Different predictive and accuracy models for sex and stature estimation from second- and fourth-digit lengths in the Kinnaur population of Himachal Pradesh,North India: Medico-legal and forensic implications

Abstract

Forensic anthropologists are not only active in courtroom testimonies but also contribute to disaster victim identification. The timely recovery and identification of victims in any disaster is crucial, and anthropologists play a vital role in this process. In mass disasters such as airplane crashes, tsunamis, earthquakes or terrorist attacks, fragments of dead bodies are usually brought for identification. Each fragment demands special tactics to fulfil any of the basic attributes of forensic identification such as age, sex, stature or ancestry. If sex and stature can be estimated from a single fragment, then it can potentially reduce the identification time and possible victim matches. In this study, an attempt has been made to estimate sex and stature from the second- and fourth-digit lengths. Different predictive and accuracy models have been devised using statistical techniques. No bilateral difference was observed in the digit lengths in either sex. Consequently, an average of the digit lengths was employed for model approximations. The fourth-digit length provided the best sex estimates (M=82.1%, F=79.2%) when binary logistic regression (BLR) statistics were applied. It was also observed that overall sex estimations improved from 78.8% (from BLR) to 79.8% when discriminant function analysis was used for sex estimation. Second- and fourth-digit lengths were used independently and together for stature estimation models for males and females individually and for the pooled sample. In stature estimation models, three major interpretations were perceived: (1) the second-digit length provided the best stature estimates; (2) the estimation models compute better stature estimates for females than their counterparts, that is, standard errors are less in females than males; and (3) second and fourth digits predict more reliable stature estimation when the sex of the digits is known than that of the pooled sample.