Why is the Queued to Initial Response interval type end time incorrect on some of these intervals?

Environment: 

Analytics, Incident Performance

Resolution:   

If there is a queue assignment and an initial response sent at the same time and a second response is sent at some point after this, the Queued to Initial Response interval end time will be recorded in the incident performance table as the start time from the second response. This is correct functionality. The incident must be assigned to the queue (or agent) before the response is sent for this interval type.

For example, if incident is assigned to a queue on 04/16/2016 at 9:50 AM and the first agent response is sent out at the same time as the queue assignment and a second response is sent on 4/16/2016 2:35 PM, since the queue assignment and first response were performed simultaneously, the end time for the Queued to Initial Response interval is the timestamp of when the second response was sent which would be 4/16/2016 at 2:35 PM.  When these actions occurred simultaneously and the incident was not assigned to the queue before that initial response was sent, the appropriate end timestamp for the Queued to Initial Response interval was the timestamp of the second response sent.  The incident must be assigned to a queue before the first response is sent.