Although homicide is a leading cause of death of infants and coddlers,
there is some suspicion that an unknown number of additional deaths a
re unrecognised homicides. The authors used California mortality data
from 1969 to 1991 to examine 12,246 injury deaths that occurred before
age 5. Characteristics of the dead child, injury event, and postmorte
m were compared for accidents, homicides, and undetermined deaths. A l
ogistic model was developed to differentiate homicides from accidental
deaths and then was used to predict whether undetermined deaths were
likely to be homicides or accidents. Unlike accidental deaths, undeter
mined and homicides had similar distribution patterns of age, race, se
x, and place of injury. The predictive model indicates that 43.8% of t
he undetermined injury deaths were similar to homicides on several cha
racteristics True rates of homicide for infants and toddlers may unfor
tunately be nearly one fifth and one tenth higher respectively if the
undetermined deaths that resemble homicides are token into account.