Reviewing Vendor Software
Vendor reviews occur when the first install of a vendor's records management system commences in Connecticut with the claim that this software contains NIBRS data elements and edits. Vendors sometimes assume that sending files we could read was sufficient, or they assumed a low error rate submitting the 33 FBI scenarios was adequate to say their software could do NIBRS. These assumptions are incorrect in our experience.
We have found no substitute for evaluating NIBRS software capabilities than submitting live data. The only alternative to submitting live data from a Connecticut agency is when the vendor is successfully submitting NIBRS data in another state with requirements at least minimally comparable to the FBI's. We confirm such claims with our counterparts in the other states.
It is amazing what data degradation can occur even when error rates are at 2% or less, and the data structure appears to be correct. When offense codes are based on another table, such as NCIC codes or state statutes, and indexed to refer to NIBRS offense codes, the structure and choices can lead to inappropriate NIBRS offense codes. The same can be said of the property codes, which may provide 700 options to users that are then indexed to the NIBRS choices out of the users view. The indexing to the NIBRS codes may not always be correct.
Categories may be incorrectly structured, such as having an age range of three characters instead of four characters. Two other instances of separate vendors were when an arrestee age was entered as Unknown, or was not entered at all, and the age submitted to the state was of a 98-year old arrestee, or of an under 10-year old arrestee.
The final caveat with vendor software is that a vendor may fix a problem at one site and leave the problem intact at all other sites. This has occurred.