The Prevalence and Effects of Stalking

Abstract

Background

In 2003, we carried out the first epidemiological study on the frequency and effects of stalking in Germany that was based on a random population sample. We repeated the study with the same design in 2018 in order to assess any potential alterations over time in the frequency of stalking and of psychological problems in the affected persons. As far as we know, this is the first replication study of this kind to be carried out anywhere.

Methods

1000 women and 1000 men were randomly chosen from the residents’ registration data of Mannheim, Germany. Each one of them received, by mail, a comprehensive questionnaire about stalking, as well as the WHO-5 Well-Being Index and the German version of the Patient Health Questionnaire (PHQ-D).

Results

In the Mannheim population samples (2003: N = 675; 2018: N = 444), the lifetime prevalence of being stalked was 11.6% in 2003 (95% confidence interval, [9.2; 14.4]) and 10.8% in 2018 [8.1; 13.7]. In both 2003 and 2018, persons who had been stalked had significantly worse mental well-being than unaffected persons (WHO-5 summated score 2003: 11.2 [9.7; 12.6] vs. 15.5 [15.1; 16.0], WHO-5 summated score 2018: 11.8 [10.1; 13.6] vs. 14.5 [13.9; 15.0]). A markedly higher percentage of persons who had been stalked also fulfilled the syndrome criteria for at least one mental disorder (PHQ-D 2003: 50.0% vs. 22.5%; odds ratio [OR]: 3.5 [2.1; 5.6], PHQ-D 2018: 46.5% vs. 24.4%; OR: 2.7 [1.4; 5.1]). In 2018, as in 2003, persons who had been stalked were dissatisfied with, or unaware of, the opportunities that they had to get help from the police and the judicial system.

Conclusion

Stalking remains a major problem that must be taken seriously. Physicians and psychologists should be well informed about it in order to help affected persons who turn to them for medical and psychological assistance.

The word “stalking,” adopted from hunting terminology, was first used in the USA in the 1990s to describe a complex pattern of behavior. Some years later, soon after the turn of the century, stalking came to be perceived as a meaningful problem for the healthcare system in Germany. Legislation followed in 2007; stalkers can be prosecuted under § 238 of the German Criminal Code. Stalking presents physicians and psychologists with many challenges, e.g., in the counseling and treatment of the victims of stalking and the diagnosis and therapy of stalkers (1). Medical/psychological expertise is also of great importance in assessing the risk of violent escalation of stalking behavior. It is not at all uncommon (>70% of cases, according to [2]) for persons who kill an intimate partner to have previously come to attention for stalking their future victim, and the homicide risk in this context is particularly high for rejected stalkers (3). Notably, physicians and psychologists are themselves at higher risk of stalking than the general population in their working life and should be in possession of the skills to enable them to react professionally in such situations (4).

Epidemiological findings

Owing to differences in definition, sampling, and survey methods, together with considerable uncertainty about how many cases go unreported, data on the prevalence of stalking vary widely (5). However, all epidemiological studies are in agreement that stalking is a widespread phenomenon and that women are stalked much more frequently than men (5– 10; Table 1). A particular problem is presented by violent behavior in the context of stalking (11). The results of a systematic literature review suggest that the lifetime prevalence of experiencing stalking ranges between 8% and 25% (12). In Germany, a recent representative study of unreported stalking in the age group 16 to 40 years found a stalking prevalence of 15% (13).

In the first population sample-based study of the prevalence of stalking in Germany, published in 2003, we found a lifetime prevalence of 11.6 % [9.2; 14.4] in the age group 18 to 65 years (14).

Potential health consequences of stalking

Stalking is a potentially traumatic experience. The victims of stalking have an elevated risk of mental illness (15). A number of studies have reported a high prevalence of depressive disorders and post-traumatic stress disorder, although they come to no general conclusions on the relevance of the type and duration of stalking in this respect (16– 20). Stalking often escalates to involve sexual and physical violence, and may in isolated cases be a risk factor for homicide (21). The economic consequences of stalking are considerable, including loss of productivity owing to time off work, treatment costs in the healthcare system, and legal costs (22). It should be noted, however, that the cited studies merely show associations, not causality, while any calculations of consequent costs rest on the assumption of a causal connection.

Study goals

In this study we set out to compare the prevalence and effects of stalking in a German population sample from 2018 with our own results from 2003. To our knowledge, no previous study anywhere in the world has explored these parameters over an extended period of time with exactly the same study design. Longitudinal investigation is important, for instance, in assessing the efficacy of the measures that have increasingly been introduced in recent years, in Germany and elsewhere, to combat stalking. The following questions were of particular interest:

Has the prevalence of stalking changed in the past 15 years? The police and criminal justice system are more active against stalkers than used to be the case, but new methods of stalking have emerged, e.g., via social media.

Has the health status of stalking victims improved between 2003 and 2018? There is now more widespread access to expert counseling and therapy, so it could be assumed that treatment has become more appropriate, with lasting positive effects.

Do victims of stalking think that the recently introduced legal powers against stalkers are adequate?

Material and methods

In October 2018, we selected 1000 women and 1000 men between the ages of 18 and 65 years at random from the residents’ registration data of the city of Mannheim. Each of these persons was sent a comprehensive questionnaire on the topic of stalking (23), together with the WHO-5 Well-Being Index (24, 25), the German version of the Patient Health Questionnaire (PHQ-D) (26, 27), and a letter explaining the goals of the study. The methods are described in detail in the eMethods supplement. For the purposes of this study, stalking had taken place if there had been multiple episodes of harassment or unwanted contact featuring multiple behavior patterns over a period of at least 2 weeks that made the victim anxious or fearful.

Results

The principal findings of the comparisons between the data from 2003 and 2018 are presented in Table 2.

Table 2. Comparison of the stalking surveys in 2003 and 2018.

	2003	2018
	N*1 (% or mean [95% CI])	N*1 (% or mean [95% CI])	Test statistic (df)	p-value
Responses Overall Female Age	679*2 (34.2 % [32.2; 36.3]) 392 (58.9 % [55.2; 62.9]) 42.5 [41.6; 43.6]	444 (23.0 % [21.1; 24.7]) 278 (64.1 % [59.4; 68.9]) 42.8 [41.5; 44.0]	Chi2  = 60.02 (1) Chi2  = 2.88 (1) t = 0.33 (1 074)	< 0.001 0.090 0.744
Stalking victims	78 (11.6 % [9.2; 14.4])	48 (10.8 % [8.1; 13.7])	Chi2 = 0.15 (1)	0.700
Sex of victims Female Male	68 (87.2 % [79.5; 94.4]) 10 (12.8 % [6.3; 21.4])	40 (83.3 % [71.4; 93.5]) 8 (16.7 % [7.1; 27.9])	Chi2 = 0.36 (1)	0.549
Sex of stalkers Male Female	65 (85.5 % [77.3; 93.2]) 11 (14.5 % [6.9; 22.1])	40 (87.0 % [76.4; 95.8]) 6 (13.0 % [4.3; 24.0])	Chi2 = 0.05 (1)	0.825
Duration of stalking at least 1 year	19 (24.4 % [15.0; 34.3])	17 (35.4 % [21.6; 50.0])	Chi2 = 1.78 (1)	0.182
Frequency of stalking More than once per month or less More than once per week Once/more than once per day	31 (40.3 % [29.1; 51.5]) 27 (35.1 % [25.0; 45.5]) 19 (24.7 % [15.7; 34.2])	11 (23.9 % [11.9; 36.6]) 20 (43.5 % [29.8; 58.3]) 15 (32.6 % [19.0; 47.5])	Chi2 = 3.44 (2)	0.179
Average number of stalking methods	5.0 [4.4; 5.6]	5.4 [4.6; 6.3]	t = 0.82 (124)	0.412
Threats	27 (54.0 % [40.8; 68.9])	31 (66.0 % [52.0; 78.2])	Chi2 = 1.44 (1)	0.230
Physical and/or sexual violence	30 (38.5 % [27.3; 50.0])	21 (43.8 % [29.8; 58.3])	Chi2 = 0.03 (1)	0.874
Stalker known to victim	59 (75.6 % [66.7; 84.8])	44 (93.6 % [85.1; 100])	Chi2 = 6.54 (1)	0.011
Stalking by ex-partner	25 (32.1 % [21.7; 42.9])	22 (45.8 % [32.4; 60.4])	Chi2 = 2.41 (1)	0.120
Charges filed	16 (20.5 % [11.8; 29.5])	9 (19.1 % [7.9; 31.8])	Chi2 = 0.03 (1)	0.854
Professional help enlisted	20 (27.0 % [17.2; 37.5])	16 (34.8 % [20.5; 49.1])	Chi2 = 0.81 (1)	0.367
Sick leave as consequence of stalking	14 (18.4 % [10.1; 28.1])	13 (27.7 % [15.2; 41.5])	Chi2 = 1.45 (1)	0.229
Self-reported health-related problems Agitation Sleep disorders Depression	44 (56.4 % [45.9; 67.2]) 32 (41.0 % [30.2; 52.5]) 22 (28.2 % [18.3; 38.3])	29 (60.4 % [45.2; 74.5]) 19 (39.6 % [26.0; 53.5]) 11 (22.9 % [11.1; 35.5])	Chi2  = 0.20 (1) Chi2  = 0.03 (1) Chi2  = 0.43 (1)	0.658 0.873 0.512
Scope for action by police and courts Not evaluable Inadequate Adequate	41 (53.2 % [41.1; 63.6]) 35 (45.5 % [34.9; 57.4]) 1 (1.3 % [0.0; 4.3])	23 (47.9 % [33.3; 61.7]) 25 (52.1 % [38.3; 66.7]) 0 (0.0 %)	Chi2  = 1.058 (2)	0.589
WHO-5 total score*3	77 (11.2 [9.7; 12.5])	45 (11.8 [10.1; 13.4])	F = 0.01 (1.113)	0.931
PHQ-D anxiety score*3	76 (5.4 [4.4; 6.3])	42 (5.2 [3.8; 6.7])	F = 0.01 (1.113)	0.935
PHQ-D depression score*3	74 (7.8 [6.4; 9.4])	43 (7.7 [6.1; 9.5])	F = 0.04 (1.112)	0.850

*¹ N reduced by missing data in some cases

*² Including four questionnaires with obviously psychotic content that were excluded from further analysis (see [14])

*³ Results adjusted for age, sex, and level of education (< qualification for university of applied sciences versus ≥ qualification for university of applied sciences) 95% CI, 95% confidence interval; df, degrees of freedom; PHQ-D, Patient Health Questionnaire (German version); WHO-5, WHO-5 Well-Being Index

n = 15 (2003) and n = 70 (2018) addresses were drawn twice or the questionnaires could not be delivered because the addressee’s current residence was unknown. These are not included in the calculation of responses.

After adjustment for double selection and undeliverable questionnaires, letters and questionnaires were sent to 1985 contactable persons in 2003 and 1930 persons in 2018. The response rate in 2018 was, at 23.0%, much lower than in 2003 (34.2%). With regard to the representativeness of the samples, there was no statistically significant difference between the percentage age distribution of the population of Mannheim and the study sample of persons born between 1953 and 2000. In both 2003 (58.9%) and 2018 (64.1%), women were somewhat overrepresented in the study sample relative to the population as a whole. The two samples did not differ significantly in average age or in sex distribution.

In 2003, the criteria for stalking used in both studies were fulfilled by 78 persons (11.6%). The proportion was similar in the 2018 study (48 persons, 10.8%).

Women formed a clear majority of the victims of stalking in both surveys, with no significant difference between 2003 and 2018. In both surveys, the victims reported that more than 80% of the stalkers were male. With regard to the duration of stalking, more than a third of the victims in 2018 stated they had been stalked for a year or more, against not quite a quarter in 2003. Overall, the victims listed many different types of persecution and harassment. On average, in both surveys each person was exposed to around five different kinds of stalking (figure).

Figure.

Reported methods of stalking

The categories “False information on Internet” and “Contact via Internet” were included only in the 2018 survey.

The relevance of the stalking behavior patterns recorded here is underlined by the finding that explicit threats were uttered in 54% of cases in 2003 and 66% in 2018. The overall proportion of cases involving physical and/or sexual violence is similar in the two surveys (over 50% in each case) (table 2).

In both surveys, (ex-)partner stalking was the main type of stalking reported (2003: 32.1%, 2018: 45.8%). The next most common category of stalkers were acquaintances or friends of the victims (2003: 20.5% [11.8; 29.9]; 2018: 20.8% [10.0; 34.0]). Less often, the stalkers were colleagues (2003: 3.8% [0; 8.4]; 2018: 4.2% [0; 10.0]) or family members (2003: 3.8% [0; 9.0]; 2018: 4.2% [0; 10.5]). There were also individual cases of stalking in, for example, the professional context.

The victims were asked for their subjective assessment of the social, mental, and medical effects of the stalking. Most of them reported not only anxiety but also other psychic and somatic symptoms perceived as direct consequences of the stalking, for example increased agitation (2003: 56.4%; 2018: 60.4%), sleep disorders (2003: 41%; 2018: 39.6%), and depression (2003: 28.2%; 2018: 22.9%). In 2018, 27.7% of the victims (2003: 18.4%) stated that their doctor had put them on sick leave due to the stalking’s impact on their health.

Only 20.5% of the victims filed charges with the police in 2003, and there was no increase on this figure in 2018 (19.1%). In contrast, 34.8% sought treatment in 2018, against 27.0% in 2003. The scope for legal action against stalkers was viewed as inadequate by 52.1% of the victims in 2018 (2003: 45.5%). A striking finding in both surveys is the high proportion of victims who are obviously insufficiently aware of the scope for legal action (2003: 53.2%, 2018: 47.9%).

With regard to their current mental wellbeing, in both 2003 and 2018 the victims of stalking showed statistically significantly lower levels of wellbeing (WHO-5) and statistically significantly higher levels of anxiety and depression, even after adjustment for sociodemographic variables (PHQ-D; all p-values ≤ 0.001 [Table 3]). Stalking victims also differed at diagnostic level, in both 2003 and 2018, from those who had not been stalked: In 2003 the criteria for a diagnosis of at least one psychiatric disorder were fulfilled by 50% of the victims but only 22.5% of the non-victims; in 2018 the figures were 46.5% and 24.4%, respectively (table 3). However, the mental wellbeing of stalking victims did not differ between 2003 and 2018 (table 2).

Table 3. Mental wellbeing of stalking victims and persons not exposed to stalking (2003 and 2018).

	2003			2018
	N*1 (% or mean ([95% CI])	Test statistic (df)	p-value	N*1 (% or mean ([95% CI])	Test statistic (df)	p-value
WHO-5 total score < 13 Stalking victims Others	44 (57.1% [46.8; 67.5]) 158 (27.1% [23.2; 30.7])	Chi2 = 28.90 (1)	< 0.001	24 (53.3% [37.8; 66.7]) 122 (32.3% [27.5; 36.8])	Chi2 = 7.89 (1)	0.005
PHQ-D Diagnosis of at least one psychiatric disorder (DSM-IV) Stalking victims Others	38 (50.0% [39.5; 60.5]) 127 (22.5% [19.1; 26.2])	Chi2 = 26.55 (1)	< 0.001	20 (46.5% [32.6; 60.5]) 88 (24.4% [20.0; 29.4])	Chi2 = 9.54 (1)	0.002
WHO-5 total score*2 Stalking victims Others	77 (11.2 [9.7; 12.5]) 557 (15.5 [15.1; 16.0])	F = 31.5 (1.627)	< 0.001	45 (11.8 [10.1; 13.4]) 360 (14.5 [13.9; 15.1])	F = 10.4 (1.398)	0.001
PHQ-D score for anxiety symptoms*2 Stalking victims Others	76 (5.4 [4.4; 6.3]) 557 (2.7 [2.4; 3.0])	F = 30.4 (1.628)	< 0.001	42 (5.2 [3.8; 6.7]) 363 (2.6 [2.2; 3.1])	F = 12.4 (1.400)	< 0.001
PHQ-D score for depressive symptoms*2 Stalking victims Others	74 (7.8 [6.4; 9.4]) 528 (3.9 [3.5; 4.3])	F = 40.7 (1.597)	< 0.001	43 (7.7 [6.1; 9.5]) 364 (4.7 [4.2; 5.2])	F = 14.9 (1.402)	< 0.001

*1 N reduced by missing data in some cases

*2 Results adjusted for age, sex, and level of education (< qualification for university of applied sciences versus ≥ qualification for university of applied sciences)

95% CI, 95% confidence interval; df, degrees of freedom; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders; PHQ-D, Patient Health Questionnaire (German version); WHO-5,

WHO-5 Well-Being Index

Supplementary analyses showed that longer duration and greater frequency of stalking events were both, independent of each other, associated with higher PHQ-D depression scores (p-values ≤ 0.05 [eBox]).

Discussion

To our knowledge, this study is the first population-based replication study on the frequency and effects of stalking. The same study design was used to investigate this topic in both 2003 (14) and 2018, enabling description of any changes in the findings. Awareness of stalking has increased greatly since 2003. The police and courts, for example, have changed how they deal with stalking: the specific offense of stalking was added to the German Criminal Code in 2007. In addition, specialized counseling centers for victims of stalking have been opened.

This replication study was intended to answer questions of scientific and practical interest: Has the prevalence of stalking changed in the past 15 years? Has the health status of victims of stalking improved? Do they think the police and courts currently have sufficient powers?

First, however, the limitations of the study should be mentioned: The response rate, although satisfactory for surveys of this type on each occasion, was much lower in 2018 than in 2003. This reflects a general international trend towards lower response rates in social science studies (28). With regard to the age distribution, both samples were largely representative, but in both cases more women than men took part. A further limitation is the fact that the samples were drawn from the population of a fairly large western German city, which, like the low response rate, restricts the generalizability of our findings. However, the prevalence closely matched the findings of studies from other countries. Furthermore, the cross-sectional nature of the survey means that no causal connection between stalking and mental health can be concluded. For example, it cannot be excluded that some of the identified stalking victims were suffering psychic stress beforehand, and persons with psychic stress are at elevated risk of becoming victims of stalking (29). Moreover, such individuals may have a lower threshold for perceiving events as threatening and reacting with anxiety. Clear conclusions can be reached only by means of longitudinal studies, which can explore to what extent the risk of mental disorders is increased by stalking among persons whose mental health is initially good. A longitudinal study of this kind by Diette et al. (30) integrated data sets on girls and women from three large studies from the USA. The authors showed that from the age of 18 years onward, becoming a victim of stalking increased the risk of later mental disorders by a factor of 2.7–3.9. However, our study’s disadvantage of being a cross-sectional survey of stalking experience and mental wellbeing is shared by the vast majority of published research. Longitudinal studies need very large data sets to be able to investigate the incidence of mental disorders prospectively. This necessarily involves dispensing with precise analysis of the characteristics of stalking and the victims’ reactions, which we were able to accomplish.

The results of our replication study support the assumption that stalking remains an important problem in Germany. The lifetime prevalence of stalking in the 2018 sample was 10.8%, close to the figure of 11.6% in 2003. The prevalence of stalking has therefore hardly decreased over a 15-year period, although the powers of the police and the criminal justice system against stalkers have been strengthened. The sex distribution of both perpetrators and victims of stalking has also changed very little, and so-called (ex-)partner stalking remains the main problem. We found no grounds for the plausible assumption of increased prevalence owing to enhanced attention to stalking in the media, which speaks in favor of the prevalence of stalking—like that of other criminal offenses—being quite stable.

The victims were exposed to many different patterns of stalking behavior. In both surveys, stalkers used an average of around five different ways of establishing unwanted contact. In 2018 we added the category of cyberstalking, i.e., unwanted contact via social media. However, the inclusion of cyberstalking did not lead to an increase in the overall prevalence of stalking between 2003 and 2018. Whenever cyberstalking was used, it was an adjunct to other methods of stalking.

Both in 2003 and in 2018, the mental disorders of stalking victims were more severe than those in persons who had not been exposed to stalking. In both surveys, the former had significantly worse scores for overall wellbeing, anxiety, and depression (WHO-5, PHQ-D). Moreover, the chance of fulfilling the criteria for a mental disorder diagnosis according to the PHQ-D was higher in victims than non-victims by a factor of 3.5 in 2003 and 2.7 in 2018. In contrast, there was no difference in the corresponding outcome measures between stalking victims in 2003 and in 2018. Characteristics of the severity of stalking, such as frequency and duration, seem to be associated with elevated depression in the victims. Thus no improvement in the situation can be identified on the basis of our results, and the data confirm the findings of studies from other countries with regard to the elevated prevalence of mental disorders in victims of stalking (16– 19). Despite the increased availability of counseling and treatment, care still needs to be optimized. The legal powers against stalkers, strengthened during the period between the two surveys, were viewed as inadequate by just over half of the victims in 2018, against 45.5% in 2003. In both surveys, a strikingly high proportion of the stalking victims had insufficient knowledge of the scope for legal action (53.2 % in 2003, 47.9 % in 2018).

In conclusion, our study has shown that the provision of counseling and information for those exposed to stalking remains insufficient. The victims still regard the legal powers against stalkers as inadequate. Whether the reform of § 238 of the Criminal Code will achieve any improvement remains to be seen, but how it is implemented in practice will be crucial.

Physicians and psychologists should be aware of the clinical aspects of stalking and familiar with risk assessment (see “The Clinical Perspective”). Those caring for victims of stalking must always bear in mind the possibility of escalation and the need for dynamic risk assessment (31– 33).

The anti-stalking rules presented in the Box are valid for all constellations and should be followed in the care and treatment of all stalking victims. The manual published by Gallas et al., for example, can be used in cases where more intensive care is needed (34).

The Clinical Perspective.

Stalking is a widespread problem. Because victims of stalking have an elevated risk of mental and physical illness and may present unspecific symptoms when they consult a physician, questioning should include the possibility of stalking. Should it emerge that stalking has taken place, the physician should not just treat somatic or psychic symptoms, but rather view themself as one element of a complex aid system. Victims should be told they can seek advice from the police and, if applicable, file charges. In many parts of Germany the police force has units that specialize in advice and prevention. Physicians should also tell stalking victims that it may be a good idea to consult a lawyer about the legal options available for use against stalkers.

When treating victims of stalking, the possibility of violent escalation must be considered. Dynamic risk estimation is essential. Crucial risk factors include concrete suicide plans by the stalker, concrete death fantasies, last resort thinking (“if I can’t have her, no-one else is going to have her”), psychopathic personality traits of the stalker, earlier violent actions by the stalker, damage to the victim’s property, access to weapons, physical proximity (verbal confrontation, gaining entry to residence), impulsiveness, low tolerance of frustration, and substance abuse. The risk must be continually reassessed, because the risk factors can change over time. The danger of violent escalation may increase, for example, when a stalker receives a court order to keep his or her distance or is banned by the family court from having contact with shared children. In such situations it may be expedient to introduce concrete protective measures for the victim, e.g., accompaniment by trusted persons.

Supplementary Material

eMethods

Material and Methods

The persons contacted by letter in October 2018 were asked to return the completed documents anonymously, using the stamped addressed envelope provided. One single reminder was sent 14 days later if required.

The stalking questionnaire we used (23) contains sociodemographic items and 51 items on the experience of being threatened, followed, or harassed. Participants who had been threatened, followed, or harassed in at least one way were requested to answer additional questions about the duration, nature, and frequency of this experience as well as their personal relationship, if any, with the perpetrator; what they thought that person’s motives might be; their own reactions to the perpetrator’s behavior; and any medical or psychological consequences. In addition we used the WHO-5 Well-Being Index, a scale for estimation of mental wellbeing which, in epidemiological studies, has also proved useful as a screening instrument for depression (24, 25). The WHO-5 comprises five items, each with six possible responses. The total score ranges from 0 to 25 points.

The German version of the Patient Health Questionnaire (PHQ-D [26, 27]) was used to assess current psychiatric morbidity. The PHQ-D enables the assignment of DSM-IV diagnostic codes (DSM-IV, Diagnostic and Statistical Manual of Mental Disorders) for the most commonly occurring mental disorders (major depressive syndrome, other depressive syndromes, panic syndrome, other anxiety syndromes, eating disorders, alcohol syndrome). PHQ-D scores for depressive symptoms and anxiety symptoms can be calculated by adding the individual scores for diagnosis-specific symptoms. International studies have credited the PHQ-D with good reliability and diagnostic validity (27).

We define stalking as present when all of the following conditions are met:

This definition is based on the criteria of our own study in 2003 and on findings from research in other countries (5– 7). Group differences between the samples from 2003 and 2018 were tested by means of Pearson’s chi-squared test for categorical variables and the t-test for continuously distributed variables. With regard to the symptom scales, group comparisons (stalking victims 2003 versus 2018, victims versus non-victims 2003 and 2018) were calculated on the basis of covariance analyses; together with the respective group factor of interest, these analyses included the covariates sex, age, and educational level to control for their effects. The impact of characteristics of the severity of stalking on the magnitude of the adverse effect on the mental health of stalking victims (supplementary analysis in the eBox) was investigated with the aid of linear regressions. All analyses were carried out using SPSS Version 24 for Windows (SPSS Inc., Chicago, IL).

• Multiple episodes of harassment take place

• These episodes cover a period of at least 2 weeks

• Multiple types of stalking behavior occur

• The experience causes the victim anxiety or fear

Table 1. Epidemiological studies on stalking.

Authors	Year of publication	Country	N	Overall prevalence of stalking	Prevalence: female victims	Prevalence: male victims
Tjaden, Thoenness (6)	1997	USA	16 000	8%	12%	4%
Budd, Mattinson (7)	2000	UK			16%	7%
Purcell et al. (8)	2002	Australia	1844	12.8%	17.5%	7.2%
Basile et al. (10)	2006	USA	9684	4.5% (95% CI [2.77; 4.90])	7%	2%
Breiding et al. (9)	2014	USA	12 727		15.2 % [13.9; 16.6]	5.7 % [4.7; 6.8]
Dreßing et al. (14) Dreßing et al. (this study)	2005 2020	Germany Germany	679 444	11.6 % [9.2; 14.4] 10.8 % [8.1; 13.7]	17.5 % [13.8; 21.2] 14.4 % [10.1; 18.7]	3.7 % [1.5; 6.2] 5.1 % [1.9; 9.0]

95% CI, 95% confidence interval

BOX. Anti-stalking rules.

1. Explain only once, but with absolute clarity, that no contact is wanted.

2. Ignore completely all further attempts at contact.

3. Make public what is going on, i.e., inform neighbors, colleagues, and friends.

4. Document all events in a stalking diary.

5. Do not delete text messages and e-mails; they constitute evidence.

6. Do not cancel your telephone number, but record the stalker’s calls on an answering machine. Use a secret number for all other calls.

7. Do not return any presents the stalker may send, but keep them as evidence. Sending them back establishes contact.

8. Talk to the police at an early stage.

9. Seek early advice from a specialized attorney.

Key Messages.

• The results of this replication study (comparison of surveys carried out in 2003 and 2018) demonstrate that stalking remains a widespread phenomenon.

• Despite higher awareness of the topic and the definition of stalking as a specific criminal offense, the prevalence of stalking has not changed in the past 15 years.

• In both surveys, the mental wellbeing of stalking victims was statistically significantly worse that that of persons from the general population who were not exposed to stalking.

• A much higher proportion of stalking victims fulfilled the criteria for at least one psychiatric disorder. Physicians and psychologists should be familiar with the topic of stalking so that they can provide proper help to victims and also react professionally in situations where there is a risk that they themselves might be stalked.

eBOX. Supplementary analysis.

In addition to the results presented in the main text, a supplementary analysis was conducted to investigate the extent to which certain indicators of the severity of stalking, such as duration and frequency, together with the presence of violent actions, affect the current mental wellbeing of stalking victims.

To this end, we calculated separate linear regressions with the dependent variables WHO-5 total score, PHQ-D score for anxiety, and PHQ-D score for depression. The independent variables used in these models were duration, frequency, and the presence of violent actions. To avoid inflation of individual tests, we used the whole samples from 2003 and 2018 and explored interactions between stalking victims and those not exposed to stalking with various independent variables in order to test whether the association between stalking characteristics and symptom intensity differed between the two groups. If an interaction term was found to be non-significant, it was removed from the model and the corresponding principal effects were interpreted.

In none of the models analyzed were the respective interaction terms statistically significant, i.e., the associations were similar in 2003 and 2018. The two principal-effect models for the WHO-5 total score (F[3.115] = 1.47; p = 0.227) and the PHQ-D anxiety score (F[3.115] = 2.55; p = 0.059) were also not statistically significant, but the PHQ-D depression score was statistically significant (F[3.114] = 4.08; p = 0.009). (eTables 1– eTables 3) show the corresponding principal effects of the investigated variables on the three symptom scales.

eTable 1. Association of stalking characteristics with the WHO-5 total score “mental wellbeing”.

Model	B	Standard error	Beta	t	p-value
Duration of stalking	− 0.29	0.85	− 0.03	− 0.34	0.732
Frequency of stalking	− 0.47	0.41	− 0.11	− 1.17	0.247
Violent actions	− 1.58	1.20	− 0.13	− 1.32	0.190

eTable 2. Association of stalking characteristics with the PHQ-D anxiety score.

Model	B	Standard error	Beta	t	p-value
Duration of stalking	0.25	0.63	0.04	0.39	0.697
Frequency of stalking	0.59	0.30	0.18	1.93	0.056
Violent actions	1.23	0.90	0.13	1.37	0.172

eTable 3. Association of stalking characteristics with the PHQ-D depression score.

Model	B	Standard error	Beta	t	p-value
Duration of stalking	1.69	0.85	0.18	1.98	0.050
Frequency of stalking	0.94	0.41	0.21	2.29	0.024
Violent actions	0.95	1.21	0.07	0.78	0.435

Therefore, there is a specific association between the duration and frequency of stalking and the severity of current depression. Longer duration and more frequent initiation of contact by the stalker are both associated with high depressivity, independently of each other, but the presence of violent actions does not explain any relevant additional portion of the depressivity.

PHQ-D, Patient Health Questionnaire, German version; WHO-5, WHO-5 Well-Being Index