WARREN S. BROWNER, M.D., M.P.H.; THOMAS B. NEWMAN, M.D., M.P.H.
To the editor: Both Simon (1) and Rothman (2) advocate the use of confidence intervals when reporting the results of scientific research. Although we agree with their basic recommendations, we want to sound a note of caution.
Contrary to common belief, determining the 95% confidence interval for a value does not imply that there is a 95% probability that the true value is contained within that interval. As an example, consider the following hypothetic experiment. Suppose an investigator performs a randomized, double-blinded trial comparing dextrose and sucrose in the treatment of 200 patients with malignant melanoma. The results show a
BROWNER WS, NEWMAN TB. Confidence Intervals. Ann Intern Med. ;105:973–974. doi: 10.7326/0003-4819-105-6-973_3
Download citation file:
Published: Ann Intern Med. 1986;105(6):973-974.
Copyright © 2018 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use