I think the point tsax meant to make is that the NCVS tracks most crimes besides homicide. For crimes besides homicide and burglary, the victimization surveys a

I think the point tsax meant to make is that the NCVS tracks most crimes besides homicide. For crimes besides homicide and burglary, the victimization surveys agree well with the arrest rates. I suppose this data doesn't rule out the possibility that the UCR is not very biased for robbery/rape/assault, but is biased for murder. If you are interested in this topic, Scott Alexander has a recent blog post on it that's well worth reading: http://slatestarcodex.com/2014/11/25/race-and-justice-much-more-than-you-wanted-to-know/ http://slatestarcodex.com/2014/11/25/race-and-justice-much-m...