TY  - JOUR
PY  - 2004//
TI  - Why Didn't We Get Them All?: Analyzing Unlinked Records in California's Linked Homicide File
JO  - Homicide studies
A1  - Trent, Roger B.
A1  - Van Court, Jason
SP  - 311
EP  - 321
VL  - 8
IS  - 3
N2  - The California Linked Homicide File links two major files containing information on homicides. The basic file includes 34,542 cases from Supplementary Homicide Reports (SHR). Using a probabilistic matching program, 32,122 (93.0%) of these cases were linked with information from Vital Statistics (VS) death records. Why weren't the remaining 7% of cases linked? What makes these 2,420 records different from the cases that were linked? Analysis demonstrated that many of the unlinked cases involved child abuse and that many deaths appearing as child abuse in the SHR were not classified as assault deaths in the VS death records. By broadening the definition of what kinds of deaths in VS death records could be possible links with SHR reports, we were able to marginally improve the overall linkage rate (from 93.0% to 94.2%) and to increase the pool of linked child abuse deaths by 7.1%.<p />
LA  - 
SN  - 1088-7679
UR  - http://dx.doi.org/10.1177/1088767904265433
ID  - ref1
ER  -