Validity and reliability of the 3-minute all-out test in national and international competitive swimmers

Purpose: To assess the concurrent and predictive validity of the 3-minute all-out test (3MT) against conventional methods (CM) of determining critical speed (CS) and curvature constant (D)′ and to examine the test-retest reliability of the 3MT in highly trained swimmers. Methods: Thirteen highly trained swimmers (age 16 [2] y, weight 64.7 [8.5] kg, height 1.76 [0.07] m) completed 4 time trials and two 3MTs over 2 wk. The distance-time (DT) and speed-l/time (1/T) models were used to determine CS and D from 4 time trials. CS3MT and D0 3MT were determined as the mean speed in the final 30 s of 3MT and as the speed-time integral above CS, respectively. Results: CS3MT (1.33 [0.06] m·s ^-1) did not differ from CSCM (1.33 [0.06] m·s ^-1, P > .05) and correlated nearly perfectly with CSCM (r = .95, P < .0001). D0 3MT (19.50 [3.52] m) was lower than D0 DT (23.30 [6.24] m, P < .05) and D0 1=T (22.15 [5.75] m, P = .09). Correlations between D0 _3MT and D0 CM were very large (r = .79, P = .002). CS and D between the two 3MT trials were not different (CS mean change =-0.009 m·s ^-1, P = .102; D mean change = 0.82 m, P = .221). Correlations between the two 3MT trials were nearly perfect and very large for CS (r = .97) and D′ (r = .87, P < .05), respectively, with coefficients of variation of 0.9% for CS and 9.1% for D′. Conclusion: The 3MT is a valid protocol for estimation of CS and produces high test-retest reliability for CS and D in highly trained swimmers.