Caliber Estimation from Cranial Entrance Defect Measurements

Abstract

Caliber estimation from entrance defects has long been rejected by forensic scientists. This appears to be a consequence of soft tissue perspective of forensic pathologists. This study examined the relation between caliber and cranial entrance defects and maximum cranial thickness. The calibers considered in this inquiry were .22, .25, .32, and .38. The sample consisted of 73 specimens obtained at autopsy (thirty-seven of .22 caliber, five of .25, six of .32, and twenty-five of .38). To test the strength of the relation between caliber, minimum diameter, and maximum thickness Pearson correlation coefficients were conducted. The strongest relationship was observed between caliber and minimum diameter. A relationship between minimum diameter and maximum thickness was also observed. To test the null hypothesis that the mean minimum diameter is not significantly different between calibers an analysis of variance procedure was performed. The ANOVA yielded a strong relationship between dependent variable minimum diameter and caliber. Multiple regression analysis measuring the association between minimum diameter, caliber, and maximum thickness was also conducted. The Pr > F .0001 suggests that the overall model is significant. Discriminant functions and canonical variables were obtained. Classification was first performed by using two values small and large calibers. The large caliber group consisted of .38, while the small caliber group included .22, .25, and .32. The correct classification rate using crossvalidation for large caliber is 86.96%, and 93.33% for the group small caliber. A narrower classification was also performed by using three values, .23 caliber (.22 and .25 calibers grouped), .32, and .38 as the criterion variable groups also using minimum diameter and maximum thickness as predictors. The correct classification rate using crossvalidation is 82.02% for .23 caliber, 73.94% for .38 caliber, and 16.67% for .32 caliber defects. The discriminant functions can be used with appropriate caution to classify observations into groups defined by caliber using minimum diameter and maximum thickness as the predictors. Caution is suggested when attempting to estimate caliber from defects that are not produced from the perpendicular entrance of a bullet.