54

WEAPONS
i’ORLDWIDE EFFECTS OF ATOMIC

e collection. Such pettravels to the foreign place to supervise the sampl
ency of sample collection
sonal contact will undoubtedly improve the effici
for the domestic cases, too.

size of the specimens
Nothing has yet been said about the probable
ing program outlined
required, statistical factors that relate to the sampl
ts are discussed in the
here, or the laboratory techniques. These subjec
following sections of this chapter.
Statistical Considerations

a pilot sample survey have
The general area of study and the outline of
carrying out of this pilot
been presented in the previous section. The
that will have to be solved
sample survey poses several statistical problems
by the survey. Basically the
before satisfactory results can be obtained
make to sampling problems:
science of statistics has two contributions to
individual observations
1. To enjoin randomness in the selection of

ss so as to ensure against
at various points in the sampling proce
bias in the estimates obtained;

standard error) of the
> ‘To increase the accuracy (minimize the

average lifetime dose
estimates of the required quantities (e.g.,

g advantage of the
given some steady-state situation) by takin
n of the collection
desig
structure of the underlying process in the
and in the analysis of the sample data.

come into play when the detailed
The first of these contributions will
collection of the various samples
prescriptions are written out for the
fact the description of the samples
described in the preceding section. In

before any sampling can be
themselves will have to be greatly expanded
nd
y is a pilot-type program, the seco
done. Since the proposed sample surve

much into prominence. Pilot procontribution of statistics cannot come

,
SPN ey ehranipar erianbermpaebeci i eisSCUCO
ned, using the
iously desig
larger sampling program can be more judic

r (2) above. At present we know
considerations that are subsumed unde

so that there is not much
very little about the processes we are studying,
ntage of now. Something, of
structure in the problem to be taken adva
the kind of analysis one might
course, can be done, and an example of

SAMPLING AND RESEARCH PROGRAM

55

bed in the following
attempt in the case of the human samplesis descri
paragraphs.

A Sample Analysis of Human Data
geographic disAll hazard calculations will have to be specific to some
from various
comes
r,
howeve
tribution, The food supply of a given city,
itself, but in most cases
regions, some of them quite distant from the city

ed. In any case
the water, milk, and fresh vegetables are locally suppli
analysis will have
differentcities have different food-supply areas and any
is to be done.
to take this into account. The only question is how this
e on which an
schem
of
The following is a crude model of the sori
exposure function
analysis may be built. The model has two elements: an

any person, the age and
E(t) and an ingestion function a(t), in which, tor

to it by transall chronological times for the community are normalized
Sr’? available
of
t
amoun
the
Jation in the time axis. By exposure is meant
of the indit
onmen
envir
in the total environmentof the individual. The
, for
Kansas
es
vidual no matter where he lives in this cou try includ
. The dose of an
example, if the wheat in his diet comes from Kansas

individual at age T ts

D(T) = fEa) at.
variables are to
The real problem for even the crudest analysis is what
The following
them.
enter into E(t) and a(t) and how are we to measure
ted in the bones of
is suggested as a first attempt to relate the Sr°° deposi

ure funchuman beings to somecalculated exposure. Let E(t), the expos

distribution of fallout
tion, be determined by (1) the geographic and time
study, and (3) the
under
of Sr°°, (2) the food sources of the area (city)

the fallout is made
gross scavenging process of nature by meansof which
7 per unit area
nonavailable. Denote by X(r) the fallout of Sr°° at time
be assumed to
over an appropriate area; the exposure E(t) might then
be equal to

E(t) =f X(z) exp [—B(# — 7) 1 4r,

ter. As a
where f represents a generalized decay and scavenging parame

Select target paragraph3