Fear of Crime and Feelings of Unsafety in European Countries: Macro and Micro Explanations in Cross-National Perspective

Abstract

In this study, we set out to explain fear of crime and feelings of unsafety, using two waves of the European Social Survey (2006 and 2008) covering 25 European countries (N = 77,674). The results of our multilevel analyses showed varying effects of contextual- and individual-level characteristics on our two outcomes. Higher crime levels in countries increase the fear of crime; however, they do not affect feelings of unsafety. Social protection expenditure proves to be an important determinant of both fear of crime and feelings of unsafety. Moreover, distrust in the police, generalized social distrust, and perceived ethnic threat induce fear of crime as well as feelings of unsafety. Finally, policy implications are discussed.

NOTES

Notes

¹ In our analyses, we include all our contextual determinants in one model to avoid spurious findings. Also, we make use of significance tests to assess the effects of country-level characteristics.

² More recently, researchers have developed new approaches to conceptualize fear of crime. Jackson (2006) argued that standard questions to measure fear of crime often reflect an emotionally assessed risk of victimization rather than fear of crime resulting from actual threatening situations. Therefore, in various studies, the frequency (how often people worry), intensity (how fearful people are), and impact (on the quality of life) of fear of crime are considered (Gray et al. 2008, 2011a, 2011b; Farrall et al. 2009; Jackson 2009). Elaborating even further on these new conceptualizations, Jackson (2011) proposes that perceived control and perceived consequence are also important in measuring the fear of crime. However, our data only allow partial coverage of the complex nature of fear of crime (Van de Vijver and Leung 1997). The ESS contains questions on fear of crime, the impact of fear of crime on the quality of life, and feelings of unsafety. Thus, how often people are fearful and the intensity of these occasions are not covered, whereas more recent studies propose to include these measures. Since we are only able to reach partial coverage, we have chosen not to incorporate the impact of fear of crime in our research. In addition, European studies that assess the differential effects of the determinants of both fear of crime and feelings of unsafety are scarce.

³ We cannot test these determinants at the neighborhood level since the ESS does not indicate in which neighborhood a respondent lives because of privacy policies. Information about the region is available, but because of a lack of validly comparable contextual information, we are not able to test determinants at this level either.

⁴ Our choice to consider feelings of unsafety as continuous is supported by linearity tests. The results of these tests show that the relationship between our mediating variables and feelings of unsafety are indeed linear. In all cases, η² is (almost) equal to R², which indicates that there is hardly any difference between feelings of unsafety as a categorical or continuous outcome variable. In addition, we considered feelings of unsafety as a dichotomous dependent variable (0 = safe and 1 = unsafe), and we ran a logistic multilevel regression analysis in MLwiN (Centre for Multilevel Modelling, University of Bristol, Bristol, UK). Our findings indicate that the results with regard to our hypotheses are substantially the same compared with the results presented in this study.

⁵ We also investigated the correlations between all the individual-level variables. These correlations do not indicate that multicollinearity affects our estimates. The correlations between the mediators are 0.362 (between the distrust variables), 0.176 (between distrust in the police and perceived ethnic threat), and 0.270 (between generalized social distrust and perceived ethnic threat). Moreover, the correlation between distrust in the police and generalized distrust is the highest coefficient of all possible combinations at the individual level.

⁶ Both distrust in the police and general distrust exert a suppressor effect on the relationship between the national victimization rate and the fear of crime. In models 2 and 3, the effect of victimization rate becomes stronger after adding the mediating distrust variables (from 2.308 to 2.500 and 3.078, respectively). This suppressor effect is due to the combination of the positive effect of victimization rate on the fear of crime, the positive effect of distrust on the fear of crime, and the unexpected negative effect of victimization rate on distrust.