Crime Insight Reporting includes the Error Rates and Common Errors report to help administrators recognize the most frequent issues being submitted by agencies, and check in on the outstanding error rate for the selected months. 


This report is available by default on the LEA and ADMIN reporting sites. 



To run the report, select the state or Agency, the Start Month, and End Month and click View Report. The first table shows the details for all months for which there is data in the system. The other tables provide specifics about the most common errors and warnings generated by the state or Agency within the selected period. 


The goal should be to eliminate all errors in Crime Insight so that the data provided to the FBI is of the highest quality, but it's very important to also provide feedback to the agencies generating the errors so they can have the situation resolved in their records management system.


Warnings may need to be addressed as well, representing possible issues with the submission. 


Technical Details


Prompts
  • Agency
  • Start Month
  • End Month
Data Sources
  • DataWarehouse
  • FBI cube database
No DataThe report only shows months when data was uploaded. Any months with no uploads are not displayed.


State-Specific Reporting


The following state-specific updates to the base report are available:


  • Error Rates and Common Errors has no state-specific customizations


Definition of the columns in the Error Rates and Common Errors report 

  • Month and Year -- based on the date when the file was uploaded to the system, not any other dates in each record (e.g. incident date, arrest date, etc)

  • Files Submitted -- total number of files submitted that contain data for this agency in the specified month.  This could be batch files or files containing individual records (e.g. from the Incident Editor or Sidearm)

  • Records Submitted -- total number of Group A Incidents, Group B Arrests and Zero reports across all uploaded files.  For this column if any record is uploaded multiple times (e.g., multiple uses of the Incident Editor to fix all errors), it is counted multiple times.

  • Distinct Records Submitted -- The same as the previous column, but where the report only counts the last time each incident, arrest or zero is uploaded in the specified month.

  • Records Rejected -- this counts the number of individual records that were rejected due to errors found when they were submitted.  As with the "Records Submitted" column, all errors are counted, even if the same incident were uploaded multiple times.

  • Outstanding Records Rejected -- like the Distinct Records Submitted, this looks only at the last submission for each incident, arrest or zero report in the specified month.  If the final version was accepted, it's not counted, if it was rejected, it is counted in this column.

Error Rate is simply "Records Rejected" / "Records Submitted" expressed as a percentage.  This is valuable is determining how an agency is doing at submitting records without errors.


Outstanding Error Rate is "Outstanding Records Rejected" / "Distinct Records Submitted" expressed as a percentage.  This is valuable in determining how an agency is doing at getting their errors corrected in the month when they're reported.


Scenario 1:
An agency is using the Incident Editor and it takes 5 tries to get one incident accepted without errors, but they succeed on the 5th try.  The number of File Submitted will be 5, Records Submitted = 5, Distinct Records Submitted = 1, Records Rejected = 4, Outstanding Records Rejected = 0, Error Rate = 80% and Outstanding Error Rate = 0%

Scenario 2:
Agency uses an RMS. They create a batch file that has 100 incidents in it, and CI rejects 5 due to errors. Then later the RMS submits a file with corrections for 3 of those errors, but one of them has a new error. In this case:
Files Submitted = 2
Records Submitted = 103
Distinct Records Submitted = 100
Records Rejected = 6
Outstanding Records Rejected = 3 (2 weren't corrected at all + 1 still rejected after the correction)
Error Rate = 6/103 (someone else can do the math, but something just under 6%)
Outstanding Error Rate = 3/100 = 3%
 


See Also