Does Hot Spots Policing Have Meaningful Impacts on Crime? Findings from An Alternative Approach to Estimating Effect Sizes from Place-Based Program Evaluations
Objectives Prior meta analyses of hot spots policing show that the approach reduces crime, but report relatively small mean effect sizes based on Cohen’s d . The natural logarithm of the relative incidence rate ratio (log RIRR) has been suggested as a more suitable effect size metric for place-based...
| Main Author: | |
|---|---|
| Contributors: | |
| Format: | Electronic Article |
| Language: | English |
| Published: |
2022
|
| In: |
Journal of quantitative criminology
Year: 2022, Volume: 38, Issue: 1, Pages: 1-22 |
| Online Access: |
Volltext (lizenzpflichtig) Volltext (lizenzpflichtig) |
| Journals Online & Print: | Drawer...
|
| Check availability: | HBZ Gateway |
| Keywords: |
| Summary: | Objectives Prior meta analyses of hot spots policing show that the approach reduces crime, but report relatively small mean effect sizes based on Cohen’s d . The natural logarithm of the relative incidence rate ratio (log RIRR) has been suggested as a more suitable effect size metric for place-based studies that report crime outcomes as count data. We calculate the log RIRR for hot spots policing studies to assess whether it changes interpretation of hot spots policing’s impact on crime. Methods Cohen’s d and log RIRR effect size metrics were calculated for 53 studies representing 60 tests of hot spots policing programs. Meta-analytic techniques were used to compare the estimated impacts of hot spots policing on crime and investigate the influence of moderating variables using the two differing effect size metrics. Results The Cohen’s d meta-analysis revealed a “small” statistically significant mean effect size favoring hot spots policing in reducing crime outcomes at treatment places relative to control places ( d = .12) of approximately 8.1%. In contrast, the log RIRR meta-analysis suggests that hot spots policing generated a more substantive 16% ( d = .24) statistically significant crime reduction. The two metrics also produced differing rank orders in magnitudes of effect for the same studies. Conclusion Cohen’s d provides misleading results when used to calculate mean effect size in place based studies both in terms of the relative ranking of the magnitude of study outcomes, and in the interpretation of average impacts of interventions. Our analyses suggest a much more meaningful impact of hot spots policing on crime than previous reviews. |
|---|---|
| ISSN: | 1573-7799 |
| DOI: | 10.1007/s10940-020-09481-7 |