Heterogeneous Impact of Victimization on Sense of Safety: The Influence of Past Victimization

ABSTRACT

It is well-established that recent victimization increases fear of crime, but less is known about the effect of past victimization. Using data from a large panel survey (N = 27,004) combined with register data on past victimization, we examined how past and recent victimization impacts a sense of safety in Denmark. Overall, we found no main effect of past victimization on sense of safety. However, for recurrent victims, past victimization impacts current sense of safety in a heterogeneous way. The results suggest that past adversity can function both as an amplifier and as an attenuator of the effects of victimization on fear of crime.

KEYWORDS:

• Fear of crime
• sense of safety
• GEE
• panel data
• persistency of victimization
• victimization

Data availability statement

The data that support the findings of this study are not freely available. Restrictions apply to the availability of these data, which were used with special permission from the Danish National Police for this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. The mixed findings are also likely a consequence of using different methodologies and data, i.e., different measures of fear of crime and victimization. Specifically, longitudinal studies have generally found larger and more significant effects of victimization when using analytic technics (like OLS) that ignore unobserved heterogeneity as compared to techniques (like FE) that do not (see Cornaglia et al., 2014; Janssen et al., 2020; Mahuteau & Zhu, 2016; Scherg & Ejrnæs, 2020).

2. Conceptual clarity regarding how to differentiate between repeated victimization (normally measured as the frequency of victimization within a year) and the persistence of victimization (which entails a long-term pattern of victimization, e.g., being victimized yearly in a given time period) is lacking. Clark et al. (2019) have differentiated between “chronicity” (the average amount of victimization with a given time period) and “persistence,” which is a measure of the continuity of victimization. Others (Heap, 2021; Korkodeilou, 2017; Stanko, 1995) have used the terms “chronic” or “endemic” to characterize an enduring pattern of victimization.

3. The survey was carried out as a computer-assisted web and telephone interview (and in 2016–2018 also as a computer-assisted personal interview). More information about the survey can be found in the yearly methodological reports (e.g., (Rigspolitiet, 2018) [in Danish]).

4. The oversampling of respondents from disadvantaged neighborhoods reflected the Danish National Police’s focus on attaining a detailed picture of the safety and security situation in those areas.

5. Given the theoretical importance of recency for the impact of victimization it might seem counterintuitive to group those victimized in either T₁ or T₂ in the same group. The reason for doing this is that from a persistency perspective there is no difference between these two groups of victims as we estimate the average level of sense of safety across the two sweeps and therefore control for the effect of recency. We report on additional analysis with four groups (i.e., where those victimized in either T₁ or T₂ are analyzed separately (see “Robustness of results”).

6. We report on additional analysis done with a victimization profile measures which combine the persistency dimension with type of crime victimization (see “Robustness of results”).

7. The drawback of using register data is that only a subset of all crimes is reported to the police and registered in official statistics.

8. As it is not well-established what constitutes “near past” or “distant past” in the context of past victimization, the motivation to conceptualizing near past and distant past victimization is mainly informed by the desire to secure equal exposure. The oldest available victimization data is from the year 2001 so in order to secure an equal extent of exposure to past victimization a moving window of the last 2–7 and 8–13 years of victimization was used. As recent (self-reported) victimization is measured within the last year, past victimization was measured from two to 13 years before. A more fine-grained measures of past victimization where considered, for example, using three time periods (2–5 years before, 6–9 years before and 10–13 years before) instead of only two. However, the consequence of using such a conceptualization is relative few respondents in each group (especially in the past personal victimization categories) which thereby weakens the analysis. As part of the test of the robustness of the results we report on additional analysis done with this alternative specification of past victimization (see “Robustness of results”).

9. The data on past victimization came from the register KROF from Statistics Denmark. As crime is reported to the police, it is coded by the police using a five-digit crime-type identifier. This identifier is then recoded by Statistics Denmark into a seven-digit identifier called “OFR_GER7” according to a specific hierarchy, which enabled us to distinguish between different types of crime – specifically, between personal crime victimization (OFR_GER7 = “12xxxxx”) and property crime victimization (OFR_GER7 = “13xxxxx”). For more information on KROF, see: https://www.dst.dk/en/Statistik/dokumentation/documentationofstatistics/victims-of-reported-criminal-offences.

10. For more information on the KRSI-register see: https://www.dst.dk/en/Statistik/dokumentation/documentationofstatistics/convicted-persons”

11. There is not a good measure of frequency of victimization in our data as the victimization questions were dichotomous (victimized / not victimized). Instead, we have used the number of different types of crime (0–6) a person has been victimized by as a proxy for frequency of victimization. Those recurrently victimized have been victimized by an average of 3.12 types of crime, whereas occasional victims have been victimized by an average of 1.28 types of crime. Supplementary analysis, which is available upon request from the authors (Supplementary table A”), shows that the difference between victim groups becomes insignificant when controlling for frequency of victimization.

12. Supplementary table A is available from the authors upon request.

13. Supplementary table B is available from the authors upon request.

Additional information

Funding

This work was supported by the Danish Victims Fund [grant number 18-610-00035.