Data from IVY gummed paper duplicate samples were compared to
a selection of pairs from UPSHOT-KNOTIDLE. Data for each operation
were also divided according to area; that is, duplicates for North
America were compared to duplicates outside of North America for

_ the IVY tests while the UPSHOT-KNOTHDLE data (which was available

Duplicate samples were used in this

_study only if each paper had at least 100 d/m on counting day.

ot be

The question to be answered, "Is the nature of the fallout

such that the differences between duplicate gummed papers varies

from one part of the world to another » OF from one test series to
another?"

.

The difference between adjacent gumed papers is described,
in this study, by a "“variabllity characteristic"” end the differences
*the “variability characteristic" issimply the arithmetic mean of

the individual Chi-squarad values (“y2) for each comparison group,
i.e., X? was computed for, each pair of gummed papers and then
averaged with the other y*

of its particular group, e.g., UPSHOT-

|

- KMOTHOLE, east of the Mississippi River.
.

d2

ed!

an M

= Iho

hae

for North America only) was divided according to location west or

east of the Mississippi River.

where M = mean of two simultaneous
papers

qe departure of first paper
‘of pair from M .
do= departure of second paper

“

of pair from M

. (Therefore, +d, ™ -d,)

- thus,
°

4

ab L

N ist

(x*),

t

where N= tote) number of pairs in

the "j-th" group

(x®), = value of Chi-square for the

i-th pair of the j-th group

The values of

x?

are different from group to group and the signi-

ficance of their difference is examined by use of the standard -

t-test for significance, of Gifference of means.

- 43 -

Select target paragraph3