Reliability and Validity of Predicted Performance in the Severe-Intensity Domain From the 3-Minute All-Out Running...

The aim of this study was to analyze the reliability and validity of the predicted distance-time relationship in the severe-intensity domain from a 3-minute all-out running test (3MT).

Twelve runners performed two 3MTs (test #1 and test #2) on an outdoor 400-m track after familiarization. Eighteen-hertz Global Positioning System data were used to estimate critical speed (CS) and distance covered above CS (D'). Time to cover 1200 and 3600 m (T1200 and T3600, respectively) was predicted using CS and D' estimates from each 3MT. Eight runners performed 2 time trials in a single visit to assess real T1200 and T3600. Intraclass correlation coefficients (ICCs) and standard errors of measurement were calculated for reliability analysis.

Good to excellent reliability was found for CS, T1200, and T3600 estimates from 3MT (ICC > .95, standard error of measurement between 1.3% and 2.2%), and poor reliability was found for D' (ICC = .55, standard error of measurement = 27%). Predictions from 3MT were significantly correlated to actual T1200 (r = .87 and .85 for test #1 and test #2, respectively) and T3600 (r = .91 and .82 for test #1 and test #2, respectively). The calculation of error prediction showed a systematic error between predicted and real T3600 (6.4% and 7.8% for test #1 and test #2, respectively, P < .01) contrary to T1200 (P > .1). Random error was between 4.4% and 6.1% for both distances.

Despite low reliability of D', 3MT yielded a reliable predicted distance-time relationship allowing repeated measures to evidence change with training adaptation. However, caution should be taken with prediction of performance potential of a single individual because of substantial random error and significant underestimation of T3600.