33 2 
EMPLOYMENT PSYCHOLOGY 
moments, may do much to counteract this hastiness and 
to promote a more just result. The same check will 
operate on opinions that are suddenly inclined to become 
too favorable. 
On the whole, then, it may be accepted that the impres 
sionistic method as outlined, while not free from the errors 
of the personal equation, is nevertheless free from these 
errors in their extremer forms. To recapitulate briefly, 
the liability of error is lessened for the following reasons: 
i. Because a periodic rather than a sporadic expression 
of opinion is required. 2. Because opinions are limited 
to a fixed number of qualities. 3. Because those qualities 
are defined in the same way for all. 4. Because it becomes 
necessary for the appraiser to check off the various quali 
ties against each other, thus producing a balanced result. 
5. Because the judgment goes on record for or against 
the man who makes it. 6. Because the judgment of one 
appraiser will serve as a check against that of another. 
Like the method of comparative productiveness, the 
method described has two general applications, the sta 
tistical and the corrective. In the first place, it offers a 
basis upon which to compute the degree of correlation 
between selection and retention. Unless such a record 
is kept, it can not be ascertained with any degree of cer 
tainty how successful the selections of the employment 
manager are. As has been repeatedly pointed out, the 
dismissal, resignation, or transfer of an employee is by no 
means a proof that he was poorly selected. Unless some 
other criterion of success exists, these incidents may be 
attributed to a hundred different reasons. The record il 
lustrated will furnish at least an approximate means for 
judging the success of selections,—or, it may be, forjudging 
the success of the supervisor in handling his subordinates.