Technical Note: Reliability of Pöch's facial shape classification system: A morphometric approach.

Area of Science:

• Forensic Anthropology
• Human Morphology
• Biometrics

Background:

• Standardized facial characteristic vocabularies are crucial for individual identification in forensic anthropology.
• Pöch's facial shape classification system is widely adopted but lacks critical reliability evaluation, especially for non-European populations.

Purpose of the Study:

• To evaluate the representativeness and reliability of Pöch's facial shape classification system in a Mexican population.
• To assess observer consistency and agreement when using Pöch's system for face shape description.

Main Methods:

• Geometric morphometrics and principal component analysis were used to assess facial diversity capture.
• Sixteen observers performed dual classifications of 60 faces using Pöch's system.
• Intra- and inter-observer agreement were quantified using Cohen's Kappa and Fleiss' Kappa statistics.

Main Results:

• Pöch's system exhibited significant morphological redundancy among categories.
• Low intra-observer (mean Cohen's Kappa = 0.203) and inter-observer (mean Fleiss' Kappa = 0.112) agreement was observed.
• Only 'round' and 'oval' facial shapes achieved high consensus; other categories showed poor agreement.

Conclusions:

• Pöch's system demonstrates inadequate reliability and representativeness for describing facial shape variation in the studied Mexican sample.
• A post-hoc k-means clustering analysis identified four distinct facial shape categories, suggesting a basis for improved classification systems.