Normalizing data: defined and an example from a prosecutor’s office

A performance audit in late 2002 of the Pierce County (Washington) Prosecuting Attorney’s Office by the American Prosecutor’s Research Institute (APRI) provides a useful example of why analysts normalize data. In the table below, the first row of data shows the number of felony case referrals in 2001 for six Washington State Counties. The second row normalizes that data in two steps: it divides the county’s population by one thousand and then divides that number into the number of felony referrals.

Clark King Kitsap Pierce Snohomish Thurston
4,101 11,859 4,061 10,008 5,654 2,350
11.6 6.8 17.4 14.0 9.1 11.2

Once you see the normalized data adjusted for each county’s population, so that sheer numbers of people is not the main determinant of felony deferrals, you see that King must be very large, because it has more felony referrals than any county, but its rate per 1,000 residents is around half or less of four counties’ rates. Or that Kitsap, which has comparatively few referrals, actually leads the group once you normalize for population. (See also my post of May 31, 2005 on normalizing city and state figures by numbers of residents.)

Whatever the law department metric, it conveys more if the analyst normalizes those metrics against some shared figure, such as revenue, lawyers, countries operating in, patents filed, lawsuits resolved, or another figure.

We welcome comments

Your email address will not be published. Required fields are marked *