Structural principal components analysis of the kinematics of the soccer kick using different types of rating scales

Abstract

Different types of rating scales have been developed to guide the qualitative analysis of sports technique and improve rating accuracy. However, it remains unclear what type of scale is more accurate and how rating accuracy is influenced by scale effects. This study aimed to: i) investigate the accuracy of rating scale type (analogue, numerical analogue, Likert and numerical Likert) for the rating of kinematic variables; and ii) explore how rating accuracy is influenced by scale effects. One experienced rater constructed the 100 mm-long scales and a group of 327 novice analysts rated the video-recorded soccer kick performance of 32 school children. Estimation error was calculated as the difference between estimated values and objective digitized values. Over-and under-estimation tendencies were assessed. The scale range was divided into equal-length intervals (e.g., low, middle, high angular range) to assess scale effects. Statistical analysis consisted of an exploratory principal components analysis with varimax rotation, where items were forced into a three-factor solution. Numerical Likert scale designs may be suggested as optimum for the instruction of novice raters. Central-tendency error adversely affected visual estimations; however, other scale effects enhanced accuracy. The findings have implications for optimum scale design and improved instruction in qualitative analysis.