engleski

Validation of glabella morphology for sex estimation on the human skull

This study aimed to test the applicability of glabella morphology for sex estimation on the medieval sample from Eastern Adriatic Coast and to test the repeatability of scoring the trait. The glabella was scored according to Buikstra and Ubelaker twice by two experienced researchers and two students. We analyzed 38 skulls of known sex, determined by amelogenin analysis in previous research.We tested intraobserver and interobserver error using Cohen’s Weighted Kappa and interclass correlation coefficient. We computed the frequencies of the scores by sex and calculated the probability of being male and female for each score. Sex of each was estimated using probabilities from the previous research by Walker and from a present study using the assumption that individuals with a probability of being male > 0.5 are males, and individuals with a probability of being female >0.5 are females.The prominence of glabella showed moderate or almost perfect levels of agreement for intraobserver and interobserver errors for most evaluators (Cohen`s kappa test = 0.55 – 0.89, ICC = 0.73 – 0.95), and if we showed that the influence of subjectivity was small for this cranial characteristic. However, the experience of the researcher impacted the reliability assessment of morphological feature. The accuracy of sexing skulls was 76% (74% males and 80% females) and was same for probabilities from the previous and present study. Finally, the study showed that the prominence of glabella is a relatively reliable morphological sex indicator, especially accurate in cases with hypermasculine scores (4 and 5)

glabella, sex estimation, croatian medieval sample

nije evidentirano