ERIC Number: ED194630
Record Type: RIE
Publication Date: 1980-Apr
Reference Count: 0
The Power of An Experiment: Potential Vs. Actual.
A relationship between experimental and statistical controls is considered in an effort to show how this relationship affects the power of an experiment. Statistical rather than experimental controls are often used to reduce the effect of subjects at the extremes of individual difference. This results in attrition of actual experimental subjects. Questionnaire data from two experiments was analyzed for subject attrition, although they were not specifically designed for this purpose. In an experiment on cues and notetaking using 54 graduate students as subjects, there was a failure to find two strong expected treatment effects (cues over no cues and notes over no notes). In the second experiment on lecture vs. reading with and without notes using 58 college juniors and seniors, the treatment group which heard the lecture and took notes had 10 non-subjects and only six subjects. In both experiments, substantial power loss from subject attrition was found. Recommendations are made for conducting experiments and summarizing research. Inadequate actual power of experiments reporting no difference may be a major reason for pessimistic research summaries. (Author/RL)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the American Educational Research Association (64th, Boston, MA, April 7-11, 1980).