Comparison of Start Detection Methods for the Countermovement Vertical Jump with Arm Swing

Abstract

Start detection methods for force-time data analyses of the countermovement vertical jump with arms akimbo (CMJ) have been applied to the CMJ with an arm swing (CMJ-AS). However, little is known regarding the implications of different start detection methods on derived CMJ-AS outcome measures. We compared CMJ-AS start locations that were identified by four detection methods: 1) (body weight [BW] ± 5 standard deviations [5SD]), 2) (maximum [max] BW + 5SD or minimum [min] BW – 5SD), 3) (BW ± 2.5%), and 4) (max BW + 2.5% or min BW – 2.5%). Twenty-two collegiate male basketball players (20.23 ± 2.00 yr, 1.99 ± 0.07 m, 93.79 ± 8.48 kg) performed 3 CMJ-AS trials while vertical ground reaction force (vGRF) data was recorded. A one-way repeated measures ANOVA (α ≤ 0.05) supplemented by Cohen’s d effect sizes compared time to takeoff (TTT), unloading phase metrics, and CMJ-AS start vGRF value amongst the methods. TTT was longer for method 1 (0.99 ± 0.16 s) than methods 2 (0.94 ± 0.17 s, p < 0.01, d = 0.29) and 4 (0.92 ± 0.19 s, p = 0.01, d = 0.43). Unloading time was longer for method 1 than methods 2 (p < 0.01, d = 0.16) and 4 (p < 0.01, d = 0.22) and longer for method 3 than method 4 (p = 0.02, d = 0.21). Unloading force and rate of force development were different between methods 1 and 2 (p ≤ 0.01, d = 0.17) and methods 3 and 4 (p ≤ 0.01, d = 0.21-0.22). There was no difference amongst methods in CMJ-AS start vGRF value. Erroneous start identification rate was lowest for method 4 (3.03%). When analyzing the CMJ-AS, max BW + 2.5% or min BW – 2.5% is recommended to improve start detection and avoid compromised unloading phase metrics.