Content uploaded by Karl Anton Roberts
Author content
All content in this area was uploaded by Karl Anton Roberts on Nov 22, 2017
Content may be subject to copyright.
The term stalking has become the label for a pattern of persistent
pursuit and intrusive behavior directed by one person towards an-
other that often continues for several months or even years (1–4).
Stalking is a significant social problem. Prevalence studies using
community samples have found relatively high rates of victimiza-
tion. In the USA (5) a conservative prevalence rate for stalking of
8% for females and 2% for males has been found and higher rates
have been found in other Western countries (6,7). Stalking of ex-
partners appears to be the most common form of stalking (3,5,8–11),
and ex-partners who stalk are more likely than strangers or ac-
quaintances to act violently towards their victims (12–16). In-
creased knowledge of this type of stalking is therefore likely to be
useful in designing relevant social policy and in the prevention of
violence against women. This study was concerned with stalking by
males of females following the breakup of romantic relationships.
Of particular interest were the characteristics of participants and
their former partners that were associated with stalking and which
differentiated stalking from other possible relationship outcomes.
There are a number of possible outcomes of failed intimate rela-
tionships that can be related to the degree one partner attempts to
maintain contact that is unwanted by the other (17). In this regard,
there appears to be a continuum ranging from a “harassment free
relationship breakup” through a breakup followed by “harassment”
to “stalking” (3). Previous research has examined the characteris-
tics of individuals and relationships that end in stalking (17,18), but
only the studies by Coleman (17) and Fremouw et al. (18) have ex-
plicitly compared these characteristics with those of the other pos-
sible failed relationship outcomes. Research of this type is neces-
sary because, in the absence of these comparisons, it remains un-
clear the extent to which characteristics apparently related to stalk-
ing are specific to stalking. In addition, such research will enable
the consideration of the characteristics of former partners that pre-
dict and differentiate between each outcome. This study attempted
to address this by identifying those characteristics that were signif-
icant correlates of and differentiated between “stalking,” “harass-
ment,” and “harassment free” relationship outcomes.
One of the difficulties for studies of stalking is that there is as yet
no agreed definition of stalking (19). Indeed, one of the problems
is that many of the common behaviors carried out by stalkers can
be considered to be routine or even harmless (19). Consider, for ex-
ample, some of the more common stalking behaviors such as mak-
ing telephone calls, sending e-mails, sending letters and gifts (20).
Most of these behaviors are socially acceptable activities; it is only
when they are unwanted and form a persistent pattern that they be-
come more sinister, especially if the victim suffers fear and distress
as a result. All definitions of stalking, therefore, stress the persis-
tent, unwanted, and fear-inducing nature of the behaviors (3). For
the purposes of this study, the definition proposed by Mullen and
colleagues (14,21) was used. Within this definition, behavior is
classed as stalking only when it involves at least ten separate intru-
sions and/or communications and continues for a period of at least
four weeks. This definition of stalking has a number of advantages;
because it relies upon observable behavior, it can be easily opera-
tionalized. Similarly, the conservative nature of this definition is
likely to ensure that behavior meeting the criteria for stalking rep-
resents a long-term pattern of intrusion and pursuit rather than
merely shorter-term innocuous or irritating behavior.
The definition of stalking used here differs from those used in
similar studies such as Coleman (17) or Lewis et al. (18). Cole-
Copyright © 2002 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
Karl A. Roberts,
1
Ph.D.
Stalking Following the Breakup of Romantic
Relationships: Characteristics of Stalking
Former Partners
ABSTRACT: This study investigated female experiences of stalking by former partners following the dissolution of heterosexual romantic rela-
tionships. It aimed to identify those characteristics of former partners that were associated with stalking as opposed to other post-relationship expe-
riences of minor harassment or no-harassment. Three hundred and five female undergraduates (all had experienced dissolution of a heterosexual ro-
mantic relationship) completed a 48-item questionnaire. This assessed characteristics of participants, former partners, and experiences of harassment
following the relationship. One hundred and five (34.4%) participants were classified as stalking victims; ninety-eight (32.1%) as suffering harass-
ment, and 102 (33.4%) as experiencing no-harassment. No differences were found between the three groups in demographic characteristics of par-
ticipants or former partners. Stalking former partners were most likely to have: a history of substance use (alcohol and/or drugs); criminal involve-
ment; violence; mental health problems; difficulties in forming relationships; reacting with inappropriate emotion and jealousy and suspiciousness
of the participant’s relationships with others. These results add to an emerging profile of former partners who are likely to engage in stalking fol-
lowing the dissolution of romantic relationships. The findings are also consistent with explanations of stalking behavior that stress the etiological
importance of attachment difficulties (45).
KEYWORDS: forensic science, forensic psychology, stalking, stalker characteristics, failed romantic relationships
J Forensic Sci, Sept. 2002, Vol. 47, No. 5
Paper ID JFS2001397_475
Available online at: www.astm.org
1
1
Senior lecturer in Forensic Psychology, School of Social Sciences, Univer-
sity of Teesside, Middlesbrough, United Kingdom, TS1 3BA.
Received 30 Nov. 2001; and in revised form 21 Feb. 2002; accepted 21 Feb.
2002; published 10 July 2002.
2JOURNAL OF FORENSIC SCIENCES
man’s definition of stalking focused specifically upon the presence
of fear-inducing threats and victim-perceived malicious intent of
the stalker, while harassment was regarded as unwanted attention
without fear or malicious intent (17). The present study did not
make threats and malicious intent central to the definition of stalk-
ing used. It is argued that explicit threats or perceived malicious-
ness, while undoubtedly inducing fear in the victim, are not strictly
necessary for a set of behaviors to be called stalking. This is sug-
gested because many innocuous behaviors that do not include ex-
plicit threats, such as sending gifts, making telephone calls, etc.,
may in themselves become threatening and induce fear if repeated
on a number of occasions (22). Hence, for the present study any one
or combination of a number of harassing behaviors (including
threats) if carried out repeatedly and causing fear in the victim
would be enough to constitute stalking. Harassment in this study
was defined as unwanted attention that did not achieve Mullen et
al.’s (14,21) criteria for stalking.
Much previous research into the characteristics of stalking per-
petrators has been carried out using forensic populations of stalk-
ers, and, as such, it may not be possible to generalize the results to
all stalkers (19,23). It is noteworthy that the criminal justice system
may become aware of only the minority of stalking incidents. It
seems that only about one third of victims actually report their vic-
timization to the police (6,24). Therefore, forensic samples may
contain a highly selected sub-sample of all stalkers, with the result
that the prevalence of various characteristics may be artificially in-
flated relative to non-forensic samples. It would not be surprising
if those individuals requested by courts to undergo forensic evalu-
ation were most likely to exhibit various psychological problems,
be more violent, or have the most extensive criminal histories (3).
Studies that consider stalking in non-forensic populations to gener-
alize findings are needed. Similarly, studies that seek to identify
those characteristics most associated with stalking in non-forensic
samples are important in attempting to identify and prevent stalk-
ing behavior. The present study attempted to address these issues.
The non-forensic participant sample in this study consisted of fe-
male undergraduate students. This is an approach that has been
used in previous research on stalking (17,18,25–28). This sample
was chosen for a number of reasons. Female students are one of the
groups most likely to have experienced stalking by a former part-
ner (6,17,25,26), and so their use maximizes the likelihood of ob-
taining a sample of stalking victims. Also, while it is recognized
that students present a potential problem for the generalizability of
the results given their restricted demographic characteristics
(17,23), it is argued that obtaining a representative sample of the
population was less important than controlling for potential con-
founding variables. Students provide a good participant sample in
this regard, as they are relatively homogeneous as to social class,
level of intelligence, educational attainment, and age. If findings
similar to those with forensic patients were found with a student
sample, this would serve to add to the generalizability of these find-
ings.
The present study stipulated the minimum duration of the ro-
mantic relationships. Following Lewis et al. (18), a romantic rela-
tionship was defined as one that lasted for a minimum of three
months; this was selected so as to eliminate short-term relation-
ships and casual dates from consideration.
Previous research has found great difficulties in identifying and
obtaining information from community samples of stalkers, as in-
dividuals appear to be reluctant to admit this activity (25). It is,
however, possible to obtain information about stalker characteris-
tics indirectly by questioning their victims (17,26,27). In the con-
text of failed romantic relationships, it is argued that most individ-
uals should be able to shed some light onto the characteristics and
background of their former romantic partner regardless of the out-
come of their relationship. Hence, data concerning former partners
were based upon the responses of the participants, although this is
also a limitation of the study.
Characteristics of Stalking Victims
Studies that have compared stalking victims with non-victims
have generally failed to find differences between them. Bjerregaard
(16) compared student stalking victims and non-victims and found
no differences between them in terms of age, education level,
household income, and marital status. Similarly, Coleman (17), in
a study that compared the demographic characteristics of stalking
victims to those who experienced other outcomes following the end
of an interpersonal relationship (harassment and no-harassment),
found no differences in age, race, and mothers’ and fathers’ level
of education. It is possible that the results of both these studies were
due to the use of undergraduate students, since they exhibit a rela-
tively narrow range of demographic characteristics (16,17). Given
the general lack of differences found in the literature between stalk-
ing victims and non-victims, no specific hypotheses were made
concerning the characteristics of the victims in this study.
Characteristics of Stalking Perpetrators
There appears to be no single demographic profile of stalkers
(3). Indeed, the only consistent finding is that men are more likely
than women to be stalkers (1,2,4,29). Due to the inconsistent find-
ings of previous research, this study made no specific hypotheses
concerning the demographic characteristics of the participant’s for-
mer partners.
Criminal Background
Studies have found that a high proportion of those convicted of
stalking-related offences have a history of involvement in other
criminal activities, and many have a history of violent behavior
(1,2,12,29–32). It was hypothesized that participants adjudged to
be victims of stalking would be most likely to report that their for-
mer partner had a history of criminal involvement and violent be-
havior.
Substance Abuse
Substance abuse has been found to be relatively common among
stalkers, although there is some variation in the results between
studies. Estimates of the prevalence of substance abuse in forensic
samples of stalkers range from 2% (29) to 70% (1). Burgess et al.
(33) found that substance abuse was more common among stalkers
compared to other groups of offenders, while other studies have
found no differences with other types of offender (1,32,34,35).
Stalking victimization of former romantic partners has been found
to be associated with frequent alcohol use by the stalker (26). It was
hypothesized that participants adjudged to be victims of stalking
would be most likely to report that their former partner had a his-
tory of non-prescription drug use and frequent and excessive alco-
hol use.
Mental Health
Studies have attempted to assess the prevalence of mental illness
and psychological disorders within samples of stalkers (23). A
study by the Canadian Department of Justice (31) examined the
records of offenders charged under Canada’s anti-stalking legisla-
tion and found that 14% (N601) had a history of “mental health
problems,” although the nature of these was not specified. Other
research has attempted to identify the prevalence of specific disor-
ders in stalker samples. The most common disorders present within
stalker samples appear to be schizophrenia (2,12,29,34), mood dis-
orders such as dysthymia, major depression or bipolar disorder
(1,2,12,34,35), and personality disorders in particular Cluster B
personality disorders (antisocial, histrionic, borderline, and narcis-
sistic personality disorders (36)). It appears that stalker samples
show a raised incidence of various forms of mental health prob-
lems, although by no means do all stalkers suffer such problems
(10). On the basis of this research, it was hypothesized that partic-
ipants adjudged to be victims of stalking would be most likely to
report that their former partner had a history of various mental
health problems.
Social Skills
Stalkers have been found to score higher on measures of inse-
cure attachment and borderline personality features (18). Lewis et
al. (18) suggested that stalkers would be likely to exhibit a general
pattern of inadequate interpersonal attachments, were more likely
to have difficulty in forming and maintaining relationships, were
likely to be emotionally labile and unstable, and would be most
likely to experience ambivalence regarding their interpersonal re-
lationships. Therefore, within the present study it was hypothesized
that participants adjudged to be victims of stalking would be most
likely to report that their former partner had difficulty forming re-
lationships with others, and that their former partners frequently re-
acted with inappropriate emotion.
Jealousy
Jealousy has been found to be a common characteristic of stalk-
ers (3,37). A number of stalking victims, particularly those stalked
by ex-partners, describe how their relationship with the stalker was
characterized by a history of jealousy-related behaviors such as be-
ing checked up, cross-examined, and accused (3,27,38). White and
Mullen (39) argue that jealousy most commonly appears in the
context of a relationship that is threatened, and it serves to intensify
the concern and increases the contact the jealous person has with
their partner (40,41). Indeed, anger and jealousy have been found
to commonly occur within stalkers who stalk former intimates (42).
Therefore, it was hypothesized that participants adjudged to be vic-
tims of stalking would be most likely to report that their former
partner was jealous and suspicious of their relationships with oth-
ers.
Method
Participants
A non-random sample of convenience consisting of 307 female
undergraduate students drawn from the School of Social Sciences
at the University of Teesside acted as participants in this study. In
order to meet the selection criteria, participants had to have been in-
volved in a heterosexual romantic relationship that had ended and
had lasted for a minimum of three months. Data from two partici-
pants were eliminated since the participants reported same sex for-
mer partners. Thus, the experimental sample consisted of 305 par-
ticipants. The mean age of the sample was 24.63 (standard
deviation 6.17 years). All of the participants classified them-
selves as white and of British nationality. Within the sample as a
whole, the mean duration of the relationships was 25.76 months
(standard deviation 24.56 months).
Stalking was defined according to the definition produced by
Mullen and his colleagues (14,21). It was operationalized by re-
quiring participants to indicate whether they had experienced any
unwanted attention following the end of a prior relationship; to in-
dicate which and how often they experienced a number of harass-
ing behaviors following the termination of the relationship (harass-
ment checklist), and whether the harassment caused them to
experience fear. For the participant to be included in the “stalked”
group, they had to satisfy all four of the criteria below:
1. Any attention from their previous partner following the end of
the relationship was unwanted.
2. The unwanted attention had to have lasted for at least one
month.
3. There had to have been at least ten separate instances of un-
wanted attention.
4. The unwanted attention had to induce fear within the partici-
pant.
Participants who experienced unwanted attention, but whose ex-
periences failed to meet all of the other criteria were placed into the
harassed group. Those participants who did not experience any un-
wanted attention were placed in the no-harassment group. Within
the sample 34.4% (N105) met the criteria for stalking victim-
ization; this is consistent with previous research using female un-
dergraduate samples (16,25). 32.1% (N98) were judged to have
been harassed by their former partner, and 33.4% (N102) were
judged to have experienced no harassment.
Materials
Participants were presented with a 48-item questionnaire. The
questionnaire was divided into several sections designed to obtain
information about participant’s experiences following the breakup
of a romantic relationship, the demographic characteristics of the
participant and their former partner, and other details concerning
the background of the former partner.
Before starting the questionnaire proper, participants answered a
screening question. The purpose of this question was to identify
those participants who had experienced unwanted attention follow-
ing a failed relationship and those who had not. All participants
were required to base their responses to the remainder of the ques-
tionnaire upon the most recent failed relationship they had experi-
enced. Those who answered in the affirmative to the screening
question were asked to base their answers upon the most recent re-
lationship they had experienced that ended in unwanted attention.
Section 1 of the questionnaire began with a control question,
which asked for the sex of the former partner. This question al-
lowed the identification of participants with same sex former part-
ners whose data were eliminated from further analysis. This was
followed by ten items, which asked the participants to consider var-
ious characteristics of their former partners. These included the for-
mer partner’s age at the start and end of the relationship and their
race. Using a yes-no response format the remaining questions in
Section 1 asked participants whether their former partner had a his-
tory of alcohol use, non-prescription drug use, mental health prob-
lems, criminal convictions, violence, difficulty in forming relation-
ships with others, and reacting with inappropriate emotion.
Participants were also asked whether the former partner was jeal-
ous or suspiciousness of their relationships with others.
ROBERTS • CHARACTERISTICS OF STALKING FORMER 3
4JOURNAL OF FORENSIC SCIENCES
Section 2 of the questionnaire asked participants to report vari-
ous biographical details about themselves including their sex, cur-
rent age, age at the start and end of the relationship, race, and the
duration of the relationship.
Section 3 of the questionnaire consisted of 27 items. The first
item asked participants if they had ever experienced fear because of
their former partner’s behavior following the end of the relation-
ship. The next 25 items made up a Harassment Behavior Checklist
(HBC) designed to operationalize the definition of stalking (14,21)
used in this study. The HBC was a self-report checklist describing
various harassing behaviors that a participant may have experi-
enced following the end of the relationship. Participants were re-
quired to indicate, using a yes-no response format, which of the be-
haviors they had experienced. For each item responded to in the
affirmative, participants were then asked to estimate the approxi-
mate number of separate occurrences of the behavior they had ex-
perienced. The behaviors described were taken from the Stalking
Behavior Checklist (SBC; 17). The SBC was designed to examine
experiences of various behaviors occurring after the breakup of a
relationship including common dating behaviors such as receiving
letters, telephone calls, and gifts through to more overtly aggres-
sive behavior such as stealing mail, issuing threats, and destroying
property. The SBC uses a 6-point Likert-type scale, anchored by
never through to once per day or more (17) to estimate the fre-
quency of the various behaviors. The SBC response format differs
from that used by the Harassment Behavior Checklist (HBC) in this
study. There are several reasons for this difference. It was felt that
the response format of the SBC did not allow adequate judgments
to be made regarding the frequency of harassing behaviors. For ex-
ample, it is possible for participants with very different experiences
to give the same response to a given item—a response of once per
day or more could equally be given by someone who had experi-
enced a behavior once per day as well as someone who had experi-
enced the same behavior ten times per day. The format of the HBC
allows Mullen et al.’s criteria for stalking to be operationalized as
it directly assesses the number and frequency of harassing behav-
iors. The yes-no format of the HBC was used to maintain consis-
tency in the response format with the questionnaire as a whole. The
final item on the questionnaire asked participants to estimate how
long any unwanted attention continued.
Procedure
This study was carried out in accordance with the ethical guide-
lines of the British Psychological Society. Participants were first
contacted during undergraduate lectures at the University of
Teesside. They were informed that the experimenter was carrying
out research into their experiences after romantic relationships that
had ended and of the selection criteria for the study. Those who
were interested in taking part and met the selection criteria were in-
vited to remain in the lecture theatre. Participants were then pre-
sented with the questionnaire and were allowed to examine it and
to ask any questions. They were informed that all responses were
anonymous and that their data would be treated as confidential.
They were then allowed to take away the questionnaire and asked
to fill it out in their own time. Completed questionnaires were to be
returned to a sealed box located in the reception area of the Social
Sciences building of the University of Teesside.
Results
Duration of Relationships, Duration of Unwanted Attention, and
Number of Separate Incidents of Unwanted Attention
Table 1 shows the mean duration of the relationships, mean du-
ration of the unwanted attention, and the mean number of separate
harassing incidents reported by the participants. Differences be-
tween the groups with respect to the duration and number of sepa-
rate incidents of unwanted attention are largely a product of the
method by which group membership was allocated. It is interesting
to note, however, that both the stalked and harassed participants
were subject to unwanted attention that lasted for over a year on av-
erage. The stalked group appears to have suffered a substantial
number of separate harassing behaviors.
Relationships ending in stalking appeared to be of longer dura-
tion than other relationship outcomes. One-way analysis of vari-
ance revealed a statistically significant effect of group membership
upon duration of relationship (F(2,304) 6341.68; p0.001).
Post hoc analysis using the Scheffe procedure with alpha set at 0.05
revealed that the duration of the stalked relationships was signifi-
cantly longer than those of harassed or no-harassment outcomes,
while harassed and non-harassed relationship outcomes were not
significantly different from each other.
Demographic Characteristics of Participants and Former
Partners
Race of Participants and Former Partners
All participants classed themselves and their former partners as
being white. This rendered analysis of the racial characteristics of
the participants redundant.
Age-Related Characteristics
Table 2 presents the age-related demographic characteristics of
the sample. The current age means were similar to those obtained
by Coleman (17). In general, the age-related means were similar
for each of the participant groups. Difference in age between for-
mer partners and the participants was calculated by subtracting the
“age of the participant at the start of the relationship” from the “age
of the former partner at the start of the relationship.” The positive
values for the difference in age reveal that on average former part-
ners were older than the participants were. The difference in age for
the stalked group appears to be greater than for the two other
groups.
TABLE 1—Mean duration of relationships, unwanted attention, and number of separate incidents of unwanted attention.
Stalked Harassed No-Harassment Whole Sample
Duration of relationship (months) 34.64* (29.21) 21.47 (21.57) 20.73 (19.11) 25.76 (24.56)
Duration of unwanted attention (months) 19.51 (20.23) 12.17 (20.93) 0 (0) 10.67 (18.58)
Number of separate incidents of unwanted attention 51.33 (55.68) 4.01 (2.37) 0 (0) 24.40 (36.17)
* Indicates statistically significant effect of relationship type on duration of relationship (F (2,304) 6341.68; p0.001).
To explore differences between the stalked, harassed, and no-ha-
rassment groups, the age-related demographic data were subject to
analysis using multivariate analysis of variance (MANOVA).
MANOVA revealed a statistically significant multivariate effect
(F(10,598) 4.77; p0.001). However, follow-up univariate
analysis of variance revealed that there were no statistically signif-
icant effects of participant group (stalked; harassed; no-harass-
ment) upon the age-related demographic characteristics.
Former Partner Characteristics
Table 3 summarizes the responses of participants concerning the
various characteristics of their former partners. Support was found
for all of the experimental hypotheses.
A minority of former partners was reported to have frequently
used alcohol to excess (49%; 150/305). The majority of these were
from the stalked group (55%), followed by the harassed group
(40%), and the no-harassment group (5%). Data analysis using
Pearson’s Chi squared revealed a statistically significant relation-
ship between group membership and frequent alcohol use (
110.55, df 2, p0.001).
A minority of former partners was reported to have a history of
frequent non-prescription drug use (41%; 125/305). The majority
of these were from the stalked group (65%), followed by the ha-
rassed group (30%), and the no-harassment group (5%). Data anal-
ysis using Pearson’s Chi squared revealed a statistically significant
relationship between group membership and frequent non-pre-
scription drug use (
2
106.28, df 2, p0.001).
A minority of former partners was reported to have suffered
from mental health problems (41/305; 13.4%). Of those suffering
from mental health problems, most were from the stalked group
(51.2%), followed by the harassed group (37%), and the no-ha-
rassment group (13%). Data analysis using Pearson’s Chi squared
revealed a statistically significant relationship between group
membership and mental health problems (
2
11.32, df 2, p
0.001).
A minority of former partners was reported as having criminal
convictions (52/305; 17%). The majority of these (46%) were from
the stalked group followed by the harassed group (40%), then the
no-harassment group (17%). Data analysis using Pearson’s Chi
squared revealed a statistically significant relationship between
group membership and former partner’s criminal history (
2
11.32, df 2, p0.01).
A minority of former partners was reported to have a history of
violence (100/305; 33%). The majority of these (86%) were from
the stalked group followed by the harassed group (8%), then the no-
ROBERTS • CHARACTERISTICS OF STALKING FORMER 5
TABLE 2—Age-related characteristics of the participants (mean ages in years, standard deviation in brackets).
Stalked Harassed No-Harassment Whole Sample
Participants
Current age 24.89 (6.68) 24.78 (5.81) 23.63 (5.95) 24.43 (6.17)
Age at start of relationship 20.59 (5.66) 21.44 (5.76) 21.09 (5.40) 21.03 (5.60)
Age at end of relationship 23.06 (6.01) 23.31 (6.93) 22.69 (5.91) 23.01 (5.97)
Former Partners
Age at start of relationship 22.21 (5.46) 22.36 (5.66) 22.02 (5.52) 22.2 (5.53)
Age at end of relationship 25.18 (5.70) 24.04 (5.53) 23.74 (5.92) 24.33 (5.73)
Age Difference
Age difference between partners 1.63 (4.05) 0.91 (4.85) 0.93 (7.02) 1.17 (5.44)
(Calculated by subtracting participants from former partners age at the start of the relationship.)
TABLE 3—Distribution of former partner characteristics by relationship outcome.
Overall
Percentage
Stalked Harassed No-Harassment Exhibiting
Former Partner Characteristics (N 105) (N 98) (N 102) Characteristic
2
p
Frequent alcohol use 55% (82) 40% (60) 5% (8) 49% (150) 110.55 0.001
Frequent non-prescription drug use 65% (81) 30% (37) 5% (7) 41% (125) 106.28 0.001
Mental health problems 51% (21) 37% (15) 12% (5) 13% (41) 11.32 0.001
Criminal convictions 46% (24) 40% (21) 14% (7) 17% (52) 11.32 .01
Violence 86% (86) 8% (8) 6% (6) 33% (100) 175.45 0.001
Difficulty in forming relationships 73% (30) 24% (10) 3% (1) 13% (41) 35.15 0.001
Frequently reacting with inappropriate emotion 73% (90) 17% (21) 10% (13) 41% (124) 136.31 0.001
Jealous of relationships with others 49% (98) 38% (80) 13% (27) 67% (205) 118.57 0.001
Suspicious of relationship with others 51% (96) 41% (78) 7% (14) 62% (188) 47.08 0.001
Note: Total sample size 305, Pearson
2
statistic was used and df 2 in all cases. The numbers in brackets in columns 2–4 represent the n numbers
for each cell.
The percentages in columns 2–4 represent the number of former partners within a subgroup exhibiting a particular characteristic (n) relative to the total
number of participants within the sample as a whole who exhibited that characteristic.
6JOURNAL OF FORENSIC SCIENCES
harassment group (6%). Data analysis using Pearson’s Chi squared
revealed a statistically significant relationship between group
membership and the former partner’s criminal history (
2
175.42, df 2, p0.001).
A minority of former partners was reported as experiencing dif-
ficulty in forming relationships (41/305; 13.4%). The majority of
these were from the stalked group (73.2%) followed by harassed
group (28.4%), then the no-harassment group (2.4%). Data analy-
sis using Pearson’s Chi squared revealed a statistically significant
relationship between group membership and the former partner’s
experiencing difficulty forming relationships (
2
35.15, df 2,
p0.001).
A minority of former partners was reported as frequently react-
ing with inappropriate emotion (124/305; 40.7%); the majority of
these (72.6%) were from the stalked group followed by harassed
group (16.9%), then the no-harassment group (10.5%). Data analy-
sis using Pearson’s Chi squared revealed a statistically significant
relationship between group membership and the former partner re-
acting with inappropriate emotion (
2
136.31, df2, p0.001).
The majority of former partners was reported as being jealous of
participant’s relationships with others (205/305; 67%); the major-
ity of these (49.6%) were from the stalked group followed by ha-
rassed group (38%), then the no-harassment group (13%). Data
analysis using Pearson’s Chi squared revealed a statistically signif-
icant relationship between group membership and the former part-
ner being jealous of participants’ relationships with others (
2
118.57, df 2, p0.001).
The majority of former partners was reported as being suspicious
of participant’s relationships with others (188/305; 62%), the ma-
jority of these (51%) were from the stalked group followed by ha-
rassed group (41%), then the no-harassment group (7%). Data anal-
ysis using Pearson’s Chi squared revealed a statistically significant
relationship between group membership and the former partner be-
ing suspicious of the participant’s relationships with others (
2
47.08, df 2, p0.001).
Discussion
Approximately 34% of the sample could be classified as having
suffered stalking victimization. A further 32% suffered harassment
from their former partner. This is a surprising finding given that
this study used a particularly strict criterion to define stalking of ten
separate incidents of unwanted attention occurring for at least four
weeks (14,21). This finding is consistent with prevalence rates
identified by other studies of undergraduate students (25). The cur-
rent findings add further to the growing evidence that stalking vic-
timization is particularly common among female student popula-
tions (6,10,25,26).
No clear demographic profile for the victims or perpetrators of
stalking in age-related characteristics were found. These findings
were consistent with other research (10,17). Age-related criteria
are of limited value in attempting to differentiate between relation-
ships that result in stalking, harassment, or no-harassment, al-
though age effects may have been masked by the relatively re-
stricted age range of the participant sample.
The duration of the relationship did appear to differentiate be-
tween the three relationship outcomes, with relationships ending in
stalking being generally longer and relationships ending with ha-
rassment and no-harassment relationships being shorter and of sim-
ilar duration. The reason for this finding is not clear. Previous re-
search has reported a range of different relationship durations
associated with stalking (3,10,42), although relationship duration
has seldom been compared for different relationship outcomes as
in this study. It is possible that relationships of longer duration give
rise to more intense attachments between the respective partners
than do shorter relationships. When such relationships end, they
may be accompanied by a greater sense of loss, anger, or frustra-
tion, which may stimulate one partner to maintain (unwanted) con-
tact with the other. This is an area for further research.
Support was obtained for all the hypotheses related to the char-
acteristics of former partners. Former partners of stalking victims
were more likely to have a history of criminal involvement, violent
behavior, frequent drug use, frequent excessive alcohol use, men-
tal health problems, difficulty in forming relationships with others,
inappropriate emotional responses, and jealousy and suspicious-
ness of their partner’s relationships with others. This over-repre-
sentation of stalkers in groups with these various characteristics is
consistent with research that has been carried out with forensic
samples (3). Therefore, some of the characteristics of stalkers iden-
tified in forensic samples do appear to generalize to non-forensic
groups.
The reported incidence of the various former partner character-
istics was not particularly high. For the sample as a whole (N
305), the incidence of any one characteristic was greater than
50% only for jealousy (67%) and suspiciousness of the partici-
pant’s relationships with others (62%). For other characteristics,
the highest incidence was frequent excess alcohol use (49%), while
the lowest incidence was for mental health problems (13%). These
generally low incidence levels are likely to reflect the relatively
low incidence of many of these characteristics within the general
population.
Given the low incidence rates of these characteristics within the
sample as a whole, it is important to examine which, if any of them,
were particularly common characteristics of the stalking as com-
pared with the other types of former partners. The data were exam-
ined to identify those characteristics attributed to the majority of
former partners within each relationship outcome (defined as char-
acteristics with an incidence of greater than 50% within each
group).
Six former-partner characteristics were identified that were at-
tributed to the majority of the stalkers. These were jealousy (93%;
98/105) and suspiciousness (91%; 96/105) of the participant’s re-
lationships with others; frequent excess alcohol use (78%; 82/105);
frequent non-prescription drug use (77%; 81/105); a history of vi-
olent behavior (82%; 86/105); and reacting with inappropriate
emotion (86%; 90/105). These findings are consistent with previ-
ous research carried out with both forensic and non-forensic sam-
ples (1,2,12,18,26,29–32,43,44,45).
Former partners from relationships ending in harassment also
exhibited a high incidence of two of these characteristics, jeal-
ousy (82%; 8/98) and suspiciousness (80%; 78/98) of relation-
ships with others, while none of the characteristics were attributed
to the majority of no-harassment partners. The high incidence of
jealousy and suspiciousness for both the stalking and harassing
former partners suggests that these are common characteristics of
partners who indulge in some form of post-relationship harass-
ment, of which stalking is perhaps an extreme example. This find-
ing is consistent with previous research stressing the importance
of these characteristics in the aetiology of stalking and harassment
(3,27,37–39,44).
Meloy (46) hypothesized that the fundamental deficit in stalking
and harassment is an insecure attachment style resulting from patho-
logical early attachments. Although this study has not explicitly as-
sessed the attachment style of the participants, it is argued that the
present findings are consistent with this view. Bowlby (47) defined
attachment as a strong enduring affectional bond between individ-
uals. Initially attachments develop during childhood and are be-
tween parent and child; attachment behavior, however, persists
throughout the lifespan, and so in later life attachments are also
formed between adults (47,48). Bowlby (47) argued that stable early
attachments contributed to an individual’s healthy development;
however, pathological early attachment patterns often lead to vari-
ous maladaptive personality traits and various forms of psy-
chopathology (49). Situations perceived to be threats to the exis-
tence of a relationship elicit various so-called attachment behaviors
that are designed to safeguard the bond between two individuals
(47). There are various individual differences in attachment behav-
iors, and some individuals exhibit maladaptive attachment styles re-
sulting from a pathological attachment history (46,47). Of particu-
lar interest in the context of stalking are individuals with an insecure
attachment style (46). Such individuals have a tendency to exhibit
maladaptive attachment behaviors in the context of interpersonal re-
lationships such as a lack of trust, jealousy, and suspiciousness of a
partner (18).
The present study found evidence of maladaptive attachment be-
havior consistent with an insecure attachment style for the stalking
and harassing former partners; this was in the form of jealousy and
suspiciousness. These findings are, therefore, suggestive of an as-
sociation between insecure attachment style and harassment of a
former partner, consistent with Meloy’s (46) hypothesis. A patho-
logical attachment history may also be an explanation for other
findings of this study. The majority of stalking former partners ex-
hibited a history of non-prescription drug use, frequent excess al-
cohol use, violence, and reacting with inappropriate emotion. In
previous research, these behaviors have been found associated with
a pathological attachment history (49, 50). Therefore, it may be the
case that various forms of anti-social behavior, including harass-
ment and stalking of a former partner, are a product of a patholog-
ical attachment history (46,49).
It is interesting to consider how attachment history may impact
upon future relationship behavior including stalking and harass-
ment. Bowlby (51) argued that individuals develop cognitive
schema that he termed working models of the attachment figure
(who they are, where they are, their expected responses) and of
themselves (perception of how acceptable or unacceptable they are
to the attachment figure). Bretherton (52) argued that the nature of
the attachment between caregiver and child directly affects the con-
tents of the working models. Secure attachments in which the care-
giver is responsive to the child’s needs for comfort and protection
while allowing the child to independently explore the environment
give rise to a working model of the self as valued and self-reliant.
Pathological attachment patterns where the caregiver consistently
fails to meet the child’s needs for comfort and independence give
rise to working models of the self as incompetent and unworthy of
the attachment figure. An individual who feels unworthy and in-
competent might be expected to have difficulties forming and
maintaining future interpersonal relationships. Within a relation-
ship, such an individual may experience regular feelings that the re-
lationship may be under threat, perhaps because they feel unworthy
of their partner. One tangible source of threat to the relationship
may be other individuals. Jealousy and suspiciousness may be
characterized as maladaptive attachment behaviors deployed in re-
sponse to the perception of threat and designed to maintain the re-
lationship (47). Such behaviors are likely to be deployed by those
with pathological attachment histories (47). Feelings and behaviors
indicative of suspiciousness and jealousy of a partner’s relation-
ships with others might therefore be a common feature of the rela-
tionship. Strong emotional reactions resulting from perceived
threats to the relationship may also be evident and would perhaps
be experienced by the other partner as inappropriate emotional re-
actions. If the relationship ends, the feelings of incompetence and
unworthiness may be compounded within an individual. In some
cases, this may lead them to attempt to contact the former partner
with the aim of maintaining proximity to the former attachment fig-
ure and of restarting the relationship (53). If these contacts are
repetitive, continue for some time, and are unwanted by the former
partner, they may be classed as harassment or even stalking. This
model might, therefore, explain why former partners of stalkers re-
port experiencing jealousy, suspiciousness, and inappropriate emo-
tional reactions from their partner during the relationship.
There are a number of issues with the design of this study that
merit consideration. The generalizability of the findings is limited
as the participant sample consisted exclusively of white female un-
dergraduate students who had heterosexual relationships with
white males. None of the participants were involved in marital re-
lationships. While there were a number of advantages in using this
sample as discussed previously, they represented a narrow range of
demographic characteristics. In some ways this may account for
the lack of differentiation in the demographic characteristics be-
tween the three participant groups. It is also likely to account for
the relatively low incidence of some of the relationship and former
partner characteristics.
The data rely upon the self-reports of participants. This may be
subject to various biases, which may impact upon the validity of
the results. The information in this study was obtained retrospec-
tively. Therefore, faulty recall is problematic in this type of design.
Some participants may have inaccurately recalled or even forgot-
ten aspects of their relationship or their former partner, especially
if the relationship occurred some years previously. Different results
may have been obtained using a prospective study of the breakup
of dating relationships.
There is an inherent potential for bias in asking participants to
consider relationships for which they experienced unwanted atten-
tion from a former partner. If this attention provoked fear and
anger, there may be a bias against the former partner, which could
result in exaggerated and overtly negative responses.
This study did not examine marital relationships, relationships
between same-sex partners, or partners of different races. Future
studies might consider stalking in relationships between same sex
and mixed race partners, or the characteristics of marital relation-
ships that end in stalking. Research of this sort would further ex-
tend knowledge of the characteristics of relationships that end in
stalking.
This research has focused upon victims of stalking. There is a
need for future research to examine perpetrators of stalking, espe-
cially using non-forensic samples. Some initial work has been done
in this regard (18, 27).
Summary and Conclusions
This study has examined various characteristics of former ro-
mantic partners who were involved in relationships that followed
by stalking, harassment, or no-harassment. Using strict criteria
(2,3) to define stalking, a surprisingly high level of stalking was
found in the sample (34%), especially as this was a non-forensic
sample. This figure is consistent with rates of stalking found in
other studies (25).
Support was obtained for all of the experimental hypotheses, and
the results of the present study are generally consistent with previ-
ous research. These results add to an emerging profile of former
partners who are likely to engage in stalking following the dissolu-
ROBERTS • CHARACTERISTICS OF STALKING FORMER 7
8JOURNAL OF FORENSIC SCIENCES
tion of romantic relationships. The findings are also consistent with
explanations of stalking behavior that stress the etiological impor-
tance of attachment difficulties (10,47).
Acknowledgments
The author would like to thank two anonymous referees for their
helpful comments.
References
1. Meloy JR, Gothard S. A demographic and clinical comparison of obses-
sional following and offenders with mental disorders. Am J Psychiatry
1995;152:258–63.
2. Mullen PA, Pathe M. Stalking and the pathologies of love. Australian
New Zealand J Psychiatry 1994;24:469–77.
3. Mullen PA, Pathe M, Purcell R. Stalkers and their victims. Cambridge:
Cambridge University Press, 2000.
4. Zona MA, Sharma KK, Lane J. A comparative study of erotomanic and
obsessional subjects in a forensic sample. J Forensic Sci 1993;38(4):
894–903.
5. Tjaden P, Thoennes N. Stalking in America: findings from the National
Violence Against Women Survey. Washington, DC: National Institute
of Justice and Centres for Disease Control and Prevention, 1998.
6. Budd T, Mattison J. Stalking: findings from the 1998 British Crime Sur-
vey. Research Findings No.129. London: Home Office, 2000.
7. Australian Bureau of Statistics. Women’s safety, Australia. Canberra:
Commonwealth of Australia, 1996.
8. Roberts AR, Dziegielewski SF. Assessment typology and intervention
with the survivors of stalking. Aggression and violence behavior 1996;
1(4):359–68.
9. Kong R. Criminal harassment. Juristat 1996;16(12):1–13.
10. Meloy JR. The psychology of stalking. Clinical and forensic perspec-
tives. Academic Press, Harcourt Brace and Company, 1998.
11. Thomas KR. How to stop the stalker: state anti-stalking laws. Criminal
Law Bulletin 1993;29(2):124–36.
12. Kienlen KK, Birmingham DL, Solberg KB, O’Regan J, Meloy JR. A
comparative study of psychotic and non-psychotic stalking. J Am
Academy Psychiatry Law 1997;25:317–34.
13. Walker LEA, Meloy JR. Stalking and domestic violence. In: Meloy JR.
The psychology of stalking: clinical and forensic perspectives. San
Diego: Academic Press 1998:139–61.
14. Mullen PE, Pathe M, Purcell R, Stuart GW. A study of stalkers. Am J
Psychiatry 1999;156:1244–9.
15. Farnum F, James D, Cantrell P. Association between violence, psy-
chosis, and relationship to victim in stalkers. The Lancet 2000;355:199.
16. Bjerregaard B. An empirical study of stalking victimization. Violence
and Victims 2000;15( 4):389–407.
17. Coleman FL. Stalking behavior and the cycle of domestic violence. J In-
terpersonal Violence 1997;12:420–32.
18. Lewis S, Frenmouw WJ, Del Ben K, Farr C. An investigation of the psy-
chological characteristics of stalkers: empathy, problem-solving, attach-
ment and borderline personality features. J Forensic Sci 2001;46(1):
80–4.
19. Sheridan L, Davies GM. Stalking: the elusive crime. Legal and Crimi-
nological Psychology 2001;6(2):133–48.
20. Mullen PA, Pathe M. Stalkers and their victims. Psychiatric Times
2001;8(4):1–8.
21. Pathe M, Mullen PE. The impact of stalkers on their victims. British J
Psychiatry 1997;170:12–17.
22. Cupach WR, Spitzberg BH. Obsessive relational intrusion and stalking.
In: Spitzberg BH, Cupach WR, editors. The dark side of close relation-
ships. Mahwah, NJ: Lawrence Erlbaum Associates 1998:233–63.
23. Westrup D, Fremouw WJ. Stalking behavior: a literature review and
suggested functional analytic assessment technology. Aggression and
Violent Behavior 1998; 3:255–74.
24. Westrup D, Fremouw WJ, Thompson RN, Lewis SF. The psychological
impact of stalking on female undergraduates. J Forensic Sci 1999;44(3):
554–7.
25. Fremouw WJ, Westrup D, Pennypacker J. Stalking on campus; the
prevalence and strategies for coping with stalking. J Forensic Sci 1997;
42(4):666–9.
26. Logan TK, Leukefeld C, Walker B. Stalking as a variant of intimate vi-
olence: implications from a young adult sample. Violence and victims
2000;15(1):91–111.
27. Langhinrischen-Rohling J, Palera RE, Cohen J, Rohling ML. Breaking
up is hard to do: unwanted pursuit behaviors following the dissolution of
a romantic relationship. Violence and Victims 2000;15(1):73–90.
28. Sheridan L, Davies GM, Boon JCW. Stalking; perceptions and preva-
lence. J Interpersonal Violence 2001;16(2):151–67.
29. Harmon RB, Rosner R, Owens H. Obsessional harassment and erotoma-
nia in a criminal court population. J Forensic Sci 1995;40:188–96.
30. British Colombia Attorney General. The report of the criminal harass-
ment unit. Part II: The nature and extent of criminal harassment in British
Columbia. Vancouver: Attorney General, 1995.
31. Department of Justice Canada. A review of Section 264 (criminal ha-
rassment) of the Criminal Codes of Canada. Ottawa, 1996.
32. Lyon DR. The characteristics of stalkers in British Columbia: a statisti-
cal comparison of persons charged with criminal harassment and persons
charged with other Criminal Code offences[dissertation]. Burnaby:
British Columbia. Simon Fraser University, 1998.
33. Burgess AW, Baker T, Greening D, Hartman C, Burgess A, Douglass JE,
et al. Stalking behaviors within domestic violence. J Family Violence
1997;12:389–403.
34. Sandberg DA, McNeil DE, Binder RL. Characteristics of psychiatric pa-
tients who stalk, threaten, or harass hospital staff after discharge. Am J
Psychiatry 1998;155:1102–5.
35. Schwartz-Watts D, Morgan DW. Violent verses non-violent stalkers. J
Am Academy Psychiatry Law 1998;26:241–5.
36. McGuire B, Wraith A. Legal and psychological aspects of stalking: a re-
view. J Forensic Psychiatry 2000;11(2):316–27.
37. Silva JA, Derecho DV, Leong GB, Ferrari MM. Stalking behavior in
delusional jealousy. J Forensic Sci 2000;45(1):77–82.
38. Dutton DG, Van Ginkel C, Landolt MA. Jealousy, intimate abusiveness
and intrusiveness. J Family Violence 1996;11:411–23.
39. White GE, Mullen PE. Jealousy: theory research and clinical strategies.
New York: Guilford Press, 1998.
40. Mullen PE. A phenomenology of jealousy. Australian and New Zealand
J Psychiatry 1990;24:17–28.
41. Mullen PE. Jealousy: the pathology of passion. British J Psychiatry
1991;158:593–601.
42. Davis KE, Frieze IH. Research on stalking: what do we know and where
do we go? Violence and Victims 2000;15(4): 473–88.
43. Brewster MP. Stalking by former intimates; verbal threats and other pre-
dictors of physical violence. Violence and Victims 2000;15(1):41–54.
44. Davis KE, Ace A, Andra M. Stalking perpetrators and psychological
maltreatment of partners: anger jealousy, attachment insecurity, need for
control, and breakup context. Violence and Victims 2000;15(4):407–26.
45. Hall DM. The victims of stalking. In: Meloy JR. The psychology of
stalking: clinical and forensic perspectives. San Diego: Academic
Press,1998:113–37.
46. Meloy JR. Stalking (obsessional following): a review of some prelimi-
nary studies. Aggression and Violent Behavior 1996;1:147–62.
47. Bowlby J. Attachment and Loss: Vol. III: Loss, sadness and depression.
New York: Basic Books, 1980.
48. Ainsworth MDS. Attachments beyond infancy. American Psychologist
1989;44(4):709–16.
49. Meloy JR. Violent attachments. Northvale, NJ: Aronson 1992.
50. Raine A. The psychopathology of crime: criminal behavior as a clinical
disorder. San Diego: Academic Press, 1993.
51. Bowlby J. Attachment and loss: Vol. II. Separation, anxiety and anger.
New York: Basic Books, 1973.
52. Bretherton I. The origins of attachment theory: John Bowlby and Mary
Ainsworth. Developmental Psychology 1992;28(5):759–75.
53. Kienlen KK. Developmental and social antecedents of stalking. In:
Meloy JR. The psychology of stalking: clinical and forensic perspec-
tives. San Diego: Academic Press 1998:51–67.
Additional information and reprint requests to:
Dr. Karl Roberts
Senior Lecturer in Forensic Psychology
School of Social Sciences
University of Teesside
Middlesbrough
United Kingdom
TS1 3BA
E-mail: K.roberts@tees.ac.uk
