Abstract
Fear of crime has become a major focus of attention both for policy makers and within the academic literature and although geography is considered relevant to fear of crime little is known about how it varies spatially. The present article applies multilevel modelling to a newly available geocoded version of the British Crime Survey to explore the feasibility of estimating fear of crime down to small area level across England. Five of the six models are considered sufficiently powerful in explaining the between-area variance to produce estimates. Estimates of fear of crime for these five crime types are produced (with 95 per cent credible intervals to indicate their likely precision) for small areas and local authorities in England and are placed in the public domain. The models also highlight, however, that most of the variance in individual's fear of crime is within rather than between small areas, questioning the logic of area-based policies to tackle fear of crime. In an era of public sector resource tightening and targeting, it is hoped that these findings will enable practitioners, policy makers and academics to continue to understand and monitor geographical variations in fear of crime.
Original language | English |
---|---|
Pages (from-to) | 48-68 |
Number of pages | 20 |
Journal | Crime Prevention and Community Safety |
Volume | 14 |
DOIs | |
Publication status | Published - 20 Jan 2012 |
Keywords
- fear of crime
- multilevel modelling
- monitoring
- synthetic estimation
- small area estimation
- area-based intervention