Social perceptions of stalking in Spain: The behaviour’s seriousness and legal response

The recent process of criminalization of stalking in various European countries has proved to be complex, due to the diffuse nature of the phenomenon. One tool for helping to identify the most appropriate legal response to this issue is social perception studies. This article presents the findings of one such study conducted with a sample of 1010 university students, one of the population segments most victimized by stalking, based on two case scenarios. The aim was to determine, first, this population segment’s perceptions of the incidence and seriousness of the stalking phenomenon and, second, their attitude regarding the proportionate legal response in cases of stalking, depending on both certain personal characteristics of the sample members and other variables related to the case scenario itself. The results show that the sample population’s attitude with regard to the appropriate legal response in such cases is relatively nonpunitive. They also reveal distortions in the sample population’s view of the incidence and seriousness of these behaviours: in contrast to the findings of stalking victimization studies, the respondents considered cases of stalking by strangers to be more serious.


Introduction
Western European countries have only recently begun to legislate stalking. Although countries such as Germany or Italy criminalised it at the end of the first decade of the 21 st century, it was not until the adoption of the Council of Europe Convention on preventing and combating violence against women and domestic violence (Istanbul Convention) in 2011 that several other European countries followed suit. This was the case of Sweden (2011), the United Kingdom (2012), or Spain and Portugal (2015).
Despite the trend detected in recent years towards criminalising this phenomenon, a legislative formula has yet to be found that provides a fully satisfactory response to it.
The complex nature of stalking has made it difficult to legislate, leading to the emergence of various legislative models aimed at offering a suitable legal response (Van der Aa, 2017). One potential tool for tackling these challenges is victimisation studies, which can provide information on the incidence and dynamics of stalking processes. Such information enables greater understanding of the phenomenon, which, in turn, can make the legislation process more rational. They have been used in the US (Tjaden and Thoennes, 1998), the United Kingdom (Budd and Mattinson, 2000), and even at the European level through the study Violence against Women: an EU-wide survey (FRA, 2014), prepared by the EU Agency for Fundamental Rights (FRA). These studies show how the most victimised population by this phenomenon are women under the age of 30 (Baum et al., 2009;FRA, 2014;Tjaden and Thoennes, 1998). As a result of this higher incidence of stalking in young people, numerous studies have analysed both the prevalence and the dynamics of these behaviours in samples made up solely of university students, resulting in rates of victimisation notably higher than those perceived in the general population. Specifically, in research with samples of university students, these percentages rate between 9.2% and 35.2% in women and 2% and 35% in men (Bjerregaard, 2000;Feltes et al., 2012;National Union of Students, 2011).
In addition to studies on the stalking victimisation, another quantitative research tool that can heighten understanding of the incidence of the phenomenon, social perspectives on how to approach it, and, therefore, the most appropriate way to deal with it, is perception studies of this reality based on case scenarios. Such studies make it possible not only to compare how closely the social perception of the phenomenon mirrors the findings of studies that analyse it directly, but also to measure the degree of social punitiveness against such behaviours. This is especially well-suited to the case at hand, as the Istanbul Convention itself allows states to respond to stalking with non-criminal legislative measures.
However, such perception analyses are not yet as widespread as studies on the victimisation by such behaviours. Although the issue of social punitiveness and social perceptions of the criminal justice system is one of the major issues in criminology todaythe works of Roberts and Hough or Maruna and King on punitive populism and social perceptions of crime are well known Maruna and King, 2009;Roberts and Hough, 2002;Roberts, 2008)research on social perceptions of gender-based violence are not so common. Such analyses of manifestations of this form of violence as specific as the one examined here are even rarer.
With regard to social attitudes towards gender-based violence, beyond the issues related to the social perception of such cases included in the European survey developed by the FRA (FRA, 2014) or in the Eurobarometers (European Commission Directorate-General, 1999;TNS Opinion & Social, 2010), there is little available information. There is even less information on social perceptions of stalking. Although relatively plentiful in English-speaking countries (Cass and Mallicoat, 2015;Duff and Scott, 2013;Gavin and Scott, 2016;Scott, Lloyd and Gavin, 2010;Scott et al., 2014;Scott, Duff, Sheridan and Rajakaruna, 2018;Uhl, Rhyner, Terrance and Plumm, 2017), the information for continental European countries is scant (De Fazio, Sgarbi, Moore and Spitzberg, 2015; Kamphuis et al., 2005;Matos and Sheridan, 2015;Modena Group on Stalking, 2005;Sheridan, Scott, Archer and Roberts, 2017), particularly measurements of perceptions of stalking based on case scenarios. Given the high prevalence of stalking on university campuses (Baum et al., 2009;Bjerregaard, 2000;Feltes et al., 2012;National Union of Students, 2011), research that explores the perceptions of stalking by university students should be merited, since they can highlight prejudices and stereotypes regarding gender and relationships and also inform about the perceived effectiveness of the criminal justice system. The aforementioned perceptions also become potential indicators of the low stalking reporting rate -since the criminal justice system is perceived as unable to respond adequately to this phenomenon-and of the high prevalence of victimisation in this population group -since college students may be less deterred from stalking others if the criminal justice system reaction is considered too lenient (Cass and Rosay, 2012).
Furthermore, a sample made up of students of law, criminology and private investigation, such as the one used here, enables to know the perceptions of citizens having some understanding of what stalking is and how the criminal justice system works, which may imply the results from the study to be even more significant.
The relatively few social perception studies that have been conducted on stalking have sought to identify the situational and personal factors that influence people's attitudes towards it. Some of the factors that have been studied include the existence of a previous relationship between the victim and offender, the sexes of the victim and offender, the seriousness of the behaviour, and the use of violence. In addition to helping to identify the variables that most influence the consideration of these cases as stalking, the comparison of the findings of these perception studies with those of victimisation surveys reveals the distorted social conceptions respondents sometimes have of stalking.
One of the most important variables in explaining the social perception of stalking, as well as one of the variables for which the greatest differences have been found between victimisation and perception surveys, is the previous relationship between the victim and offender. In this regard, most of the research on perceptions has found that the acts are regarded as more serious when they are perpetrated by a stranger (see, amongst others, Hills and Taplin, 1998;Scott, Lloyd and Gavin, 2010;and Scott and Sheridan, 2011; in the opposite sense, see Dennison and Thomson, 2002). In contrast, victimisation surveys indicate that being stalked by a former partner takes a higher psychological and emotional toll on the victim (Aucoin, 2005, p. 40;Dovelius, Öberg and Holmberg, 2006, p. 18). Similarly, in those cases in which the victim and offender have previously been involved in a relationship, the victim is attributed greater coresponsibility for the acts than when the stalker is a stranger, in which case the victim is viewed as innocent. Accordingly, in cases in which the victim is stalked by a stranger, the need for police intervention is considered greater (Sheridan, Gillett, Davies, Blaauw and Patel, 2003). In short, cases involving strangers are viewed as 'classic stalking' and, therefore, are considered more credible and potentially more likely to be taken up by the police (Jordan, Wilcox and Pritchard, 2007). The explanations offered by the literature to understand this distortion are based, amongst others, on the idea of fear of the unknown (Hills and Taplin, 1998, p. 145), the false belief that it is easier to control the behaviour of someone the victim knows (Cass, 2011, pp. 9-10), the difficulty of distinguishing stalking from normal behaviours aimed at establishing a relationship (Dunn, 2002) or following a break-up (Dennison, 2007, p. 363), or on recourse to the just-world hypothesis (Scott, Lloyd and Gavin, 2010, pp. 1186-1187, whereby it is easier to mitigate the stalker's behaviour and assign responsibility to the victim if the parties have previously been involved in a romantic relationship. Second, the sexes of the parties involved in the case scenario also significantly influence respondents' attitudes: women are more strongly affected by these behaviours, in keeping with the findings reported by stalking victimisation studies. This belief can be attributed to the social and cultural construct of gender, according to which men are considered more aggressive and dangerous than women. Accordingly, police intervention is regarded as more necessary when the victim is a woman and the offender a man than when the roles are reversed (Phillips, Quirk, Rosenfeld and O'Connor, 2004).
The sex of the survey respondents themselves also influences perceptions of stalking.
These studies confirm that women are more likely than men to believe that stalking situations will cause fear or concern in the victim (Dennison and Thomson, 2000).
Given the near lack of quantitative analyses of attitudes towards stalking in Western Europe, and the complete lack thereof in Spain, the present study aims, first, to identify the perceptions of university studentsone of the groups most victimised by this phenomenon of the incidence and seriousness of stalking. To this end, the following items are measured in relation to two stalking case scenarios: the identification of the case as stalking, the assessment of the potential impact on the victim, and the evaluation of the victim's responsibility in causing the acts. Second, it aims to determine the study population's attitude with regard to the legal response considered proportionate for cases of stalking, first, measuring whether police intervention in such cases is viewed as necessary and, second, asking the sample to identify the most appropriate way of managing them from a legal point of view. At the same time, it aims to measure the influence of certain situational factorsi.e. factors related to the case scenario in questionand personal factors on these attitudes. Specifically, amongst the situational factors, it will focus primarily on the incidence of a previous relationship between the victim and offender, as well as the sexes of the parties involved in the stalking.
Amongst the personal variables, it will assess the influence of the sex of the survey respondents, as well as their degree of familiarity with stalking, whether because they themselves have been victims of this phenomenon or because they know people who have.

Methodology
The present analysis of perceptions of stalking amongst the university-student population was carried out within the framework of a survey on stalking victimisation conducted with a total sample of 1,162 university students pursuing degrees in law, criminology or private investigation. The participants in the total sample were selected from amongst 8 universities located in the Spanish regions of Catalonia and the Valencian Community and were stratified according to the total number of students enrolled in each province. Given that the total university student population in these regions was 237,426 1 and 140,171, 2 respectively, the sample had a level of confidence of 95% and a margin of error of ±5%.
The findings presented here, however, refer to a subsample of 1,010 participants from the total sample who did not self-identify as stalking victims when completing the questionnaire, even though 314 of them acknowledged having repeatedly experienced one or more of the behaviours included in the definition of stalking used by the researchers.
The data-gathering process, carried out between January and April 2015, was conducted in person at the respondents' universities. Participation was voluntary for all respondents and students were not rewarded for taking part in the survey. The day the questionnaire was completed at each of the selected universities, the researchers administering the paper survey explained the content and purpose of the research to the university students, clearly informing them about its voluntary, anonymous and confidential nature. Before completing the questionnaire, participants were provided with an oral and written definition of stalking, where it was defined as an insidious and unwanted course of conduct that can cause certain effects in the victim, such as fear or unease (Villacampa, 2009). The surveys were handed out to all students attending the lecture. The students were then informed that anyone not wishing to complete the survey could turn it in blank to the researchers at the end of the session. Students were supervised by the researchers themselves to ensure that they completed the questionnaires individually and that anyone unwilling to complete it could do something else. The data were processed using SPSS v. 20 software.
The instrument used in the study consisted of 32 questions and was divided into two parts. The first part (Questions 1 to 16) was intended for self-identified victims and, therefore, focused on questions aimed at characterising the victim's experience. The second part (Questions 17 to 32) was designed to be answered by the rest of the participants, including both those who had been victimised by such behaviours despite not having self-identified as victims (i.e. researcher-identified victims) and those who had not experienced stalking. The findings reported here correspond to the data from this second part of the questionnaire, which primarily focused on students' perceptions of two case scenarios of stalking.
This questionnaire was developed based on the instrument used by Scott, Lloyd and Gavin in a study with a sample of 315 university students in the United Kingdom and Australia (Scott, Lloyd and Gavin, 2010). It was included to meet the need to determine university students' perspective on stalking behaviours and the legal response that they feel they merit. At the same time, a similar instrument was chosen to that used in the aforementioned study so that the results obtained could be compared, even though the comparison can only be partial due to the methodological differences between the two analyses, which will be discussed below.
The instrument used in the present study began with a question aimed at determining whether the respondents had experienced stalking behaviours, even if they did not selfidentify as victims. To this end, they were asked whether they had repeatedly experienced one or more of a list of 10 possible behaviours included as possible manifestations of stalking on the survey on violence against women conducted by the FRA in 2014. 4 Next, they were asked if they knew anyone who had been stalked and, if so, how many people. The data obtained from the response to this question, together with the personal data requested for the survey, was considered essential to enable comparisons between the judgements made with regard to the proposed case scenarios of stalking depending on the respondents' personal exposure to stalking processes. These initial questions were followed by 12 additional questions corresponding to two different case scenarios (see Appendix). The two case scenarios differed only in the existence, or lack thereof, of a previous relationship between the victim and offender. In both cases, the stalking behaviour was the same; however, whilst in the first situation the stalker was described as a stranger to the victim, in the second the stalker was described as a former partner.
With regard to the two hypothetical situations, the students were asked to score the four questions shown below on an 11-point Likert scale: -To what extent do you consider the behaviour to be stalking? (Stalking) ('definitely not stalking' to 'definitely stalking') -Do you think police intervention is necessary to solve the situation? (Police intervention) ('definitely not' to 'definitely') -How likely do you think it is that the stalker's behaviour causes fear or concern in the victim? (Impact) ('not at all likely' to 'extremely likely') -To what extent do you think the victim is responsible for encouraging the offender's behaviour? (Responsibility) ('not at all responsible' to 'extremely responsible') Two additional questions were included in which the students: -Which of these outcomes would be the most appropriate if the stalker were to be tried for these actions? (possible responses: (1) shelving of the case; (2) the imposition of a protection order in the context of civil proceedings; or (3) a criminal conviction with the imposition of a protection order in the context of criminal proceedings).
-If the perpetrator of the actions had been the woman instead of the man, you would consider the facts... (possible answers: (1) equally serious; (2) more serious; or (3) less serious).
Unlike the methodology used in the study on which it drew, the present research included two, rather than three, case scenarios. Specifically, the case in which the victim and offender were merely acquaintances was omitted. This was because it was considered that formulating two cases of identical intensity in which the only difference was the prior relationship would make it possible to determine and compare the participants' opinions regarding two situations that were very different from a relational point of view. Second, given the close link established between stalking and genderbased violence, the present research sought to assess the influence of the gender of the parties in the case scenarios on the identification of the acts as stalking; hence the inclusion of the question on the seriousness of the acts had the stalker been a woman and the victim a man.
The data-gathering process, carried out between January and April 2015, was conducted in person at the respondents' universities, and the respondents completed the questionnaire individually and anonymously in one of their classes. The data were processed using SPSS v. 20 software.

Results
As noted, the results presented here refer not only to the university students' perceptions of stalking itself, but also to what legal response they consider appropriate for this reality and to the influence, on both points, of certain situational, or case-scenariorelated, variables (i.e. the existence of a previous relationship between the offender and victim and the sexes of the parties) and personal variables (the sex of the respondent and his or her degree of familiarity with stalking cases).
First, with regard to the personal variables examined here, the descriptive statistical results for the sex of the respondents and their degree of familiarity with cases of stalkingwhether as a victim of such behaviours or because they knew someone who had been a victim of themare shown in Table 1.
As for the results for the questions concerning the two stalking case scenarios, the descriptive statistics for which are shown in Table 2, a 2 (sex of the respondent: man/woman) x 2 (victim status: victim/non-victim) multivariate analysis of variance (MANOVA) was performed to determine the influence of gender on the assessed stalking indicators (F(1, 1010) = 2.68; ρ < .05; η 2 =.01) (see Table 4). Additionally, univariate analyses of variance (ANOVAs) using Bonferroni correction for alpha values of .05 were performed, as well as post-hoc analyses based on Tukey's procedure using alpha values of .05 for each individual question. The F ratios and significance values are shown in Table 4. To see the differences in the assessments depending on the existence of a previous relationship between the victim and offender, Student's t-tests were performed for related samples, yielding significant correlation values for each pair of variables and a t-statistic value that supports the existence of differences between the mean scores depending on the existence of a previous relationship, as shown in Table 5.
The results of such analyses are explained below for each of the measured items with regard to the perception of stalking.

Identification of the case as stalking
With regard to the first objective, i.e. to determine perceptions of stalking situations, in terms of the identification of such cases as stalking, the same hypothetical situation was more easily labelled stalking when the stalker was a stranger than when the stalker was a former partner, as has been reported elsewhere (Phillips et al., 2004(Phillips et al., , p. 1191Scott et al., 2014, pp. 227-228;Sheridan et al., 2003, pp. 92 and 94). In this regard, the distribution of the scale for strangers showed a large concentration of scores at the top end, revealing a leptokurtic distribution expressing a clear tendency to consider the situation an example of stalking (see Chart 1). In contrast, the distribution for the scale for the former partner peaked around 7 before declining at the top of the scale, indicating a lesser tendency to consider the acts stalking (see Chart 1). The t-tests performed revealed the existence of statistically significant differences with regard to the respondents' perceptions depending on the victim-offender relationship. They were more likely to consider cases in which the parties had not previously been involved in a relationship to be stalking (t(1009)=12.78, ρ <.001) (see Table 5). In contrast to the findings of some previous studies (Phillips et al., 2004, pp. 82-83, according to whom women are more likely than men to describe the acts as stalking) but consistent with those of others (Cass, 2011, pp. 4-5), the factorial ANOVA indicated that neither the respondent's sex nor the fact of having been a stalking victim him or herself influenced the likelihood of identifying the acts as stalking (see Tables 3 and 4). Nor was the fact of having known someone who had undergone this type of victimisation process significant (strangers: t(1008)=1.32, ρ=.188; former couple: t(1008)=.29, ρ =.774).

Seriousness of the facts according to the sexes of the parties
Given that the existing literature indicates that there is a close relationship between stalking and gender-based violence (Baldry, 2005;Burgess et al., 1997;Douglas and Dutton, 2001;Van der Aa, 2012), this study sought to measure the variance in the perceived seriousness of the acts when the sexes of the parties in the case scenario were reversed. Specifically, the respondents were asked whether they thought that the behaviours depicted in the case scenarios would have been more, equally or less serious if the offender had been a woman and the victim a man. In the case scenario in which the victim and offender were strangers, 87.8% of the respondents thought the facts would be equally serious, 11% less serious and 1.2% more serious. A slightly higher percentage of participants (90%) thought the facts would be equally serious if the roles had been reversed in the case perpetrated by the former partner, to the detriment of those who believed they would be less serious (8.6%) (see Chart 2).
No evidence was found of a statistically significant relationship in the assessment of the seriousness of the acts depending on the existence of a previous relationship between the victim and offender. Thus, contrary to what might be expected, the formulation of the case scenarios according to a pattern not expressive of a situation of gender-based violence (because the woman was not the victim, but the offender) did not lead most respondents to consider the acts to be less serious.

(Chart 2 around here)
Although the sexes of the parties in the case scenarios did not significantly influence the perceived seriousness of the behaviour, the sex of the respondents themselves was determinant in this regard. However, it was only statistically significant when the stalker and victim were strangers (χ 2 (2, 1010) = 9.88; ρ = .007, Φ=.099). Thus, whilst 90.3% of the women considered that the acts would be equally serious if they had been perpetrated by a woman, only 83.9% of the men did. In contrast, 14.2% of the men thought the acts would be less serious in that case, compared to 9% of women. No statistically significant differences were found, however, between the sexes in the case in which the parties were a former couple (χ 2 (2, 1010) = 5.35; ρ = .069). This notwithstanding, more men (10.9%) than women (7.3%) thought that the acts would be less serious if perpetrated by a woman.
Respondents' own status as a victim also significantly influenced their assessment of the seriousness of the acts when the genders of the victim and offender were reversed (strangers: χ 2 (2, 1010) = 9.55; ρ = .008, Φ=.106; former couple: χ 2 (2, 1010) = 8.35; ρ = .015, Φ=.099). In the case of strangers, stalking victims were more likely to consider that the actions would be less serious if committed by a woman (16.7% versus 9.1%). These respondents were also more likely to indicate that the same behaviours would be less serious were the victim a man and the offender a woman in the case of stalking by a former partner (13.6% versus 7.6%). However, whether or not the respondent knew someone who had undergone such a victimisation process was not relevant (strangers: χ 2 (2, 1010) = .095; ρ = .954; former couple: χ 2 (2, 1010) = .151; ρ = .927).

Assessment of the possible impact on victims
With regard to the sample members' perception of the potential impact of this stalking behaviour on the victim, once again a clear difference was found depending on whether the victim and offender had previously been involved in a relationship. The distribution of the data in the case where they were strangers showed a clear trend towards scores at the top of the scale. This indicates that the participants considered that being subjected to this behaviour was very likely to generate fear or concern in the victim. However, in the case of the former partner, the score for how likely the behaviour was to have a real impact on the victim was around 7-8. Above this score, the frequency of responses declined. Therefore, the respondents considered it less likely that the victim would suffer such negative consequences (see Chart 3).

(Chart 3 around here)
Specifically, the respondents thought that being stalked by a person the victim did not know was significantly more likely to cause fear or concern in the victim than the same behaviour carried out by a former partner (t(1009)=16.93, ρ <.001) (see Table 5). Whilst 70.4% of the participants rated the likelihood of such an impact in the case of strangers between 7 and 10 on a scale of ten, only 52.8% gave such scores in the case of the former partner.
Contrary to previous studies (Lambert, Smith, Gestman, Cluse-Tolar, Jiang, 2013, pp. 199-202), which indicated that both being a woman and being a victim made respondents more likely to state that these behaviours could be harmful for the victim, in the present study, these personal variables were not significant in explaining the sample's assessment of the effects of stalking on the victims (see Tables 3 and 4). Nor was having known someone who had undergone a similar type of victimisation relevant in this regard (strangers: t(1008)=.28, ρ =.777; former couple: t(1008)=-.67, ρ =.501).

Victim's responsibility for the acts
To conclude with the findings related to the studied population's perception of stalking, with regard to the responsibility attributed to the victim for causing the stalking process, once again statistically significant differences were found depending on the existence of a previous relationship between the parties. In this regard, the respondents considered the victim more responsible in cases in which the victim had previously been involved in a relationship with the stalker than in those in which they were strangers (t(1009)=-9.48, ρ <.001). In fact, whilst in the case of strangers, 56.3% of the participants thought that the victim bore no responsibility (i.e. rated the victim's responsibility 0 on a scale of 0 to 10), in the cases of the former couple, only 43.8% of the respondents said so. In this regard, the present findings confirm those of previous studies (Sheridan et al., 2003, pp. 92 and 94).
With regard to the influence of personal variables on the results, also in keeping with previous studies (Lambert et al., 2013, pp. 198-202), male respondents were more likely to assign greater responsibility to the victim. Specifically, men were significantly more likely than women to consider that the victim bore some of the responsibility for encouraging the stalker's behaviour both in the case in which the stalker and victim were strangers (F(1.1010)= 4.66; ρ =.031; η 2 =.04) and in the case where they were a former couple (F(1.1010) = 10.98; ρ =.001; η 2 =.06) (see Tables 3 and 4).
Continuing with the influence of personal variables on this point, contrary to what might be expected, as one would think that stalking victims would have greater empathy for people who had undergone similar processes, those respondents were more likely than those who had never been stalked to hold the victim in the case scenarios responsible. Respondents who had been victimised by stalking considered that, both in the case where the stalker was a stranger and in the case where the stalker was a former partner, the victim had been responsible for encouraging the stalker's behaviour; in the latter case, the difference was moreover statistically significant (F(1, 1010) = 15.40; ρ < .001; η 2 =.01) (see Tables 3 and 4). However, no significant differences were found in this case either depending on whether or not the respondents knew someone who had been stalked (strangers: t(1008)=-.87, ρ =.386; former couple: t(1008)=-.419, ρ =.675).

Need for police intervention in cases of stalking
With regard to the second objective of the present study, i.e. to determine the sample population's views on the appropriate legal response to stalking, let us first look at whether the respondents considered that formulated cases necessitated police intervention. The need for police intervention received a score of around 5 in the case of both stalking by a stranger and stalking by a former partner, although the said need was significantly perceived in the case in which there was no previous relationship between the stalker and the victim (t(1009)=2.17, ρ =.031). These findings corroborate those reported for previous empirical studies (Scott et al., 2014, pp. 227-228;Sheridan et al., 2002, pp. 92-94), which have noted that police intervention is perceived as more necessary when the stalker is a stranger. Despite these differences, the participants did not consider police intervention essential in either of the presented cases of stalking: only 3.4% of the respondents in the case of stalking between strangers, and 2.5% in the case of the former couple, considered that police intervention was absolutely necessary, i.e. gave it a score of 10 out of 10 on the Likert scale.
With regard to the incidence of the personal variables, the respondent's sex was the most important in explaining the need for police intervention. Thus, statistically significant differences were detected depending on the sex of the respondent in the case in which the victim and offender were presented as strangers (F(1.1010) = 5.37; ρ = .021; η 2 =.09), with women being more likely than men to request police intervention. In contrast, neither the fact of having been stalked (see Tables

Appropriate legal response to stalking
Finally, in order to determine the sample's opinion with regard to the appropriate legal response to this phenomenon, participants were asked what they thought would be the most appropriate legal response to the described acts. Most of the respondents proved to be relatively nonpunitive, as the literature indicates will happen in cases in which social punitive attitudes are measured based on case scenarios (Diamond and Stalans, 1989, p. 87;Hutton, 2005, pp. 243-258). The majority of respondents considered that the most appropriate solution should the case be brought to trial would be the imposition of a protection order in the context of civil proceedings (60.7% in the case of stalking between strangers and 57.6% in the case of the former couple). The second most frequent choice was the shelving of the case due to its scant seriousness (32.9% in the case of stalking between strangers, and 31.5% in the case of the former couple). The option of responding with a criminal conviction ranked third and last, with only 6.4% of respondents choosing it in the case of stalking between strangers and 10.9% in the case of stalking by a former partner (see Chart 4).

(Chart 4 around here)
Significant differences were observed in the legal response considered appropriate depending on whether the victim and offender had previously been involved in a relationship (χ 2 (4. 1010) = 711.36; ρ < .001, Φ=.593). Thus, the respondents were more likely to request a criminal conviction in those cases in which the victim and offender had previously been a couple, with 41% of respondents who had favoured a criminal conviction in the case of strangers considering this measure to be inappropriate when the stalker was the victim's former partner: specifically, 35.5% chose the imposition of a civil protection order and 5.5% the shelving of the case. These findings contradict those reported in previous studies, which indicate that respondents take a more punitive attitude when the victim and offender are strangers (Scott, Lloyd andGavin, 2010, p. 1191).
With regard to the incidence of the personal variables on the choice of the appropriate legal response in cases of stalking, statistically significant differences were found depending on the sex of the respondents (χ 2 (2. 1010) = 17.23; ρ < .001, Φ=.131).
Specifically, women were more punitive than men, as 10.3% of women chose a criminal conviction compared to only 5.7% of men. Likewise, whilst 39.8% of men considered that the most appropriate response would be to shelve the case, only 28.4% of women did. However, these differences could not be confirmed in the case scenario of the former couple (χ 2 (2, 1010) = 5.53; ρ = .063), although more women (12.3%) than men (8.8%) chose a criminal conviction.

Conclusions
The first objective of this research was to determine the perceptions of Spanish university students regarding the incidence and seriousness of stalking and the importance to that assessment of both situational and personal variables related to the respondents themselves. The findings show that the most decisive dependent variable of a situational nature was the existence of a previous relationship between the offender and victim. The other situational variable studiedthe sexes of the parties in the stalking casewas not as important in determining the seriousness of the case, at least not in those cases in which the parties involved had previously been a couple. Less determinant with regard to the assessment of the incidence and seriousness of stalking were the personal variables of the respondents studied, i.e. their sex and their degree of familiarity with such cases.
In short, the findings of this study suggest that, with regard to university students' assessment of stalking, stalking by a stranger is more easily identified as such, largely regardless of the offender's sex, although the acts would sometimes be regarded as less serious if the stalker were a woman. Thus, this population seems to downplay the seriousness and effects of stalking by former partners, presumably in accordance with the logic of holding the victim him or herself responsible.
The second objective of this research was to determine the analysed population's attitude towards what would constitute a proportionate legal response in such cases depending on the aforementioned variables. This was done, first, by measuring the extent to which they felt the situations necessitated police intervention and, second, by asking the respondents to indicate what they thought would be the most appropriate legal response. In general, the sample's response was largely nonpunitive and not oriented towards the intervention of the criminal justice system as the first resort for solving these situations. This confirms the conclusions drawn from the literature on social punitiveness, which indicate that attitudes become less punitive when the questions are formulated based on case scenarios, as already noted. Here again the situational variable related to the relationship existing between the parties was decisive in explaining the attitudes of the sample members. First, police intervention was clearly perceived as more necessary when the stalker and victim were strangers than when they were a former couple, although women were more likely than men to indicate the need for such an intervention, regardless of how familiar they were with the stalking phenomenon. Second, the respondents were not particularly in favour of the use of criminal sanctions, which they relegated to third place, choosing the imposition of a civil protection order or the shelving of the case as the preferred solutions in such cases.
Notwithstanding the above, in the present study, inconsistently, a criminal conviction was more often chosen when the stalker was a former partner than when it was stranger.
In light of the foregoing conclusions, it remains to be seen whether public opinion on this reality is subject to certain distortions, as has been found, as noted, in other cases in which the results of perception surveys have been compared with those of victimisation surveys. Indeed, that is also the case here. Whilst it can be deduced from this study that the studied population considers stalking by a stranger to be the most impactful and purest form, victimisation surveys indicate that stalking behaviours are usually carried out by a partner or former partner (see, in all cases, FRA 2014, pp. 85-86). They further indicate that the impact on the victim is greatest when the victim and the offender are or have been involved in a relationship (Aucoin, 20015, p. 40, and Dovelius, Öberg, and Holmberg, 2006, p. 18). To cite just one example, the only stalking victimisation survey of the university-student population conducted in Spain so far found that most victims had been stalked by people from their inner circle (37% of cases) or with whom they had been involved in a romantic relationship (25%) (Villacampa and Pujols, 2017a, p. 16). That same study showed that, although the experiencing of feelings such as fear was clearly correlated with cases of stalking by strangers, the suffering of adverse psychological effects by victims was closely related to stalking by people from the victim's inner circle (Villacampa and Pujols, 2017b, pp. 14-15).
Given the effects of stalking on the victims, in terms of both the emotions it causes and the psychological consequences, depending on various variables that come together in the stalking process, two behavioural models were identified in that victimisation survey as the most harmful in relation to these processes (Villacampa and Pujols, 2017b, p. 28). The first model involved stalking by a family member or sentimental partner seeking physical proximity to the victim, which would correspond to the model of stalking related to domestic and gender-based violence. In the second model, the stalker was a stranger, also seeking physical proximity to the victim, such that the most harmful cases of stalking would correspond to the pattern of a stranger pursuing the victim, in accordance with the earliest conceptualisations of the phenomenon.
Whilst the reality of the victimisation measured in the university-student population points to these two models as the most common, the findings of the present perception survey seem to identify the stalking phenomenon with only one of them, specifically, the second, which is hardly the most prevalent. Indeed, the findings of the present research show that, whilst the variable concerning the victim-offender relationship is the most determinant in the perception of these behaviours, the stalking seems to be minimised as a phenomenon related to domestic and gender-based violence, contrary to what the literature shows. It would seem that cases of stalking occurring between a couple are considered to fall within what is identified as everyday interpersonal conflict or as an outcome of a romantic relationship that ended badly (Miglietta and Acquadro Maran, 2017). To explain this perception, it has already been noted that previous analyses of the issue have cited, amongst other things, the false belief that it is easier to control the behaviour of someone you know, the difficulty of discerning socially normal behaviours from socially abnormal ones in certain sentimental relational contexts, and the just-world hypothesis. However, the use of other explanatory hypotheses not far removed from those focused on the idea of fear of the unknown could also be tried, such as the idea that situating the essence of these behaviours in stalking by a stranger better fits the pattern of stalking depicted in films, which may have influenced public opinion.
It may also coincide better with the position defended by constructions based on the idea of stranger danger, which have been used to explain various victimisation phenomena, such as sexual victimisation (Finkelhor, 2009;Yung, 2010), that, like stalking, are usually perpetrated by someone from the victim's inner circle.
In any case, with regard to the first objective of this study, it can be concluded that the studied population segment's perceptions of stalking do not correspond to actual stalking victimisation. It would thus be desirable, no doubt as with other forms of victimisation, to disseminate the results of victimisation analyses that could help to shape a better informed public opinion with fewer distortions.
As noted, precisely this type of meditated public opinion tends to be less punitive than simple public opinion. This aspect could be considered to have a bearing on the second objective of this study, concerning the punitiveness shown with regard to stalking.
However, in Spain, in the specific case of gender-based domestic violence, the mere dissemination of research on this type of victimisation is unlikely to be enough to counteract the potential effects on public opinion of years of application of a clearly punitive criminal policy. In this regard, the passage of Organic Law 1/2004, which undoubtedly entailed an undeniable regulatory milestone and helped to thrust into the public light something that had, until quite recently, been a private matter, namely, domestic violence, also conveyed the message that the response to this reality should primarily come from criminal law. This same message may also have taken root amongst young people in the case of stalking by a partner, for whilst in this study stalking by strangers was considered more seriouseven though not even in that case was intervention by the criminal justice system advocatedinconsistently and contrary to the findings of similar studies (Scott, Lloyd, and Gavin, 2010), the application of criminal sanctions as a means of solving the problem received a higher score in the case of stalking by a former partner than by strangers. This result would be impossible to understand compared to similar studies were it not for the fact that the criminal justice policy pursued by Spain in matters of gender-based violence is clearly more punitive than that pursued by other Western legal systems, in which, together with the criminal protection of victims of gender-based violence, it is common to implement protection mechanisms of a civil nature, too. This study could thus constitute an example that a markedly punitive criminal justice policy may not always be preceded and driven by a punishment-oriented public opinion, but rather, on the contrary, public policy itself helps to shape public opinion. This would be possible through the transmission of messages oriented in a certain direction that, in the present case, would be that violence of any intensity occurring within the context of a couple must merit an almost exclusively criminal justice response. In a context such as the one described, the containment of social punitiveness does not depend solely on how informed public opinion is, but also largely on a change of strategy in legislative policy.   Strangers 7.5 1.8 12.7*** 5.0 2.5 2.1*** 7.8 1.8 16.9*** 1.1 2.0 -9.4*** Former couple 6.8 2.2 4.9 4.9 6.7 2.4 1.7 1.7 *p < .05. **p < .01. ***p < .001.

Chart 3. Histograms showing the normal curve of the attitude scales related to the potential impact on the victim in the case of strangers (left) and a former couple
(right).