Moscow University
Anthropology
Bulletin
Anthropology
Bulletin
Peter the Great Museum of Anthropology and Ethnography (Kunstkamera), Russian Academy of Sciences, Saint Petersburg, Russia
Shirobokov I.G., e-mail: ivansmith@bk.ru
In 2015 Peter the Great Museum of Anthropology and Ethnography (Kunstkamera) organized a workshop on estimation of inter-observer consistency in craniometry. Workshop participants measured two cranial series of Russians and Buryats. Each series consisted of 10 adult human sculls. The program of cranial measurements included traits widely used by anthropologists (N according to Martin: 1, 8, 17, 9, 11, 45, 48, 51, 52, 55, 54, SC, SS, MC, MS, 72, 73, 75(1)). The lowest consistency rates among participants were obtained for traits describing nasal bones and vertical facial angles, followed by the form of the orbit and piriform opening. The highest consistency rates were obtained for basic diameters and horizontal facial angles. Most probably, discrepancies in measurements are due to accidental errors in small size traits and methodical differences owing to difficulties in localizing some cranial landmarks (particularly maxillofrontale, ektokonchion and nasospinale). Mistakes made due to lack of attention are no less significant than methodological differences between anthropologists. The effect of inter-observer errors on results of intragroup analysis was evaluated using principal component analysis, on results of intergroup analysis – with discriminant canonical analysis. Results of intragroup analysis (homogeneity of sample) are comparable among researchers when there are no discrepancies in sex estimation. In intergroup analysis cumulative effect of inter-observer error and sample effect is very significant. It is not correct to evaluate similarity between given populations on the basis of rapprochement of corresponding samples within the first canonical vector space. Curtailment of craniometric program by means of removal of unreliable traits don’t increase validity of research results because of inter-observer discrepancies in sex estimation. Solution to these problems may consist in introduction of compulsory practice of repeated measurement of all morphometric traits in each cranial series and in using comparative data of those colleagues whose method of measurement is known to be the same.
craniometry, inter-observer consistency, intergroup analysis, intragroup analysis
Цит.: Shirobokov I.G. How serious is an effect of inter-observer errors on conclusions of craniological studies? (some results of the workshop on craniometricconsistency in Peter the great Museum of Anthropology and Ethnography) // Moscow University Anthropology Bulletin (Vestnik Moskovskogo Universiteta. Seria XXIII. Antropologia), 2016; 3/2016; с. 36-48
Download text