A1560
Title: Odds ratio for assessing the risk of crime associated with darkness
Authors: Ezgi Erturk - University College London (United Kingdom) [presenting]
Jemima Unwin Teji - University College London (United Kingdom)
Peter Raynham - University College London (United Kingdom)
Abstract: The aim is to explore whether manipulating the control period does not affect the results in estimating the crime risk associated with darkness. The Metropolitan police services dataset for London between 2013 and 2019 was used. Comparing crime rates that took place in light in one week and dark in the other week at the same clock time (spring and autumn clock change) or vice versa is defined as the "Test period." Different control periods were created to compare the changes in the crime counts in the test period and to minimise other factors that may encourage crime. Therefore, 45-minute, 1-hour, and 2-hour lengths of control periods on the odd ratio calculation were explored. Although, in the 45-minute duration, the odds ratio of criminal damage, robbery, and motor vehicle offences increased by 7\%, 11\%, and 10\% more compared to the one-hour period, the significance tests showed that all p-values were well above the conventional threshold of 0.05. The only noticeable change was in the odds ratio for robbery from the person between the 45-minute and 2-hour control periods, with a p-value of 0.089. However, this result was still not statistically significant. It shows that even through the control period, there are no significant statistical changes. All variables demonstrate that darkness has an impact on crime. It is important to understand the crimes and areas in which electric lighting design is likely to have an effect.
