Prevalence, dynamics and characteristics of forced marriage in Spain

Forced marriages were criminalized in Spain in 2015, both as a specific form of coercion (offence against liberty) and as a type of human trafficking. However, the criminalisation was carried out with limited data on existing cases and without empirical research questioning whether forced marriages were taking place in Spain and what specific individuals were at risk. This paper presents the main findings of an empirical study aimed at detecting the existence of victims of forced marriage in this country and at describing the dynamics and characteristics of the processes which led to their victimisation. A total of 150 support organizations operating in Spain reported their knowledge of the extent and dynamics of forced marriages, revealing the existence of people forced to marry against their will, including some legal residents and even Spanish nationals. The research identifies at least 57 victims who have received assistance from support organizations. People most at risk were very young women, usually from North African, Sub Saharan African or Central Asian, as well as, to a lesser extent, Roma, and with a large percentage of Spanish nationals and legal residents. Mechanisms used to compel the victims are usually subtle, far from those more incisive active means required in the Spanish legislation. The results contribute to the debate on what is tolerable in today’s globalized society and underscore the need to understand this reality before regulating it.


Introduction
Forced marriage refers to marriages performed without the consent of one or both parties. Despite the lack of an international consensus regarding the definition, the practice is widely rejected and has been denounced by international organizations and NGOs, especially when it involves minors [1][2][3]. Researchers, government agencies and NGOs have recently published several reports on the extent of forced marriage at the global level. The findings underscore the harm that the practice entails for the victims' fundamental rights [4,5]. They also highlight the challenge of distinguishing between forced marriage and the practice of arranged marriages, found in certain cultural traditions and characterized by the intense involvement of the family and community in proposing candidates for marriage and conditioning the decision [6][7][8][9]. Finally, they suggest that forced marriage is a global phenomenon that is most common in certain regions of Africa and Asia, but of which evidence can also be found in the West [3,[10][11][12][13][14][15][16][17][18][19]. However, the scant official data and the fact that it is usually practiced in the domestic and family spheres, often between members of cultural minorities who are less likely to turn to the formal control system, have exacerbated the lack of knowledge about the phenomenon.
The European Union Agency for Fundamental Rights [10] has called for more research on the practice, noting that there are three ways to determine the prevalence of incidents of forced marriage, all of which should be supplemented with qualitative research models: population-based surveys, data from police and judicial records, and civil-society data collected by NGOs and victim service providers. In fact, most of the empirical research aimed at determining the magnitude of forced marriage, the profile of its victims and the commissive dynamic use the quantitative and qualitative data collected by the staff of organizations that deal with victims of the phenomenon [13,14,18]. The information provided by these organizations has enabled the verification of the existence of victims in the West.
The available data on the practice of forced marriage in Spain are very limited. They have mainly been taken from certain police records in Catalonia and, more recently, from the Matrifor project on human trafficking and forced marriage [20]. The lack of reliable information on the reality of this phenomenon did not, however, prevent Spanish lawmakers from deciding to classify forced marriage as a criminal offence under Organic Law 1/2015. The criminalization entailed the country's adhesion in this matter to the mandate contained in Article 37 of the 2011 Istanbul Convention and the almost direct adaptation of the precept into Article 172 bis of the Spanish Criminal Code (CP). Forced marriage constitutes a specific modality of coercion, typified among liberty offences, which requires severe intimidation or violence to compel another person to marry. Deception is also included as a comissive mean but only for the conduct of forcing someone to leave Spain or no to return to it with the purpose of compelling her to marry. The result has not been particularly well received by scholars, who have highlighted both the ascendant nature of this offence over other previously established offences in the Criminal Code [21][22][23] and the unsuitability of prioritizing recourse to criminal law despite the difficulties of intervening in a matter closely linked to the family and community spheres without providing for the deployment of other protective measures [24,25] increasing the risk of othering certain cultures [26,27]. Moreover, this legal concept of forced marriage is linked with the traditional understanding of it, focused on the sharp difference between arranged and forced marriages, and distances itself from more contemporary conceptions of this reality. According to these, less attached to the binary conceptualisation of coercion and consent [7], a more comprehensive notion of forced marriages appears that makes more fluid the limits between these categories. It describes a process of leading women to marry that can also encompass the application of more subtile means to force them, such as coercive control [28].
Besides its criminalization as a coercion offence, unlike common practice in the international arena, Spanish lawmakers also recognized forced marriage as a form of human trafficking, following the preliminary considerations of Directive 2011/36/EU on preventing and combating trafficking in human beings and protecting its victims. This meant including it as one of the purposes of exploitation provided for under Article 177 bis CP and thus opening the doors to the application of much harsher penalties than those provided for under Article 172 bis CP.
This paper aims to contribute to the debate at the European level on the incoherence between the criminalizing option pursued as a priority by some countries, including Spain, and the dearth of research on the reality of this phenomenon required both to properly define the criminal offence and to develop the necessary instruments to adequately protect and assist the victims.

Objectives
The main objective of this research was to determine whether forced marriages exist in Spain by confirming whether social or support organizations have dealt with current or potential victims of the phenomenon. In addition to this main objective, the research also aimed to establish a profile of the victims of forced marriages, i.e. to determine which groups are most at risk of being victimized, whether for reasons of gender, age, nationality or membership in certain groups or communities. Second, it sought to identify the dynamics of victimization entailed by forced marriage, primarily by identifying the mechanisms used to pressure the victims, as well as by determining where the marriages are performed. Finally, it aimed to determine the conditions under which the aforementioned organizations learn of such cases, identifying the organizations most effective at bringing these victimization processes to light, determining the circumstances of time and place in which they are reported, and identifying the characteristics of the victims who turn to support organizations.

Methodology
In order to contact organizations that could potentially be sources of information, a database was developed of local, regional and national organizations which it was thought might come into contact with victims of forced marriage who might turn to them for advice or assistance. This resulted in a total of 518 organizations, all of which were invited to be part of the sample. The list was built by searching the websites of organizations that provide counselling and support services for victims of domestic and gender-based violence or human trafficking, victim assistance offices, cultural associations of immigrants, and secondary schools. An initial classification of these organizations was performed based on how likely they were thought to be to come into contact with victims of forced marriage.
Thus, a purposive sampling technique was used in which the sampling units were rationally chosen based on their characteristics, i.e. depending on whether, based on the types of services provided, the organization was considered likely to have come into contact with victims of forced marriage.
At the same time, an online questionnaire was created using the Google Forms platform. Because it was an instrument developed specifically for this research, it was pre-tested with a sample of 10 people. Once the necessary adjustments had been made, it was successively e-mailed, along with a letter explaining the research, to the organizations from the database every two weeks for a period of five months, from May to September 2016. More efforts were made with those organizations that the researchers deemed most likely to encounter victims of forced marriage.
In light of the low response to the online questionnaireonly 19 responses had been received by September 2016the methodology was modified. Between October 2016 and February 2017, the 499 organizations that had not responded to the initial online request were contacted by phone. Once they had been recontacted by phone, and the content of the research had been explained, the online questionnaire was sent to them again. This increased the total number of responses to 150, which was the actual sample for the research.
The questionnaire used consisted of three parts. The first two concerned the organization, whilst the third contained files to be completed on possible victims it might have assisted. The first part of the survey comprised four general questions about the organization, intended to identify it and its scope of activity, as well as to determine whether it considered forced marriages a priority. If the organization had come across a case of forced marriage, it was asked to continue completing the questionnaire; otherwise, the questionnaire ended there.
The second part of the questionnaire, which was solely for those organizations that had come across a case of forced marriage, included five questions. Respondents were asked to indicate the specific number of cases that they recalled, how the facts had come to their attention (i.e. whether they had been reported as cases of domestic/ gender-based violence, directly as cases of forced marriage, not known, or other), who had brought them to their attention (the victim, a relative, a friend, an acquaintance, one of the victim's teachers, other, or the respondent did not know), and whether the victims were already married or at risk of being married when the organization was contacted. Finally, the organizations were asked whether they recalled specific details about one or more of the assisted victims and, if so, to complete the third part of the questionnaire.
The third part consisted of individual victim files. Each organization could complete a maximum of 3 files, consisting of 11 questions each. These questions concerned the victims' sex, their age at the time they were married or promised in marriage and at the time of the organization's intervention, whether they had any children in their care at the time of the organization's intervention, and their nationality if they were not Spanish. Additionally, the following questions were included with the stated possible responses: -Victim's marital status at the time of the organization's intervention. Once the responses had been compiled and anonymized, they were processed with the statistical software SPSS v24.

Sample
Of the invited sample of 518 support, social and educational organizations, 150 ultimately responded to the online questionnaire. This group constituted the real sample.
With regard to their geographical location, nearly two thirds of these organizations were headquartered in Catalonia. This is because the mailing of questionnaires was more exhaustive in this region. Nevertheless, one third of the organizations were located in other regions (see Fig. 1). Amongst those located in Catalonia, in addition to those with a regional scope (n = 3), by province, the majority were located in Barcelona (n = 38), followed by Lleida (n = 24), Tarragona (n = 20) and Girona (n = 14).
With regard to the type of activity in which they engaged, most of the organizations (Fig. 2) provided assistance to victims of domestic and genderbased violence (n = 74), victims in general (n = 51), victims of human trafficking (n = 34) and victims of forced marriage (n = 32), as well as to immigrants (n = 34). Of the entire real sample, 39 organizations (30.5%) included forced marriage as a priority versus 84 (65.6%) that did not.
Finally, with regard to the degree of completion of the questionnaire by the sampled organizations, 58% (n = 88) responded only to the questions on basic information about the organization and indicated that they had not encountered victims of forced marriage. A total of 42% (n = 62) had come across victims of this phenomenon and thus continued with the questionnaire. Of these organizations, 13% of the total sample (n = 19) completed only the second part of the questionnaire, without filling out any victim files, whilst 29% (n = 43) completed the entire questionnaire, yielding a total of 57 specific victim files.

Results
The fact that 62 of the organizations that completed the questionnaire reported having detected victims of forced marriage, completing a total of 57 specific victim files, confirms that forced marriage also happens in Spain. In this regard, the present study corroborates the findings of previous research [3,[10][11][12][13][14][15]18], insofar as it verifies the existence of victims of this process in the first world.

Socio-demographic characteristics of the victims of forced marriage
Once the existence of forced marriages in Spain had been verified, it was possible to infer a victim profile from the 57 victim files completed by the organizations.
With regard to sex, all of the identified victims were women; none of the organizations reported having assisted any male victims of such processes. This finding is  consistent with previous research that has highlighted the strong gender bias in the practice of forced marriage [10,13,29].
As for age, Fig. 3 compares the victims' ages at the time they were married or promised in marriage and when the organization intervened. As can be seen, the victims were younger when they were married or promised in marriage than when they sought assistance. Furthermore, of the 33.9% of the victims who were minors when they were married or promised in married, 1.8% were 15 or younger and 32.1% were between the ages of 16 and 17. A total of 76.8% of the victims were 21 or younger. These findings also corroborate those of previous studies [14,15,30].
Given the relationship established in some political-criminal discourses between forced marriage and membership in certain ethnic, cultural, religious and even national minorities [31,32], in addition to sex and age, the organizations were also asked about the victim's nationality, community or ethnicity, and professed religion.
By nationality, identified with the status of being citizen of a particular country, along with the percentage of victims of unknown nationality, the most frequent nationality was Moroccan, followed by a considerable percentage of victims of Spanish nationality and, at a greater distance, Pakistanis and Gambians (see Fig. 4). These results are coherent with the migration patterns in Spain shown by data from the Spanish National Statistics Institute, according to which Moroccan is the most widespread foreign nationality in Spain. Chinese and Pakistanese nationals constitute also two considerable groups of foreign nationals in the Spanish society, even if the Asian population as a whole is smaller than that of African and Soth American natives among our citizens.
By community or ethnicity, the majority of the victims were North African (42.9%), sub-Saharan African (25%) or Central Asian (8.9%), although 7.1% were Roma (Fig. 5). However, the victim's community or ethnicity did not explain the age at which she was married or promised in marriage, as the relationship between the two variables was not statistically significant (χ2 (27, N = 57) = 26.88; ρ = .470; Φ = .419). With regard to the victims' professed religion, according to the respondents, the overwhelming majority of the identified victims were Muslim, although a considerable percentage of victims were not identified as practitioners of any religion (see Fig. 6). The percentage of Catholic victims was 3.6%, whilst the percentages of Sikh, Eastern Orthodox and Hindu victims were even more negligible. Nor was a statistically significant relationship found between the victim's religion and the age at which she was married or promised in marriage (χ2 (15, N = 57) = 20.17; ρ = .165; Φ = .350). With all religions, including the one to which most of the victims belonged, the women were married or promised in marriage after coming of age.  The variables were correlated to determine which ones were the most relevant to explain victimization by forced marriage. Those related to nationality, religion and community or ethnicity had the greatest explanatory weight.

Victimization dynamic
As noted, the forced marriage dynamic that this study aims to describe refers, first, to the means used to force the victim to marry and, second, to the place in which the marriage was performed or expected to take place.
According to the descriptive statistics (Fig. 7), the means most often used to force victims to marry were family pressure and attachment to the victim's tradition or culture. The use of force or intimidation was less common, and the use of deception or violence even less so. In other words, it is more common to resort to subtle mechanisms to compel the victim to marry that are not included as active means of  Prevalence, dynamics and characteristics of forced marriage in Spain committing the criminal offence of forced marriage provided for in Article 172 bis CP than to use means, such as threat or intimidation, deception or violence, that are provided for in the criminal offence.
Although one might think that the means used might depend on the age of the victim at the time they are married or promised in marriage, in fact no statistically significant relationship was found between the means variables provided for under the CP and the victim's age when married or promised in marriage (χ2 (3, N = 57) = 2.494; ρ = .476; Φ = .211). However, the means used with regard to minor victims were found to be subtler than those used with victims who were of age, who were more often threatened. Thus, when the means are grouped into those provided for in the CP in the offence of forced marriage and those that are not, criminally relevant means were used in 10 cases with victims under the age of 18 versus 20 cases with victims of legal age. Nevertheless, subtler mechanisms were more commonly used than drastic ones with all age brackets.
The means used likewise did not depend on the victim's nationality or the community or ethnicity to which she belonged, at least not to the extent in which such a relationship might be expected based on the link that some discourses establish between the practice of forced marriage and certain minorities or cultural communities. In fact, no statistically significant relationship was found between the variables referring to the victim's nationality or community and whether the means used to compel her to marry were included in the offence of forced marriage. 1 However, the most insidious meansviolence or intimidationwere used with victims of Spanish nationality, amongst whom the most common means of commission was threats. This differed from other nationalities, such as Moroccan, with which the use of family pressure or attachment to tradition was more common. In contrast, with sub-Saharan African victims, the use of more insidious means was not as limited. Although such results are not statistically relevant, they were confirmed by cross-checking the variables referring to the victim's legal status and the means used, which, again, showed that the use of means such as violence or intimidation was more common in the case of victims who were Spanish nationals (11 cases of the use of means that are included in the CP versus 9 that are not) or legal residents in Spain than in cases of illegal residence.
Nor did the victim's religion explain the means used in the forced marriage dynamic. Again, the relationship between the variables referring to means provided for in the CP and the victim's religion was not statistically relevant (χ2 (5, N = 57) = 5.344; ρ = .375; Φ = .312). Nevertheless, according to the available data, with Muslim victims, means not provided for in the CP were most often used, with Catholic and Hindu victims, only means not provided for in the CP were used, and with Eastern Orthodox and Sikh victims, means that are and are not provided for in the CP were used equally.
As for where the marriage had taken place or was expected to take place, the most frequent response was in the victim's country of origin, although in 30.4% of the cases, the marriage was held or expected to be held in Spanish territory (Fig. 8). 1 Statistical values for the correspondence between the variables "means included in the CP" with the variables: "victim's nationality" (χ2 ( To explain this finding, no statistically relevant correlation was found between this variable and that referring to the victim's age when married (χ2 ( 439; Φ = .304). However, the data show that the oldest victims were most often married in Spain or another EU country, whilst the youngest ones were most often married in their family's country of origin. Likewise, in addition to the victims of Spanish nationality, who were logically more likely to be married in Spanish territory, victims who were legal residents in the country were also more likely to marry in it. Therefore, the victims most likely to go to their country of origin to be married were those who lacked this legal residence. By community, North African and sub-Saharan African victims were more likely to be married in their country of origin, whereas Central Asian victims were equally likely to marry in Spain and their country of origin and Roma victims were mainly married in Spain. By nationality, aside from the Spanish victims, Moroccan and Guinean victims were more likely to be married in their country of origin, whereas Gambian victims were more likely to be married in Spain. However, due to the small number of victims in each category, no patterns could be established.

Conditions in which these victimization processes come to light
The final objective of this research was to determine how these types of processes were brought to the organizations' attention. To this end, it focused on three factors: the characteristics of the organizations that proved most effective at identifying them, the circumstances of time and place in which the cases were reported, and the types of victims who turned to them for assistance.

Characteristics of the organization
The organizations that proved most effective at detecting or learning about cases of forced marriage (see Fig. 9) were those dedicated to assisting victims of domestic and gender-based violence (scope of activity of 65% of the organizations with knowledge of cases), assisting victims of human trafficking (scope of activity of 38.3% of the organizations) and assisting victims of forced marriage (35%), in that order. In all three cases, the correlation between the variables referring to the organization's scope of Fig. 8 Place where the marriage was performed or expected to be performed activity and to the detection or knowledge of cases of forced marriage was statistically significant. Respectively, for organizations that assist victims of gender-based and domestic violence, the results were χ2 (2, N = 150) = 13.62; ρ = .001; Φ = .303; for those assisting victims of human trafficking, they were χ2 (2, N = 150) = 14.45; ρ = .001; Φ = .312; and for those assisting victims of forced marriage, they were χ2 (2, N = 150) = 13.62; ρ = .003; Φ = .281. Additionally, a logistic regression analysis was performed to determine the likelihood that the group of organizations whose scope of activity included assisting victims of domestic and gender-based violence would detect victims of forced marriage. According to the results, they were 2.597 times more likely to identify them than an organization whose scope of activity did not include assisting such victims.
These organizations were also the ones to complete the most victim files, as can be seen in Fig. 10.
The finding that those support organization whose scope of activity might lead their staff members to be familiar with such processes were more effective at identifying cases of forced marriage was consistent with those of previous studies that have highlighted the close relationship between professional training and the ability to detect cases of victimization in areas related to the one under study here, such as human trafficking [33,34].
However, the fact that, in the present study, no statistically significant positive correlation was found between having forced marriage as a priority and having knowledge of more cases (χ2 (4, N = 150) = 11.32; ρ = .023; Φ = .211) calls this finding into question. Indeed, organizations that have forced marriage as a priority were not even the ones to detect the most cases in percentage terms: they detected 44% of the cases versus 54% by organizations whose priorities do not include forced marriage. Nevertheless, according to the results of the logistical regression analysis, organizations that prioritize forced marriage were 2.823 times more likely to have knowledge of cases of it than those that do not. Furthermore, whether or not an organization prioritizes forced marriage did explain the number of cases that it recalled having dealt with, as it was precisely those organizations that do prioritize it that recalled having handled more than 5 cases. Accordingly, they were also the organizations to complete the most victim files (51.8% of the organizations that completed them had forced marriage as a priority).
Although the analysed organizations' scope of activity was determinant in explaining their ability to detect cases, it was not especially relevant to explaining how the cases had been brought to their attention, i.e. whether as cases of forced marriage, domestic or gender-based violence or something else (see Fig. 11). According to the bivariate analysis, the organizations that were most effective at detecting cases of forced marriage usually learned of them as cases of domestic or gender-based  violence. This finding is understandable for organizations dedicated to assisting victims of domestic and gender-based violence, but it also held true for organizations dedicated to assisting victims of forced marriage and of human trafficking. In contrast, in the case of organizations dedicated to assisting immigrants, although the difference was not statistically significant, cases were more often reported directly as cases of forced marriage than as cases of gender-based violence (χ2 (1, N = 150 Circumstances in which the cases were reported to the organizations Whilst most of the cases were reported to the organizations as cases of gender-based violence, in terms of when they came to light, they were slightly more likely to be reported once the victims were married than when they were at risk of being married (see Fig. 12). When the organizations were asked whether the victims were married or at risk of being married when the case was brought to their attention, they indicated that a slightly higher percentage of the victims were already married.
The facts were more likely to be brought to the attention of the organizations most effective at detecting cases of forced marriage once the marriage had already taken place rather than when the victims were still at risk of being married. In numerical terms, of the total number of organizations dedicated to assisting victims of genderbased and domestic violence, 20 indicated that some of the victims had already been married versus 18 that said they were still at risk of being married. These figures were 15 versus 10 for organizations dedicated to assisting victims of human trafficking, and 14 versus 11 for organizations dedicated to assisting victims of forced marriage.
As a general rule, it was the victims themselves who reported the cases to the organizations, followed by teachers, acquaintances or friends of the victim, i.e. third parties, and only very rarely one of the victim's relatives (Fig. 13). Grouping the Fig. 12 When was the case brought to the organization's attention? variable concerning who reported the case to the organization into three categories showed that it was most often the victims themselves who did so (63.6%), followed by third parties (29.9%), with cases reported by relatives coming in a distant third (3.9%).
Victim's status upon requesting assistance from the organization With regard to the victims' statuses when they turned to the surveyed organizations for help, i.e. when their situation of victimization came to light, as will be shown below, their personal and family situations afforded them a certain degree of empowerment in terms of certain variables compared to the permanent variables they had at the time they were married or promised in marriage.
First, with regard to the victims' age at the time they sought assistance, the results of the research (see Fig. 3 above) show that the victims were older at the time the organization intervened than they were when they were married. Just over 23% of the victims were minors when the organization intervened, whilst the remaining 77% were aged 18 or older, and nearly 27% were over the age of 25.
As for the victims' marital status at the time of the organization's intervention, most were single, probably because they did not ultimately marry; however, the second largest group was married women, presumably also in this case because they did marry (see Fig. 14). Although the relationship was not statistically significant (χ2 (9, N = 57) = 19.28; ρ = .023; Φ = .339), the younger the victims were at the time they were promised in marriage, the less likely they were to be married when the organization intervened. This suggests that the marriages of women who were of age were more likely to be performed prior to the organization's intervention than those of minors.
Additionally, according to the respondent organizations, most of the women who turned to them did not have children in their care when they sought assistance (67.9%), although 32.1% of them did. Furthermore, a statistically significant positive correlation was found between whether or not they had children at that time and the age at which they were married, such that the younger the woman was when she was married or promised in marriage, the less likely she was to have children in her care when the organization intervened (χ2 (3, N = 57) = 15.97; ρ = .001; Φ = .534).
Finally, most of the women who sought assistance were legal residents in Spain at the time they turned to the organization (53.6%), when not directly Spanish nationals (21.4%) (Fig. 15). The bivariate analysis of this variable with that referring to the age at which they were married revealed a trend for women with Spanish nationality or who were legal residents in the country to be married or promised in marriage at older ages, although the correlation was not statistically significant (χ2 (9, N = 57) = 10.78; ρ = .291; Φ = .253).
The corresponding correlation of variables was performed to determine which ones had the greatest explanatory relevance in enabling the cases to come to light. The victim's nationality and legal status at the time of the organization's intervention were found to be the most relevant.

Limitations of the research
The methodology used for this research had two main limitations that should be considered when assessing the results. The first is the difficulty of obtaining direct access to victims of forced marriage, arising from the difficulty of contacting any type of victim at all. In the case of forced marriages, this difficulty is further compounded by the fact that there is still relatively little knowledge of this reality in Spain. The information about these victims was thus accessed through support organizations that may have assisted them. Consequently, the data on the victims presented here were not gathered directly from the victims themselves, but rather were provided by the surveyed organizations.
Second, as is often the case with studies that use this methodology, the response rate for the online questionnaire was not high. The reasons for the low initial response rate in this particular case were threefold. First, the questionnaire was sent from a Gmail address rather than an official account of the university coordinating the research. Some of the recipients may thus not have recognized the email as pertaining to an official research project, despite the accompanying letter of presentation. Second, some respondents had technical problems accessing the questionnaire because some official servers block access to certain websites, such as the one from which the questionnaire was sent. Finally, in some cases, the person who received the questionnaire at the organization was not the person best suited to complete it, but rather simply the person responsible for the account listed as the contact address on the institutional website searched to create the database. The detection of these limitations led to the change in methodology, whereby the organizations were first contacted by telephone and then sent the online questionnaire. Nevertheless, the actual sample consisted of 150 organizations, making it small for drawing inferences.

Discussion
The research presented here confirms the existence of victims of forced marriage in Spain, as well. Specifically, the research identified at least 57 victims who have received assistance from the surveyed support organizations.
With regard to the second objective of the research, i.e. the establishment of a profile identifying the victims of forced marriage, although the sample was small, it can be concluded that the people most at risk of being victimized are very young womenunder the age of 21or even minors and are thus usually still under the care of their parents or other adults in their lives. These women are usually North African, sub-Saharan African or Central Asian, as well as, to a lesser extent, Roma. Additionally, a large percentage have Spanish nationality, and most are Muslim. This profile does not match the political and media discourse that associates the process exclusively with certain migrant cultural minorities [26,27], ignoring the multiple factors giving rise to the phenomenon. The evidence of victims with Spanish nationality and who are legal residents in Spain aligns the present findings more closely with those of studies that relativize the influence of ethnic or religious factors [14,32] and that, on the contrary, highlight the link between forced marriage and patriarchal family structures [8,26,27,31] or situations of poverty. Nevertheless, the variables with the greatest weight in explaining victimization in this study were the victim's nationality, religion and community or ethnicity.
As regards the typical dynamics of these processes, the mechanisms used to compel the victims to marry are usually low-intensity, subtle or all-encompassing, related to the victims' membership in certain communities, as they mostly seek to leverage family pressure or attachment to tradition. Forms of coercive control and conducive contexts as those described in Chantler and McCarry [28] do also apply for most cases in our research and the use of more incisive active means, such as intimidation, violence and deception, is clearly secondary. However, these latter means are the only ones provided for in relation to the criminal offence of forced marriage introduced in the Spanish Criminal Code, which evidences the incoherence between legal provisions and the reality experienced by victims. More incisive means are also used with older victims and victims who are Spanish nationals or legal residents in the country. The marriage is usually performed in the victim's family's country of origin, except with victims who have Spanish nationality, in which case it is usually performed in Spain.
Finally, in terms of how these victimization processes come to light in a support service setting, the organizations that have proved most effective at detecting cases of forced marriage are those dedicated to assisting victims of domestic and gender-based violence, human trafficking and forced marriage, as well as those that make forced marriage one of their priorities. This finding confirms that the more information staff have about a victimization process, the better able they are to identify people who are subjected to it [14,34]. The cases are more likely to be brought to the attention of these organizations after the marriage has been performed as opposed to before. Genderbased violence organisations seem to detect more historic cases of forced marriage whilst offering support for other issues (e.g. domestic abuse), while forced marriage organisations work mostly with current cases. Such cases are usually exposed or reported by the victims themselves or by third parties and only rarely by close relatives of the woman being married. The women who turn to the organizations for assistance are mostly single, have no children in their care, and have Spanish nationality or are legal residents in Spain. The latter two factors are the ones that best explain whether or not the victim turned to the support organizations for assistance.
In short, the results of this research suggest that forced marriage is a practice consented to, when not directly promoted by, the victims' families that is usually brought to the support organizations' attention when the victims are in a personal and family situation involving a certain degree of empowerment, either because of their age, their residential status or the obtention of personal or social support that allow them to reveal the current or past situation. In contrast, women who are still married, have children in their care, and who are not legal residents are not only unlikely to report the case, but may not even turn to support organizations at all.
In view of this situation, the use of criminal law as the sole recourse to address this reality, far from eradicating it, may contribute to rendering it invisible, preventing the victimization processes experienced by these women from coming to light. This is not only because the introduction of Article 172 bis CP created a superfluous, highly symbolic offence that is ascendant over other pre-existing ones and serves to otherize certain cultures. It also because it criminalizes the use of commissive means that are rarely used in these processes and, especially, because it places victims in the difficult position of having to separate themselves from their community of origin in order to obtain institutional protection.
Rather than this option, a more intersectional approach to the legal response to forced marriage should be pursued, taking into consideration all the variables that can explain the victimization and discrimination processes to which these women may be subjected [7,28,35,36]. Such an approach would emphasize strategies for secondary prevention of victimization aimed at at-risk groups, as well as training to produce professionals equipped to detect these situations. At the same time, the design of legal mechanisms to protect victims should include solutions that do not necessarily require the use of the criminal justice system, such as the implementation of civil protection orders which, for the time being, have not been foreseen in Spanish legislation, even if they could better help victims unwilling to report their family [28]. All this without prejudice to better equipping the care sector, which has so far been largely responsible for developing prevention activities, as well as providing psychological and social assistance and even residential resources to victims when necessary. Furthermore, even in those cases in which recourse to the criminal justice system is deemed necessary, preference should be given to the use of restorative justice mechanisms that do not require the victims to pursue criminal proceedings against family members who are sometimes quite close to them in order to obtain criminal protection.