EMPLOYMENT  PSYCHOLOGY

II 4

again.  The  experimenter  believed  that  an  operator  could
do  this  more  rapidly  and  successfully  at  the  beginning  of
a  day  when  fresh,  than  at  the  end  of  a  day  when  fatigue
had  set  in.  He  further  believed  that  if  this  could  be  demonstrated, ­
  it  would  also  follow  that  machine  operators
were  more  liable  to  accidents  toward  the  end  of  the  day
than  in  the  morning,  for  the  same  reason.
The  manner  in  which  this  experiment  was  conducted
and  the  merits  of  its  results  are  not  of  particular  interest
here.  However,  the  apparatus  seemed  very  promising
for  the  purpose  of  separating  the  slower  from  the  faster
operators.  A  similar  apparatus  was  therefore  constructed
and  tried  out.  To  begin  with,  tests  were  given  to  twenty
operators  in  order  to  familiarize  the  experimenter  with  the
use  of  the  apparatus,  the  giving  of  directions  for  it,  and
scoring  the  results.  As  a  result  of  these  preliminary  tests,
certain  difficulties  were  discovered.  First,  the  construction ­
  of  the  apparatus  was  such  as  to  make  accuracy  of
performance  impossible.  For  instance,  each  time  the  arm
struck  the  cube  on  the  square  it  would  knock  it  off  onto  a
different  part  of  the  table.  This  condition  became  worse
as  the  speed  of  the  rotating  arms  was  increased,  and  frequently ­
  the  cube  would  be  knocked  entirely  off  the  table
and  would  have  to  be  picked  up  from  the  floor.  This
factor  might  have  been  a  desirable  one,  and  a  measure
of  the  operator’s  resourcefulness,  had  it  not  been  for
a  second  difficulty.  It  was  practically  impossible  to  score
the  results.  The  operator  hardly  ever  placed  the  cube
exactly  on  the  painted  square,  but  each  time  the  angle
or  degree  of  displacement  was  different,  making  it  extremely ­
  difficult  for  the  experimenter  to  make  any  score
on  the  mistakes  of  the  operator.  Even  if  the  observer
had  been  able  to  estimate  the  misplacement  in  every