Living in a foreign country: the meaning of place of origin and gender for risk perceptions, experiences, and behaviors

Abstract Previous research has convincingly proven that perceptions, experience, and exposure to risks vary among certain groups in society. By drawing from a unique combination of Swedish survey data and interviews, this study aims to investigate perceptions and experiences of risks as well as in relation to behavior by analyzing the cleavages related to interactions between place of origin and gender. Theoretically, we see individual risk perception as part of situated hierarchical power relations where an individual’s position (which is an intersection of, for example, gender, race, age, and place of origin) structures action and thought. Findings verify that foreign-born men and women perceive risks to a greater extent than those born in Sweden. However, no direct pattern of ethnicity is apparent in exposure to risks, but since predictors measuring experience of discrimination are shown to be significant, the effect can be mediated by such circumstances. In terms of how risks have affected behavior, women, irrespective of their ethnicity, are affected. Exposure to tragic experiences among those who are foreign born can pose risks that are perceived to a greater extent. Furthermore, more vulnerable material conditions can also affect how risks are perceived, and uncertainty due to a lack of resources and as an inherent ingredient of living in a foreign country seem to enhance perceptions of risk and feelings of unsafety. Last, the sense of discrimination appears to influence exposure to certain risks, which might capture an interaction between racism and violence.


Introduction
You feel unsafe. You are in Sweden and it is their country. (I-6) Previous research has convincingly proven that perceptions, experience, and exposure to risks vary among different groups within society. From a European perspective, Olofsson and € Ohman (2015) have found, for example, that heterogeneity factors (ethnic background, gender, etc.) influence how individuals perceive and act in relation to risk. Hence, social structures, such as gender and ethnicity, mediated as individual characteristics, seem to matter. Much of the previous understanding has famously been composed in the general understanding of the 'white male effect' (Flynn, Slovic, and Mertz 1994;Finucane et al. 2000). This theoretical assumption quantitative analyses of survey data as well as epidemiological and other health-related statistics dominate. There is no unambiguous pattern in the results when it comes to risk perception and migration status, which can be explained by the great variety of risks analyzed, and how migrant status is defined and measured. However, many studies have indicated what we previously pointed out regarding the white male effect, that ethnic minorities and people with a foreign background rate risks as higher and/or are more vulnerable to particular risks and risk behaviors (Macias 2016). Many also point out that studies concerning migrants' risk perceptions specifically and those concerning vulnerability to various risk in general are scarce (Yong et al. 2017;Yan, McManus, and Duncan 2018). Few have considered inequality or discrimination in the analysis and as a way to better understand how migrants experience and perceive risks, although migrants often are considered vulnerable to risks because of limited economic, social, and language resources.
To change this, we depart from critical risk theory that reveals how an individual's understanding of risk and societal risk management and policy are closely linked to gender and race norms as well as other social categorizations related to inequality and discrimination, such as sexuality, age, and functionality (Giritli Nygren, € Ohman, and Olofsson 2017;Hankivsky and Cormier 2011;Hannah-Moffat and Maurutto 2012). Furthermore, it shows how the concept of risk is associated with normative values and beliefs; for example, to term something a 'risk' does not only imply the calculation, management, and mitigation of the risk, but it also calls for political action (Olofsson et al. 2014;Wilkinson 2012). Applying such an approach to the perception of risk accounts for the ways race, gender, and class intersect to position individuals in relation to risk (Crenshaw 1989) and how cultural-identity protection (Kahan et al. 2007) can be seen as the self-regulation of the individual to adhere to social norms. Since risk calculation and management have colonialized not only policy and governance in terms of risk regimes (Hood, Rothstein, and Baldwin 2001), although in conjunction with an uncoupling of individuals from social structures in the process of individualization (Beck 1992), there is responsibilization of the individual, who is expected to have knowledge of how to manage all kinds of risks. This means that the processes and dynamics of risk and causation are often translated into direct personal lifestyles and/or behaviors of 'at-risk' groups. The change, revealed in the individualized concept of being 'at risk', is that the responsibility is individualized and, thus, renders structural inequalities invisible. Instead, identified 'at-risk' and 'risky' populations, including, for example, migrants, long-term unemployed, and young people, become subject to governance and interventions (since risk creates a demand for changes in both society and people (Cebulla 2005;Giritli Nygren, Fahlgren, and Johansson 2015). Therefore, risk offers a means to promote special interests and to influence behavior in order to attain a specific goal or, by extension, to bring about specific forms of social change.
Based on this, we see individual risk perception as part of situated hierarchical relations of power where the individual's position (which is an intersection of, for example, gender, race, age, and place of origin) structures action and thought. On this individual level, inequality is expressed in terms of heightened risk perceptions due to being identified by others as 'at risk' or even 'risky'; lack of control of and confidence in one's precarious situation; and experiences of adverse life events, disadvantage, and discrimination.

Research design, data, and applied methods
The purpose of this study requires a mixed-methods design in which both quantitative and qualitative materials are used and analyzed (Creswell 2014). These two parts will be described in turn.
The quantitative material was gathered through a Swedish national survey, Society and Risk 2011. The survey raises the question of how risks and values are viewed among Swedish citizens. The data were collected in 2011 through a random sample (n ¼ 3000) of the Swedish population aged 16-75. The total number of respondents was 1094, and of these, 210 respondents answered that they were born in a country other than Sweden.
To assess perceptions of risks and experienced varieties of these, logistic regressions were applied. Empirically we employed several different dependent variables to capture the dimensions of perception, experience, and behavior. The perception of risks was measured in terms of how respondents graded their risk for personally being harmed by violence and aggression. The actual experience of risks was measured through a question that examined whether respondents had been exposed to violence and aggression. Finally, we also examined the behavior among respondents in relation to issues of risks by analyzing how often they stated that they walk home alone at night.
In relation to previous research (Olofsson and € Ohman 2015; Olofsson and Rashid 2011), a number of theoretically assumed predictors were applied as explanatory variables in the models, including age, gender, marital status, education, and income levels. In line with the key argument of the paper, foreign background was included as the main independent variable. To add more precision and to capture the intersection of foreign background and gender control, variables were constructed that divided respondents in different categories. These accounted for if respondents were female or male and if they were born in Sweden or not. In addition to logistic regressions, descriptive statistics are displayed in relation to the three different dependent variables and the main independent variable.
The empirical material underlying the qualitative analysis was gathered as part of the Migration to Work research project, funded by the Swedish Agency for Economic and Regional Growth and the County Administrative Board of V€ asternorrland (Br€ annstr€ oom et al. 2016). Within this project, 28 individuals who were born abroad were interviewed. Although they were all residing in the County of V€ asternorrland, the ambition was to ensure a variation of different potentially influential characteristics, such as gender, educational level, ethnicity, residence in a city or in the countryside, and time spent in Sweden.
The participants of the study were partly approached through recommendations from various municipalities in the county and partly through private networks and non-profit associations. Prospective interviewees were initially contacted by telephone, and a presentation of the project was given. After that, written information about the project was distributed, mainly via e-mail, but also via traditional mail when requested. The interviews were conducted between November 2015 and May 2016 in private conversation rooms at a location easily accessed by the participant. Before the meeting, the participant was once again informed about the project and about his or her right to decline participation and withdraw from the project at any time during or after the interview. Interviews were tape-recorded and transcribed, with about approximately 20 to 80 minutes of conversation for each participant. In Table 1 the diversity of our informants based on a number of different categories is described. Although interviews aimed to explore immigrants' narratives of reception, institutional support, and working life experiences, this study particularly focuses on questions raised in relation to issues of perceived safety and security among informants. Questions were raised regarding whether they experienced their situation as safe, and informants were thereafter encouraged to elaborate on their answers in relation to their own interpretation of the question. The analysis of the material had the ambition to uncover meaningful passages in relation to the purpose of the study and the concepts it involves. These concepts have also been discussed and elaborated in relation to previous literature. Therefore, narratives connected to issues of risks, safety, fear and violence were identified and closely analyzed with the ambition to find common themes (Kvale and Brinkmann 2009). This strategy made it possible to not only grasp narratives relating to relevant issues, but also to conduct a contextual examination of each narrative with the ambition to categorize such material based on if it related to perceptions, experiences or actions of risks. This process was not instrumental but also contained elements that made it possible to address narratives from interviews that either enriched or contrasted quantitative data with the ambition to juxtapose different data with each other (Bryman 2004). The coding of the material was done by one of the two authors but narratives and passages from the interviews were discussed jointly. The analyses of this material were assisted by NVivo, and the software was used to identify and structure the occurrence of the themes analyzed.

Caveats
The study has limitations. One of the main concerns is the low level of explained variance. Even though this is not unusual in survey research about social phenomenon, and one could argue that it is also important to identify patterns and not only explain particular phenomenon (Finucane and Holup 2005), the results should be interpreted with this in mind. Another concern is the low number of respondents with foreign backgrounds. This may reduce the representativeness of the study while also possibly reducing our chances of observing effects.
Although the combination of different types of data creates advantages, it also emphasizes some caveats. First, the quantitative material precedes the qualitative by some years. Although this could mean that there are substantial differences in the composition of the different migrant groups, the qualitative material was not selected for being representative but rather to reflect the depth of experiences. Moreover, almost half of the informants had been in Sweden for at least seven years during the time of the interviews. Secondly, the informants were all residing in one Swedish county when interviews were conducted, whereas respondents to the survey were collected nationally. Even if the ambition of the qualitative material was to contextualize and enrich the qualitative findings, it should be kept in mind that such narratives are geographically limited. Third, the quantification of risk is not thorough but does still make it possible to measure one of the core risks that can affect people.

Quantitative findings
Regarding the purpose of the study, the first ambition was to investigate the cleavage in risk perception related to place of origin and gender in terms of quantified measurements. Table 2 presents the three dependent variables. The first variable examines respondents' perceptions of risks in terms of violence and aggression. The foreign-born respondents had significantly higher perceptions of risks compared to those of the native population. Through another question, the actual exposure to violence and aggression was measured. The results show similar tendencies. Although the differences between foreign-and native-born individuals are of a smaller magnitude, they must still be regarded as statistically significant. Finally, how risks influence behavior is displayed through the question regarding if respondents walk home alone at night. It was revealed that foreign-born individuals are somewhat more inclined to do so. However, since differences compared to the native population are not significant, that aspect of risk cannot be verified as being more salient among the foreign-born population.
We dwelled deeper into this question, and in relation to ideas of intersectional beliefs, analyses were conducted that examined the difference between gender (see Table 3). In regard to risk perception, this is most strongly estimated among foreign-born females. In terms of gender, the pattern was not consistent between the foreign-born and native populations, since males in the latter group, rather than females, see this as more problematic. Turning to the actual exposure to risks, both foreign-born females and males have been more exposed to violence and aggression than the corresponding groups of natives. Differences, though, are not that substantial. Concerning how risks influence behavior, it was revealed that foreign-born females are more inclined to walk home alone at night than native-born. When looking at the situation among males, it was found to be is the opposite.
These differences in perceived risk among foreign-born and Swedish-born respondents were analyzed further with control for what previous research has pointed out as important predictors, together with discrimination and the size of the community in which the respondents live (see Table 4). The latter is particularly interesting since it comprises two variables that the interview study points to. Logistic regressions were performed that examine to what extent the intersection of place of origin and gender will influence risk perception, previous experience of violence, and behavior in relation to other important predictors. The analyses were performed in two steps, or blocks, where the first block included the intersections of place of origin and gender together with a number of control variables (age, family situation, income, and education), and the second block included, in addition to the already mentioned variables, experience of discrimination (related to ethnicity and gender) and the size of the city/place of residence.
The results from the analyses show some interesting patterns. Starting with the variables that aim to be a proxy for the intersection of gender and ethnicity, we find that both men and women born in countries other than Sweden have higher risk perceptions than men born in Sweden, but that women born in Sweden have lower risk perceptions. This effect remains when control variables are added (Block 2). Turning to experiences of violence, in the first analysis (Block 1), men born outside Sweden reported more experiences than men born in Sweden, and women in general reported less experiences of violence. However, when experiences of discrimination due to ethnicity and gender and when size of place of residence were added to the analysis (Block 2), the only clear difference between these groups was that women born in Sweden had less experiences of violence than men born in Sweden. Of the additional variables, the two measuring experiences of discrimination were both positively associated with experiences of violence. Last, we found that more women, regardless of place of birth, avoid walking home alone at night than men born in Sweden. This effect holds when control variables are added (Block 2). Gender discrimination was also associated with such behavior, but interestingly, the experience of gender discrimination was associated with sometimes walking home alone. Also, the size of the respondent's residence (city/town/area of living) was associated with behavior; for example, respondents living in smaller places were less keen on walking alone at night. Note: Choices are different for the three questions: (i) very small, small, moderate, and very large; (ii) no never, yes one time, and yes several times; (iii) never, rarely, sometimes, often, and very often/every time. The two questions regarding violence and aggression reveal a statistically significant difference in perception and exposure between foreign born and natives. The scale is ordinal, ranging from 1-5; 1-3; 1-5.
Turning to the control variables, we found that older respondents perceived and experienced risks to a lower degree than younger respondents. However, older respondents answered that they never, or almost never, walk home alone at night. Furthermore, we found that people who live alone tend to perceive risk of violence as higher, have more experiences of violence, and expose themselves more by walking alone more often than respondents who are cohabiting. Income was not associated with any of the dependent variables, and education was only partly associated, indicating a negative relation to perception, thus, the higher the education, the lower the perception of risks.

Qualitative findings
The quantitative findings derived from the three applied dimensions of risks both corroborate prior knowledge within the field and also show some intriguing new patterns. Although identifying statistical patterns of that character are necessary to get us closer to the effects that influence the feelings of risks and safety, our general understanding can improve if such material is complemented with analyses of qualitative material. Hence, we proceeded by analyzing the 28 qualitative interviews conducted with individuals born abroad. A majority of the informants migrated to Sweden from countries that have been torn by war or other catastrophic events.
The interviews represent a wide spectrum of views on how risks and safety are perceived among individuals with migrant backgrounds. The interview materials were analyzed from the angle of the three different dimensions of risks exploited in the study.

Risk perceptions: Influence of social and material conditions
In the interviews, social relations and how these work are intimately related to perceived risks and feelings of safety. Obviously, to experience a sense of safety is important for everyone in society, but one of the interviewees in the material expressed why it could be of particular importance for people with foreign backgrounds: 'If you have escaped from persecution and violations when you come here and face that … It will create a huge catastrophe for people' (I-4). Thus, having a traumatic experience in the past might make you more sensitive to perceiving risks in your everyday life, something previous research also confirms (Olofsson and € Ohman 2015;Olofsson and Rashid 2011). Reasonably, this is one of the mechanisms that explains why people with foreign backgrounds tend to perceive risks to a greater extent, which was found in the quantitative material.
The feelings of insecurity among the women was often gendered, associated with female clothing but also a sense of uncertainty in relation to others: 'Not even in Sweden are women free … Unfortunately, they have no freedom. Therefore, I am more cautious now … This world is not good for women, but I also think that Sweden can be the first [country] for a "women's way"' (I-24). All things considered, an individual's potential vulnerability as a migrant, considering their family and networks have been left behind, is much more apparent compared to others. The analyzed data also indicates that risk perceptions can be associated with claims of more material aspects. More concretely, such notions have been closely related to two aspects of personal life. First, senses related to living conditions appear to be of importance for how one perceives risks. The centrality of having a place to call home is illustrated by a female informant residing in the largest city in the region who emphasized how important it is to get a home in one's new country: 'Your home is the most important … It is much more important than the language, than the job. If you have left all behind and are on a search for a new home, when you have a new home, then you feel safe and can think of other things' (I-19).
Moreover, for people with foreign backgrounds to establish themselves in a new society takes time and depends on a number of more or less individually controllable aspects; for example, the outline of the integration policy has a great influence, particularly for the groups of immigrants with more challenging preconditions needed to provide for themselves. Swedish integration policies, over the past decade, have gone through a number of important changes, of which the most significant one has centralized responsibility from municipalities to the Public Employment Service (Lid en, Nyhl en, and Nyhl en 2015), which has caused distress among the informants. One of the women in the material reacted to how this has altered preconditions and also illustrated how support systems, such as settlement support, is an important factor for feeling safe: 'You need safety, a job, and a full-time job … Refugees had a settlement at once, furniture, and other necessary things in the apartment. You could instantly get … allowance, and children could go to school, and they had introduction plans ready' (I-4).
The implication of this statement is that, when the municipality was in charge of the integration policy, the reception was better organized from the beginning, but it was also experienced as creating better preconditions to be able to provide for yourself (Br€ annstr€ om et al. 2018). The material also reveals another trajectory that is not uncommon for migrants who decide to start their own businesses. As is common for many who are self-employed, it creates financial exposure due to mortgage, which often will require a heavy work load in order to be successful. Still, as was stressed by one informant (I-26), the fact that, in Sweden, it is possible to start your own business without having a lot of funds accumulated creates additional options for the individual, even if it creates financial pressure.

Experiences of risks: different types of risk and the capacity to handle them
The quantitative material captures that experiences of discrimination can be linked to experiences of risks. Although conducted interviews consist of few explicit references to exposures to violence and aggression, they are indeed capable of invoking a wider experience of how risks can strike people. One example is the following narrative from a young man: 'For example, you get on a bus and they will not sit with you, and if you sit next to them they just get off … You feel unsafe. You are in Sweden and it is their country' (I-6). This man had moved from the countryside to the city to deal with this feeling of being both unsafe and not accepted in the small community in which he used to live. Another trait of discriminatory behavior is when such actions affect one's working conditions, such as the sense that 'Immigrants cannot advance [at their job]' (I-21), which is mainly expressed as people without foreign backgrounds getting better treatment than the deliberate discrimination directed toward immigrants.
Other narratives contain histories of violence. One woman who has gone through a divorce with her former husband and was assaulted described such problems: 'I am still afraid for the future. I am alone; I have no family. I cannot go back' (I-25). The experience of violence that she portrays is enhanced by a lack of family or networks in Sweden. In such situations, the welfare state, obviously, has a crucial role both to provide reassuring material needs but also, as is expressed by the woman, to function as a social support in situations in which one has no other. Theoretically, this is a plausible linkage of how individuals experience risks and the intersection of gender and foreign background. Statistical models do verify that those living alone are, to a greater extent, more exposed to risks.
Such preconditions to being affected by different risks can also be related to varying living conditions. Even if an individual feels satisfied with their own home, how it is located can also pose both positive and negative perceptions in relation to safety. The following quote is from a woman living in Sundsvall: 'There is an invisible border between foreign people and Swedes … There is a small area where the foreign people live' (I-25). This expression of segregation is a wellknown situation in the Swedish society. Quantitative research has shown a significant correlation between ethnic segregation and income segregation, which also has increased over the years (Hedman and Andersson 2016). From that perspective, perceptions of safety in economic terms are also inevitably related to an individual's living conditions and area of residency.
Drawing on a similar logic, an individual's new environment will also inevitably be compared to what they left behind. For several of the interviewed informants, the contrast was stark, fleeing from warzones of absolute danger. During such conditions, an individual's previous home loses the sense of safety that it may have had before: 'In Syria we were in the house all the time, afraid that the bomb would come' (I-17). With such experiences, it can be a struggle to rebuild the confidence for safeness, even if one has escaped to a completely new environment that, as several of our informants described, is safe. Therefore, it can still create an increased sense of being exposed to risks.

Risk-changing behavior: gender and geographical patterns
The quantitative findings found that modifying behavior in relation to potential risks was a common strategy taken by women, irrespective of their ethnic background. Motives for such behavior did not form a uniform pattern in the qualitative material. Still, two women living in the largest city in the county, Sundsvall, added some key understanding to their own behavior: You do not risk going from here at night. It used to not be like that. I was with friends until 1 or 2 at night, and then, I walked home by myself. But I do not dare doing that anymore … Nothing has happened, but you hear about it [things that have happened]. (I-21) Yes, I go out and do things. I am not afraid. I used to be, after what happened in Europe and how these idiotic extremists have acted … But then, I thought, "I will not let them win because that is how they want us to feel. (I-24) These narratives indicate opposite practices, but they both relate the sense of safety to the awareness that other people in the local community see them as different. The latter of these two informants experienced the fear of being associated with religiously motivated terrorists because she herself wore a veil. It is commonly known that people wearing clothes and symbols of that character, particularly if they can be interpreted as religious symbols for minority groups, risk being exposed to hatred from others (The Swedish National Council for Crime Prevention 2016). The quote from the first woman shows an increased sense of uncertainty and risk in relation to others, which has affected how she lives her life. Other interviewed women underlined the same tendencies and have also changed their behavior.
In other words, the informants found ways to manage the social relations that made them feel unsafe. Also, one of the women quoted above (I-24), who used to be afraid to go outside because of being harassed or being seen by others as different or as a terrorist, has found strategies of her own to change her situation. Although the number of potentially important conditions that could affect these different experiences are beyond reach for this study, the informants gave some clues in their stories that also stressed some parallels between them. While the first woman pointed toward just opening the newspaper in the morning as sparking fear due to all the horrors happening in society, the other woman has had similar experiences and, as a consequence, has stopped consuming that kind of media. Similarly, the young man who had moved to the city (I-6) referred to how moving and becoming better at Swedish gave him more confidence to handle the actions of xenophobia he experienced in busses and other interactions with locals. For this individual, living in a city also appeared to bring him into contact with a society that, in this case, was experienced as more tolerant. However, such experiences were not concurrent across the material. A young woman, for example, lives in the countryside and commutes to the nearby city for work, but her sense of security is in the local community: 'I know everybody here … and we socialize with all our neighbors' (I-27). Experiences can apparently differ, but the result of the quantitative models shows some effect of individuals residing in smaller societies as being more inclined to alter their behavior than those living in larger cities.
Putting the pieces together: combining quantitative and qualitative data Quantitative findings indicate that foreign-born individuals perceive a heightened risk of being harmed by violence and aggression compared to native individuals. Material from interviews provides some clues about how this could be established, and based on this evidence, two main effects could be identified. One potential effect is the importance of an individual's background. A considerable share of migrants who have reached Sweden during the latest decades have done so due to problematic situations in their home countries. As previous studies have indicated, having a traumatic experience in the past might make an individual more sensitive to risk in everyday life (Olofsson and € Ohman 2015;Olofsson and Rashid 2011). Hence, it can be assumed that at least some proportion of foreign-born individuals in Sweden can express that kind of sensitivity. In the interview material, the mechanisms that can be at play are described in more detail, showing how experiences of living in and fleeing from war zones also can affect an individual's sense of safety in a new country and home. Another potential effect is related to material resources. The qualitative data reveals how the individual, as a migrant, requires resources that, for many natives, probably are taken for granted for feeling safe. The most obvious example is the housing issue in which migrants will, to a larger extent, find themselves living in more unsafe communities. Since it is often more challenging for migrants to accumulate such resources, this poses additional stress for the individual. One concrete example of this is how the significant transformation of policies for immigrants, which have been ongoing since 2010 in Sweden, risk creating additional uncertainties (Br€ annstr€ om et al. 2018).
In terms of experiences of risks, no patterns were found that draw from differences in place of origin. Actual exposure to violence and aggression were more common among younger men, irrespective of their ethnicity. However, experiences of discrimination appear to have a positive influence on individuals' experiences of risks. When including the predictor of discrimination in the analysis, the effect of being born outside of Sweden diminishes or becomes insignificant. As an immigrant, an individual risks exposure to discriminatory behavior (Rydgren 2004) that, even if it does not lead to a physical conflict, can influence the individual's perception of safety. The interview material in this study reflects such circumstances in different sectors of society and gives examples of how they are present in both housing and labor markets. Facing that in everyday life can create a kind of uncertainty that will produce senses of risk. As demonstrated, interviews contained narratives of informants who have been exposed to such actions, but such material cannot be used for assessing the likelihood of such exposure. At first sight, though quite puzzling, foreign-born individuals perceive higher risks of violence and aggression but do not find themselves as exposed to such activities as other groups in society. This finding might seem to nuance prior studies that have found a linkage between perceived risks and experiences of risks for individuals with foreign backgrounds (Macias 2016). However, it is important to remember that, in the quantitative analysis, we did not differentiate between different foreign origins, and it is plausible that the discrimination variable captured not random but structural discrimination of people from certain regions of the world and/or particular appearances, something that the interviews confirm.
It should finally be noted that women appear to take specific concerns for lowering their risk exposure to environments or situations that they associate with risks. Women who were born in Sweden avoid walking home alone at night even though they have relatively few experiences of violence and, even more importantly, do not perceive the risk of violence for them to be personally as high. Women born in another country, however, respond according to previous studies; that is, they have a high-risk perception and avoid walking home late at night. Hence, they are more prone than men to changing their behavior. This could be interpreted as one example of concerns taken by women, functioning as a mediating effect that also lowers their feelings of being exposed to violence and aggression. The quantitative material does not prove that such behavior is more common among foreign-born women but that such refraining behavior is more common in less-populated societies than in larger ones. Interviews corroborate that such strategies are found among foreign-born women, referring to narratives of how women refrain from walking home alone late at night. In other words, differences were found between genders but not between ethnic groups. It is apparent that such exploiting strategies are one example of how women's behaviors are constrained, but more fine-grained data needs to be used to examine if it also have an effect on how they perceive risk.

Concluding remarks
This study set out to examine the cleavage in risk perceptions, exposure, and behavior based on place of origin and gender. The purpose was delimited through two research questions.
Concerning the first question, cleavages due to intersections of place of origin and gender show an intriguing pattern in relation to how risks are perceived, experienced, and affect behavior (Giritli Nygren, € Ohman, and Olofsson 2017; Olofsson et al. 2014). Quantitative findings verify that foreign-born men and women perceive risks to a greater extent than those born in Sweden. No direct pattern of ethnicity, though, is apparent in exposure to risks, but since predictors measuring experiences of discrimination are shown to be significant, the effect can be mediated by such circumstances. In terms of how risks have affected behavior, women, irrespective of their ethnicity, are affected. With the help of the conducted interviews, we can propose several reasons for these outcomes. Exposure to tragic experiences among foreign-born individuals can pose that risks are perceived to a greater extent. Furthermore, more vulnerable material conditions can also affect how risks are perceived, and uncertainty due to a lack of resources and as an inherent ingredient of living in a foreign country seems to enhance perceptions of risk and feelings of unsafety. Last, senses of discrimination appear to influence an individual's exposure to certain risks, which might capture an interaction between racism and violence.
Concerning the second question, patterns between the different dimensions of risk does not play out as expected. There is no regular connection between perception, experience, and behavior. Hence, those who perceive risks are not always the same as those who are exposed to them. We have already touched on several reasons for this outcome. For example, perceptions of risks can have strong psychological motives that are enhanced by an individual's background and feelings of vulnerability. But as noticed, there are also actions taken to diminish exposure to risks which, which isolated as an example of an awareness that creates a perception of risks, will reduce exposure to such risks. However, women appear to find strategies to lower their potential exposure to risks. This is not only an example of the perceived vulnerability in public places of and by women but is also an example of how such public discourses, translated into individual perceptions, restrains the behavior of women.
In general, future studies should continue to inquire about these sequences of risk and do it with more fine-grained data and in other contexts.

Disclosure statement
No potential conflict of interest was reported by the authors.