|
If we
form totals over both periods for the patients in an AB/BA design,
these totals reflect the effect of both periods and both treatments.
They thus cannot differ between patients in any way that reflects
these effects. However, if there is carry-over, the order in which
treatments are given will affect these totals. We may thus compare
the totals between sequences to form a test of carry-over.
Question
3.1
How
does the power of the test for carry-over compare to the test
for the treatment effect?
|