Correlations, Sample Size, and Practical Significance: A Comparison of Selected Psychological and Medical Investigations

Abstract A common interpretation of correlation coefficients is that values of r that are less than .30 are considered “low.” Correlations at about this value are quite common in psychological studies, and this situation often leads to questions concerning the validity of the findings. In this article, some comparative correlational findings in medical research are presented, showing that many important medical and societal decisions have been made on the basis of similar or lower calculated correlation coefficients. The influence of sample size and issues of practical significance are also discussed.