Summarizing the results of multiple strength trials: Truth or consequence

The purposes of this study were: (1) to examine whether the distribution of repeated within subject strength measurements obtained under theoretical no change conditions, are consistent with that of a normal distribution; and (2) to examine the consequences of using the maximum, compared with the average trial value, as a criterion measure of strength. Four studies were undertaken to examine grip strength, knee flexor and extensor strength, shoulder abductor strength, and hip extensor strength. A deliberate aspect of the design was to select measurement systems that did and did not require a rater to be actively involved with the measurement process. Also, one of the designs allowed the comparison of intra and inter-rater measurements. Procedural concerns, such as warm-up and rest intervals, were addressed in order to achieve a stable state. The results were consistent with the hypothesis that within subject strength measurements are normally distributed. This finding would sustain the premise that measurement error is normally distributed and lend support to the practice of averaging stable trials to increase the reliability of a particular strength measurement. From a practical perspective it appears that using the average value becomes more attractive (because of the increased time required) as the magnitude of the reliability coefficient diminishes.