A DOJ Scandal of a Different Sort

By: Kathy Gerber
Published On: 5/14/2007 10:26:33 AM

While the Department of Justice is embroiled in scandal after scandal, how difficult is it for personnel to carry out their jobs?  What's going on with crime in this country anyway?

Crime data is reported from the various states.  The states in turn get their data from the locales.  It doesn't appear that DOJ altered the data, but they do the next best thing - spin, spin, spin.
A quick perusal of the DOJ Bureau of Statistics Key Crime and Justice Facts at a Glance page delivers numerous reassuring messages.  Crime is going down.  Drug arrests are going up.  And by the way some versions of these charts are proudly featured on the Whitehouse web site.

But  a closer look at some of the details on that page reveals summaries absolutely loaded with convenient omission, "look over here" commentary, and occasional falsehoods.  Here's an example.  Click on the graph to reach the DOJ page.


The increase was far from slight; it was massive.  Here's my plot of the same data - just a bit larger.


And this clarifies the point that describing the increase in 2005 as "slight" is  plainly wrong.


OK, so a slight increase good mean just a little random variation or noise.  It happens.  But the truth is there's nothing slight about the increase in 2005.

Not every responsible citizen has the background to dig through these numbers.  It is the reponsibility of our public servants to inform and clarify these matters to the general public.  Taking advantage of our collective lack of experience with matters statistical means is a failure on the part of DOJ.

Someone made the decision to cast a positive light on recent trends in the nation's crime rates.  Who?  Why?  This is the Department of Justice, not The Martin Agency.  When it comes to reporting data summaries creativity is not among the core skills needed.

Congress needs to find out who is making these decisions at BJS.  And there's no need to wait.  Thoroughness, accuracy and transparency are essential when it comes to crime trends.  The public needs to be informed when crimes of a particular nature are on the rise.

For reference here is the data for the report.

                                                                                                                                                                                                                                                                                                                                                       
YearsFirearm incidents

Firearm victims
Firearm crime rates  (Victims per 1,000  residents)



Firearm crimes  as a percent
of all violent incidents
1993 1,054,820
1,248,250
5.9

11



1994 1,060,800
1,286,860
6

11



1995  902,680
1,050,900
4.9

10



1996845,220
989,930
4.6

10



1997680,900
795,560
3.6

9



1998557,200
670,480
3

8



1999457,150
562,870
2.5

7



2000428,670
533,470
2.4

7



2001467,880
524,030
2.3

9



2002353,880
430,930
1.9

7



2003366,840
449,150
1.9

7



2004280,890
331,630
1.4

6



2005419,640
477,040
2

9
















Source:  National Crime Victimization Survey







For more information about  the purposes and advantages of the UCR and the NCVS, see The Nation's Two Crime Measures.


Comments