5°

EMPLOYMENT  PSYCHOLOGY

had  been  applied,  seventeen  were  definitely  unsatisfactory,
while  nine  were  not  described.  And  of  those  above  the
standard  who  would  have  been  hired,  thirty-four  were
definitely  described  as  satisfactory,  only  one  as  unsatisfactory, ­
  and  thirty-two  could  not  be  designated.  These
results  were  such  as  to  vindicate  the  application  of  the
tests,  even  at  a  time  when  workers  were  urgently  demanded. ­

Later  on,  however,  it  was  possible  to  make  a  much  more
satisfactory  comparison,  based  upon  the  actual  production ­
  or  output  record  of  the  girls.  Forty-three  inspectors
who  had  succeeded  or  almost  succeeded  in  becoming  pieceworkers, ­
  and  who  had  been  at  work  for  two  months  or
more,  were  considered  in  this  comparison.  The  hourly
production  of  each  girl  was  averaged  for  a  period  extending ­
  from  one  to  four  months.  In  this  way  a  consistent
production  record  was  obtained  for  each  girl,  one  which
expressed  the  dominant  characteristics  of  the  girl  rather
than  her  particular  ability  on  a  particular  day.  The  ranking ­
  of  the  girls  in  production  was  then  compared  with
their  rankings  in  the  cancellation  and  number  group
checking  tests,  tests  number  8  and  6.  (Test  number
2,  the  card  sorting  test,  had  not  been  given  because  of
the  length  of  time  it  required  and  because  of  the  large
number  of  girls  who  were  applying  for  work.)  The  correlation ­
  for  these  tests  was  found  to  be  very  good,  being
forty-four  for  number  8,  forty-seven  for  number  6,  and
fifty-eight  for  both  tests  when  taken  together.  These
correspondences  are  not  as  high  as  some  of  those  found  in
the  first  experiment,  but  it  will  be  remembered  that  at  the
time  of  this  comparison  a  good  many  extremes  had  already
been  eliminated  by  their  own  failure.
The  striking  thing  about  this  comparison,  however,  is