DETERMINING OPTIMAL TRIAL SIZE USING SEQUENTIAL ANALYSIS

Abstract

This study examined the response of sequential analysis (SA) results to conditions of different trial size to ascertain the optimal number of trials for mean stability in kinematic data. Thirty overarm throws were performed by twenty participants. Discrete and time series kinematic data were submitted to SA in three main trial size conditions (first 10, 20, 30 trials) and three comparative conditions (mid and last 10, mid 20 trials) to derive the number of trials (SA score) to a stable mean for each condition. Results showed that the SA score is dependent on the trial size of the condition but not on the position where the subsample was drawn. As the first 10 condition responded differently to the other conditions, we recommend the use of the SA score of the 20 trial conditions to ascertain stable means for parsimonious data collection.