Developing probability limits for longitudinal physiological data

Abstract

The objective of this study was to determine what amount of change in heart electrophysiology is necessary to yield a significant difference between a baseline condition and an experimental treatment condition. In a single-subject longitudinal design, a modification of a t test was applied to baseline data, and the performance levels for both .05 and .01 probability levels were obtained. It was concluded that it is both ethically sound and physiologically safe to permit a subject in this type of experiment to physiologically respond at levels that would reach both the .05 and the .01 levels of significance. © 1987 Psychonomic Society, Inc.