Effects of method of subject selection (volunteer vs random) and treatment variable on motor performance

Arnett, B.; Rikli, R.

Possible interaction effects between method of subject (human) selection (volunteer vs. random) and a selected treatment variable (verbal encouragment) were tested. The presence of such an interaction would suggest that generalizations regarding verbal encouragement could be made only to specific subject populations. Volunteer (10) and randomly selected (40) subjects were randomly assigned to either a verbal encouragement condition or to a control group with no verbal encouragement. Dependent variables were a tapping task and a leg lift task. It was predicted that volunteer subjects, due to their higher need for approval and greater desire to accommodate, would perform significantly better under the verbal encouragement condition than would the randomly selected subjects. Multivariate and univariate data analyses indicated that the prediction was supported on the tapping task (female subjects only) but not on the leg lift task. At least on certain tasks, use of volunteer subjects may result in a selection bias .times. treatment interaction effect that could limit the generalizability of research results.