Data mining of key anthropometric variables in elite and non-elite wrestlers using principal component analysis

Abstract

Objective. To identify the key anthropometric variables that characterise elite and non-elite wrestlers by means of principal component analysis, and to compare the anthropometric profile of both competitive levels.

Method. Sixty male Iranian wrestlers (30 elite, 30 non-elite) underwent an anthropometric assessment including height, body mass, body mass index, eight skinfolds, two bicondylar breadths, eleven girths, body fat percentage, body fat mass, skeletal muscle index and total-body skeletal muscle mass. Principal component analysis was applied to the 27 variables in order to identify the most representative ones. Independent t-tests compared both groups and Cohen's d was calculated as effect size.

Results. Anthropometric features were reduced to two principal components explaining 76.3% of the total variance. The first component (50.7%) showed high loadings for girth measurements and muscular properties, whereas the second (25.5%) was associated with skinfold thicknesses and body fat properties. Elite wrestlers were taller, showed larger girths (shoulders, chest, biceps, forearm, thigh, knee and wrist), thinner skinfolds (triceps, calf, chest and front thigh), lower body fat percentage, lower skeletal muscle index and greater total-body skeletal muscle mass (p < .05). No differences were found for body mass index, bicondylar breadths or body fat mass.

Conclusions. Principal component analysis is a useful tool to detect representative anthropometric variables in wrestlers. Girth measurements and muscular properties emerged as the most informative dimension, which may support a more efficient variable selection process for talent identification and strength and conditioning programming in wrestling clubs.