from the 1973 National Health Interview (a sample of the U.S. popula-
tion) were used to examine the relationship between Pap testing and
some socioeconomic factors. Table E1.21 provides the percentages of
women who reported never having had a Pap test. (These are from met-
ropolitan areas.)TABLE E1.21
Age and Income White Black
25–44
Poor 13.0 14.2
Nonpoor 5.9 6.3
45–64
Poor 30.2 33.3
Nonpoor 13.2 23.3
65 and over
Poor 47.4 51.5
Nonpoor 36.9 47.4(a)Calculate the odds ratios associated with race (black versus white)
among
(i)25–44 nonpoor
(ii)45–64 nonpoor
(iii) 65 þnonpoor
Briefly discuss a possible e¤ect modification, if any.
(b)Calculate the odds ratios associated with income (poor versus non-
poor) among
(i)25–44 black
(ii)45–64 black
(iii) 65 þblack
Briefly discuss a possible e¤ect modification, if any.
(c)Calculate the odds ratios associated with race (black versus white)
among
(i) 65 þpoor
(ii) 65 þnonpoor
Briefly discuss a possible e¤ect modification.1.22 Since incidence rates of most cancers rise with age, this must always be
considered a confounder. Stratified data for an unmatched case–control
study are given in Table E1.22. The disease was esophageal cancer
among men and the risk factor was alcohol consumption.
(a)Calculate separately for the three age groups the odds ratio asso-
ciated withhighalcohol consumption.
42 DESCRIPTIVE METHODS FOR CATEGORICAL DATA