concurrent negative emotions that do not involve anger are related to legitimate coping strategies for both males and females.

Data and Methods

Data used for this study come from the 2006 Supplemental Victimization Survey (SVS) of the National Crime Victimization Survey (NCVS; for further details on the NCVS data collection and methodology, see Dugan, 1999). The 2006 SVS was a one-time supplement to the annual NCVS and was designed to measure the prevalence, characteristics, and consequences of nonfatal stalking. While NCVS interviews are normally conducted with each household member age 12 and older, only household members aged 18 or older were given an SVS interview. One notable feature of the SVS is that the term “stalking” did not appear in either the title of the survey or anywhere in the survey until the very last question in which the respondent was asked if he or she considered the experiences of unwanted contact or harassing behavior to be stalking. The SVS was administered to approximately 65,270 individuals with a response rate of 83 %. The SVS contains questions about victims’ experiences of unwanted contact or harassing behavior during the previous 12 months, victim-offender relationship, other crimes and injuries committed against the victim in conjunction with the unwanted contact or harassing behavior, victim and criminal justice responses, and any costs incurred by the victim (for further details on the SVS data collection and methodology, see Baum et al., 2009).

Sample

For the present study, individuals were identified as victims of stalking if they indicated that they experienced any of the following activities on more than one occasion in the past 12 months of the 2006 SVS: being followed or spied on; someone waiting outside/inside their home, school, workplace; someone showing up at places even though that person has no business being there; receiving unwanted phone calls or messages; receiving unwanted letters, emails, or other forms of communication; receiving unwanted items, presents or flowers; and someone posting information or spreading rumors about them both online and offline. It is noteworthy that this study opts for a fairly broad definition of stalking that encompasses the willful and repeated following and harassing of another person (National Criminal Justice Association, 1993).

A total of 1,599 respondents met the above criteria and they comprised the sample for this study. Table 1 shows the demographic characteristics of the sample. According to Table 1, the sample consists of mostly females and the mean age of the sample was 40 years. Further, the majority of the sample was white and approximately one-fifth were divorced or separated.
Table 1

Descriptive statistics and mean-difference test

Variables

Full Sample

Females

Males

N

Mean

SD

N

Mean

SD

N

Mean

SD

Gender

1,599

0.68

0.47

      

Age

1,599

40.43

15.01

1,068

39.63

15.11

531

42.03**

14.70

Race (1 = nonwhite; 0 = white)

1,561

0.12

0.33

1,040

0.13

0.33

521

0.12

0.32

Marital Status (1 = divorced/separated; 0 = married/widowed/never married)

1,599

0.22

0.42

1,068

0.24***

0.43

531

0.19

0.39

Surveillance stalking

1,599

0.68

0.46

1,068

0.68

0.47

531

0.70

0.46

Approach stalking

1,599

0.72

0.45

1,068

0.77*

0.42

531

0.64

0.48

Anxious/Concerned

1,202

0.44

0.50

813

0.50*

0.50

389

0.31

0.46

Annoyed/Angry

1,202

0.71

0.45

813

0.68

0.47

389

0.77**

0.42

Frightened

1,202

0.29

0.46

813

0.38*

0.49

389

0.10

0.30

Depressed

1,202

0.13

0.33

813

0.15*

0.36

389

0.08

0.27

Helpless

1,202

0.18

0.38

813

0.21*

0.41

389

0.12

0.32

Physically Ill

1,202

0.12

0.32

813

0.14*

0.35

389

0.07

0.25

AngerXDepression

1,202

0.10

0.31

813

0.12**

0.33

389

0.06

0.24

AngerXFear

1,202

0.18

0.38

813

0.23*

0.42

389

0.07

0.26

AngerXAnxious

1,202

0.29

0.45

813

0.32*

0.47

389

0.21

0.41

AngerXHelpless

1,202

0.14

0.35

813

0.16*

0.37

389

0.09

0.28

AngerXSick

1,202

0.10

0.30

813

0.12**

0.33

389

0.06

0.24

DepressionXAnxious

1,202

0.10

0.30

813

0.12*

0.33

389

0.05

0.22

DepressionXFear

1,202

0.09

0.28

813

0.11*

0.32

389

0.03

0.18

DepressionXHelpless

1,202

0.08

0.28

813

0.10*

0.31

389

0.04

0.20

DepressionXSick

1,202

0.07

0.26

813

0.10*

0.30

389

0.03

0.16

FearXAnxious

1,202

0.20

0.40

813

0.25*

0.44

389

0.07

0.26

FearXHelpless

1,202

0.11

0.32

813

0.15*

0.36

389

0.04

0.20

FearXSick

1,202

0.08

0.28

813

0.11*

0.31

389

0.03

0.18

HelplessXSick

1,202

0.08

0.28

813

0.11*

0.31

389

0.04

0.19

Change daily activities

1,232

0.36

0.48

832

0.41*

0.50

400

0.27

0.45

Take protective measures

1,232

0.32

0.47

832

0.35**

0.46

400

0.25

0.42

Enlist help of others

1,232

0.59

0.49

832

0.66*

0.47

400

0.44

0.50

Move

1,201

0.11

0.31

813

0.13*

0.34

388

0.05

0.22

Report to police

1,203

0.29

0.45

814

0.33*

0.47

389

0.21

0.41

*p < 0.001; **p < 0.01; ***p < 0.05

Measures

Stalking Strains

The crime of stalking involves a wide range of unwanted or threatening behaviors (see Baum et al., 2009), however, Fisher (2001) suggested that the most common types of stalking behaviors experienced by victims can be categorized into two general types: “surveillance” or “approach.” Surveillance stalking encompasses behaviors such as waiting outside the victim’s residence or workplace or following/driving by where the victim is located, while approach stalking involves behaviors such as leaving unwanted phone messages or spreading rumors about the victim on the Internet. Accordingly, two general categories of stalking experiences, surveillance and approach, were created for the study. Respondents were asked if they had experienced any of the following unwanted or threatening behaviors in the last 12 months: 1) being followed and spied on; 2) someone waiting outside or inside their home/school/workplace; 3) someone showing up at places where they were even though the person had no business of being there; 4) someone leaving unwanted items, presents, flowers; 5) receiving unwanted phone calls and unwanted phone messages; 6) receiving unsolicited letter/e-mails/other form of written communication; and 7) someone posting information or spreading rumors about them online and offline. Each of the responses for the above items were coded 1 = Yes and 0 = No.

Items 1 through 3 were combined to create a variety measure of approach stalking and this measure was then recoded as a dichotomous variable with 1 = the respondent experienced the category of approach stalking and 0 = the respondent did not experience this category of stalking. Similarly, items 4 through 7 were combined to create a variety measure of surveillance stalking and this measure was also recoded as a dichotomous variable with 1 = the respondent experienced the category of surveillance stalking and 0 = the respondent did not experience this category of stalking. The descriptive statistics for the above measures of stalking and all other measures used in the analysis are reported in Table 1.

Negative Emotional States

Six measures of negative emotions conditional on experiencing a stalking event were constructed for the study.4 Respondents were asked if they had experienced any of the following emotions as a result of the unwanted or intrusive behaviors: 1) anxious/concerned (i.e., feeling uncomfortable, uneasy, worried, nervous, or trouble); 2) annoyed/angry (i.e., feeling upset, aggravated, mad, or furious); 3) frightened (i.e., feeling scared, panicked, paranoid, alarmed, or terrified); 4) depressed (i.e., feeling hopeless or sad); 5) helpless (i.e., feeling powerless, frustrated, couldn’t do anything, or no one could help); and 6) physical sickness (i.e., feeling physically ill). The responses for the above question were coded as dichotomous variables with 1 = Yes and 0 = No (see Table 1).

Concurrent Negative Emotional States

Two-way interaction terms were created representing all combinations among the measures of negative emotional states (e.g., anger times depression, anger times anxious, anger times fear, etc.). A total of 13 interaction terms were thus created (see Table 1).

Coping Strategies

Five legitimate coping mechanisms were included in the study. Respondents were asked if they had to change their day-to-day activities, take protective measures, move, enlist the help of others, or report the incident to the police. For the variable “change daily activities,” respondents were asked if they engaged in any of the following activities: take time off from work/school; change or quit a job/school; change the way they went to work/school; avoid relatives, friends, or holiday celebration; change their usual activities outside of work/school; and stay with friends or relatives. “Change daily activities” was coded as a dichotomous variable with 1 = the respondent indicated that s/he engaged in at least one of the above activities and 0 = the respondent reported that s/he did not engage in any of the above activities (see Table 1). The Cronbach’s alpha for this scale is 0.72.

To measure the variable “take protective measures,” respondents were asked if they did any of the following: alter their appearance to be unrecognizable; take self-defense or martial arts classes; get pepper spray; get a gun; get any other kind of weapons; change their social security number; change their email address; change their telephone numbers; install caller ID or call blocking systems; and change or install new locks or a security system. Take protective measures was coded as a dichotomous variable with 1 = the respondent selected at least one of the above items and 0 = the respondent did not select any of the above items (see Table 1). The Cronbach’s alpha for this scale is 0.59.

Eleven items were used to measure the variable “enlist help of others.” Respondents were asked if they did any of the following: enlist the help of friends and family; ask people not to release information about them; hire a private investigator; talk to an attorney; contact victim services, a shelter or help line; obtain a restraining, protection, or stay-away order; talk to a mental health professional; talk to a doctor or nurse; talk to their clergy or faith leader; talk to their boss or employer; contact their building or office security person. This variable was coded as a dichotomous variable with 1 = the respondent selected at least one of the above items and 0 = the respondent did not select any of the above items (see Table 1). The Cronbach’s alpha for this scale is 0.73.

The variable “move” was measured using the question, “Did you move to …” and the response options included 1) a different house/apartment but in the same area, 2) a different city or state, 3) a shelter or safe house, and 4) some other place. This variable was coded as a dichotomous variable with 1 = the respondent selected one of the above items and 0 = the respondent did not select any of the above items. Finally, the variable “report to police” was measured using the question, “During the last 12 months did you or someone else call or contact the police to report any of these unwanted contacts or behavior?” “Report to police” was coded as a dichotomous variable with 1 = Yes and 0 = No (see Table 1).

Control Variables

According to extant evidence, young individuals, women, non-whites, and individuals who were divorced or separated display greater risks of experience stalking (see Baum et al., 2009). Accordingly, age, race and marital status were included as control variables. Age was measured in years and race was also coded as a dichotomous variable with 1 = Non-White and 0 = Caucasian. Marital status was also coded as a dichotomous variable with 1 = Divorced/Separated and 0 = Not Divorced/Separated (see Table 1).

Results

Mean-Difference Tests

The analyses for the present study begin with a series of mean-difference tests across gender for all of the variables in the study. This analysis determines if females experience higher levels of concurrent negative emotions than males (Hypotheses 1). As shown in Table 1, with the exception of two measures, (surveillance stalking and race), all of the variables exhibited a significant mean difference across gender. In particular, on average, male victims in our sample were slightly older than female victims, but female victims were more likely than male victims to be divorced or separated. With respect to gender differences in GST variables, male victims reported higher levels of anger than female victims, but consistent with prior research, female victims expressed higher levels of non-angry emotions than male victims. Similar to non-angry emotions, female victims reported higher levels of concurrent negative emotions relative to male victims. Additionally, female victims were significantly more likely than male victims to have used each of the legitimate coping mechanisms: changing their daily activities, taking protective measures, asking others for help, moving, and contacting the police in response to the unwanted or intrusive stalking behaviors. Finally, the results in Table 1 also provided support for Hypothesis 1 in that the prevalence of all concurrent negative emotions were significantly higher among female victims than male victims (see Table 1).

Impact of Concurrent Emotions on Legitimate Coping Strategies

To determine if concurrent negative emotions that involve anger (e.g., anger times fear, anger times depression) are related to legitimate coping resources for both males and females (Hypothesis 2) as well as if concurrent negative emotions that do not involve anger (e.g., depression times fear; depression times anxious) are related to legitimate coping strategies for both males and females (Hypothesis 3), we estimated two sets of logistic regression models (one for males and the other for females) in which each of the five measures of legitimate coping mechanisms was separately regressed on each of the five measures of concurrent emotional states involving anger, and each of the eight measures of concurrent negative emotions that do not involve anger. The models for female victims are presented in Table 2 and those for male victims are shown in Table 3.
Table 2

Logistic regressions of legitimate coping mechanisms on concurrent negative emotions for females

 

Model 1

Model 2

Model 3

Model 4

Model 5

Change daily activities

Take protective measures

Enlist help of others

Move

Report to police

Annoyed/Angry

0.263 (1.301)

−0.021 (0.979)

−0.001 (0.999)

−0.673 (0.510)

0.517 (1.677)

Anxious/Concerned

−0.296 (0.744)

−0.445 (0.641)

−0.248 (0.781)

0.449 (1.567)

0.606 (1.834)

Frightened

1.240** (3.456)

0.479 (1.614)

1.401**(4.061)

0.471 (1.601)

1.755* (5.785)

Depressed

0.559 (1.750)

1.004 (2.730)

1.030 (2.801)

0.121 (1.128)

1.195 (3.305)

Helpless

0.149 (1.161)

−0.039 (0.962)

0.578 (1.782)

2.081** (8.010)

−0.752 (0.472)

Physically Ill

1.480 (4.394)

0.039 (1.040)

0.910 (2.485)

0.050 (1.051)

0.212 (1.236)

AngerXDepression

0.487 (1.628)

0.837 (2.309)

−0.345 (0.708)

1.084 (2.956)

−0.958 (0.384)

AngerXFear

−0.541 (0.582)

−0.290 (0.749)

−1.065*** (0.345)

−0.842 (0.431)

−6.46 (0.524)

AngerXAnxious

0.097 (1.102)

0.274 (1.315)

0.956*** (2.602)

−0.379 (0.685)

−0.311 (0.733)

AngerXHelpless

0.544 (1.723)

0.846 (2.329)

−0.290 (0.748)

−0.372 (0.690)

1.254***(3.504)

AngerXSick

0.069 (1.071)

−0.198 (0.820)

0.881 (2.414)

−1.534 (0.216)

−0.625 (0.535)

DepressionXAnxious

0.153 (1.165)

−0.569 (0.566)

−0.279 (0.757)

1.765***(5.840)

0.138 (0.871)

DepressionXFear

1.480*** (4.395)

0.256 (1.292)

0.802 (2.230)

−0.786 (0.456)

0.038 (1.039)

DepressionXHelpless

−1.608*** (0.200)

−1.129 (0.323)

−1.208 (0.299)

−0.657 (0.518)

0.474 (1.606)

DepressionXSick

0.099 (1.104)

−0.718 (0.488)

−0.261 (0.770)

0.393 (1.481)

0.769 (2.157)

FearXAnxious

−0.323 (0.724)

0.285 (1.329)

−0.274 (0.761)

1.643**(5.173)

−0.296 (0.744)

FearXHelpless

0.136 (1.145)

−0.002 (0.998)

0.695 (2.003)

−1.254 (0.285)

−0.378 (0.685)

FearXSick

−1.118 (0.327)

−0.080 (0.923)

−1.373 (0.253)

0.659 (1.934)

0.492 (1.635)

HelplessXSick

−0.372 (0.689)

0.465 (1.592)

0.602 (1.826)

−0.130 (0.878)

−0.777 (0.460)

Age

−0.033* (0.968)

−0.029* (0.972)

−0.024* (0.976)

−0.074* (0.929)

−0.019** (0.981)

Non-white

0.186 (1.205)

0.153 (1.166)

−0.189 (0.828)

0.350 (1.420)

0.271 (1.311)

Divorced/Separated

0.642** (1.900)

0.433*** (1,541)

0.334 (1.396)

1.264* (3.540)

0.089 (1.093)

Constant

0.056

0.165

1.106

−0.597

−1.239

Model X2

178.245

88.684

131.996

130.917

101.760

Pseudo-R2

0.20

0.11

0.15

0.15

0.12

Entries are unstandardized coefficients; odds ratio are in parentheses

*p < 0.001; **p < 0.01; ***p < 0.05

Table 3

Logistic regressions of legitimate coping mechanisms on concurrent negative emotions for males

 

Model 1

Model 2

Model 3

Model 4

Model 5

Change daily activities

Take protective measures

Enlist help of others

Move

Report to police

Annoyed/Angry

0.071 (1.073)

−0.007 (0.993)

0.273 (1.315)

−0.549 (0.577)

0.985 (2.679)

Anxious/Concerned

−1.110** (0.330)

0.045 (1.046)

−0.327 (0.721)

−1.532*** (0.216)

0.390 (1.477)

Frightened

0.962 (2.618)

−0.648 (0.543)

1.505(4.505)

0.807 (2.241)

2.163*** (8.695)

Depressed

-------a

1.700 (5.474)

1.624 (5.072)

−0.215 (0.807)

−0.639 (0.528)

Helpless

0.726 (2.068)

0.343 (1.409)

0.884 (2.420)

1.065 (2.900)

1.684*** (5.388)

Physically Ill

-------a

-------a

-------a

−1.265 (0.282)

-------a

AngerXDepression

−1.939 (0.144)

−2.434 (0.088)

−0.285 (0.752)

-------a

-------a

AngerXFear

1.439 (4.216)

-------a

0.671 (1.957)

1.823 (6.191)

−0.031 (0.969)

AngerXAnxious

0.830 (2.293)

0.457 (1.579)

−0.033 (0.968)

1,761 (5.821)

−0.268 (0.765)

AngerXHelpless

0.342 (1.408)

−0.562 (0.570)

−0.075 (0.928)

1.237 (3.446)

−1.648 (0.192)

AngerXSick

2.330**(10.234)

0.006 (1.006)

1.436 (4.203)

-------a

0.099 (1.104)

DepressionXAnxious

0.481 (1.617)

0.078 (1.081)

−1.249 (0.287)

−0.652 (0.521)

0.886 (2.425)

DepressionXFear

-------a

-------a

-------a

-------a

-------a

DepressionXHelpless

1.665 (5.286)

1.479 (4.388)

−0.055 (0.946)

-------a

−1.290 (0.275)

DepressionXSick

0.847 (2.332)

−0.262 (0.770)

0.529 (1.697)

−18.519 (0.000)

−0.925 (0.397)

FearXAnxious

−1.436 (0.238)

-------a

−0.789 (0.454)

−2.654 (0.070)

−0.963 (0.382)

FearXHelpless

−1.924 (0.146)

-------a

−1.185 (0.306)

0.308 (1.361)

−0.034 (0.967)

FearXSick

−0.659 (0.517)

1.306 (3.690)

-------a

-------a

-------a

HelplessXSick

-------a

0.517 (1.677)

0.340 (1.406)

-------a

0.305 (1.357)

Age

−0.015 (0.985)

0.008 (1.008)

−0.007 (0.993)

−0.043 (0.958)

0.001 (1.001)

Non-white

0.171 (0.843)

0.246 (1.279)

−0.228 (0.796)

- 0.120 (0.887)

0.217 (1.242)

Divorced/Separated

0.716*** (2.046)

0.340 (1.405)

0.539 (1.715)

−0.142 (0.867)

0.868** (2.382)

Constant

−0.200

−1.694

−0.103

−0.722

−2.411

Model X2

71.849

18.786

49.372

23.746

38.670

Pseudo-R2

0.17

0.05

0.12

0.06

0.10

Entries are unstandardized coefficients; odds ratio are in parentheses

*p < 0.001; **p < 0.01; ***p < 0.05

aThis variable was excluded from the model due to insufficient number of cases

Pertaining to Hypothesis 2 (concurrent negative emotions involving anger are not related to legitimate coping strategies), the results reveal that among female victims, three of five measures of concurrent negative emotions involving anger (anger times fear, anger times anxious, and anger times helpless) were significantly associated with two legitimate coping resources (enlist help of others and report to police; see Models 3 & 5 on Table 2). However, contrary to our hypothesis the nature of the relationships are inconsistent in that the concurrent emotion of anger and fear tended to decrease the likelihood of legitimate outcome (Model 3 of Table 2) while the concurrent emotions of anger and anxiousness and anger and helplessness tended to increase the likelihood of legitimate outcome (Models 3 and 5 of Table 2). Specifically, whereas female victims expressing anger and fear were less likely (about 65 %) to enlist help of others, female victims expressing anger and anxiousness were more likely (over 150 %) to adopt this type of coping mechanism (Model 3 of Table 2). Still, female victims expressing anger and helplessness were more likely (over 250 %) to contact the police about their victimization relative to female victims who did not express this concurrent emotion (Model 5 of Table 2). It is noteworthy that according to the results in Table 2, anger alone was not significantly related to any of the legitimate coping mechanisms (see Models 1 through 5 of Table 2). However, when anger co-occurred with other negative emotions—in the present case fear, anxious, and helpless—these concurrent emotions exhibited a significant relationship with the legitimate coping resources. These findings are consistent with the interpretation that while anger may provide motivation for action, it requires an additional emotion to provide direction in order to craft a solution to stalking strain.

Among male victims, only one measure of concurrent negative emotions (anger times sick) was significantly related to one legitimate coping mechanism (change daily activities; Model 1 of Table 3). Further, male victims reporting feeling angry and sick (physically ill) were almost ten times more likely to change their daily activities to cope with their victimization relative to male victims who did not experience this type of concurrent emotion (Model 1 of Table 3). Also, similar to the results for female victims, the results for male victims reveal that anger alone was not related to any of legitimate coping mechanisms (see Models 1 through 5 of Table 3b). Yet, when anger co-occurred with the emotion of feeling physically ill (sick) this concurrent emotion demonstrated a significant relationship with one of the legitimate coping mechanism (see Model 1 of Table 3). Given the evidence that several measures of concurrent negative emotions involving anger were significantly related to several measures of legitimate coping mechanisms among male and female victims our Hypothesis 2 appears not to be supported.

Pertaining to the control variables, among female victims, age and marital status exhibited a significant association with the legitimate coping mechanisms in that age significantly decreased the likelihood of adopting all five types of legitimate coping strategies (Models 1 through 5 of Table 2) while marital status significantly increased the probability of adopting three of the five measures of legitimate coping mechanisms (change daily activities, take protective measures, and move). Specifically, relative to older female victims, younger female victims were less likely to change their daily activities (by about 3 %), take protective measures (by about 3 %), enlist help of others (by about 2 %), move (by about 7 %), and contact the police (by about 2 %) in response to their victimization. Conversely, female victims who were divorced or separated were significantly more likely to change their daily activities (by 90 %), take protective measures (by over 50 %), and move (by over 250 %) to cope with their victimization relative to female victims who were not divorced or separated (Models 1 through 5 of Table 2). Among male victims, the only control variable that demonstrated a significant association with the legitimate coping resources was marital status. Male victims who were divorced or separated were significantly more likely to change their daily activities (by over 100 %) and contact the police (by over 130 %) as a result of their victimization (Models 1 and 5 of Table 3).

Discussion and Conclusion

In this paper, we sought to contribute to the growing scholarship on GST by focusing on a type of strain that has to date not been examined within the GST framework, stalking. We also sought to fill the gaps in the literature by highlighting the other side of the theory—that individuals may respond to strain with conventional or legitimate coping strategies—and examining gender differences in the concurrent emotional responses to strain. To the best of our knowledge, our study is the first to employ a nationally representative sample of stalking victims and incorporate a general swath of negative emotional states, co-occurrence emotional states, and conventional coping strategies to strain to assess GST.

Pertaining to the first hypothesis (females experience higher levels of concurrent negative emotions than males), our results provided support for this hypothesis in that female victims in our sample experienced higher levels of all concurrent negative emotions relative to male victims (Table 1). Further, the relationship between gender and concurrent negative emotion was significant for all 13 co-occurrence emotions. Our results appear to lend support for the core proposition posited by Broidy and Agnew (1997) that the co-occurrence of emotions, such as anger and depression, is more typical among females than males.

With regard to hypothesis 2 (concurrent negative emotions that involve anger are not related to legitimate coping resources for both males and females) our results did not provide support for this hypothesis in that we found several measures of concurrent negative emotions involving anger were significantly related to several measures of legitimate coping mechanisms for both male and female victims. More importantly, however, we uncovered that for both male and female victims, while the emotion of anger by itself was not associated with any of the legitimate coping resources, when anger co-occurred with other emotions (i.e., fear, anxious, helpless, and sick), these concurrent emotions prompted corrective action (see Tables 2 and 3). Although our interpretation bears validation with additional research, anger alone may provide motivation for criminal and delinquent offending but in order for a more conventional solution to strain to be constructed anger must co-occur with another emotion that provides guidance and direction to the heightened state of action produced by anger. Hence, pending further investigation, our results seem to confirm the relevance and salience of concurrent emotions in testing GST.

We also discovered that the impact of concurrent emotions on the strain/non-crime relationship appear to affect female victims more than male victims. That is, among female victims, while three of the five measures of concurrent negative emotions involving anger (anger and fear, anger and anxiousness, and anger and helplessness) were significantly associated with two legitimate coping resources (enlist help of others and report to police; Models 3 & 5 on Table 2), among male victims, only one of the five measures of concurrent negative emotions involving anger (anger and sick) was related to one legitimate coping strategies (change daily activities; Model 1 of Table 3). Given the evidence that males and females tend to express their emotions in accordance with definitions of masculinity and femininity and that males are more likely than females to express the dual emotions of anger and depression with delinquency (see De Coster & Zito, 2010), we also encourage future research to continue exploring the effect of concurrent emotional responses to strain across gender, particular the co-occurrence of anger with other negative emotions. Additionally, since our data unfortunately do not include a measure of illegitimate outcome (i.e., crime), future research should attempt to assess the impact of concurrent negative emotions on criminal and non-criminal coping outcomes across gender.

With respect to hypothesis 3 (concurrent negative emotions that do not involve anger are related to legitimate coping strategies for both males and females), our results provided some support for this hypothesis in that several of the measures of concurrent negative emotions that do not involve anger were related to two legitimate coping mechanisms among female victims. However, among male victims, none of these concurrent negative emotions that involve emotions other than anger was associated with any of the legitimate coping strategies. Although our results involve concurrent negative emotions, they seem to correspond with prior findings that females are more likely than males to experience non-angry emotions in response to strain as well as that non-angry emotions tend to increase the likelihood of legitimate outcomes (see Broidy, 2001). We recommend that future research continue to examine the link between strain, concurrent negative emotions, legitimate and illegitimate outcomes across gender.

Another notable finding that emerged from our results was that the co-occurrence of emotions can have both proscriptive and precipitating effects on legitimate outcomes. For instance, among female stalking victims, while the combined effects of depression and fear increased the likelihood of victims changing their daily activities, the combined effects of depression and helplessness decreased the likelihood of them adopting this conventional coping strategy (see Model 1 of Table 2). This finding appears to parallel findings in prior research (see Ganem, 2010) but since our study does not contain measures of illegitimate coping mechanisms, we suggest future research attempt to validate this particular finding with both illegitimate and legitimate outcomes.

It is noteworthy that our study is not without its limitations. Most notably, we have noted that our research does not include a measure of illegitimate coping mechanisms (i.e., criminal behavior). In addition, the measures of negative emotions employed in our study were not assessed in response to particular stalking incidents; rather, they captured the emotional responses to any of the seven stalking experiences. Further, the measures of legitimate coping resources included in our research were all behavioral strategies. Hence, future research should consider incorporating cognitive and emotional coping strategies in tests of GST. Our research also lacks measures of personality traits and since recent editions of GST specify that that individuals high in negative emotionality and low in constraint are more disposed to a criminal coping strategy relative to individuals without such traits conditional variables, future research should strive to include personality characteristics as conditioning variables. Another limitation of our study is the employment of cross-sectional data and hence, we could not unambiguously make an inference about the causal direction of the relationships among stalking strains, negative emotions, and legitimate coping resources. Lastly, since the responses included in the 2006 Supplemental Victimization Survey (SVS) are linked to the National Crime Victimization Survey (NCVS) instrument responses, the SVS possesses the same shortcomings associated with the NCVS; including proxy interviews, false reports, over reporting and/or under reporting, telescoping, and memory failure and decay. In spite of these limitations we hope that our efforts will provide an impetus for additional research on the general strain theory of crime.

Footnotes
1

Note that since its introduction in 1992, GST has been extended and revised several times. For a current description of GST, see Agnew (2006b, 2011).

 
2

For instance, Agnew outlines the specific types of strain likely to lead to crime, incorporates personality traits as conditioning variables, introduces a macro-level component to the theory, and proposes ways to apply GST to study crime and deviance over the life course (Agnew, 1999, 2001, 2002, 2006a, 2006b; Agnew, Brezina, Wright, & Cullen, 2002).

 
3

The Model Antistalking Code for States was mandated by Congress to assist the States in their efforts to respond to stalking. Specifically, in 1992, Congress directed the National Institute of Justice to develop model antistalking legislation that would be both enforceable and constitutional. For a more comprehensive discussion on the development of the Model Antistalking Code for States, see National Institute of Justice (1996).

 
4

Our measures of negative emotions were not assessed in response to particular incidents. Rather, they capture the emotional responses to any of the seven stalking experiences.

 

Acknowledgment

The data set examined for this article was made available by the National Archive of Criminal Justice Data. The National Archive of Criminal Justice Data does not bear any responsibility for the analyses presented here. The authors would like to thank Lynn Addington, Tom Zelenock, and Tim Bynum for their data assistance, and the reviewers for their helpful comments on earlier version.

Copyright information

© Southern Criminal Justice Association 2012