Abstract
Previous research on the relationships between age, sex, ethnic group or previous victimization experience and fear of crime is controversial, as inconsistent results have been obtained, many of them attributable to a lack of methodological and conceptual consensus. We propose the study of self-protective behaviours motivated by fear of crime as a useful alternative that pays attention to the consequences of fear for urban quality of life. A survey with a representative sample of residents in a major city in Spain provides evidence about sex and age differences in self-protective behaviours against crime. Results from the ‘classical’ vulnerability and victimization models are discussed and the suitability of the evolutionary and the routine activities perspectives is also considered.
Introduction
Fear of crime has been widely studied in criminology since the 1970s. But it represents a multidisciplinary field of study; the factors that produce fear of crime and the consequences of that fear both involve a variety of disciplines. Despite intensive research, results have often been disparate, making it difficult to reach the consensus required to propose social, judicial and political measures. In addition, there has been a certain level of stagnation in the research. Some topics have become recurrent (Farrall and Lee, 2009; Fernández-Ramírez, 2008). Of all the explanations for this stagnation, a methodological issue may be the most significant: there is no unanimously agreed-upon instrument for measuring fear of crime. A large part of the research deals with variability in the conception of fear and the operationalization of related factors, and the search for more reliable and valid measures has remained a constant concern throughout the years (Farrall et al., 1997; Ferraro and LaGrange, 1987; Gray et al., 2008; Jackson, 2005; Vozmediano et al., 2008). Vanderveen (2006) characterizes the fear of crime as an ‘umbrella’ concept, under which different constructs (such as fear, risk and victimization) have been included. In this study we approach the fear of crime research from a different perspective, by examining the question ‘what do you do to avoid the risk of being the victim of a crime?’
Informed self-protective behaviour as a relevant focus of interest
The behaviours related to the perceptions of insecurity have been addressed in the literature from different points of view. Insecurity has been perceived as a consequence of fear (Ferraro, 1995; Garofalo, 1981) and as a part of a tri-dimensional ‘fear of crime’ construct (Gabriel and Greve, 2003). It has also been perceived as an antecedent, in an escalating loop where fear leads to avoidance and these measures influence the fear of crime (Liska et al., 1988). Precautionary or protective behaviours would be, therefore, closely linked to what we consider to be indicators of subjective security. The Liska, Sanchirico and Reed (1988) study provides a very interesting alternative, because the results can be contrasted with results from studies focused on the fear of crime or on variables of a more cognitive nature (such as risk estimation).
Proposing crime self-protection behaviours as the focus of analysis emphasizes an especially relevant aspect of subjective evaluations of crime: their effect on urban quality of life. Frequently, fear-motivated behaviours imply that citizens’ right to move freely through public spaces in urban areas cannot effectively be satisfied. Evaluating and, when necessary, proposing solutions for this loss of urban quality of life should be one of the objectives in this area of research. Furthermore, as Sacco and Nakhaie (2001) point out, paying attention to the behavioural aspects of fear of crime makes it possible to move beyond debates about the rationality of citizens’ attitudes, emotions and perceptions of crime, while establishing an area of study that leads to intervention. For example, the fact that a large number of citizens avoid walking through a specific zone is a clear indicator of the need for intervention to convert this space into a place where people want to walk. This type of intervention would be more appropriate and less ethically questionable for public institutions than one that attempts to modify people’s perceptions and emotions.
Vulnerability and victimization models applied to self-protective behaviour
In literature reviews of fear of crime (Hale, 1996; Miceli et al., 2004; and explicitly in Van Beek, 2004), three different perspectives are used to analyse, understand and predict fear of crime: (1) the socio-demographic perspective, related to vulnerability models; (2) experience of crime; and (3) the ecological perspective, which includes variables such as urbanism, lack of civility and community life. When elaborating crime prevention or urban quality of life improvement policies, the complexity of the studied phenomenon and the limited power of explanation of individual theoretical perspectives must be acknowledged. But in order to successfully identify relevant variables that will be useful in the search for social, judicial and political solutions, we believe it is appropriate to develop a more focused approach. We apply the classic vulnerability and victimization models to the study of the behaviours that citizens adopt in order to protect themselves from crime, although we are aware that ecological factors play a role in self-protective behaviours (Vozmediano et al., 2010) and we intend to address influences of level-two variables in future work.
The vulnerability perspective proposes that certain groups are more vulnerable than others to crime, with this being the underlying key to fear. Any model proposing to explain fear of crime will include some notion of vulnerability (Hale, 1996). Age, sex, income and social class are some of the variables that will establish differences in this sense. According to the study by Killias (1990), vulnerability is related to exposure to risk (for example, the case of taxi drivers), the damage incurred (the consequences of an attack involving physical harm are more serious for an elderly person), or the inability to deal with an attack (women or physically weak people can feel defenceless against possible attack). From this perspective, it has been commonly accepted that women and the elderly suffer from fear of crime to a greater degree, even though not all studies have found support for the existence of this relationship. Sex and age might produce an interaction effect: in women, the fear would be more intense regardless of age, whereas in men it would increase over the years, so that by a certain age there would be very little difference between the sexes (Van Beek, 2004).
The so-called ‘fear of crime paradox’ states that those groups with the greatest feelings of insecurity, specifically women and the elderly, are those who experience less victimization. Meanwhile, men and young adults, in spite of experiencing the most victimization, feel less fear. In sum, it can be said that a consolidated line of study has focused on certain population groups.
The victimization perspective points to how prior victimization, information about victimization in the social environment, or crime news in media can increase the perception of insecurity (Dowler, 2003; Winkel and Vrij, 1990). International victimization surveys have shown that victimization is not consistently related to fear of crime (Van Dijk et al., 2007b). Early studies, however, have reported a consistent relationship between victimization and fear of crime (Balkin, 1979; Dull and Wint, 1997; Garofalo, 1979; or Stafford and Galle, 1984). Further, there is recent research to suggest that methodological problems could have limited our understanding of the impact that victimization has on those who have suffered from a crime (Farrall et al., 2008). New evidence in this area could help clarify this relationship between victimization and fear of crime.
New interpretations of old problems: The evolutionary psychology perspective
Recently, two articles have suggested that evolutionary psychology offers new perspectives for interpreting the results of research in the area of fear of crime. Fetchenhauer and Buunk (2005), who try to find an explanation for women’s greater fear of crime, rule out the existence of a greater real threat. They found that fear of rape, the tradition outlook of the family and the perception of a social norm according to which boys must expose themselves to more risks than girls did not explain the paradox. They suggest that the differences between men and women are the result of sex selection, which favours risk-taking and fighting for status in men, and cautiousness and protecting their offspring in women. Sidebottom and Tilley (2008) propose that evolutionary psychology offers a new framework for interpreting the frequently discussed ‘irrationality of fear of crime’ in which the level of fear does not correspond to true crime rates. The history of the species would have left in humans a series of predispositions towards fear of predators, which would readily be activated in certain situations. Some of these fear-activating situations would be, of course, socially constructed. But evolutionary psychology offers new lines of thought that, in the authors’ view, are worth considering. One example would be to study scenarios that clearly activate this perception of threat, in order to make improvements in urban design.
Other disciplines that could offer additional information should be taken into consideration as well. If applied to fear of crime, the path initiated by Smith and Konik (2011), who discuss the feminist reappraisals of evolutionary psychology, and reflections on how evolutionary biology and behavioural ecology could help to better understand our reproductive behaviours (Liesen 2012) could add to our understanding of the differential patterns of fear and behaviour shown by women and men. The behavioural flexibility proposed by West-Eberhard (2003) is a challenging concept and offers new ideas to continue studying gendered fear of crime from a more complex point of view. Our phylogenetic past casts a long shadow, but the behaviours of women and men in modern societies have ‘evolved’ so quickly that it raises more questions than answers. The social sciences will have to address these questions in the future. We have applied these new perspectives, along with more ‘classic’ hypotheses, in the present study.
Method
Sample
The sample was composed of 1700 adults residing in the city of Bilbao (Spain). We carried out a probabilistic sampling procedure; the sample was stratified by district, sex and age. Regarding the stratification by district of residence (the city is divided into eight administrative districts), the number of people interviewed in each district was proportional to its population. With regard to sex, a balanced distribution was established. Finally, for age, quotas were fixed to ensure a sufficient representation of all the age segments in the sample. Because the target population included all people residing in Bilbao, with a sample of 1700 interviewees and an established confidence level of 95 percent, the sampling error was limited to ±2.4 percent. As a result of the criteria established, the sample was balanced at 50 percent with regard to sex. The age of the interviewees ranged between 18 and 95 years, with a mean of 47 and a standard deviation of 15. Of the people interviewed, 3.2 percent were foreigners.
Materials and procedure
The instrument used was designed ad hoc by the research team, who took into account variables from the vulnerability and victimization models. In most cases, closed-ended questions were used that were to be answered on Likert-type scales. A pilot study confirmed that the reliability and validity of the main measures were adequate. 1
Socio-demographic variables
Placed at the beginning and end of the questionnaire, questions were included to measure sex, age, place of birth and residence, number of people living in the household, income and education.
Victimization in the past year
In the case of a crime experienced by the person or someone in the family unit, details about the crime and the situation were collected.
Self-protective behaviours
Social and geographical contexts with different rates of actual crime undoubtedly present different typical precautionary and protective behaviours. To measure what we have called ‘self-protective behaviours against crime’, we identified two of the subjects’ most common self-protection measures from a longer list: ‘Avoiding passing through certain areas or streets of the city’ and ‘Taking additional security measures for the home, such as installing alarms, reinforced doors, security locks, etc.’. We considered these measures to be two prototypical behaviours of avoidance and active security, respectively. It was specifically asked whether the behaviours were performed owing to fear of being a victim of a crime. The subjects could adopt one of these behaviours, both or neither, so that four possible behaviour styles were established whose percentage of use in the sample is shown in Table 1.
Crime self-protection behaviours.
We used a telephone survey method because of the anonymity it provided for respondents and because it had advantages of cost and efficiency. Various authors have argued that qualitative studies are ideal for offering a framework that facilitates understanding of the quantitative results (Tulloch, 2000). But the use of a questionnaire applicable to a large number of subjects is still valid, because it offers advantages such as the possibility of diagnosing the extent of the phenomenon and constructing explanatory models that can be generalized to the population.
Trained interviewers telephoned the homes that were randomly selected in the specific geographical units (districts) and asked for the collaboration of an adult living in the household, verifying that each subject who agreed to collaborate in the study resided in the target district and was 18 years of age or more. All of the interviewees were offered the possibility of verifying the identity of the interviewer and the truthfulness of the data presented about the institution responsible for the study and its objectives. Data collected were recorded in SPSS software and analysed to study the relationship between the self-protective measures and the groups of age, sex and previous victimization.
Results
Self-protective behaviours and victimization
The questionnaire used included questions on the interviewee’s prior experience as a victim (in the past year). The question referred to everyone living in the household and provided us with an indicator of objective safety in the city where the research took place. This allowed us to measure the percentage of people who had been victims of a crime in the past year, similarly to the European Crime and Safety Survey (EUICS). 2 For this percentage, the target sample would not consist only of the 1700 people studied; the total number of people referred to in the victimization data consisted of all the people residing in these 1700 households. There was a total of 4648 individuals residing in these homes and 195 people had been a victim of crime in the past year: thus the percentage of people who experienced a crime in that time period was 4.16 percent. Expressed in terms of household nuclei, this meant a rate of approximately 11 percent of homes exposed to the impact of victimization.
The EUICS presents the prevalence of 10 common crimes, whereas in our study any crime of which a member of the household unit had been a victim is taken into account. However, the interviewees described the crimes experienced in detail, and in all cases they were common crimes that can be assimilated into the international survey. We think the comparison can be made. Contrasting our results for Bilbao with the data from the latest edition of the international survey (Van Dijk et al., 2007a; Van Dijk et al., 2007b) confirms that it is an objectively safe city.
Was this relatively small proportion of subjects who have recently experienced victimization adopting self-protective behaviours to a greater extent? The data indicate that this is the case. The Chi-squared test (χ2(3) = 24,969; p = .000) found a statistically significant relationship: among victims of crimes, not taking precautions was more infrequent. There was also an increase in the frequency of the use of avoidance measures only and in using both measures, the two categories in which avoidance is present. The results can be verified by observing the corrected standardized residuals in Table 2.
Self-protective behaviours among victims vs. non-victims of crime: Contingency table.
It is worth pointing out that the experience of victimization was not distributed in the same way across the sexes. The women reported more victimization in their family unit in the past year (χ2(1) = 6.30; p = .012), and, overall, 13 percent of the women reported being victims compared with 9 percent of the men.
When analysing the role that sex can play in the relationship between victimization and self-protective behaviours, the contingency table by layers offered interesting results. Among the men (χ2(3) = 12.97; p = .005), the experience of victimization was associated with a greater frequency of adopting both measures, whereas the women (χ2(3) = 14.57; p = .002) opted for avoidance more frequently than would be expected in the same situation, as shown in Table 3. In contrast, the women who had no crime victims in their household unit were more likely not to protect themselves using any of the measures identified.
Self-protective behaviours by victimization and sex.
Self-protective behaviours and vulnerability indicators
In order to analyse the relationship between subjects’ sex and the self-protective measures adopted against crime, we carried out a Chi-squared test (χ2(3) = 45.572; p = .000) (see Table 4). Our analysis of the corrected standardized residuals revealed that the men did not take measures (27.5 percent) or took active safety measures (31.6 percent) more frequently than expected, and they adopted both measures simultaneously much less than would be expected by chance. On the other hand, the women adopted both measures more frequently than expected (35.3 percent), but they engaged only in active safety or did not take any measure less frequently than expected.
Sex differences in self-protective behaviours against crime.
Considering another key variable from the vulnerability models, we analysed whether differences existed in the adoption of one type of strategy or another depending on the individual’s age group. The Chi-squared test showed statistically significant differences in this regard (χ2(9) = 114.415; p = .000). Thus, as the corrected standardized residuals show (see Table 5), in the 18–25 age group, exclusively adopting the avoidance measure was more frequent than expected; in the 46–65 age group, the combination of both measures was more frequent; and, among those over 65, active safety was used exclusively more frequently than expected.
Self-protective behaviours by age group.
The fear of crime literature reports age effects in the level of fear depending on sex, and has recommended considering the hypothesis that sex and age could jointly influence the self-protective measures used by the subjects (Van Beek, 2004). In order to test this hypothesis, an analysis was performed on contingency tables by layers, whose results are shown in Table 6.
Self-protective behaviours by sex and age: Contingency table by layers.
The Chi-squared test showed statistically significant differences (χ2(9) = 50.452; p = .000 and χ2(9) = 86.74; p = .000). The corrected standardized residuals revealed that, in the 18–25 age group (the youngest group), both sexes used the avoidance measure more than expected, although the women did so to a greater degree (46 percent compared with 32 percent among the men). The males took no measure more frequently, and they adopted both measures less frequently. The categories involving active safety were less frequent among the women, in favour of avoidance, given that, as we pointed out, it was the measure preferred by women in this age range.
For the 26–45 age group, the only noteworthy result was an increase in the frequency of the use of both measures by the women (up to 41.7 percent). Between 46 and 65 years of age, among men, active safety was used more frequently and the use of avoidance measures or no measure was less frequent; the women used both measures more frequently than expected, while their use of avoidance as the only strategy decreased. Among older men, the use of active safety alone stood out, a strategy used by almost 40 % the women also chose active safety more often than expected, with the adoption of both measures being less frequent. Curiously, and more frequently than expected, 28 percent of these older women did not use any measure.
These data can be summarized as a preferential use of the avoidance measure among the youngest members of the sample, although the women used this strategy to a greater degree, and more young men did not use any measure at all. In the successive age groups, both sexes used both measures more frequently; in the case of the women, this meant not engaging in avoidance as the only measure. In retirement age, we found an important use of active safety, and it is worth noting that this is the only age group in which the percentage of women who took no protective measures equalled that of the men.
Finally, and taking into account other variables related to the vulnerability models, we were able to verify that the foreign subjects in our study did not use self-protective behaviours to a greater degree than subjects born in Spain, nor did they manifest greater victimization. Similarly, neither income level nor education level had any relationship with self-protective behaviours against crime.
Discussion
The purpose of this study was to initiate a line of research in self-protection measures against crime and to examine the data collected in the light of classic vulnerability and victimization models that have been proposed to explain the fear of crime. Choosing self-protection measures as the focus of interest and analysis emphasizes one particularly relevant aspect of what we refer to as fear of crime: its consequences for the urban quality of life. According to Dolan and Peasgood (2006), fear of crime is, above all, a reason to be concerned about people’s self-protective and avoidance measures in response to it, which can have a significant impact. This is especially true, we would add, when behaviours motivated by fear imply that the right of citizens to move freely through public spaces in the city where they live is not protected in an effective way. Evaluating and, if necessary, proposing solutions for this loss of urban quality of life is the ultimate objective of the line of research. We found that a majority of people in Bilbao adopted self-protective measures, and Bilbao is an objectively safe city.
The present study included self-protective behaviours to study subjective security in different contexts, based on the vulnerability and victimization hypotheses. We were able to obtained interesting results. Starting with the victimization hypothesis, our results found the victimization experiences to be relevant, in a broad sense, because they include the entire household nucleus. Thus, our results support what Sutton and Tseloni (2009) pointed out: in recent studies in which the construct was operationalized to better respond to theoretical perspectives, fear draws closer to victimization experiences, just as occurs here with self-protective behaviours.
In relation to the so-called ‘fear of crime paradox’, we did not find that women and the elderly experienced greater victimization than men and young people. No statistically significant differences were found in the victimization reported by the age groups. With regard to sex, 60 percent of the people who reported having been victimized were women. This result does not necessarily imply that women are victims of crimes to a greater degree than men. Regardless of whether that is the case, the women empathized with members of the family unit who had been victims more than the men did. This empathy seems to be decisive in the experience of victimization. From a psychological point of view, not only is the direct experience important, but also what we could call the ‘level of empathy with the victim’. This hypothesis of indirect victimization proposes that the experience of crime through friends, relatives or acquaintances who have been victims could play a role in the genesis of insecurity. This hypothesis is related to the true concept of vulnerability, because the fact that people who are similar or significant to us are victims would reinforce this feeling. It seems difficult to speak about victimization without once more encountering the concept of vulnerability, or at least the gender dimension, when victimization and self-protective behaviours present a differential relationship depending on sex (Beck and Travis, 2004; Mesch, 2000; Snedker, 2006; Warr, 1992; Warr and Ellison, 2000). Men and women have somewhat different reactions in terms of self-protection strategies after an experience of victimization that is not necessarily their own but may have been experienced by a member of their household unit.
In relation to the vulnerability model, one of the variables frequently included in the models is nationality. Specifically, being a foreigner can be a vulnerability factor, depending on the degree of acceptance by the host society (Arikan, 2009). Our results, however, found no differences between the Spanish and foreign populations with regard to self-protective behaviours and victimization. Given the small proportion of foreigners in the sample, these results should be interpreted with caution, but our findings suggest an interesting research project for the future. Further work should examine differences between ethnic subgroups. In the Bilbao context, there are important differences in the social integration processes of Latin Americans, sub-Saharans or populations from Eastern Europe. It might be necessary to study these groups separately, keeping in mind that, in the USA, Parker et al. (2001) found higher fear of crime indices in Latinos than in African-Americans.
We have focused our analysis on two other key variables in the vulnerability models: sex and age. We found sex differences in self-protection, but it would be simplistic to merely draw the conclusion that women protect themselves more. We have to keep in mind that only 27 percent of men do not take any protective measures, which means that the majority of men protect themselves from crime in their daily lives in one way or another. The two sexes use different strategies, but the interpretation of frightened women who protect themselves and of men who take no measures at all does not hold. These results are consistent with those of Gilchrist et al. (1998), who found striking similarities between men’s and women’s fear of crime and the steps taken to avoid crime. Survey data could have failed to capture men’s fears owing to social desirability; Sutton and Farral (2005) even argue that, in reality, men might have a greater fear of crime than women but it is a private experience. In our case, studying behaviours instead of fears seems to address the social undesirability problem quite well: the majority of the men admitted that they protected themselves owing to fear. However, social desirability could still have an influence on the protection strategy they reported using: men may be reluctant to admit that they use avoidance, which would imply recognizing their vulnerability to a greater degree. Future research will have to examine this aspect more closely.
On the other hand, age is undoubtedly a relevant variable in all the vulnerability models of urban insecurity. Some of the literature points to the elderly having limited life experience owing to a high fear of crime. Our results do not support this. Our findings follow those of Tulloch (2000), who said that elderly people, in spite of what they reveal in surveys on urban danger, do not ‘live in fear’. This argument would also agree with the Sacco and Nakhaie (2001) proposal that statements about generalized fear in the elderly are exaggerated, and that, moreover, the fears and behavioural restrictions of older people are rational.
In our study, we found avoidance to be more typical of younger people, a combination of measures more likely for middle-aged people, and the exclusive use of home security more frequent among people over the age of 65. But the most interesting results are found by looking at the difference between self-protective measures for women and men. As we have seen, not adopting any measure is more frequent among men than among women, although it is also frequent among elderly women. The exclusive use of the avoidance measure is more frequent among women than among men, especially the youngest women, whereas exclusively using active safety is more typical of adult and elderly men. Finally, the combination of the two measures is most frequent among women aged 26–65.
Can these results be interpreted in terms of the evolution of fear with age? It has been proposed that, among women, fear would be more intense regardless of age, whereas it increases with time among men, so that by a certain age there would hardly be any gender differences (Van Beek, 2004). Interestingly, our data showed that, at retirement age, similar percentages of women and men do not protect themselves, but, among those who do protect themselves, the strategies differ. However, caution keeps us from concluding that fear evolves with age: with data taken at only one point in time it is not possible to make such statements. Nonetheless, in light of the interesting and complex results obtained on the interaction between age and sex, it would be advisable to explore this route using longitudinal methodology.
It is worth noting that the findings from the study are quite consistent with lifestyles in our society. Therefore, they could easily be interpreted based on the routine activities theory (Felson, 2008), which could provide us with clues about experiences of fear. Young people spend more time on the street and, owing to the threat of sexual violence, women in particular use avoidance measures. They do not protect their home as much, because they probably do not have one of their own. Active safety is typical among older adults, in order to protect a home for which they are responsible. In the case of women, this situation coexists with worry about their own safety outside, and so they combine the two measures; in contrast, the men only protect the home. The elderly report less frequency in the categories that involve avoidance, which is consistent with routines in which they go out less.
In relation to sex, the vulnerability models applied to protective behaviours could be enriched with the contributions of the evolutionary perspective. This field of study examines sex differences from the point of view of men’s and women’s relationship with risk, the night, enemies, the herd and survival, and its significance in the adaptation of our species. Obviously, we are not questioning the equality of the sexes, but the shadow of our philogenetic past is long. The studies by Fetchenhauer and Buunk (2005) and Sidebottom and Tilley (2008) look for keys in the natural history of the species, which would have left in humans a series of predispositions to fear of predators that would readily be activated in certain situations. Without denying that these situations can be socially constructed, it is possible to interpret some of the results of this study in terms of behaviours that favour one’s own survival and that of one’s offspring. The results on self-protection in older women could be interpreted in this way, because their percentages of taking no precautions only approach those of men when they no longer have families to take care of. Another example would be the fact that men protect themselves less and choose active safety more frequently rather than avoidance. The answer probably cannot be provided by only one model; rather, the right combination of this complexity of elements will finally shed more light on the fears of the inhabitants of our cities.
In summary, we aimed to go beyond the permanent crossroads of methodological relativism that has characterized the measurement of the fear of crime (Vozmediano et al., 2008). We have regarded fear as something more than expressed perceptions and emotions, and have examined actual self-protective behaviours. People can report that they do not feel fear for various reasons, but there is only one reason for taking a self-protection measure: to reduce the risk of being the victim of a crime. When reporting on these measures, men are not so distant from the women in self-protection, although they choose different alternatives at different moments in life. And the elderly are not so distant from young people, who show their need for safety by using avoidance in public spaces. Vulnerability seems to intertwine with routine activities to offer a coherent and comprehensible image of the self-protective behaviours, far from the irrational fears of the more classic literature. We find that these are good reasons to continue studying these behaviours in more depth. The first task for future work should be the development of more complex and detailed measures of self-protective behaviours, to better evaluate possible consequences on the urban quality of life. Another necessary step should be to broaden the theoretical perspective to include second-level variables in the analysis in order to consider the effects of neighbourhood and city on behaviours. Following Sachs (1994), ultimately cities are like people: they belong to an urban species but they have their own unique personalities. It is important to take into account their particular configurations, natural and social factors, and pasts and traditions, to better understand the criminal phenomenon and its psychosocial consequences.