Page 199 - Automated Fingerprint Identification Systems (AFIS)

184 AUTOMATED FINGERPRINT IDENTIFICATION SYSTEMS

8.5.3.5 Is the Hit Rate Calculated by the Number of Unique Cases,
Multiple Searches of the Same Case, or All Latents Received?
If the hit rate is calculated by the number of unique latent print cases, it will
be a much higher rate than a hit rate based on the number of idents against
the number of latent prints collected. A hit rate of 10% is considered standard.
In the example above, if any one if the eight latent prints is identified on the
AFIS system, the case has an ident.
If the hit rate is calculated by using the number of all latent prints received,
this would necessarily include those prints determined not to be “of value.” If
the agency has crime scene investigators who could eliminate images which
were not “of value,” the number of latent prints would be smaller, but they
would be of better quality than the images collected by investigators without
fingerprint training.

8.5.3.6 Are Data Management System Reports Used
to Calculate a Hit Rate?
While very advantageous in a tenprint processing environment, statistics pro-
duced by data management systems have to be reviewed for duplication, testing,
and other search factors. The three functional areas of latent print capture and
entry practices, database characteristics, and identification statistics provide a
continuum from gathering latent images at a crime scene to searching, identi-
fication, and tabulation of the search results. While each agency might perform
these tasks in a slightly different manner, the combination of these individual
differences has a significant impact on the final search results and how they are
reported.
To reiterate a statement from the beginning of this section, there is no right
way to report idents, nor is there a national standard for reporting idents. Until
there is, opportunities for improvements will be missed because the technical
language and concepts are not fully developed.

8.5.4 NEW YORK STATE SURVEY
In the late 1990s, the New York State Division of Criminal Justice Services (NYS
DCJS) conducted a survey of latent print procedures of other identification
agencies based on the three functional areas of latent print capture and entry
practices, database characteristics and identification statistics. Under the direc-
tion of NYS DCJS Deputy Commissioner Leo Carroll, the goal of the survey was
to determine why some agencies were reporting such a wide range of identifi-
cation rates, and whether information could be gathered that would increase
the total number of latent print identifications. The findings of this survey were
presented to the Sagem Users Group, published in the Sagem newsletter, and