Greenwood press

heart attack
no heart attack
TOTAL
aspirin
104
10,933
11,037
placebo
189
10,845
11,034
TOTAL
293
21,778
22,071
The results from the aspirin–physician heart study.
Statisticians assume that the medication and heart attacks are independent. If
so, then P (aspirin and heart attack) = P (aspirin) • P (heart attack). Using the rel-
ative frequencies from the total row and total column gives the following prod-
uct:
11037
22071
•
293
22071
. That probability times the number of participants tells how 
many doctors receiving aspirin would have had heart attacks if heart condition
were independent of medication. That frequency is 147. As you can see from the
table below, the almost equal separation of physicians into aspirin and placebo
treatments indicates that the expected values for heart attacks should have been
separated into almost equal proportions.