Sex determination in Thai skulls by using craniometry: Multiple logistic regression analysis.

ABSTRACT

Objective: To evaluate sexual dimorphism in the cranium and mandible of Thais by using Jorgensenûs craniometry and to develop a statistical model to determine sex from craniometrical measurements and indices. Methods: One hundred and one Thai skulls (66 males and 35 females) which ranged in age from 18 to 86 years were studied. Results: According to craniometry, the skull of a male is larger and higher than that of a female. Considering each individual measurement, although 26 of 30 measurements and 5 of 14 indices showed a statistically significant difference between males and females, they had some overlaps. To predict gender more accurately, a multiple logistic regression model based on 4 skull measurements (mm) i.e., nasion-basion length (M5), maximum breadth of the cranium (M8), facial length (M40), and bizygomatic breadth of the face (M45) was developed as follows In (odds) = Z = -52.5312 + 0.27 M5 - 0.1867 M8 + 0.1268 M40 + 0.319 M45 The probability of being males (P) is then ez/(1+ez). Using a cut off point for P of 0.5, this logistic model could correctly predict 82.9% (95% CI 66.4%, 93.4%) of females and 92.1% (95% CI 82.4%, 97.4%) of males respectively with the overall accuracy of 88.8% (95% CI 80.8%, 94.3%). Conclusion: Jorgensenûs craniometry of the cranium and mandible can be used to determine gender among Thais via a multiple logistic regression model on M5, M8, M40 and M45.