A Total Score of Athleticism to Estimate the Amount of Variance Explained in On-Field Performance Within Collegiate American Football Players

Abstract

The identification of sport and position-specific key performance indicators has been of increased interest to practitioners working within high performance sport settings. The aim of this study was to create statistical models, across a spectrum of position groups, that explain the largest amount of variance in on-field performance over four collegiate American football seasons. A total of four position groups were analyzed, including Wide Receiver/Tight End (n = 29), Running Backs (n = 8), Linebacker/Defensive Line (n =41), Defensive Backs (n = 28). Different tests of physical performance (back squat, bench press, hang clean, 40-yard dash, pro-agility, L-drill, vertical jump, and broad jump) were set as the independent variables, while a Total Score of On-Field Performance, consisting of position-specific game statistics was set as the dependent variable. Linear mixed models, using backwards model selection were used to identify models that best fit the position-specific data. Results revealed that models consisted of at least three different physical performance test, were position-specific, and explained the variance in on-field performance to different degrees. Significant correlations were found between physical performance models and on-field performance. Results from our study may be of interest to practitioners working within American football that are interested in maximizing success through and individual approach towards the implementation of training.