In artistic gymnastics, competition results are determined by judges, whose task is to evaluate the perceived quality of movement as accurately as possible. A comparative analysis of measurable motion data and judges' scores is intended. For this study, video-recorded stretch jumps with turn motion data were analysed and evaluated by judges. A total of 174 videos from 29 female participants, each performing 6 trials, were assessed by three expert judges (national licence, at least 10 years' experience). We hypothesise that, in the comparative analysis of motion analysis data and judges' scores, various movement aspects have different influences on the evaluation. The results show several significant correlations between the judges' scores (deductions for jump height, position, landing) and the parameters of the motion analysis. The strongest Kendall’s tau-b correlation between judges' deductions and motion analysis data was observed for the number of steps (τb = 0.600, p < .001), and a strong negative correlation for jump height (τb = -0.547, p < .001). The results of Krippendorff’s alpha test show a lack of agreement between the experts for the judges' scores, but there is a significant positive correlation between the experts for the judges' scores. We conclude that deductions have a varying influence on the total score, depending on the number of parameters considered and whether the competition rules clearly define the gradation of deductions.
Veit et al. (Sat,) studied this question.