In choosing the final Subgroup sample sizes the statistician involved will need prior data on measurement variability and should be encouraged to apply U.S.
Moss et al
2.
(1969).
Planning study of health effects.
| There are several difficulties,
tion,
inherent in the situa-
which this planning phase is designed to meet.
Ologic methodology
for
countries
the
(e.g.,
in England,
Japan).
location of
Epidemi-
study of chronic disease has developed
quite rapidly since 1950
in
results such as
the U.S.
and a few other
Such methods have not been pursued
interest,
however.
And
in addition,
the
specific conditions of this study are not well suited to
epidemiologic field work.
Our collaborator's
and in
"monitoring"
interest
leads,
dotal medical records well
in the
health of
unfortunately,
residents
to the sort of anec-
known to result in serious bias,
which can be dangerously misleading.
It would be my hope
that by focussing substantial energy and attention,
funds,
Gotal
on planning a modern epidemiologic survey,
these anec-
collections can be either avoided or counterbalanced.
An example
study
and some
concerns
choice of controls
for a
health
(the control group in the PROPOSAL is adequate for a
study of contamination levels)
ning phase.
Step one
is
the choice of
parable
community.
Step two is
k Moss,
W.D.
(1969).
et al.
9009u3 I
ii
which can start in the plana
"matched"
or com-
recommended to be mere
Health Physics 17,
571-8.