Maximal strength measurement: A critical evaluation of common methods—a narrative review

Measuring maximal strength (MSt) is a very common performance diagnoses, especially in elite and competitive sports. The most popular procedure in test batteries is to test the one repetition maximum (1RM). Since testing maximum dynamic strength is very time consuming, it often suggested to use isometric testing conditions instead. This suggestion is based on the assumption that the high Pearson correlation coefficients of r ≥ 0.7 between isometric and dynamic conditions indicate that both tests would provide similar measures of MSt. However, calculating r provides information about the relationship between two parameters, but does not provide any statement about the agreement or concordance of two testing procedures. Hence, to assess replaceability, the concordance correlation coefficient (ρ_c) and the Bland-Altman analysis including the mean absolute error (MAE) and the mean absolute percentage error (MAPE) seem to be more appropriate. Therefore, an exemplary model based on r = 0.55 showed ρ_c = 0.53, A MAE of 413.58 N and a MAPE = 23.6% with a range of −1,000–800 N within 95% Confidence interval (95%CI), while r = 0.7 and 0.92 showed ρ_c = 0.68 with a MAE = 304.51N/MAPE = 17.4% with a range of −750 N–600 N within a 95% CI and ρ_c = 0.9 with a MAE = 139.99/MAPE = 7.1% with a range of −200–450 N within a 95% CI, respectively. This model illustrates the limited validity of correlation coefficients to evaluate the replaceability of two testing procedures. Interpretation and classification of ρ_c, MAE and MAPE seem to depend on expected changes of the measured parameter. A MAPE of about 17% between two testing procedures can be assumed to be intolerably high.