Automation of Instructions in Human Experiments

Abstract

This study examined the effect of automation of instructions on error rates in human experiments. Both repetition of important information (no repetition versus one repetition) and mode of presentation (live experimenter, videotape, audiotape, and audiotape plus written instructions) were studied using a between-subjects design. Data for seven different tasks indicated that instructions stated twice were more effective than instructions stating key information only once, but mode of presentation was not important for 165 college students.