|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 32
6
The Referent Cohort
To detains whether participants at the five nuclear test series had differ-
ent mortality experience than a comparable group of nonparticipants, we built a
comparison referent cohort of nonparticipants. Using records kept by the De-
partment of Defense and the National Archives and Records Administration, we
used frequency matching to assemble a referent cohort that would be similar to
the Defense Threat Reduction Agency (DTRA)-provided five-series participant
cohort according to branch of service, time period, location (Pacific, western
United States, other), age, type of unit, and paygrade. We did this by creating a
pool of units such as ships and battalions from among those selected by
DTRA as likely to be comparable. DTRA selected reference units by consider-
ing their similarity to the participating units. The similarity was defined by
function, size, paygrade distribution, and time period. Units assigned to states
downwind of the Nevada Test Site, to operations in Korea, or to participation in
any atmospheric nuclear test were not eligible for selection.
From the eligible units, we selected individuals to fit the paygrade distribu-
tion of the participants in the related unit. Those participants without reasonably
close referents in terms of paygrade were pooled by type of unit, within branch
of service and series, as were the excess individuals in referent cohort units who
were not selected. Individuals in each service were selected from the larger ref-
erent cohort pools to be similar in paygrade and unit type to the individuals re-
maining in the participant pool.
MFUA, with DTRA assistance, assembled a 64,781-member military refer-
ence cohort. Reference individuals were selected using frequency matching on the
following criteria: (1) service during the 12 months immediately preceding or fol-
lowing the date of the participant's selection series, (2) service in a similar unit in
the same branch of service as the participant during the test period, (3) the same or
32
OCR for page 33
THE REFERENT COHORT
33
similar paygrade at the time as the participant during the test period, and (4) no
participation in any atmospheric nuclear weapons testing program.
Since no single source document can provide all of the information neces-
sary for assembling the referent cohort, the assembly procedure was divided into
three phases. The first phase involved selecting reference units; the second phase
involved building a referent pool by identifying names, service numbers, and
paygrades of all individuals in the units; and the third phase involved selecting
individual referents and further obtaining identifying information concerning
those individuals.
The NTPR team identified referent units through a review of Station Lists,
which specify all units according to their numerical designation in each calendar
year; these unit specifications can be cross-referenced with the unit's physical
location.
The similarity between the participant and referent units was determined by
considering their function, size, and paygrade distribution. Since unit names are
usually consistent and can provide a basic understanding of these characteristics
within the unit, the reference unit selection was based on unit names. The geo-
graphic area of the station was also considered in selecting reference units. Units
stationed in Utah, Arizona, New Mexico, Colorado, and Nevada within 2 years
of any atmospheric continental test were excluded from referent unit selection
since they may have been exposed to test fallout. Units stationed in Korea dur-
ing the Korean War (1950-1953) were also excluded. Units participating in any
atmospheric nuclear weapon tests within a defined time period were excluded as
well: for the Army and Marine Corps, units within a 2-year window of any test
period; for the Air Force, units within a 3-year window; and for the Navy, units
within a 4-year window. These time frames were chosen based on typical unit
rotation periods within the services.
More than one reference unit was selected for each participating unit: two
units for each participating unit in the Navy and Marine Corps; six units for each
participating unit in the Army and Air Force. The principal purpose of obtaining
multiple reference units was to provide a referent population pool large enough
for frequency sampling on branch, series, and paygrade. These multiple units
were ordered and sampled according first to geography and then to time, as indi-
cated below. For continental tests, units stationed within the continental United
States had a higher priority than those stationed outside the continental United
States. For Pacific test series, similar units stationed within the Pacific theater
had a higher priority than those outside the Pacific theater. Units stationed
within the 6-month window of the test series period had a higher priority than
those within the 12-month window. Therefore, for the continental tests, for ex-
ample, units stationed in the continental United States within the 6-month win-
dow had the highest priority, followed by those stationed in the continental
United States within the 12-month window.
The degree of difficulty in identifying reference units ranged from minimal
to extreme. Finding reference units such as ships, battalions, and standard squad-
rons was relatively simple. However, finding counterparts of temporary units
OCR for page 34
34
THE FIVE SERIES STUDY
such as provisional, special project, and observer units, was difficult. The struc-
ture of these temporary units did not follow the established standards (e.g., ta-
bles of distribution and allowances, tables of organization and equipment);
therefore, the unit names do not provide a basic description of their size, func-
tion, and paygrade distribution.
Once the reference units had been chosen, organizational records of each
branch of service were reviewed to identify the names, service numbers, and
paygrades of all personnel in these units. For the Navy, all enlisted men aboard a
particular ship can be found in the ship's muster rolls (on microfilm); officers
are listed in the ship's deck logs (in log books) and on post-1955 muster rolls;
both sources are available through the National Archives. For Navy shore units,
which do not have muster rolls and deck logs, unit diaries were reviewed to as-
certain the identifiers of the unit members. For the Marine Corps, muster rolls
(on microfilm) and Station Lists were searched to accurately identify military
service numbers or names.
For the Army, monthly personnel rosters and morning reports that are avail-
able in the National Personnel Records Center (NPRC) in St. Louis were used.
The morning reports have been completed each day since World War II, usually
at the company level. They list persons who experienced a change in duty status,
showing their names, service numbers, and ranks. A change in status could be
discharge, temporary duty, absence, return from absence, reassignment, promo-
tion, etc. Since the morning reports may list the same person multiple times and
omit certain persons depending on their changes in duty status, the monthly per-
sonnel roster was used as the primary information source. For the Air Force,
morning reports are the only available source.
Once the referent pool was constructed for each service, the roster was
matched to the NTPR database on name and service numbers, to exclude those
who are known to be participants in any atmospheric nuclear weapons test.
The members of the referent pool were grouped by unit and paygrade, as
were members of the participating units. For each paygrade, the same number of
reference subjects as participants was selected according to the unit priority or-
der described earlier. When there were insufficient numbers of referent pool
members in a specific paygrade, an adjacent paygrade was used.
Table 6-1 shows the closeness of matching. Selection categories 1 and 2 in
which service, series, paygrade, and type of unit are all exact matches account
for 79.5 percent of the study population. Another 15.5 percent was selected using
one of six close, although not exact, combinations of characteristics outlined in the
table. The pool of potential referent personnel did not include, however, sufficient
numbers of certain participant characteristic combinations for there to be equal
numbers in each cohort. Each combination, however, is represented in both co-
horts. The referent cohort has 3,388 fewer members than the participant cohort.
The referent cohort acquisition process yielded a group of individuals with
distributions similar to the participant cohort for the desired and available char-
acteristics. This balance within the overall and series-specific cohorts is illus-
trated in Chapter 10.
OCR for page 35
35
~ o c t4
.~ 'e ~
o ~ ~ ~ ~o
an
a' ~ ~ O
x
To v)
v
o
o
.~
c)
ad
A
ad
~ -
c)
cd
at
at
At
o
at
at
at
o
v
- ~
'e
u,
~ -
v)
at
.-
A
at
;.-
~ o~
'e
o v
A,
~ .-
o at
id ~
to
o'
z
o
a,
c)
.- .
o o
as 'e
v) To
*
0
o m ~ ~
~0 E 3 ~ ~ c
~,, ~i j~24
~ ~O ~ ~ O 0 0
x xx ~x x _ ~
~ ~ ~v v
o~ ~ == ~
0 ~ O ° ~ 0
z ~ ~ ~ ~ ~
~ ~x ~
*~ ~4.~ }
c~ ~c ~c' ~ c ~c' ~
ctct ct ct ct ct ct
x xx x x x x x
(~ d" ~- ~O
O ~] ~ O O OO
°
_
~ ~ O ~00
O ~ ~_
_ ~
oo ~_ ~
~_
~C~
_' O
_ ~- , ~ ~oo
--
=0
._
.
t4 S:
^:O
-= ~
~ C;:
S:
Ct ,=
_
·_ ._
U~ C~
Cd
s~ ~
~ Ct
4-
C~
._
~ O
O
S:~ ~
·= C)
C)
O
·< O
O ~
CD
a, 4_
C~
=0
. - ~
~ ,s:
.= ~
O.=
~q
U, ~
·- C.)
_
. C)
U) ~
.9
s~ U~
(4-o, 'e
Ct
C) ~
X Ci,
*¢ =O
Representative terms from entire chapter:
reference units
