Have new street lighting schemes reduced crime in London?

Abstract

Crime counts published by the Home Office for the Metropolitan Police Crime and Disorder Reduction Partnership areas have been collated across the years 2003-2009. The crime counts over time have been modelled taking into account the ‘multilevel’ (years within areas) nature of the data. The key variable of interest, as a predictor of within-area change of crime, is the proportion of a Core Investment Period of new Private Finance Initiative street lighting which had been completed up to the given time point as a predictor of within area change of crime. The final model gave a 95% confidence interval for the multiplier by which the number of crimes is increased of (0.87, 1.11), for a fully implemented lighting programme, consistent with zero effect.

The full text can is available here.