I need to evaluate the impact of an intervention that was applied at different times in a number of police units to improve the effectiveness of the police. We need to control other covariates such as rotation of the commissioners, and available resources.
The response variable is the number of offenses (10 different offenses) and we have 60 observations. These correspond to the monthly number of complaints and arrests and (48 pre-and 12 post-intervention).
Hello Juan,
As your DV is a count, you may use either Poisson regression or a negative Binomial regression. Here is an excellent resource from the UCLA, which mentions 4-5 methods of analysing count data using a regression analytical tool. The analysis used STATA but you should be able to do the same analysis in SAS or SPSS.
http://www.ats.ucla.edu/stat/stata/seminars/count_presentation/count.htm
Hope this helps.
Andrea
Did you measure multiple respondent within each police unit? Then, you need to correct for the multilevel data structure in your analyses.
- Badges
- Science topic