Research Reports / Tackling UCR’s missing data problem: A multiple imputation approach

Tackling the Problem of Missing Data in the FBI’s Uniform Crime Reporting Program

The Uniform Crime Reporting Program (UCR) is an FBI initiative to collect crime and arrest statistics from law enforcement agencies across the country. However, reporting to the UCR is voluntary and many agencies do not participate, leaving the data looking like swiss cheese. If you want to know the crime and arrest rates in your county, you’re likely to find an incomplete picture.

The Research Team at MFJ developed a procedure to help fill in the gaps. To do this, we use a technique called Multiple Imputation by Chained Equations, or MICE. You can read more about our methods in this paper: Tackling UCR’s missing data problem: A multiple imputation approach. 

We also invite you to our GitHub repository to check out our code or experiment with the data.  Trouble getting the code to work?  Thoughts on improving the code, or general questions?  We welcome any feedback, just create an issue or pull request, or email us.  We’re happy to help!