Determination of sex from hand and finger dimensions in a regional eastern Indian population using discriminant function analysis

Abstract

Sex determination is one of the important facets of human identification. It becomes especially difficult, when fragmentary remains  are brought for forensic examination or when bare/gloved hand prints are recovered from a crime scene. In this study we try to  determine mathematical models to find out sexual dimorphism from hand and finger dimensions among adult Eastern Indian  population. 170 2nd Year medical undergraduate students (125 males and 45 female) between the age group of 20-25 years were  examined, and their hand and finger dimensions estimated. The average hand and finger dimensions were greater in males as  compared to females. Discriminant function analysis was used to analyse the data, which could estimate sexual dimorphism from  right hand finger lengths in 74.7% cases, based on cross-validated results, with a Wilk's Lambda value of 0.703 while from left hand  finger lengths, the same was 72.4%, with a Wilk's Lambda value of 0.762. While, the model for the right-hand dimensions had a  predictive accuracy of 90 %, based on cross-validated results, with a Wilks's Lambda value of 0.455, and the same for left hand  dimensions was 87.6% with a Wilk's Lambda value of 0.468. Since anthropometric variations exist between different ethnic groups  residing in different geographical areas, the results obtained in this study can be exclusively applied to the Eastern Indian regional  population. Although it has a moderately high predictive accuracy, this should act as an adjunctive in forensic case work in future.