Crime rates vary across space and time. The reasons crimes are committed in some places but not others can be difficult to detect because of complex socio-economic factors, but policymakers still need to understand how crime rates are changing from place to place and from time to time to inform their policies.
Many government statistics, such as crime rates, come from nested datasets. Most US States are divided into counties (Alaska has “burrows,” and Louisiana has “parishes”), and counties and county-level governments can vary within the same state. For example, one county might have a high population density and be urban, whereas a second county might have a low population density and be rural.
In this project we will use a form of regression called hierarchical modeling to capture and explore crime statistics collected by the State of Maryland to see if there is a linear trend in violent crime across the state between 1975 and 2016. These data come from the Maryland Statistical Analysis Center.