Exploring places of street drug dealing in a downtown area in Brazil: An analysis of the reliability of Google Street view in international criminological research

This study assesses the reliability of Google Street View (GSV) in auditing environmental features that help create hotbeds of drug dealing in Belo Horizonte, one of Brazil's largest cities. Based on concepts of “crime generators” and “crime enablers,” a set of 40 items were selected using arrest data related to drug activities for the period between 2007 and 2011. These items served to develop a GSV data collection instrument used to observe features of 135 street segments that were identified as drug dealing hot spots in downtown Belo Horizonte. The study employs an intra-class correlation (ICC) statistics as a measure of reliability. The study showed mixed findings regarding agreement on some features among raters. One on hand, the observer's lack of familiarity with the local culture and street dynamics may pose a challenge with regards to identifying environmental features. On the other hand, factors such as image quality, objects that obstruct the view, and the overlooking of addresses that are not officially registered also decrease the reliability of the instrument. We conclude that a combination of tools and strategies should be applied to make the use of GSV truly reliable in the field of international criminological research.