Data Quality Assessments with Statistical and Database Tools
NIBRS is a statistical data set managed in a database environment. Those of us who want to be more confident in data quality than we are today must strive to use both sets of tools to manage, manipulate and analyze the data for quality purposes. Connecticut has automated the tests listed in the Certification: Referential Integrity Reviews section.
In keeping with the discovery mode of our data quality work up to this time, we are trying out a range of other possibilities not covered by FBI edits. As we become more adept with the database tools, a variety of database methodologies will be applied to data quality issues.
A few of the many prospects for investigation are as follows:
- Runaways: We know this is inconsistently reported in the UCR program. Do we see the same inconsistency in NIBRS? Does all software contain this field?
- Vehicles Stolen Locally and Recovered in Other Jurisdictions: Compare pre-NIBRS/UCR data with the NIBRS data from the same jurisdictions. There is no required NIBRS field. Is this entered as Recovered Locally, or not entered at all? This cannot be answered with data analysis alone. Communication with jurisdictions will be necessary to clarify practices.
- Compare Aggravated Assault, Simple Assault and Intimidation proportions with the standards used for quality checks on similar UCR data.
- Compare ratios of Disorderly Conduct and Simple Assault arrests in UCR with NIBRS by jurisdiction to find changes and understand their origins.
- Compare ratios of Robberies, Burglaries, Motor Vehicle Thefts and Larcenies to see if these are roughly comparable, uncover changes and determine whether lasting changes result from errors or are justifiable. There are examples where coding errors have varied these numbers significantly in both UCR and NIBRS.
- Are juvenile arrests in UCR comparable to NIBRS? Agencies certified in the last year will have this issue addressed before certification, but not the previously certified jurisdictions. This can be a problem with NIBRS when it was not a problem in UCR due to the computerized medium of NIBRS stimulating juvenile staff to avoid entering the data to maintain confidentiality.