INTIL
STRY
; --~~
.ONVENTION
\->
UAL
N
MAY 31st, 1920
NEW YORK, N. Y.
May 31st, June 1st, 2nd, 3rd, 4th, 1920
COMMITTEE ON THE
APPLICATION OF PSYCHOLOGICAL TESTS AND
ADVANCE COPY
|-
G
2.
Z
þær!
Ç/)
©
3
Q/Q
CD
Z
}=e^{
±
<!
£4
-** -- ~~~
, ...”
OF CORPORATION SCHOOLS
EIGHTH AN,
NOT RELEASED FOR PUBLICATION
COPYRIGHT, 1920 -
THE NATIONAL ASSOCATION OF CORPORATION SCHOOLS
THE NATIONAL ASSOCIATION
：¿º，
ſae
¿№ №.

ſya， *******£
************#\*：\;， º^º！
###3€.§§§）:
±，±，±，±，±，±，±），
#！ſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſſ!!!!!!!!!!!！!!!!！!!！!!!!!!!!!!!!!!!
|----• ·ÑĢĒ§！ -ºr：！º
ſºſºſ，
ŠŲJŲ
Eſ įĶ[\\[\\]\\
=）-'.' ;' ~
Eſ”. __ . ' ' .. ！?！ '
t-ae)· · · =.=.
ģēģ©ae;Zaeſſ:
źģZ
##ãą?:%ſ，%ſ，%ſ%

OUTLINE
THE USE OF THE RATING SCALE
PART I—-USE AND ABUSE OF Rating SCALES
A. The Rating Scale.
B. Kinds of Rating Scales in Use.
C. General Intelligence Tests and the Rating Scale.
D. Methods of Devising Rating Scales.
E. Character Analysis and the Rating Scale.
F. Value of the Rating Scale, Negative and Positive.
Summary
A. Objections to the Rating Scale Method.
B. Advantages of the Rating Scale Method.














„-ewwaeººru
****
* -20 - 17 27 5
PART I
THE USE OF RATING SCALES IN INDUSTRY
A—The Rating Scale
Out of the 117 companies from whom information was ob-
tained, 36 were making a more or less extensive use of rating
Scales. Fifty per cent more companies were using rating scales
than were making use of psychological tests.
B—Kinds of Rating Scales in Use
The use of the rating scales was found to be very diverse.
Twelve companies used rating scales to apply to all positions;
seven used a rating scale for executives only; five for salesmen;
three for machinists; three for men in training; five for foremen
and assistant foremen; two for agents; three for telephone oper-
ators; one for all employes on a salary basis. Consequently,
there was little uniformity in the character of the rating scales
in use. Some of the rating scales showed distinctly the influence
of the army rating scale. In fact, whatever uniformity appeared
in the various scales could with reasonable certainty be traced
to the army scale.
In spite of the variety of jobs to which rating scales were
applied and the differences between them, there were no evidences
of any fundamental method which could account for these dif-
ferences. There were no consistent variations in the qualities re-
quired by a foreman, an employe, an executive, a salesman, etc.
For example, the explanation on one rating scale was in part as
follows: “Even though the characteristics named in the rating
scale are used for foremen, sub-foremen, etc., it should be borne
in mind that the higher the position, the greater the need for
these characteristics, and they will have a different value for po-
sitions of greater or lesser responsibility. The significant fea-
ture of this system is the comparison of each man with men of
equal and superior rank, thus supplying a concrete standard of
6
judgment.” The rating scale which contained this explanation
had selected seven characteristics which were to be applied as
a universal measuring rod to men whether in higher or lower
positions. The scale was used on a quantitative basis, or after
what we might call “linear fashion.” It is true that different
qualities have a different value in higher and lower positions, but
it seems equally true that the higher the position, the greater the
need for some qualities and the less the need for others. Trade
knowledge, for instance, is much more important to the fore-
man or sub-foreman than it is to a production manager; while
in a production manager a much higher degree of organizing
ability is essential than in the case of a foreman. For different
types of work, it must be assumed that different types of ability
are necessary, and that no one scale or set of qualities can with
ahy degree of accuracy or concreteness measure men for wide
variety of jobs. We might compare this method to an attempt
in the realm of physics to measure pressure, acceleration weight,
specific gravity, tension, friction and magnetism all by means of
a yardstick.
The Mutual Rating Scale
One of the most interesting developments of the rating scale
was brought to light by a paper read before the Taylor Society
of New York by Henry Wood Shelton in December, 1919. This
paper describes a unique method of rating by which executives
and foremen not only rate their workers but are in turn rated
by them. The ratings are mutual. Another distinctive point in
the method is the fact that the ratings are made by secret ballot.
It is also interesting to note that in his paper, Mr. Shelton calls
attention to the fact that the rating scale is not a distinctively
psychological product and gives a very fair estimate of the re-
lation between the psychological method and the layman method.
He also gives a very fine discussion of the theory underlying the
use of the rating scales. |
More details of Mr. Shelton's paper are not given, because
it is being published by the Taylor Society and will probably be
in print before this report appears. This paper and the chapters
on the production scale and rating scale in Part III of “Employ-
ment Psychology” are the only comprehensive discussions of the
theory and practice of rating scales available. §
The Graphic Opinion Record
One company uses a method which lends itself to graphic
presentation. On the following page is a copy of the form used,
showing how it is filled out and the graph drawn.
It will be seen that the qualities are checked, and a line (usu-
ally in red crayon) run through the points checked. In addition,
a numerical result is attached to the record as a whole. This is
obtained by giving a value of three points to each quality checked
under “good,” two points for all under “fair” and one point for
all under “poor.”
The company using this record is the only one found which
is following up its results with any approach to scientific accu-
racy. The record for each foreman is made out by three dif-
ferent executives. These records, taken for a considerable num-
ber of foremen, are then correlated according to formulae well
known in statistics, to determine the degree of correspondence
which exists between the ratings made by the three different
executives. Naturally, if there is no consistent agreement, there
must be something wrong either with the method or with those
making the ratings. So far the correspondence has been very
striking.
It should be particularly noted that this form is not called a
RATING SCALE but an OPINION RECORD. This expresses more truly
the real nature of the method and its relation to mathematical or
scientific measurements.
C–General Intelligence Tests and the Rating Scale
In this respect there is a lesson to be learned from the use
of general intelligence tests. It was originally assumed that a
general intelligence or alertness test could measure the degrees
of intelligence of any group of men. In a sense it could, but after
the degree of intelligence had been determined the next question
was: What is each degree of intelligence good for 2 Here is
where the difficulty arose, because it was found impossible to
say that a man of grade A intelligence would make a good execu-
tive or production manager, a man of grade B intelligence would
make a good foreman, a man of grade C intelligence a good
conductor or mechanic, or clerk; a man of grade D, a good
sweeper or watchman, etc. The correlation in specific cases be-
tween degrees of general intelligence and ability at a particular
Name John Smith Date February 20, 1920
(*Note 2) t
OPINION RECORD
GooD FAIR Poor
Energy . . . . . . . . . . . . . . . . . . . . . . . .
Health . . . . . . . . . . . . . . . . . . . . . .
Appearance . . . . . . . . . . . . . . . . . . . .
Openmindedness to new ideas . . . .
Education . . . . . . . . . . . . . . . . . . . . .
Ability to issue clear instructions. . N-
T-
Initiative . . . . . . . . . . . . . . . . . . . . . . _T
~T

* * * * * * * * * * * *
Disposition . . . . . . . . . . . . . . . . . . . .
Co-operation . . . . . . . . . . . . . . . . .
Professional knowlege of experience
Organizing ability . . . . . . . . . . . . . .
Results. . . . . . . . . . . . . . . . . . . . . . . .
Total number of points 38
REMARKs:
Signed F. P. A.
9
job was not high enough to warrant much emphasis on the
straight intelligence test for industrial purposes. This fact seems
now fairly well recognized. The rating scale at present seems
to be in a position similar to that from which the general intelli-
gence scale has just emerged. That is, it is being used to meas-
ure a large variety of men for a large variety of jobs on the
basis of a single set of qualities.
D—Methods of Devising Rating Scales
There were no evidences of any fundamental method in the
development of rating scales, or of any established technique
which was followed. As stated above, the army rating Scale
seemed in some cases to be the basis upon which many subse-
quent rating scales were built up, with such additions as seemed
in particular cases to be necessary. In fact, it seems that one of
the principal virtues of the rating scale—and this may account
in some measure for the extensive use to which it is put—is the
ease with which a rating scale can be devised and put into prac-
tice. A single individual or group of individuals can get to—
gether, discuss the various personal qualities, which they can
muster up, select those which they consider most important, as-
sign to them an arbitrary value, compile a set of directions, and
then apply the result to any number of jobs or individuals. All'
this can be done in a few sessions and with a comparatively
small amount of experimental work or perhaps none at all.
E—Character Analysis and the Rating Scale
Considerable criticism has been leveled at the so-called sci-
ence of character analysis for its vague and indiscriminate use
of a large number of personal characteristics. Strangely enough,
the rating scales now in use are highly reminiscent of the quali-
ties usually found on a character analysis sheet. To be sure, the
terms mental, motive, vital, texture, color are absent, but the
terms such as leadership, moral fiber, mental ability, perception,
personality, character, discrimination, disposition, alertness, adap-
tability, unselfishness, interest, etc., frequently appear on rating
scales now in use. These qualities, considered in themselves,
are very difficult to determine or measure in any reliable man-
ner. They are altogether too general and too vague, and where
an attempt is made to define a quality the definition consists only
10
of synonyms which render the meaning of the quality more gen-
eral rather than more concrete. For instance, leadership may
be defined as: tact, initiative, cooperation, judgment, ability to
supervise and direct, a definition which does not seem to make
the task of estimating it much easier.
That this vagueness is being recognized is indicated by some
of the later rating scales found in use. The Dennison Manufac-
turing Company, for instance, uses one rating scale for foremen
and another for salesmen. Each rating scale has five different
qualities, but these qualities are selected with specific reference
to the job and furthermore are defined in terms of the concrete
and measurable activities of that job. For example, on the rating
scale for salesmen one of the five points listed is “Orders.”
Under “Orders” the estimator is asked to “Consider number of
orders, grade of goods sold, variety of items on orders, number
of new accounts, unusual orders, new uses for goods, selling
without overstocking, responsiveness to ‘selling drives’.” This is
evidently a very concrete and measurable set of items about a
salesman. The estimates are made by comparing salesmen with
other salesmen, and, in the case of foremen, foremen with other
foremen. The items according to which foremen are measured
have been selected with distinct reference to the essentials of fore-
manship. The qualities on the salesman rating scale have been
selected with reference to salesmanship. In short, the rating
scale is in each case a job-man specification, in distinction from
the more general type of rating scale still quite prevalent which
may be called the man rating scale.
F—The Value of the Rating Scale, Negative and Positive
A number of serious objections to rating scales as used thus
far can be raised and indeed should be. understood by all who
intend to make use of any rating scales. In the first place, as
was pointed out above, the ease with which rating scales can be
compiled has led to their sudden and extensive use in many cases.
Because of this extensive use, it is difficult to make any careful
and intensive estimate of their value in specific instances. In
the second place, although called a rating scale and heralded as
an accurate measure of a person's chracteristics, there is nothing
about a rating scale which can be called scientifically accurate.
One of the best rating scales found begins with the statement:
11
“The Rating Scale is a practical method of gauging a foreman's
capacity and fitness for promotion quickly, accurately, and with
uniformity and justice.” The first claim made in this statement
may be taken without question, but the claims for accuracy, uni-
formity, and justice must be taken with a generous tolerance.
The fact that there are numbers assigned to the various degrees
of different qualities by no means makes the rating scale numeri-
cally or quantitatively accurate; for, in the first place, the num-
bers are arbitrarily asigned, and, in the second place, the num-
bers selected by the person who makes the estimate are not
necessarily just. There is considerable latitude for the human
equation in making out any rating scale. When the claim is
made that “the rating scale itself is a numerical expression of
the degree in which a foreman possesses the industrial qualifica-
tions most essential,” these facts must be borne in mind. In the
third place, because of a certain unavoidable vagueness in the
rating scale, it is difficult to follow up the results with accuracy
and pin mistakes in rating down to any one definite cause or
factor.
On the other hand, there is no question that the rating scale
has certain very definite values. Its first and greatest value is
its educational effect. Whether the rating scale is accurate or
not, it does require a certain amount of thought on the part of
those who use it. A careful and systematic study of individuals
such as is at least suggested by a rating scale is the first valuable
step in the direction of producing more accurate results. More-
over, if the use of the rating scale educates its users to a realiza-
tion of the fact that human nature and its measurement is not
as simple as they thought, a great deal will have been gained.
There are already evidences that this result is being achieved,
and that personnel workers are realizing the importance of a
more fundamental study than that which is presupposed by the
rating scale theory.
The rating scale may be of value not only to the person who
uses it, but to the one who is rated. At the Larkin Company,
where a rating scale for foreman has been in use for four years
or more, it has been the custom to give, each Christmas, a bonus
based upon the rating scale. If a man failed to receive a bonus
and asked for the reason, he was shown the rating scale and
informed that he had fallen short in certain qualities. This
12
process was found mutually enlightening, for it informed the
foreman where he stood in the eyes of the management, and
gave the management an opportunity to hear the foreman's opin-
ion of their judgment. A rating scale kept in secret or unknown
to the person being rated loses half its educational value.
The great question which still remains is: Does this method
eliminate that prolific source of trouble, prejudice? The unhesi-
tating answer to this question must be: NO. No method which
rests in the final anlysis upon personal impressions or opinions
can be free from prejudice. However, it can be stated with
equal emphasis that the method described will tend to minimize
prejudices most decidedly.
In the first place, although prejudices can never be avoided,
they can to a large extent be neutralized. A prejudice, as is
well known, often rests upon only one quality in a person. Con-
sequently, when the attention is called to qualities possessed by
the worker other than the offensive one, the tendency is toward
a more equitable and balanced opinion. For example, a super-
visor may have conceived an intense dislike for one of his workers
because of a strange and affected manner of speech. However,
on glancing down the card, he may be compelled to admit that
this worker is very industrious, intelligent, orderly, accurate,
reliable, fast and regular in attendance. At the end of this com-
plete survey, he must admit, even though reluctantly, that after
all the man has his good points. On the other hand, he may be
compelled to admit that another subordinate for whom he has
a distinct liking because of his amiable manner and ready wit,
is lazy, slow, and unreliable, and not at all the man he had been
impressed with. *
Moreover, the fact that these opinions must be recorded and
signed will do much to reduce the degree of prejudice expressed.
The appraiser will be conscious of the fact that, in expressing
his judgment, he is putting himself, as well as his subordinates,
On record and that any mistake or exaggeration on his part may
rebound to his own discomfiture. Again, the fact that the esti-
mates are made at stated periods will also lessen the degree of
prejudice expressed. One of the most fruitful sources of preju-
dice is ill-temper. Ill-temper may be induced by any number of
causes, but its inevitable result is to produce a prejudice against
anything or anybody it comes in contact with. In such moments
13
it is very easy to conceive an opinion of a person which is quite
likely to result in an extremely unfavorable estimate. However,
a comparison with estimates of that person made in previous,
calmer 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.
There are two factors which to a large extent determine the
value of the rating scale method. First, the definiteness with
which it applies to the qualities required by a specific job; sec-
ondly, the definiteness with which it defines these qualities. When
attempts are made, however, to meet these two conditions in a
thoroughgoing fashion, it seems to lead to results which become
increasingly confusing. Much remains to be done in this di-
rection.
G—Psychology and Rating Scales
The rating scale should not be confused with psychological
tests or with any psychological technique. (See part two of this
report.) The rating scale has been and is being used by some
psychologists, but it is not based on any procedure which is char-
acteristic of or peculiar to psychology. A person who is not a
psychologist and has had no training in its technique may devise
and use rating scales as successfully as a highly trained psy-
chologist or even more so. To devise a rating scale requires in-
genuity rather than a knowledge of psychology, as is evidenced
by the rating scales now in use. The difference between rating
scales is largely a difference in the ingenuity with which they are
devised, and a difference in the knowledge of industrial condi-
tions on the part of those who desire them. By the psychologist
the rating scale is regarded as a necessary practical makeshift
which he must use because of his inability as a psychologist to
provide something which meets the more rigid demands of his
science.
Summary
A—Objections to the Rating Scale Method
1. The lack of any fundamental or consistent method by
which rating scales are devised.
2. The fact that they are often applied indiscriminately to a
wide variety of jobs and persons.
14
3. The lack of any close relationship between rating seales
and job specifications.
4. The vague and indeterminable manner in which the quali-
ties on the rating scale are usually defined.
5. The diffiéulties which the inherent vagueness of the rating
Scale places in the way of an accurate estimate of the results of
its use.
B—Advantages of the Rating Scale Method
1. It tends to educate executives to observe more closely and
accurately the character of their workers.
2. If properly used, it educates the worker in the knowledge
of his own shortcomings. At least, it informs him of the opin-
ions which his supervisors hold.
3. If systematically used, it promotes a periodic, instead of
sporadic expression of opinion.
4. Because it makes it necessary for the appraiser to check
off the various qualities of an individual against each other, thus
producing the balanced result.
5. Because the appraisal made goes on record for or against
the man who makes it.
6. Because the ratings by one appraiser can be systematically
checked against those of another.
OUTLINE
THE USE OF PSYCHOLOGICAL TESTS
PART II—PRESENT STATUS OF PSYCHOLOGICAL TESTs
IN INDUSTRY
. Plan of Report.
. Progress made during the past year.
Comparison with Previous Report.
. Better Refinement and Understanding of Tests.
Some Practical Applications of Tests.
i
1. A Typical Manufacturing Concern.
2. A Large Department Store.
3. A Large Public Utility Company.
4. Bureau of Personnel Research.
F. “Scarcity of Labor” and the Use of Tests.
G. Some Methods to be Avoided.
PART III—PRACTICAL ASPECTS OF INTRODUCING AND APPLYING
TESTS
A. Steps in Developing and Applying Tests.
1. Selecting Job to be Tested.
2. Study and Analysis of the Job Selected.
3. Developing and Testing the Tests.
4. Follow-up of Tests.
B. Cost of Installing.
C. Who Shall be Chosen to Do the Work?
D. Place in Organization.
Summary
THE USE OF PSYCHOLOGICAL TESTS
PART II—PRESENT STATUS OF PsychologICAL TESTs
IN INDUSTRY
A—Plan of Report
The purpose of the committee's report this year is to furnish
the members of the Association, first with a concise report of
just what use is being made of psychological tests in industry,
and, second, what are some of the practical aspects involved in
introducing and applying tests. As a basis for this report, 117
companies were either visited or interrogated by members of the
committee, both in and out of the Association. We have at-
tempted to make an analysis of this data, so that it will be of
most help to companies desiring to introduce tests as a part of
their personnel program.
B—Progress Made During the Past Year
Except in a limited number of cases, there has been but little
progress made in the development and application of psycho-
logical tests within the past year. Strangely enough, the widest
application of tests has been made by a large public utility cor-
poration, about which there has been little publicity. For the
past five years this organization, with employment and training
branches in all the principal cities of the country, has been using
tests to select applicants for training as operators. These tests,
or “examinations,” as they are termed, were systematically de-
veloped to meet their particular needs. Through the examina-
tion of thousands of applicants, definite standards of perform-
ance have been established. Their success, to a large extent, is
attributable to the fact that they were planned to meet a specific
need within that organization. The company was spending an-
nually a considerable amount of money to train these people
before they were put upon productive work. Naturally, it was
to their advantage to see that only those applicants most likely
to succeed were selected.
17
C—Comparison with Previous Report
In comparing the result of our investigation with that of a
year ago, several interesting facts were brought out. In 1919,
out of the 118 companies either visited or interrogated, 29 com-
panies, or 25 per cent, were using tests; 18 companies, or 15
per cent, expressed their intention to apply them, and most of
the rest were very much interested. A similar study this year
of 117 companies shows that 24 companies, or 20 per cent, were
using standardized written or performance tests in connection
with employment and training. There were 23 per cent using
“general intelligence” tests, while only about 16 per cent were
making use of “trade” tests. “Occupational” tests, devoted
mainly to the selection of stenographers, typists and clerks, were
being used by 28 per cent of the companies.
D—Better Refinement and Understanding of Tests
While the extravagant claims of the efficacy of tests are still
heralded by the pseudo-psychologists, propagandists and popular
publications, still a great deal of the “psychic” has been eliminated
from the word “psychology.” The infallibility of tests has not
been established by any of the companies visited, nor is it likely
to be before the millennium.
A better understanding of what is meant by “psychological
tests,” as well as a systematic, scientific and judicious application
of tests in industry, will do more to enhance their value than the
extravagant claims and propaganda mentioned above. If this is
So, there has been a distinct step in advance, because heretofore
ideas about psychology have been extremely nebulous. One of
the theories quite prevalent in several companies is the idea that
standardized performance tests do not fall within the range of
psychology. These companies, as well as many others, are still
laboring under the impression that tests of intelligence are the
only tests which involve any psychological principles. One com-
pany reports that they use “performance tests for stenographers,
typists and adding machine operators, but felt that these tests
are least important from the standpoint of the activities of your
committee.” This statement shows an obvious misconception of
the meaning of psychological tests.
It might be well to repeat again the definition arrived at by
the committee on page 334 of the Proceedings for 1919:
18
“A psychological test is a set task which the person being
tested is asked to perform. The nature of the task depends on
the purpose in mind. Any task whatsoever may be used as the
basis for a psychological test. For instance, a task on the type-
writer, or a series of problems in arithmetic, or a task in crossing
out all the numbers on a certain page, or a number of directions,
or a series of questions—any one of these tasks may be used as
the nucleus of a psychological test. Psychological tests are
closely allied to physical or mechanical tests. Most machines are
tested by being required to perform certain set tasks. The same
practical principle is applied to psychology to human beings.”
The contributions made by the army tests have been, at least,
to promote a more definite idea of what tests really are. On the
other hand, it has been found that the army work has resulted
in a false impression, namely, that “general intelligence tests”
were psychological tests, but that “trade tests” were not psycho-
logical tests. This same idea has been carried over in the use of
Occupational tests, i.e., for stenographers, accountants, etc.
What most companies are interested in, in employment and
training work, is “specific abilities” and “kinds of intelligence”
for a particular job, and not in degrees of intelligence, per se.
All of these tests, in so far as they are properly devised and
applied, are psychological tests.
E—Some Practical Applications of Tests
The committee found that a great many companies are still
skeptical as to the value of tests for their particular business.
They always advance the theory that “their business is differ-
ent”; that “their work requires special abilities which differenti-
ate it from other work,” thus making the use of psychological
tests impracticable. They contend that “there is no floating sup-
ply of the kind of people that they want for their particular
industry.” Imbued with these ideas, they have no use for tests
of any kind. These companies are still thinking primarily of
what has been done, and not of what the possibilities are of adapt-
ing tests to meet their particular needs.
Several typical and concrete cases of just how tests are being
used today are outlined below to correct this idea as far as it is
possible to do so at this time.
19
I—A Typical Manufacturing Concern
One of the most extensive applications of tests known to the
committee is being carried out by a large New England manu-
facturing concern employing some fifteen thousand men and
women. The tests were first introduced in this company some
three and a half years ago as an experiment. The tests were
first tried on workers already employed before being tried on
new applicants. As soon as their practical value was established,
they were extended to some ten or twelve different types of
work, including assemblers, inspectors, apprentices, typists, ste-
nographers, comptometer and machine operators, and clerical
workers. Through the testing of some eight thousand employes,
these tests have been thoroughly standardized. Instead of the
usual method of applying general tests to all kinds of work at
Once, each type of work was dealt with as a unit.
At the present time the work is in charge of a trained ex-
aminer. A section of the employment office has been partitioned
off, where the tests are being performed. In making the tests,
standard office appliances and equipment are used. The time re-
quired for giving the series of tests varies from fifteen to twenty
minutes each. The results of each applicant’s performances are
recorded and furnished to the interviewers and departments in-
terested. A definite “follow-up” is made after one month of
employment, and every three months thereafter. The records of
performance are also used as a basis for both promotion and
transfer. There is nothing spectacular about the testing, which
has come to be looked upon as one of the essential parts of the
employment procedure in this organization.
2—A Large Department Store
In New York City a large department store has attacked this
problem in a very systematic way and with considerable success.
Since the spring of 1919, when the experiment began, over eight
thousand employes have been examined. The work is under the
supervision of a graduate of Columbia University, who was first
allowed to carry out some experiments among the salesgirls of
the training classes. As a result of her work, she has been per-
manently engaged in extending the use of texts. Among the
most useful tests which she has developed is a pre-employment
test, which is given to every prospective employe before being
considered for any job. It is a modification of Trabue's Com-
pletion Test, simplified and adapted to meet their particular
20
needs. There is a time limit of ten minutes allowed to complete
the test. As many as 250 applicants a day have taken this test.
Through the examination and close follow-up of several thou-
sand applicants and employes, standards of performance for dif-
ferent classes of work have been definitely established.
In selecting people for particular types of work, there are
being developed occupational and specific ability tests. The meth-
ods of attack used in this particular instance are worthy of imi-
tation. The tests themselves are conducted in the employment
office, and form a part of the regular procedure for entrance
into the company. For carrying out further experiments, a room
has been set aside where additional experiments and studies may
be made without interruption. The work here is closely allied
with both the training and employment departments.
Similar work is being done with equal success by several
large department stores in Philadelphia and Pittsburgh. Tests
have been used in the personnel departments of these stores dur-
ing the past year. From reports of their work it is evident that
these tests have been carefully adapted and are demonstrating
unquestioned worth.
3—A Large Public Utility Company
Since 1915 a large public utility corporation, which has em-
ployment and training offices in most of the principal cities of
the country, has been using tests to examine applicants. These
tests have been given to telephone operators, automatic telegraph
operators, clerical workers and messengers. Some sort of an
examination was necessary, bec use the prospective employes
come into the organization as students and are taught to become
operators on different machines.
The different jobs were first carefully analyzed. Their spe-
cific tests were devised to determine if the prospective employe
possessed the ability to operate the machines. During the five
years several thousand applicants have been tested and the tests
have been thoroughly standardized.
The administration of the test is comparatively simple. Sepa-
rate tests are assembled together and given to the applicant. The
direction for each group of tests is clearly written out. There
is no time limit placed on these examinations, the average time
usually consisting of about one hour and fifteen minutes.
For different grades of clerical workers there have been dif-
21.
ferent series of tests developed. One-half of the test is devoted
to testing their education and the other half to predicting whether
or not the applicant has sufficient mental ability to follow simple
directions required in the performance of their work.
The developing of these tests is under the direction of a man
trained in psychology and who is permanently engaged in this
work. The grades made in the examinations are closely fol-
lowed up. The correlation between performance in the exami-
nations and later on the job have satisfied this company that
these tests are meeting their needs.
4—Bureau of Personnel Research
The Carnegie Institute of Technology, through its Bureau of
Personnel Research, has been doing some very systematic work
in the field of psychological tests. A number of department
stores in and around Pittsburgh, as well as other large corpora-
tions, have been making use of psychologists trained by this
bureau. These people are given a thorough training in the tech-
nique of developing and applying tests. While the work in a
great many places is still in the experimental stage, still there
has been enough progress made to insure its future success.
Most of the tests used in companies cooperating with the bureau
are not available for general distribution, but their method of
attack is similar to that outlined above.
F—“Scarcity of Labor” and the Use of Tests
Coupled with the idea that the nature of their business is
different, we find a number of companies maintaining that the
“scarcity of labor” will not waſ rant psychologi ests. A large
automobile manufacturing company says that “there is at present
a very serious shortage of labor, and unless the conditions permit
a selection of three to four men for every job that is open, we
see but small advantage in the expense of an elaborate outfit of
psychological tests. We are accomplishing very satisfactory
results by using a very superior type of selectors in our labor
employment offices. These men have had a broad experience in
factory work, are natural judges of men, and know how by ques-
tioning to determine the man's fitness for the job he claims he is
able to hold. When labor conditions change and we have a
large supply available, we contemplate going more into this mat-
ter, hoping that by that time much of the present camouflage
will have been eliminated.” Another company writes that “as
22
to factory employment, there are more places vacant than there
are applicants, and we cannot justify these forms of inquisition
under these conditions. Some of these intensified methods of
inquiry may operate in cities where there are many applicants
for the same position and where discrimination may be used in
Selecting from the applicants the most desirable one for the
work.” A large company manufacturing steel products reports
that “in our business labor has been so scarce for the past five
years that a man with reasonable physical vigor, who has a nor-
mal appearance, has been given employment without further
question.” The committee has quoted the above replies to show
that there is still a mistaken idea prevalent as to what is the
chief value of psychological tests. Companies are still going on
the mistaken theory that tests are simply to reject or accept ap-
plicants, and not for a more economical classification of labor.
The tests, to be of any economic value in industry, must be used
for better classification, placement and training of employes, and
also as an aid in setting rates. Selection and rejection is the
negative aspect of their use.
G–Some Methods to be Avoided
In connection with the use of tests there are a great many
things to be avoided, and it is hoped that by calling attention to
a few of these both time and money, as well as the good will of
the applicants, may be saved.
Psychological tests should not be considered a panacea for
all labor troubles. Neither should they be relied upon as the
sole criteria for selecting applicants for a job. They are merely
to supplement and not to supplant the usual employment pro-
cedure. Thus, one writer was correct in believing that “such
tests supplied only a portion of the information desired when
they were conducted upon practice work or any other work
which is not a true representation of the actual operating con-
ditions.” For we must agree that employment, after all, except
in the more routinized jobs, is a matter of compromise and ad-
justment. Naturally, tests will never conform to all the exigen-
cies arising in an organization.
A great many companies have condemned the tests after only
a very superficial attempt to use them. It is true that there are
a great many pseudo-psychologists, vocational specialists, propa-
gandists and popular publications advocating the wholesale use
23
of psychological tests. There is even a group of psychologists
who are advertising tests for sale at so much per hundred. This
is something which should be avoided as much as possible.
The reason why tests have not been successful in a great
many cases is not inherent in the principles of the tests them-
selves, but in the procedure and method of attack. The experi-
ence of one company in this connection is well worth noting:
“We did give our entire office one of these tests, and the results
were very unsatisfactory. We found that some of our very best
employes received the poorest rating, on these tests, while some
of our very mediocre employes, because of having more recently
come from school, were able to make much better marks.” The
tests used in this particular instance were developed largely under
laboratory conditions without any particular reference to the
work in question. Naturally, when applied indiscriminately to
all classes of clerical workers under new conditions they were
destined to fail. Such a wholesale procedure is untenable and
will, in the end, destroy the entire value of psychological tests.
One of the first lessons to be drawn from the experiences of
others is the danger of a sudden and non-experimental applica-
tion of tests. An organization or an enthusiast who takes a
group of tests and says: “Now such and such tests apply to the
selection of clerks, these to the selection of machine operators,
these will select people with manual dexterity, these will do for
the choice of executives,” etc., is inviting failure at the very out-
set. For, according to the psychological technique as it is com-
monly understood, as well as according to the dictates of common
sense, no tests can be safely applied without a preliminary trial
or series of trials to discover whether the tests proposed actually
apply to the particular jobs for which they are intended. Even
though the tests have been tried out and used elsewhere, it does
not follow that they will do equally well in a new situation or
under new conditions. In introducing tests or adapting them to
a particular industry, the usual course is to take such tests and
to try them out on a few kinds of work at a time, adapting them
to the particular needs of the particular jobs of the particular
industry. Such a procedure reduces the possibilities of failure
to a minimum, and insures results that are based on a solid
foundation. Experience has shown that no test can be safely
used that has not been tried out on at least 25 employes under
local conditions.
24
PART III—PRACTICAL ASPECTs of INTRODUCING AND APPLYING
TESTS
Although any test may serve as the basis for a psychological
test, there is a further factor which enters in, namely, the factor
of Scientific accuracy. A test is not strictly psychological until
it has been carefully tried out and standardized. Many com-
panies, for instance, give tests of one kind or another, which
they themselves do not call psychological, because they realize
that the tests have been imperfectly worked out and are not
always applied correctly.
It is a comparatively simple matter to give a test, just as it
is comparatively easy to take a doctor's instruments and test
another person's temperature, pulse, blood pressure, etc. But
having done so, the chief problem remains, namely, arriving at
a conclusion. This matter of testing the test to discover its
practical value and significance is so important that it will be
summarized briefly in the following section.
A—Steps in Developing and Applying Tests
1. Selecting the Job to be Tested.
Wherever possible, begin with a job in which the labor turn-
over warrants a specific study. Abundant evidence is found
where tests have been applied to all sorts of jobs rather than
being applied directly to one or more specific jobs that required
study.
2. Study and Analysis of the Job Selected.
There is no better way to study the job than to do the work
itself. Various forms of job analyses have been developed to
secure this information, but at best they are only a poor sub-
stitute.
3. Testing the Tests.
On the basis of this first-hand study of the job, specific tests
are selected or developed. There are thousands of tests already
in use that might be modified under the hands of a skilled tech-
nician to fit this particular type of work. Fully 95 per cent of
the companies using standardized written or performance tests
report that they have tried them out beforehand on twenty-five
or more people in their own organization before putting them
º
25
into use. If such a practice is continued by other companies
still greater results may be expected.
4. Follow-up of Tests.
The practicability of tests will rest in the last analysis on
whether or not the applicant performs the job satisfactorily to
which he is assigned. To determine this impartially requires a
very close follow-up. Eighty-three per cent of the companies
using tests reported that someone else besides the psychologist
checked up the results of the tests and estimated their practical
value.
The results of selection tests may be followed up in a variety
of ways: (a) by means of obtaining production records of those
workers engaged on piece work, (b) by calculating the labor
turnover, (c) by a periodic follow-up made at the end of every
month for three months and every third month thereafter, (d)
by the use of periodic rating scales.
B—Who Shall be Chosen to do the Work &
This question is one that every company asks itself when
considering the advisability of introducing tests. The committee
found that 83 per cent of the companies using tests employed
psychologists in their personnel departments, fourteen of whom
were in a permanent capacity and six as consultants. Experience
has shown that to successfully meet the requirements outlined
above requires someone familiar with psychological technique,
in addition to a first-hand knowledge of the jobs for which tests
are sought.
C–Cost of Installing
It was rather a disappointment to the committee to find that
only 12 per cent of the companies using tests has attempted to
compute the cost of the psychological work in terms of cost per
person examined. This may be accounted for by the fact that
the work in many of the companies is still in its experimental
stage. A careful estimate of the cost of the work in one com-
pany showed that it was about $1.50 per person examined for
the first year, which included a considerable amount of experi-
mental work, and was likely to be about $1.30 per person for the
second year, when the experimental work would be proportion-
ately less. *
. . . " " : " - B
. . . . . . . . . . .
26
D—Place in Organization -
There are many uses for tests besides their use for accepting
or rejecting applicants for particular jobs. Tests and examina-
tions have always been used in connection with training work,
but their value was limited to a great extent, because of the
failure to standardize the results. There is still a broader field
of application, namely, in the classification and economical place-
ment of labor, as well as one of the basis for rate setting and
bonuses. Whenever an organization is willing to base wage pay-
ments on impartially standardized tests and examinations, we will
see them more widely used in industry. # * ,
The status of psychologists in the business world is still prob- i.
lematical. At present they perform all varieties of duties which
their close relationship to the personnel departments require. In
one company visited the psychologist was separated entirely from .
the employment and training department routine and was given
considerable latitude in development of her work. Other com-º:
panies were reluctant to give the experiment a fair showing andź
oftentimes condemned the work after a very superficial trial.
Summary
1. The number and variety of instances in which tests have º
been successfully applied under industrial conditions warrants!
the belief that they can be successfully extended to other indu
tries. - ! -
2. One distinct step in advance is the prevalence in industry
of a better understanding of what psychological tests really are :
3. The failures of tests, in a majority of cases, are usually
attributable to superficial methods of giving the tests and follow-
ing up the results, and is not inherent in the principles of the
tests themselves. -
4. Tests should be used to supplement and not to supplan.
the ordinary employment interview. &
5. Tests should be first applied to jobs which have the larges
turnover or in which the turnover is most costly. • . . . ;
6. One of the most important features in connection with the
use of tests is the maintenance of a careful and systematic follow:
up to determine the value of the results. Such a procedure wi
not only enhance the value of the tests, but will shed a welcon
light on employment methods in general. -
\
#
# .
* .
!
§
**. -- . . .
























_3 9015 07324 7044
| z
| <ſ
| GD
| III
| C
! >
! LL-
| O
' >
H-
iſ (f)
: OC
į Li
| >
2
| IrDA
N. Y.
. 21, 1908
Gaylord Bros.
JAN
Syracuse,
PATſ.
§
¿??¿№
§§-aerff;&#*Ř®******、、**ș，
¿?
• → ）， *****************
*-_-） \，;

& .
:::::
§:
jº
#
§
§3
- §:
§ j;
ø º
§
º ;
ſº § º
º §
# §§§ § *.
§ §
§
§§ §§
§
**
&
#3
§
t
º:
yº
§ § § #:
; §:
§
#
§:
!.
* * **
*
º
º
ºr
i
§
É;
§§
§ § º:
# º #
§§
§
i
§
§
{
º
;§;
r
$
s
-$º
§
º
}
§
º
t;
i
tºy
;
;
s
i
#:
§
i
i.
i
;
i
$º
- §
i
;
H
|
i
§
#
;
f
;
#
º:
#
§
§:
§
.
;
i