This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
As Reed and Slaichert point out in an article in this issue (see p 1307), two types of errors may be made when an experiment is designed and carried out that involves the testing of a hypothesis. One may erroneously reject a true hypothesis, or one may equally erroneously accept a false one. In most studies using this design, the type I error (the rejection of a true hypothesis) is usually discussed and limited to a risk of 5% or 1%. Few researchers in the medical field ever look at or discuss the probability of making a type II error (the acceptance of a false hypothesis). For example, if one is testing the hypothesis that drug A is equal to drug B, in most cases, the researcher would simply perform an experiment, and if the difference observed between the two drugs was "statistically significant" at some preassigned level of risk,
Schor S. Statistical Proof in Inconclusive 'Negative' Trials. Arch Intern Med. 1981;141(10):1263-1264. doi:10.1001/archinte.1981.00340100019005