w 
122 EMPLOYMENT PSYCHOLOGY 
came into the employment office and those who finally 
made good as operators. Secondly, an even finer dis 
tinction was made between the groups of operators who 
had survived at fast machines and those who were working 
at slow machines. Thirdly, in every group tested, whether 
slow or fast, there was a significant correlation between 
piece-work earnings and the performance in the test. 
On the basis of these conclusions, therefore, it was decided 
to use this test in the employment office in order, first, 
to eliminate the slowest and most clumsy operators, and 
secondly, to classify the better applicants for machine 
work of this kind into two groups, those for work on fast 
machines, and those for work on slower machines. This is 
now being done.