KINE 2049 Lecture 16: Experimental Research
Document Summary
Measurements for things that are easy to measure (e. g. jump distance) are more reliable. Some error is introduced anytime a measurement is made. Any obtained score is composed of true score component and an error component: o o o o o o. Examples of measurement errors that can reduce reliability: One of main differences between reliability coefficients is interpretation of reported reliability. Must have scores from 2 or more trials. Conditions of experiment when scores were obtained and time between trials brings some more error components so you must use various methods of estimating reliability. Value of the component will be b/w 0. 0 and 1. 00, 0. 0 no reliability and. Generally 0. 80 and above is desired but depends on the study. Correlation coefficient is computed between results of two sets of scored from a test being administered twice to the same group of subjects. High correlation is possible if their scores remain consistent from one set of results to the next.