|
Now consider
a more interesting case and suppose that the situation is as above
but that in addition there is a secular downward trend of three
between periods so that values in the second period are three
lower than they would have been otherwise. The table of expected
values then looks like this.
| Table
2.3: AB/BA Cross-over trial comparing a verum and a
placebo
|
|---|
| Sequence
| Period
1
| Period
2
| Difference
|
|---|
| PV | 6 | 7 | 1 |
| VP | 15 | 8 | 7 |
|
Mean | | | 4 |
|
All that
has happened is that three has been subtracted from each of the
entries in the previous table under period 2. Now the expected
difference within each sequence is no longer four.
Question
2.2
Does
this period effect affect our estimate of the treatment effect?
 
|