32

EMPLOYMENT  PSYCHOLOGY

Table  of  Ranks  and  Resulting  Correlations

Name  of

Rank  in

Rank  in

Rank  in

Girl

Production

First  Test

Second  Test

A.  .  .

I

I  .  .  .  .

8

B...

5

C...

3

3  •  •  •  •

3

D...

4

4-••  •

2

E.  .  .

s

S  ---4



F.  .  .

6

6....

7

G...

7

7-••  ■

9

H...

8

8....

I

I....

9

9-••  •

IO

J-...

IO....

6

Correlations.  i.  oo

-19
per  hour  is  ranked  as  number  one,  the  one  who  turns  out
next  most  as  number  two,  and  so  on  down  the  line  to
number  ten.  Every  girl  having  been  ranked  on  the  basis
of  output  from  1-10,  the  next  step  is  to  take  one  of  the
tests  in  question  and  rank  each  of  the  ten  girls  in  the  same
fashion  according  to  her  work  in  that  test.  We  are  now
ready  to  compare  the  output  of  the  girls  with  their  rank
in  the  test.  For  the  sake  of  explanation,  let  us  suppose
that  the  first  girl  in  production  was  first  in  the  test;  the
second  in  production,  second  in  the  test;  the  third  in
production,  third  in  the  test,  etc.,  right  through  the  group,
the  tenth  or  lowest  in  output  being  tenth  or  lowest  in  the
test.  It  is  evident  from  this  that  we  should  have  a  perfect
agreement  or  correlation  between  the  performance  of  the
girls  in  the  test  and  their  performance  in  their  regular
work.
Now,  let  us  suppose  that,  in  another  test,  there  was  not
such  an  agreement.  Suppose  that  the  girl  who  was  high ­