Comparing Methods Of Individual Training Response With Countermovement Jump Metrics In Air Force Personnel

For assessing individual training responses, several methods exist to calculate typical error (TE) including test-retest reliability, reported coefficient of variation (CoV) in literature, and gross variability from pre- to post-testing. PURPOSE: To compare 3 methods of assessing individual response to a training program using countermovement jump (CMJ) metrics of jump height (JH), peak force (PF/BM), and modified reactive strength index (RSImod). METHODS: 32 active-duty Airmen (24 M, 8F) performed 3 CMJs before, weekly, and after a 10-week training study. Three methods were used to calculate TE; reliability method used the standard deviation (SD) of the pre-test and 1st week differences, literature-based method used an average of reported CoVs, and the gross variability used pre-and post-test differences. Individual confidence intervals (CI) were calculated at 90% for reliability and literature-based and 60% for gross variability. Smallest worthwhile change (SWC) was calculated with the SD of the pre-test CMJ. RESULTS: Reliability at zero threshold showed successful changes for JH (46.9% improve), PF/BM (28.1% improve, 6.3% worse), and RSImod (53.1% improve, 3.1% worse). Reliability responses at SWC threshold lowered for JH (18.8% improve), PF/BM (25% improve, 6.3% worse), and RSImod (25% improve, 3.1% worse). Literature-based at zero threshold showed successful changes for JH (18.8% improve), PF/BM (31.3% improve, 6.3% worse), and RSImod (68.8% improve, 3.1% worse). Literature-based responses at SWC threshold lowered for JH (12.5% improve), PF/BM (28.1% improve, 6.3% worse), and RSImod (37.5% improve, 3.1% worse). Despite the expected overall change, gross variability showed similar responses to reliability, with only RSImod zero threshold being lower (46.9% improve), and JH at SWC threshold being higher (21.9% improve). CONCLUSIONS: These methods yielded different individual response estimations; thus, we recommend using them according to their strengths. The reliability method may be useful when a control group is unavailable, and test-retest is practical. The literature-based method is plausible when CoVs are low and experimental conditions are close to those reported. Gross variability may be comparable at a lower CI, accepting that it cannot separate random error from the true score.