ABSTRACT
Responses in the concurrent control represent the background state of the experimental units – including possibly a vehicle control or sham exposure. Inclusion of the concurrent control allows for proper comparability among the study groups. Pairwise comparisons of any single treatment with the control remain the basis for assessing differences from the control group, but adjustments and corrections for multiplicity are presented that allow the investigator to control the familywise false positive rate within the family of multiple comparisons being performed. Emphasis is placed on multiple comparisons with a single, concurrent control group, but extensions to other forms of multiple comparisons also are noted. Specific questions can include whether any or all of the various treatment groups differ significantly from the concurrent control group, or whether significant differences exist in the responses among adjacent treatment groups, etc.
