Discriminant function analysis for sex determination using landmark coordinate data from three-dimensional mandible models

Euna Hwang, Chenghe Lin, Benzheng Jiao, Nak Eun Chung, Seung Ho Han, Jin Kim, U. Young Lee

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


In this study, we aimed to investigate the sex-discriminating potential of mandible dimensions, which we hypothesise can be measured easily and can indicate various sex discriminant functions in a simple manner. Statistical analysis was performed using data from five linear measurements and 10 indexes acquired from 14 landmark coordinates on 240 three-dimensional Korean mandible models (120 male and 120 female). The Korean mandible showed sex dimorphism in all the measurements and in four indexes (p < 0.01). Discriminant functions were generated by univariate, bivariate, and multivariate analyses using direct and stepwise methods, with accuracy ranging from 51.3% to 89.6%. Moreover, when comparing the dimensions of Korean mandibles with other populations, morphological differences were observed, thus implying that the sex discriminant functions can be generated according to each population. Therefore, in the present study, we noted that the various discriminant functions assessed using three-dimensional models of the mandible in this study indicated that the mandible dimensions alone had sex-discriminant potential, and our findings thus provided useful information for forensic examiners to accurately predict sex in a Korean population.

Original languageEnglish
Pages (from-to)332-344
Number of pages13
JournalAustralian Journal of Forensic Sciences
Issue number3
StatePublished - 3 Jul 2015

Bibliographical note

Publisher Copyright:
© 2014 Australian Academy of Forensic Sciences.


  • Koreans
  • Sex determination
  • discriminant function analysis
  • mandible
  • population specificity
  • three-dimensional model


Dive into the research topics of 'Discriminant function analysis for sex determination using landmark coordinate data from three-dimensional mandible models'. Together they form a unique fingerprint.

Cite this