and menstrual characteristics, exogenous estrogen use, and prior pelvic
surgeries (Table E6.28). For each of the two criteria, duration of un-
protected intercourse and history of infertility, test to investigate the
relationship between that criterion and ovarian cancer. State clearly your
null and alternative hypotheses and choice of test size.6.29 Table E6.29 lists data from di¤erent studies designed to investigate the
accuracy of death certificates. The results of 5373 autopsies were com-
pared to the causes of death listed on the certificates. Test to confirm the
downward trend of accuracy over time. State clearly your null and
alternative hypotheses and choice of test size.
TABLE E6.29
Accurate Certificate
Date of Study Yes No Total
1955–1965 2040 694 2734
1970–1971 437 203 640
1975–1978 1128 599 1727
1980 121 151 2726.30 A study was conducted to ascertain factors that influence a physician’s
decision to transfuse a patient. A sample of 49 attending physicians was
selected. Each physician was asked a question concerning the frequency
with which an unnecessary transfusion was give because another physi-
cian suggested it. The same question was asked of a sample of 71 resi-
dents. The data are shown in Table E6.30. Test the null hypothesis of no
association, with attention to the natural ordering of the columns. State
clearly your alternative hypothesis and choice of test size.
TABLE E6.30
Frequency of Unnecessary Transfusion
Very
Type of Frequently Frequently Occasionally Rarely
Physician (1/week) (1/two weeks) (1/month) (1/two months) Never
Attending 1 1 3 31 13
Resident 2 13 28 23 5EXERCISES 245