Show Summary Details
Page of

(p. 37) A Forensic Investigation of Those Who Stalk Celebrities 

(p. 37) A Forensic Investigation of Those Who Stalk Celebrities
(p. 37) A Forensic Investigation of Those Who Stalk Celebrities

J. Reid Meloy

, Kris Mohandie

, and Mila Green

Page of

date: 21 February 2018

Violent attacks against celebrities by a few stalkers—most notably Arthur Jackson and Robert Bardo1—ushered in the criminalization of stalking in California almost 20 years ago. Despite the growing body of stalking research during the past decade (Meloy, 1998, 2006a; Mullen, Pathé, & Purcell, 2008), there remain very few published scientific efforts to discern the behavior of celebrity approachers (Dietz et al., 1991), those who worship celebrities (McCutcheon, Lange, & Houran, 2002), and those who otherwise contact celebrities (Leets, de Becker, & Giles, 1995). There is one published psychological autopsy of a celebrity stalker who committed suicide after he mailed a bomb to an Islandic female pop singer living in Britain (Schlesinger, 2006). Mohandie, Meloy, McGowan, and Williams (2006) published preliminary data on a non-random sample of celebrity stalkers within a larger North American sample of stalkers (n = 1,005) to validate a new stalking typology, RECON, defined by the relationship between the stalker and the victim and the context within which the stalking occurred. The typology yielded four groups—intimate, acquaintance, public figure, and private stranger—with both acceptable interrater reliability and extensive discriminant validity. The public figure group was the least violent and most frequently mentally ill of the four groups. This was a study of a sample of stalkers of celebrities, because virtually all the “public figure” group was drawn from a Los Angeles entertainment corporate security database.

Our intent in this study is to explore this sample of celebrity stalkers in detail, focusing upon demographic characteristics, pursuit motivations and patterns, clinical indicators, victim characteristics, threats, the nature and (p. 38) frequency of their violence, predictors of approach, and recidivism. Inferential comparisons will be made to a group (n = 103) of private stranger stalkers from the same study (Mohandie et al., 2006) on select variables.


Between March 2003 and June 2004, two trained researchers reviewed over 2,300 files dealing with instances of stalking, criminal harassment, menacing, terrorist threats, or domestic violence behaviors. The cases represented non-random samples of convenience and were gathered from six sources: three prosecutorial agencies in California, one police department in Canada, a security department for a large entertainment corporation in Los Angeles, and the first two authors’ case files. All data were archival, and therefore the subjects’ permission for inclusion in the study was not required. Forty-four percent of the overall reviewed cases (n = 1,005) met our operational definition of stalking behavior: two or more unwanted contacts by a subject toward a target that created a reasonable fear in that target. Twenty-three percent of the entertainment security department cases met our definition of stalking, yielding 248 cases. These cases were defined as celebrity cases because the victims were closely associated with the entertainment corporation and were targeted because of their prominence in the popular media (television, newsprint, and/or cinema). Eleven additional celebrity cases were drawn from the authors’ consultation files, six celebrity cases were from two district attorneys’ offices, and six from a law enforcement agency (n = 271). Four of these cases—a well-known police sergeant, an investigative reporter, a columnist for a large East Coast metropolitan newspaper, and an entertainment corporate executive—were included in this celebrity sample because they were widely known for what they did for a living but were not involved in entertainment as performers. Each case represented a closed, in-process, or open but dormant case of a stalker. Each file was originally created for threat assessment/management purposes, law enforcement investigation, or prosecution. Sampling was not done through a mental health or forensic mental health database. Intraclass correlation coefficient for assignment to one of the four RECON types, including the celebrity group, was .95. Dependent variables were recorded in a 50-variable codebook. Further details concerning methodology can be found in Mohandie et al. (2006).


Demographic Characteristics

Seventy-four percent (n = 200) of the celebrity stalkers were male, and 26% (n = 71) were female. Their average age was 39 years (range, 12–81). Forty-nine percent (n = 133) were Caucasian, 15% (n = 43) were African American, 7% (n = 18) were Hispanic, 2% (n = 7) were Asian, 1% (n = 2) were Native (p. 39) American, 2% (n = 4) were Middle Eastern, and 25% (n = 64) were unknown by ethnicity.

Forty-six percent were single (n = 124), 8% (n = 21) were separated or divorced, 9% (n = 24) were married, 3% (n = 7) were cohabitating, and the relationship status of 34% (n = 95) was unknown. Fifty-eight percent (n = 158) were identified as heterosexual, 6% (n = 16) as homosexual, 1% (n = 1) as bisexual, and 35% (n = 96) were of unknown sexual orientation.

Twenty-nine percent (n = 79) of the celebrity stalkers were unemployed; 7% (n = 18) were part-time, contract, or self-employed; 13% (n = 34) were fulltime employed; 2% (n = 5) were underemployed; and the employment status of 40% (n = 135) was unknown. Ten percent (n = 27) were high school graduates, 8% (n = 22) were college graduates, 6% (n = 15) did not complete high school, 1% (n = 3) received a graduate equivalency degree, 1% (n = 2) graduated from a technology school, and the education of 73% (n = 202) was unknown.

The criminal history of 39% (n = 108) was unknown, whereas 8% (n = 21) had a violent criminal history, 21% (n = 56) had a nonviolent criminal history, and 32% (n = 86) had no criminal history. Seven percent (n = 18) had previously been arrested for stalking or harassment, and 51% (n = 137) had had no such previous arrests. When a celebrity stalker did harass another target (30%, n = 80), it was another celebrity, public figure, or stranger 84% of the time (n = 67). Twenty-four percent (n = 66) had not harassed anyone else, but whether or not there was harassment of another was unknown in 46% (n = 125) of the cases. At the time of the stalking event, 43% (n = 117) had no legal status, 8% (n = 22) were in custody, 4% (n = 11) were in a hospital, 3% (n = 9) were in community supervision, 2% (n = 5) were legally restrained, and there was no information about 40% (n = 107).

Clinical Indicators

Although a psychodiagnostic evaluation was completed in only 10% (n = 27) of the sample and was generally not archived, sufficient data were available in the files to suggest the mental status of the majority of subjects. Twenty-seven percent (n = 73) were psychotic at the time of the stalking, indicated by documentation of hallucinations, delusions, or both in the case record. Sixty-seven percent (n = 182) were not psychotic at the time of the stalking, and in the remaining 6% (n = 16), there were insufficient data to determine presence or absence of psychosis. Consistent with these findings, 17% (n = 46) had a psychiatric medication history. At least 7% (n = 20) of the subjects had attempted suicide, but suicidality (ideation, intent, or attempt) was unknown for 92% (n = 251) of the sample.

In 16% (n = 43) there was suggestion of a thought-perceptual disorder diagnosis, such as schizophrenia; 5% (n = 13) suggested a mood disorder diagnosis, such as major depression; and there was a suspected major mental disorder diagnosis in 31% (n = 84) additional cases. Documented evidence precluded a major mental disorder in 41% (n = 111) of the cases. Drug use—most often (p. 40) alcohol or stimulants, or both—was evident in 77% (n = 208) of the cases, absent in 11% (n = 30), and unknown in 12% (n = 33) of the cases.

Pursuit Characteristics and Motivations

The stalking began following a precipitating event—the ending of a relationship or rejection—in 14% (n = 38) of the cases. There was no such known event in 86% (n = 231) of the cases. The duration of the stalking of the celebrity averaged 16 months, with a range of 1 week to 180 months. The modal duration was 1 month. Contact frequency was daily in 15% of the cases (n = 40), weekly in 34% (n = 93), monthly in 26% (n = 70), every 2 to 3 months in 17% (n = 45), twice yearly in 5% (n = 14), annually in 2% (n = 5), and no pattern in 1% (n = 4). Seventy-one percent of the stalkers (n = 192) used indirect means of contact, such as writing to the celebrity figure or sending him or her a gift. Twenty-three percent (n = 62) were proximity based and made a direct approach, and 86% of these approachers (n = 51) also used other direct or indirect means of contact. Six percent (n = 17) communicated to others at the same time, usually the President, FBI, or another governmental agency. There was escalation in 29% (n = 77) of the cases, meaning that the frequency or intensity of the stalking behavior noticeably increased. Frequency was defined as the amount of contact or attempted contact with the celebrity figure over time. Intensity was defined as the level of intrusiveness or disruption the contact behavior could cause; for example, an increase of intensity would be the stalker’s change from sending just e-mails to sending both e-mails and telephone calling.

The majority of the stalkers (52%, n = 141) sought a relationship (affectional or sexual, or both), and 20% (n = 53) sought help from the celebrity victims. Thirteen percent (n = 36) wanted to only communicate, whereas 8% (n = 21) insulted and 6% (n = 16) offered help.

Predicting an Approach

Based upon previous stalking research, 15 variables were selected to attempt to predict a physical approach to the celebrity: police intervention, gender, threats, quality of communication, recidivism, criminal history, mental health diagnosis other than thought disorder, presence of psychosis, age, pursuit patterns, escalation, duration, drug use, target gender, and whether the target was stalked previously. The 62 approach cases and 62 randomly selected non-approach cases were initially selected for the analysis.

Only 10 of the 15 variables had sufficient data to be included as predictors in a logistic regression. The other five—recidivism, criminal history, age, drug use, and target stalked previously—were individually tested for significant difference between the two groups, and none was found. Because of missing variables within each case, 53 approachers and 57 nonapproachers were finally considered in the logistic regression. A significant result was found (–2 log (p. 41) likelihood, 84.941; df = 22; χ‎2 = 67.402; p < .001). The 10 variables entered into the logistic regression were able to correctly classify 83% of the 110 cases as approach or nonapproach, an improvement of 31% over the base classification rate. They accounted for 40% to 60% of the variance in the approach/nonapproach variable:

  • Police intervention

  • Male gender

  • Presence of threats

  • Quality of communication

  • Mental health diagnosis other than thought disorder

  • Psychosis

  • Pursuit patterns

  • Escalation

  • Duration

  • Female target

However, only two of these variables, escalation and police intervention, were significant predictors of an approach: (a) if the stalking escalated (frequency, intensity, threatening, or violent), there was a ninefold increase in likelihood of an approach; and (b) if the subject was not reported to the criminal justice system, or there was nothing more than a police report taken and no other police response, the likelihood of an approach decreased by almost half (44%). A priori comparison utilizing χ‎2 or ANOVA before entry into the logistic regression indicated that another 3 of the 10 variables were significantly more frequent (p ≤ .01) among the approachers—threatening, having a mental health diagnosis other than thought disorder, and a longer duration of stalking—but were not significant predictors of an approach.


One hundred and fifty nine celebrity victims were pursued by 271 stalkers: 93 (58%) women, 49 (31%) men, and 17 (11%) unknown by gender. Forty-one victims (26%) had more than one stalker, ranging from 2 to 20 stalkers. The mode was 2 stalkers and the mean was 3.66 stalkers. Further analysis indicated that most of the multiple stalking cases focused upon female newscasters or female television show actors. Those with more air play, such as television show main characters, were targeted more often. Sympathetic characters or emotionally “softer,” more vulnerable characters were also targeted more often, especially by subjects who identified with or wanted to bond with them. When aggressive or strong women were targeted, the themes were more sexual. The two female victims with by far the most stalkers—10 pursuers and 20 pursuers, respectively—were television series main characters who portrayed professional women who were sexually attractive, not in a sexual pair bond, and at times, emotionally vulnerable. The one male victim with the most stalkers, 14, was the CEO of the entertainment corporation.

(p. 42) Threats

Threats were defined as “a written or oral communication that implicitly or explicitly states a wish or intent to damage, injure, or kill the target” (Meloy, 1999, p. 90). Threats were categorized according to manner (direct, indirect, implied, and conditional) and target (property, others, self, loved ones of victim, and victim). An example of a direct threat from the study: “I’m going to kill you and your new husband,” was sent in a letter after the celebrity victim publicly announced that she had married another public figure. An example of an indirect threat from the study: A letter was sent to a television show’s producer threatening the show’s star that he would “come and find her because she has not returned my calls.” An example of a conditional threat from the study: An inmate sent a letter to a newscaster telling her that “you need to keep wearing that pink suit and not the red one, it makes you look slutty, or else.” An example of an implied threat from the study: A stalker sent roses to the home address of a celebrity after only contacting her at work and subsequent to a cease-and-desist letter sent to him. Eighteen percent of the stalkers (n = 48) threatened and made a total of 61 threats. There were no threats in 82% (n = 223) of the cases. If a threat was made, it was a direct threat half the time (9%). Four percent were indirect, 4% were implied, and 1% was conditional. Among the 61 threats, 30 targeted the celebrity, 11 targeted loved ones of the celebrity, 14 targeted others, 1 targeted property, and 5 targeted the stalker himself. Of the 48 subjects who threatened in any manner, four were violent, with a true-positive rate of 8% and a false-positive rate of 92%. Of the 23 subjects who directly threatened the target, four were violent, with a true-positive rate of 17% and a false-positive rate of 83%. One person was violent who did not threaten, with a false-negative rate of 0.4% and a true-negative rate of 99.6%.

Nature and Frequency of Violence

Violence was defined as an act of intentional physical aggression toward a person (personal violence) or object (property violence). In this sample, there were five acts of violence toward persons and one act of violence toward property, with an overall frequency of 2%. Four out of five (80%) personally violent stalkers communicated a threat before they attacked. Two of the violent stalkers were female and three were male. The average age was 37.6 years with a range of 31 to 46 years. Four had a psychiatric history with a psychotic diagnosis. Four of the celebrity victims were female and all were arguably younger than the attacker. One of the stalkers had a weapon, a knife, in his possession at the time of arrest. In three of the cases, the attack resulted from a successful breach of the celebrity’s security. In no case was there physical injury to the victim requiring medical care.

Case 1

A 40-year-old male stalked a celebrity actress. His criminal history was replete with bizarre public behavior and a prior arrest for criminal (p. 43) threats. He was diagnosed with delusional disorder, erotomanic subtype, by a forensic psychiatrist. He believed the target was in love with him, knew how to reach him, had called him numerous times, and had agreed to marry him. The harassment was on a daily basis for 3 months. He started with letters and then traveled to Los Angeles. He threatened three times, both direct and conditional. He threatened to kidnap her, and then escalated at a local gym, saying he would hurt someone there if they did not get him the actress. He assaulted security when confronted at the entertainment corporation gate while attempting to find her. A large knife and survival equipment were found in his truck. He was eventually released into the custody of Canadian probation and hospitalized. There has been no recidivism to date.

Case 2

A 38-year-old male stalked a female television news reporter. His criminal history indicated both violent and nonviolent crimes, and he was diagnosed with schizophrenia, paranoid type. He delusionally believed that people were cutting out his eyes at night while he slept, poisoning him, and putting listening devices in his head. He pursued the victim for 3 years with telephone calls, then e-mails, then surveillance of her work site, and finally an approach at the television station. He never threatened, and only appeared to be seeking help. The violence occurred when he grabbed her arm while she was leaving work and she escaped his grasp. The subject’s mother intervened and he was briefly hospitalized. Following his inpatient stay, he harassed another on-air talent for a few months and then ceased.

Case 3

A 33-year-old male stalked a female investigative television reporter who was frequently on air. The harassment was of a short-term with telephone calls and on-location approaches. He was angry with her, and accused her of scheming and not reporting the truth. There were no formal psychiatric diagnoses. He directly threatened her once during an on-air call, and also implied other threats. The violence occurred when he showed up on location to confront her and she tried to walk away. He grabbed her, and was charged with aggravated assault. The case was dismissed and a restraining order was put in place. There has been no recidivism to date.

Case 4

A 31-year-old female stalked a well-known male police sergeant. She was diagnosed with a major affective disorder and alcohol abuse/dependence. The harassment continued for 2 years for no apparent reason; he had never had contact with the subject through a case or an incident as an officer. The stalking began with third-party contacts and telephone calls, and escalated to following the police officer and appearing at the police station. She made four direct death threats toward the target, and carried out two planned assaults on him and the police station property. She was always intoxicated during the threats and violence. She was eventually charged with criminal threat, criminal harassment, and violation of a restraining order. She was convicted and placed on probation. Four months later, she appeared at the police station on three occasions to ostensibly collect her belongings while the sergeant was working.

(p. 44) Case 5

A 46-year-old female stalked a female celebrity. She was diagnosed with a “thought disorder” but no specific mental disorder. She previously targeted another celebrity she believed was “part of the conspiracy,” suggesting some type of paranoid disorder. She had no prior criminal history. She harassed the victim for 5 months with telephone calls, third party contacts, letters, e-mails, and surveillance, and by showing up at the female celebrity’s home. She issued at least 10 written or oral threats to the target. The violence occurred when she assaulted a member of the celebrity’s family who came out of the house to confront her. She was convicted of criminal threats and stalking following a hospitalization to restore her to competency to stand trial. There has been no recidivism to date.

Meloy (2001, p. 1211) hypothesized, “Private targets appear to be most likely victimized by affective violence. Public targets are most likely to be victimized by predatory violence,” following a comparison of violent stalkers across several studies who attacked victims with whom they had a previous relationship, usually a sexually intimate one, and those who near-lethally approached, attacked, or assassinated a public figure (Fein & Vossekuil, 1999). The former violence appeared to be unplanned, emotional, impulsive, and a reaction to a perceived threat; the latter violence appeared to be planned, purposeful, and emotionless, usually carried out over a period of days, weeks, or months. The distinction between affective and predatory violence has a lengthy and substantial scientific basis (Meloy, 2006b).

The violence in this celebrity sample, although quite small, raises questions about the validity of this hypothesis because the majority of the acts of personal violence were affective. The RECON study (Mohandie et al., 2006) in which this sample is embedded, moreover, allows for empirical testing of this hypothesis. Four hundred and sixty-seven acts of personal and property violence occurred in the overall sample of 1,005 cases of stalking. Among the total acts, 337 involved personal violence toward another individual, usually the object of pursuit (Mohandie et al., 2006). The violent acts were classified as either affective (unplanned, emotional, reactive, and impulsive) or predatory (planned, purposeful, emotionless). Table 2.1 illustrates the findings. Type I (p. 45) stalkers are individuals who had a prior relationship with the victim before stalking, either as an intimate or an acquaintance, and represent the first two RECON types. Type II stalkers are individuals who had no prior relationship with the victim before stalking, either a public figure (celebrity) or a private stranger, and represent the second two RECON types. Three hundred and two of the personally violent acts were able to be classified as either affective or predatory (89.6%). The remaining acts of violence contained elements of both (n = 27) or there were insufficient data to determine mode of violence (n = 8). Comparison across Type I and Type II stalkers who were violent indicates a significantly higher frequency of affective violence among those stalkers who had a prior relationship with the victim than those who did not (χ‎2 = 13.376, p < .001).

Table 2.1. Comparison of affective and predatory violent acts (n = 337) among stalkers

Mode of violence

Type Ia

Type IIb


172 (54%)

4 (19%)


110 (35%)

16 (76%)


26 (8%)

1 (5%)


8 (3%)

0 (0%)


316 (100%)

21 (100%)

a Violent acts of stalkers who had a previous relationship with the victim.

b Violent acts of stalkers who did not.

* χ‎2 = 13.376, p < .001.


The criminal justice system (law enforcement and/or prosecution) was involved in 58 cases (21%). Criminal charges were filed in 23 cases (9%), usually crimes of stalking or assault. The most common sanction following the criminal litigation was hospitalization (n = 8, 35%). The reoffense rate among those stalkers of celebrities when there was criminal justice involvement was 50%. The mean time to reoffense was 7.9 months, with a range of 1 day to 6 years.

Comparison to Stalkers of Private Strangers

Mohandie et al. (2006) identified a fourth group of stalkers who pursued strangers not in the public eye (n = 103). This group provides a comparative context for the stalkers of celebrities because the absence of a prior relationship with the victim between the groups does not differ, yet the degree to which the victim became known to the stalker through the public media does. Table 2.2 lists selective variables that significantly differ between these two groups of stalkers.

Table 2.2. Significant differences between stalkers of public figures (celebrities) (n = 271) and stalkers of private strangers (n = 103) (Mohandie et al., 2006)


Public figure (celebrity) (%)

Private stranger (%)


  • Male



  • Female



Any drug use**



Any mental disorder*



Daily contact**



Proximity-based contact**



Precipitating event**









Weapons possession/use**



Criminal justice involvement**



* p < .01,

** p < .001; all comparisons by either χ2 or ANOVA.


Stalkers of celebrity figures are the sensation of the popular media, particularly if they have threatened or been violent toward the object of their pursuit. Such anecdotal cases, however, often skew the perspective of the lay person and the professional because a few cases are generalized to all stalkers of celebrities, and scientific research has lagged far behind the conveyed “wisdom” of television experts. Ironically, there appears to be a strong negative correlation between an expert’s energy invested in seeking the television limelight as an “entertainment profiler” and the number of scientific publications in the purported area of special knowledge claimed by the same expert.

This study of those who stalk celebrities both confirms and disconfirms certain areas of conventional beliefs. The vast majority of the offenders are (p. 46) men, although one in four are women, a higher proportion than is seen in most stalking samples (Spitzberg, 2007). Although the average age is comparable to other stalkers—most pursuits occur in the fourth decade of life—it is notable that those who homicidally attack celebrities in the United States, although few and far between, are often males in their early twenties. Robert Bardo, Mark Chapman, John Hinckley Jr., Sirhan Sirhan, and Lee Harvey Oswald were all young men when they lethally or near-lethally approached their targets.2 The testosterone of the young male should be seriously taken into account when considering the likelihood of an extremely aggressive act toward a celebrity target (Fixated Research Group, 2006; Meloy et al., 2004). However, contemporary attacks on western European politicians (James et al., 2007) and historical attacks on the British Royal Family (James et al., in press) evidence a greater variability in the age of the attackers.

The inability of those who stalk celebrities to establish a stable history of both work and love is also apparent in the data. The absence of a history of stable attachments to other objects such as a sexual intimate or even a career is common among stalkers (Mullen, Pathé, & Purcell, 2000; Mullen et al., 2008), and suggests that a fantasized relationship with the celebrity figure—what Meloy (1999) termed a “narcissistic linking fantasy” in which one consciously imagines a special and idealized relationship with another—compensates for the actuality of a rather blighted personal life. “Dreams of love and fateful encounters” (Person, 1988) may come to preoccupy the stalker, negatively (p. 47) impacting his ability and desire to improve his lot in life and furthering his retreat into fantasy.

Although stalkers of celebrities have less frequent criminal histories than other groups of stalkers (Mohandie et al., 2006), almost 1 out of 10 had a violent criminal history, and twice this number had a nonviolent criminal history. It was also common for these men to stalk or harass another target, either historically or contemporaneously, and one out of five of the subjects were legally constrained in some fashion before or during their stalking. The obvious motivation of such stalkers, even if judged to be benign at the time, should not dissuade the threat assessor from a thorough investigation of the criminal background of the subject of concern. A propensity for habitual aggression may be masked by an initially loving and idealizing approach to the celebrity.

Major mental disorder emerged in our study as a clinical problem for the majority of the subjects, a finding which is similar to the presence of mental disorder in other public figure studies (Meloy et al., 2004), and most recently, in a large study of those who fixate on the British Royal Family (Fixated Research Group, 2006), wherein more than 80% of the subjects showed evidence of serious mental illness. One out of four of our subjects were psychotic at the time of their stalking. Although major mental disorder may reduce the risk of violence among stalkers as a whole (Mohandie et al., 2006)—parenthetically, this general finding may not be true when psychiatric disorder interacts with other violence risk variables (Rosenfeld & Lewis, 2005)—the threat assessment task is to determine the exact relationship between the symptoms of major mental disorder and the celebrity stalking behavior in any one particular case. For example, delusional beliefs concerning the relationship between the stalker and his celebrity object may bring a determination and resolve to his pursuit that defeats any intervention other than arrest and eventual forensic hospitalization. On the other hand, belief that one is being persecuted by the celebrity—fortunately, an unusual motivation (see also Dietz et al., 1991; Fixated Research Group, 2006)—may compel an act of paranoid violence that is not suggested at all by the subject’s nonviolent history.

The pervasiveness of mental disorder in this sample, as well as other public figure samples that involve pursuit of politicians, celebrities, and royalty, bodes well for treatment interventions that address the needs of the mentally ill in the larger community, and the symptoms of those charged with the stalking of a celebrity. Most major mental disorders are currently treatable with medications if diagnosis is accurate and compliance can be monitored, the latter ensured through various forms of involuntary outpatient treatment.

A most unusual finding in this study is the absence of a precipitating event in the life of the subject in the months preceding the onset of stalking. Even when compared to stalkers of strangers who are not public figures (see Table 2.2), the ending of a relationship or rejection did not play a large role in precipitating these pursuits. Such an acute disruption, identified in other stalking research (Kienlen, Birmingham, Solberg, O’Regan, & Meloy, 1997), (p. 48) may be supplanted by a more chronic deterioration in social and emotional functioning among stalkers of public figures, including celebrities:

Many subjects evidenced a downward spiral in their lives in the months or year before their approach or attack, usually a combination of social failure and personal vulnerability to chronic anger, depression, or psychosis. These failures and the subject’s poor adaptation to them often marked a decision point wherein the public figure was identified as an object of salvation or persecution, and a plan was born to contact, approach, or attack the public figure.

Meloy et al., 2004, p. 1092

Drug abuse is endemic in this sample and should be treated as a dynamic risk factor for violence when coupled with the presence of a major mental disorder (Monahan et al., 2001; see Table 2.2). The typical drugs of concern are alcohol and stimulants, which may physiologically fuel stalking behavior given their serotonergic antagonism and their dopaminergic agonism, respectively (Meloy & Fisher, 2005).

The pursuit characteristics of this sample are similar to those of other stalking studies. A recent meta-analysis found the duration of stalking to average 22 months (Spitzberg, 2007), compared with 16 months among the stalkers of celebrities. Contact is frequent, but notable is the modal duration of stalking, which was 1 month. Teasing out the mean from the mode in stalking duration has never been done in other studies, and it underscores the degree to which extraordinarily long stalking cases can skew the data if only the average is computed. Long-duration stalking cases may also be an artifact of a sample, reflecting data gathered immediately subsequent to the criminalization of stalking (Zona, Sharma, & Lane, 1993). The Zona sample likely had a heavy loading of cases that predated antistalking legislation and specialized techniques for intervention, thus artificially inflating the apparent duration of obsessions and pursuit behavior. Our data suggest that most celebrity cases will begin with a high frequency of indirect approaches, and then will be done within a month for a variety of reasons not fully understood.

The data further confirm that the motivation for stalkers of celebrities is usually sexual, affectional, or help seeking. Dietz et al. (1991) had similar findings in their 20-year-old data set of 214 subjects who wrote letters to celebrities: Most were motivated by a desire for face-to-face contact, and not violence. Among those who approach members of the British Royal Family, the most prominent members being celebrities in their own right, such as Queen Elizabeth and Prince William, one third of the approachers were motivated by a desire for friendship (Fixated Research Group, 2006). These findings, however, do not preclude the fact that motivations can suddenly change, particularly when faced with rejection by the celebrity or his or her representative, often security personnel. In such cases, rejection can suddenly usher in feelings of humiliation, which in turn can quickly become rage toward the object of desire or displaced onto his or her protector (Meloy, 1998).

(p. 49) Although most stalkers in our sample did not attempt to approach the celebrity, one out of four did. We were able to correctly classify 83% of the subjects who approached and did not approach. Two variables significantly predicted an approach in the logistic regression—police intervention and escalation—and the most important was the escalation of the stalking over time, which increased the risk of an approach ninefold. Physical approach is commonly used as an analog for risk of violence, because the latter is rare in public figure stalking, and physical proximity is usually necessary for a violent act toward a public figure.

Victims of these stalkers were usually women, consonant with a meta-analysis across 58 studies of stalking victims that indicated that 60% to 80% of victims are females (Spitzberg, 2007). Celebrity status does not appear to influence the fact that stalking is largely a crime against women perpetrated by men. Multiple stalkers hounded one out of four victims, and this study gives for the first time a glimpse of some of the quantitative and qualitative characteristics among female celebrity victims that may increase the risk of stalking:

  1. a. The more frequent the appearance in the media, the greater the likelihood of stalking. This was most pronounced among female newscasters or female television show actors. This finding was also recently reported in the study of contemporary approachers of the British Royal Family (Fixated Research Group, 2006): the more frequent the public appearance, or the more publicized the event, the greater the likelihood of a subsequent pursuit.

  2. b. A quartet of personality characteristics or fictional fact patterns that may almost guarantee celebrity stalking is a television series female actor who portrays a professional woman who is sexually attractive, not in a sexual pair bond, and at times, emotionally vulnerable. Portrayed aggressiveness by a female actor seems to correlate with increased sexual interest by a male stalker.

Although female television and film actors fictionally portray such personal characteristics in contrast to female newscasters, the latter exercise behaviors in their daily appearances that may uniquely increase the risk of stalking: They have direct eye contact with the viewer; they want to establish a one-on-one relationship with the viewer and stimulate pleasant feelings so that he or she watches the next night; they are not portraying a fictional character; and they are increasingly encouraged to use their personalities and reveal bits of personal data to stimulate an affectional connection with the audience and build market share. All these behaviors are quite benign when experienced by the mentally healthy viewer. However, for the lone (and likely lonely) male viewer with a history of failed relationships, a strong sense of entitlement, and difficulty discriminating between reality and fantasy, such an invitation, although communicated electronically and not meant specifically for him, can become a compulsion to pursue, as both erotic feelings and affectional longings are stirred amidst ideas of reference and delusions of grandeur.

(p. 50) The threat data in this study confirm general findings from other threat studies (Meloy, 1998, 2001) and specific findings from threats toward public figures: most stalkers who threaten do not act on their threats; and most threats, if made, are direct threats toward the object of pursuit. The public figure threat data, of which there is very little (Dietz et al., 1991; Fein & Vossekuil, 1999; Fixated Research Group, 2006; James et al., 2007; Meloy et al., 2004; Scalora et al., 2002), also indicate that the vast majority of public figure approachers and stalkers do not directly threaten their target. This finding is in striking contrast to those of studies of prior sexually intimate and acquaintance stalkers that indicate that the vast majority of subjects do directly threaten their target (Mohandie et al., 2006; Meloy, 2001). In this celebrity study, there were no threats in 82% of the cases. However, if a direct threat was made to the target, the true positive rate doubled, from 8% to 17%. Another surprising finding, particularly when compared to the Fein and Vossekuil (1999) data, was that four out of the five violent stalkers in this study did communicate a direct threat before they attacked the object of pursuit. We are, however, comparing apples and oranges when we consider mode of violence. The Fein and Vossekuil (1999) data, which found that less than 10% of their sample of attackers and assassins of public figures communicated a threat to the target or law enforcement beforehand, were acts of predatory violence: planned, purposeful, and carried out over days, weeks, or months. Most of the acts of violence toward our celebrity figures were affective: emotional, reactive, unplanned, and impulsive. This contrasting pattern of association in two public figure data sets—acts of predatory violence correlate with no direct threats and acts of affective violence correlate with direct threats—lends further empirical support to one aspect of Meloy’s (2001) hypothesis but challenges another one: acts of violence toward public figures are all predatory rather than affective.

The findings concerning mode of violence question Meloy’s (2001) assertion that “public figure” stalkers engage in predatory violence, while “private figure” stalkers engage in affective violence. Table 2.1 indicates that this formulation was both too simplistic and based upon an incorrect categorization. The reality, as usual, is much more complex. If stalkers are classified as pursuing someone with whom they had a prior relationship (sexual intimate or acquaintance) or someone they did not (public figure or private stranger) (Mohandie et al., 2006), then the data suggest a refinement in formulation: affective violence toward person and property is significantly more likely if there has been a prior relationship between the stalker and his victim, while predatory violence is significantly more likely if there has been no prior relationship between the stalker and his victim, whether a public figure or not. We would also be remiss to not underscore the remarkable disparity between frequency of violence toward these celebrity figures (2%) and frequency of violence toward prior sexual intimates (=50%) in a number of published studies (Mohandie et al., 2006). This is likely a result of both heightened security on the part of celebrities and the intensified emotion that will fuel affective violence toward a person who was once intimately known to the stalker, and now has rejected him.

(p. 51) Affective violence toward public figures who are celebrities may also be more likely when compared to public figures who are politicians because of the greater likelihood of a fantasized sexual or affectional relationship with the former type of target: perceived rejection by celebrity figures may carry a more intense emotional charge when an ideal relationship has been nurtured in fantasy, whereas perceived rejection by political figures who are being pursued for other reasons, such as beneficence, does not carry the risk of such an intense emotional reactivity.

The violence in our study was rare (2%), and when it occurred, was likely to be an affective attack without a weapon by a psychotic male stalker in his fourth decade of life targeting a younger female celebrity whose personal security had been breached. There was no physical injury requiring medical care; however, the psychological or emotional trauma was unknown. Anecdotal data from the authors’ files suggest that celebrity victims of stalking will suffer the same degree of depression, anxiety, and traumatic symptoms that other members of the public do when stalked (Spitzberg, 2007), especially if they are fully apprised of the stalking events by their security personnel or are personally attacked. The violence typically followed a pattern of multiple and various means of contact and escalation, and the most likely victims of attack if the celebrity could not be accessed were members of his or her private security detail.

Half of this large sample of stalkers of celebrities went on to stalk again. The latency period was 8 months. This finding is virtually identical to the recidivism rate (49%) of a sample of stalkers in New York City (Rosenfeld, 2003), and underscores the chronicity of this behavior. The substrate of severe mental disorder in these celebrity stalkers may be the same pathology of “preoccupied” attachment found in other stalking studies (Meloy, 2007) wherein an intense, yet tumultuous bond is quickly formed with an initially idealized object. Among celebrity stalkers, this pathology of attachment, which is biologically rooted (Bowlby, 1958), is all the more remarkable because it is solely based in fantasy and there is no actual relationship with the object of desire, nor will there ever be. One stalker of a well-known female celebrity was asked at the end of his day-long forensic evaluation if he had any questions for the examiner. He politely requested that the examiner go visit the celebrity after he finished, administer the same psychological tests, and see if they matched as a couple. This same individual, living alone in his Midwestern apartment and severely mentally ill, would receive an occasional phone call late at night and be greeted with silence on the other end of the line. He was convinced that his celebrity love, the object of his heart’s desire, was communicating with him in silence because her mother did not want them to actually speak. He eventually pursued her to Los Angeles after sending a slew of gifts, including sexual paraphernalia, and was arrested for stalking on his second approach to her property. He was diagnosed with delusional disorder, with both erotomanic and grandiose features, and personality disorder not otherwise specified with narcissistic, histrionic, compulsive, schizotypal, and (p. 52) dependent features. He was found legally insane, and committed to a forensic hospital until he was released in February 2008.

Limitations and Conclusions

This is an archival study subject to several limitations. Selection bias is present in the nonrandom sample of convenience utilized, and limits the generalizability of our findings because virtually all the subjects were drawn from the security department of one large entertainment corporation. Observation bias may be present in the differential recall and data-gathering strategies among investigators in the security department over time. Nonresponse bias is evident in the relatively large proportion of unknown data, such as relationship status, employment status, and education. Inferential statistics were kept to a minimum to reduce the influence of confounding variables unknown to the researchers.

This is the second study of those who stalk celebrities to be published in the scientific literature (Dietz et al., 1991, being the first). The independent variable in the Dietz study was threatening or otherwise inappropriate communications. The independent variable in our study was stalking. Although other databases of stalkers of celebrities exist, they are considered proprietary information by those who own them, and therefore have not yet contributed to, or been scrutinized by, the scientific community. Perhaps this will change and a fuller understanding of those who stalk celebrities will be known. As the great 19th century American novelist Nathanial Hawthorne wrote in Fanshawe, “If his inmost heart could have been laid open, there would have been discovered that dream of undying fame; which, dream as it is, is more powerful than a thousand realities” (Hawthorne, 1983).


Bowlby, J. (1958). The nature of the child’s tie to his mother. International Journal of Psychoanalysis, 39, 350–373.Find this resource:

Dietz, P. E., Matthews, D., Van Duyne, C., Martell, D., Parry, C., Stewart, T., et al. (1991). Threatening and otherwise inappropriate letters to Hollywood celebrities. Journal of Forensic Sciences, 36, 185–209.Find this resource:

Fein, R., & Vossekuil, B. (1999). Assassination in the United States: An operational study of recent assassins, attackers, and near-lethal approachers. Journal of Forensic Sciences, 44, 321–333.Find this resource:

(p. 53) Fixated Research Group (2006). Inappropriate communications, approaches, and attacks on the British Royal Family. London: The Home Office.Find this resource:

Hawthorne, N. (1983). Collected works. New York: Penguin Putnam.Find this resource:

James, D., Mullen, P., Meloy, J. R., Pathé, M., Farnham, F., Preston, L., et al. (2007). The role of mental disorder in attacks on European politicians 1990–2004. Acta Psychiatrica Scandinavica, 116, 334–344.Find this resource:

James, D., Mullen, P., Pathé, M., Meloy, J. R., Farnham, F., Preston, L., et al. (in press). Attacks on the British Royal Family: The role of psychotic illness. Journal of the American Academy of Psychiatry and the Law.Find this resource:

Kienlen, K., Birmingham, D., Solberg, K., O’Regan, J., & Meloy, J. R. (1997). A comparative study of psychotic and nonpsychotic stalking. Journal of the American Academy of Psychiatry and the Law, 25, 317–334.Find this resource:

Leets, L., de Becker, G., & Giles, H. (1995). Fans: Exploring expressed motivations for contacting celebrities. Journal of Language and Social Psychology, 14, 102–123.Find this resource:

McCutcheon, L., Lange, R., & Houran, J. (2002). Conceptualization and measurement of celebrity worship. British Journal of Psychology, 93, 67–87.Find this resource:

Meloy, J. R. (Ed.) (1998). The psychology of stalking: Clinical and forensic perspectives. San Diego: Academic Press.Find this resource:

Meloy, J. R. (1999). Stalking: An old behavior, a new crime. Psychiatric Clinics of North America, 22, 85–99.Find this resource:

Meloy, J. R. (2001). Communicated threats and violence toward public and private targets: Discerning differences among those who stalk and attack. Journal of Forensic Sciences, 46, 1211–1213.Find this resource:

Meloy, J. R. (2006a). The scientific pursuit of stalking. San Diego: Specialized Training Services.Find this resource:

Meloy, J. R. (2006b). The empirical basis and forensic application of affective and predatory violence. Australian and New Zealand Journal of Psychiatry, 40, 539–547.Find this resource:

Meloy, J. R. (2007). Stalking: The state of the science. Criminal Behavior and Mental Health, 17, 1–7.Find this resource:

Meloy, J. R., & Fisher, H. (2005). Some thoughts on the neurobiology of stalking. Journal of Forensic Sciences, 50, 1472–1480.Find this resource:

Meloy, J. R., James, D., Farnham, F., Mullen, P., Pathé, M., Darnley, B., et al. (2004). A research review of public figure threats, approaches, attacks, and assassinations in the United States. Journal of Forensic Sciences, 49, 1086–1093.Find this resource:

Mohandie, K., Meloy, J. R., McGowan, M., & Williams, J. (2006). The RECON typology of stalking: Reliability and validity based upon a large sample of North American stalkers. Journal of Forensic Sciences, 51, 147–155.Find this resource:

Monahan, J., Steadman, H., Silver, E., Appelbaum, P., Robbins, P., Mulvey, E., et al. (2001). Rethinking risk assessment: The MacArthur study of mental disorder and violence. New York: Oxford University Press.Find this resource:

Mullen, P., Pathé, M., & Purcell, R. (2000). Stalkers and their victims. London: Cambridge University Press.Find this resource:

Mullen, P., Pathé, M., & Purcell, R. (2008). Stalkers and their victims (2nd ed.). London: Cambridge University Press.Find this resource:

Person, E. (1988). Dreams of love and fateful encounters: The power of romantic passion. New York: Norton.Find this resource:

Rosenfeld, B. (2003). Recidivisim in stalking and obsessional harassment. Law and Human Behavior, 27, 251–265.Find this resource:

(p. 54) Rosenfeld, B., & Lewis, C. (2005). Assessing violence risk in stalking cases: A regression tree approach. Law and Human Behavior, 29, 343–357.Find this resource:

Saunders, R. (1998). The legal perspective on stalking. In J. R. Meloy (Ed.), The psychology of stalking: Clinical and forensic perspectives (pp. 25–49). San Diego: Academic Press.Find this resource:

Scalora, M., Baumgartner, J., Callaway, D., Zimmerman, W., Hatch-Maillette, M. A., Covell, C. N., et al. (2002). An epidemiological assessment of problematic contacts to members of Congress. Journal of Forensic Sciences, 47, 1360–1364.Find this resource:

Schlesinger, L. B. (2006). Celebrity stalking, homicide, and suicide: A psychological autopsy. International Journal of Offender Therapy and Comparative Criminology, 50, 39–46.Find this resource:

Spitzberg, B. (2007). The state of the art of stalking: Taking stock of the emerging literature. Aggression and Violent Behavior, 12, 64–86.Find this resource:

Zona, M., Sharma, K., & Lane, J. (1993). A comparative study of erotomanic and obsessional subjects in a forensic sample. Journal of Forensic Sciences, 38, 894–903.Find this resource:


1. Jackson stabbed Teresa Saldana multiple times in 1982 and was convicted of attempted murder. Bardo shot and killed Rebecca Shaeffer in 1989 and was convicted of murder. Both men used a private detective agency to locate their victims and attacked them at home. In subsequent forensic examinations, both men were found to have paranoid schizophrenia (Saunders, 1998).

2. Although both John and Robert Kennedy were politicians, we believe that both men had achieved celebrity status at the time of their assassinations.