A study in which the majority of the subjects die calls into question whether the treatment itself is safe.
b. Subjects who drop out may differ from those who stay in the study, in terms of an important extraneous variable.
c. If subjects drop out of the control group, and not the experimental group, it strongly implies that there is some benefit to participation of which the researchers may not be aware.
d. Subject mortality may result in a sample that is so much smaller than anticipated that type II error may result.
e. Type I error is almost guaranteed with very unequal sample sizes.
