To the Editor.—
Tjio et al (210:849, 1969) have made an important methodological advance in the study of the effects of lysergic acid diethylamide (LSD) on human chromosomes in vivo by measuring chromosome aberration rates in subjects both before and after LSD administration. The authors were unable to conclude that LSD had a harmful effect. Using more powerful statistical methods, we conclude that LSD has a significantly harmful effect in both dose groups.
In testing hypotheses, we have strict control of the (type 1) error of rejecting the null hypothesis when true and little control of the (type 2) error of not rejecting the null hypothesis when false. To minimize the probability of failing to detect a harmful effect and to ensure valid conclusions, one-tailed tests are appropriate.1
From the data in Table 2 of the communication, the number of normal and aberrant cells in each sample can
. . . [Full Text PDF of this Article]
