THE  MEASURE  OF  COMPARATIVE  PRODUCTIVENESS  317

Standing  in

Production  in  terms

Attendance

Men

Tests

of  t  earned  per  hr.

%

A

98

70

97

B

96

69

95

C

95

68

80

D

92

66

99

E

87

64

93

F

81

63

90

G

76

6l

92

H

69

59

99

I

62

55

84

J

53

53

87

IO

809

628

916

Av.  80.9

Group  Av.  62.8

Av.  91.6

It  is  evident  from  the  above  figures  that  there  is  a
satisfactory  degree  of  correlation  between  the  records  in
the  tests  and  in  production.  The  men  who  were  highest
in  the  tests  were,  on  the  whole,  highest  in  production,
and  vice  versa.  If  we  apply  the  formula,  familiar  to
statisticians,  and  explained  in  Chapter  II,  we  would
obtain  plus  .60  as  the  index  of  correlation,  and  this  is  a
rather  high  index.  If  there  were  no  close  correspondence
between  the  test  and  production  records,  the  correlation
would  be  expressed  in  a  much  lower  figure,  for  example,
plus  .08  or  even  a  minus  quantity  such  as  minus  .62.
Thus  it  becomes  possible  to  determine  with  mathematical
exactness  just  what  degree  of  correspondence  exists  between ­
  these  two  factors,  and  to  tell  what  the  success  of
the  selections  has  been.
If,  now,  it  becomes  necessary  to  lay  men  off  for  lack  of
work,  unsuitability,  or  any  other  reason,  the  employment
manager  can  determine,  by  referring  to  the  employee’s
activity  record,  whether  or  not  the  right  men  are  being
laid  off.  For  example,  let  us  suppose  that  A,  C  and  J  on