Copyright
Herbert Wesley Rogers.

Some empirical tests in vocational selection online

. (page 2 of 4)
Online LibraryHerbert Wesley RogersSome empirical tests in vocational selection → online text (page 2 of 4)
Font size
QR-code for this ebook


them.



Opposites


object


Agent-
action


Action-
agent


Mixed


relations


north


read


baby


gallops


Eye-see


Ear-


sour


tear


fire


bites


Monday-Tuesday April-


out


throw


dog


boils


Do-did


See-


weak


paint


laborer


sleeps


Bird-sings


Dogs-


good


mail


pencil


floats


Hour-minute


Minute-


after


light


army


growls


Straw-hat


Leather-


above


sail


heart


sails


Cloud-rain


Sun-


sick


spin


pin


roars


Hammer-tool


Dictionary-


slow


lock


gun


scratches


Uncle-aunt


Brother-


large


wash


eyes


stings


Dog-puppy


Cat-


rich


bake


bird


shoots


Little-less


Much-


dark


spill


wind


melts


Wash-face


Sweep-


front


kiss


lungs


swims


House-room


Book-


love


polish


bell


explodes


Sky-blue


Grass-


tall


sweep


musician


aches


Swim-water


Fly-


open


fill


parrot


glows


Once-one


Twice-


summer


sharpen


clock


news


Cat-fur


Bird-


new


write


ax


cuts


Pan-tin


Table-


come


chew


broom


flies


Buy-sell


Come-


male


drive


mosquito


burns


Oyster- shell


Banana-



B. Individuals serving as subjects.

a. Group 1.

Seventy-seven young men and women, who were studying
typewriting, stenography, and grammar in the Extension De-



16 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION

partment of Columbia University, served as subjects in this
first group. These people were divided among four different
sections, three evening classes and one day class. In all of
the evening classes it was impossible to obtain any adequate
measure of the subjects' abilities, for they either dropped out
in short time or failed to take the prescribed tests and exami-
nations. The results worked up for the day class of forty-
five students, forty-three female and two male are the only
ones used in this investigation. The subjects made an in-
tensive study of typewriting and stenography devoting two
hours in class to each subject each day for five days a week.
Some outside study was also done. The course lasted from
the first part of October, 1915, until the middle of May, 1916.
In group 1 the performances in the tests were correlated with
abilities in typewriting, stenography and grammar.

b. Groups 2 and 3.

One hundred and eighteen typists, all female, in a large
retail commercial concern in New York City were tested.
Of these ten left the employ of the concern before any ad-
equate measure of their abilities could be determined. The
remainder were divided into two groups. Group 2 consisted
of thirty-eight typists who had been working in the same
division for at least ten months. Group 3 consisted of sixty-
five typists who had been working in this same division for
at least one month and a half and for not longer than six.
months. All of the subjects in both groups had been doing
the same kind of typing. The records of five typists who had
been working between six months and ten months were dis-
carded. This division was made in order to allow, as well
as possible, for the effect of practice in the particular kind
of typing which the subjects in these two groups were per-
forming, as it would have been unscientific to put all of the
subjects in one group.* In groups 2 and 3 the performances
in the test were correlated with ability in typewriting only.

c. Age, education and experience of groups 1, 2 and 3.
The figures in the following table are in per cent of the

number of individuals in each group.



SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION 17

AGE

Group 1. Group 2. Group 3

16-19 years old. .46 .68 .91

20-24 .32 .32 .08

25-29 .06 .00 .01

30-34 .11 .00 .00

35-38 .04 .00 .00

*See section 3, C, c, 2, (description of the work of groups 2 and 3).

EDUCATION

Grammar school education .02 .61 .47

1-3 years high school .02 .36 .47

High School graduates .51 .03 .05

1-3 years college .22 .00 .00

College graduates .22 .00 .00

EXPERIENCE

Never worked previously .85 .00 .31

Worked at typing 1-5 months .00 .00 .10

Worked at typing 10-12 months .00 .66 .26

Worked at typing more than 1 year .00 .34 .23

Worked at other work than typing .15 .23 .20

In the age table the mode for all groups is the group 16-
19 years old. In the education table the mode for group 1
is the full high school education, for group 2 it is the grammar
school education, and for group 3 it is either the grammar
school education or the 1-3 years high school class.

C. Methods of obtaining measures of abilities.

a. Stenography.

In February a mid-year examination was given in stenog-
raphy which was graded by the A, B, C, D, F, method; A,
B, C, being the order of satisfactory grades, D a condition and
F failed. In each of these grades the mark was further
qualified so that the A group, for instance, came to consist
of A plus, A, and A minus. The individuals were then ranked
in an order of merit in each of the three sub-divisions of
each grade, for instance if there were three subjects in the
A plus group they were ranked, according to their abilities,
1, 2, and 3. From these last rankings an order of merit
series, from one to forty-five, was secured.

b. Grammar.

In February a mid-year examination was given in gram-
mar, spelling, punctuation, paragraphing, letter writing, etc.
This was a thorough three hour examination and the results
were probably a good index to the subjects' abilities. The



18 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION

grades in this examination were returned in per cent terms
which were readily transformed into an order of merit series.

c. Typewriting.

More objective and reliable grades were obtained in type-
writing than in either stenography and grammar, the sub-
jects being measured in actual quantity and quality of output.

I. Group 1.

Each month a typewriting test was given by the instruc-
tor, in which the subjects copied a given piece of material
on the typewriter, accomplishing as much as they could in
ten minutes. The total number of words written in ten
minutes, less five words for each error made, divided by ten
gave the net number of words written per minute. This is
an arbitrary method of scoring adopted as the best relation-
ship between speed and accuracy by the International Type-
writing Committee and has been incorporated into the inter-
national rules. The following table gives the number of
subjects taking the examinations, the average net number of
words written per minute, standard deviations, and ranges.

No. of Av. no. S. D. in Range in
subjects words per min. words per min. words per min.

October 42 16.4 6.2 6. -37.

November 40 22.5 7.1 9. -51.

December 40 27.4 6.5 15.6-51.2

February 40 29.7 7.1 15.0-48.1

March ' 29 34.1 5.0 23.7-44.5

April 27 38.9 4.8 30.4-53.9

II. Description of the work of groups 2 and 3.

The commercial concern, in which the subjects of groups
2 and 3 worked, was composed, in part, of a number of stores
which retailed merchandise and which received all of their
orders by mail. The customers' orders, in the form received
by the concern, were sent to the division in which these sub-
jects worked, to have a sheet typed for each store in the
concern from which the customer might have ordered a com-
modity. Thus an order would require anywhere from one to
forty-six sheets written for it, as the customer might order
articles, in one letter, from one to forty-six of the stores.
A single order contained an order for a commodity from but
one store and required but a single sheet to be written for
it. Each sheet had typed on it the customer's name and ad-
dress and specifications of the commodity ordered, such as



SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION 19

amount, size, shape, style, pattern, etc. This was a very
particular kind of typing since there were practically no
long phrases or sentences written and many figures, symbols
and abbreviations were used. Any hierarchy of higher type-
writing habits which the subjects might possess appeared to
be of little advantage since it could not be made use of directly.
All of the subjects in these two groups were working ex-
clusively on the single orders.

It was ascertained, from the total number of words writ-
ten and the total number of strokes made in several samples
of three hundred sheets each that the amount of work done
in a day could be expressed in terms of the number of sheets
written in the course of a day's work. Thus the number of
typed sheets per day furnished the measure of the amount
of work turned out by the subjects. A group of "checkers"
read each sheet typed by the subjects, compared them with
the customer's original order and sent back to the subjects
every sheet on which an error was made; and the subjects
corrected the error in the course of the day's work, every error
being thus penalized. A system was put into effect in which
a girl circulated among the subjects and kept each subject
supplied with orders, never permitting any subject to run out
of orders. By this means no subject was able to select orders
and thus receive easy ones.

A weekly bonus on the number of sheets written was in
operation in which all of the subjects of group 2 shared al-
most every week and in which an average of approximately
forty per cent of the subjects of group 3 shared. There was
much evidence that each girl was urging herself to her best
efforts.

III. Group 2.

Records of the average number of sheets written per day
were kept for ten months, from these there were selected the
records of what appeared to be the best successive five weeks,
the selection being based upon; 1. Attendance records; no
one subject being absent more than once during this period,
in which case she was credited, for the day absent, with the
average number of daily sheets that she had written for the
rest of the week in which the absence occurred ; 2. Season of
the year; the season being selected when the business was
heaviest and when the subjects were being stimulated, by



20 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION

their supervisor, to turn out as many sheets as possible. What
appeared to be the best week's work was selected in like
manner, the best week being a part of the five weeks which
were selected. The best day's work was simply that day,
within the five weeks' period, in which each individual
subject did her best work. The five weeks' measures
and the best week's measures were for the same calendar
weeks for all subjects, but the best day's measures were
not the same for all subjects. Thus three measures of
ability were obtained: 1, daily average for five weeks;
2, daily average for one week; 3, best day's performance.
The number of subjects, thirty-eight, was constant for
each measure. These three cross sections of the subjects'
abilities for different lengths of time were selected simply to
ascertain whether the tests would correlate to the same degree
with each. The following table gives the average number of
sheets typed, standard deviations, and range of the average
number of sheets typed.

No. of S. D. of Range of

sheets typed sheets typed sheets typed

Daily average for five weeks 450.3 64.8 360.2-662.0

Daily average for best week 509.5 68.0 402.5-735.2

Best day's performance 612.0 68.2 531. -812.

IV. Group 3.

Not all of the subjects of this group were working for
the concern at the same time. Some subjects entered the
employ of the concern two or three months after others and
some of these latter had separated from the concern before
the former had entered. The constant flux of the coming
and going, of the arrival and departure of the individuals in
this group, where the labor turnover was over three hundred
per cent per year, made it impossible to take any one series of
weeks' output, which weeks would be the same for all, as
measures of abilities, as was done in group 2. A study of
the records showed that many of the subjects began doing
very good work about the fifth week after entering the con-
cern. Consequently, the average number of sheets written
per day during the fifth and sixth weeks of each individual's
term of employment with the concern was taken as the
measure of abilities. These weeks, then, were not the same
calendar weeks for all individuals. The best day's work was
also selected as a measure of abilities, the best day being that



SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION 21

particular day in which each subject did her best work. The
following table gives the average number of sheets typed,
standard deviations, and ranges. The number of subjects,
sixty-five, was constant for both measures.

No. of S.D. Range

sheets typed

Daily average for 5th & 6th weeks 311.8 62.3 198.4-420.3

Best day's performance 361.9 65.3 258. -500.

D. Some remarks on the procedure.

The problem of securing groups in which the output of the
workers is accurately comparable. In typing it is difficult to
obtain a large group of typists in which all of the individuals of
the group are performing the same kind of work which can be
accurately measured, quantitatively and qualitatively. The
difficulty lies in the fact that the group chosen at random
is usually a heterogeneous one; that is, one individual does
specific work which differs from the specific work of another
individual, although both are performing the same generic
act typing. This difficulty has not entered into this investi-
gation since all of the girls were a part of a homogeneous
group and all were doing exactly the same kind of typing.
The measures obtained in this investigation were not subject
to the indeterminable error of a supervisor but were ob-
jectively obtained so that it was possible to state that one girl
was so many points better or poorer than another girl.

Testing at various periods of the development of practice.
In group 1 all of the subjects were tested when they had
practically no practice or knowledge of typing. In group 2
all of the subjects were tested after they had acquired con-
siderable practice and skill in typing. In group 3 the subjects
were a mixed group being made up of individuals possessing
varying degrees of practice and ability in typewriting.

Care taken to secure accurate measures of typing ability.
More than one measure of typing ability was used in each
group. In group 1 six measures, in group 2 three measures,
and in group 3 two measures of ability were used. In group
2 approximately 9100 records of daily output, and in group 3
approximately 3000 records of daily output were tabulated
and studied in order to obtain what appeared to be the best
measures of typing ability.

Wide mental differences of the groups of subjects. The
wide mental differences between group 1 and groups 2 and 3



22 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION

in educational training indicate that two widely different
classes in these respects were tested. It appears that the
highest and lowest mental grades of girls in the typing pro-
fession were included in this experiment.

E. Administering the tests.

a. General procedure.

The subjects in group 1 were tested within two months
after they had started their course. The subjects of groups
2 and 3 were tested after they had acquired considerable speed
in typing. All of the tests in groups 2 and 3 and all
of the tests but the number checking and the form substi-
tution in group 1 were given individually. The number
checking and the form substitution tests in group 1 were
given as class tests. The method of conducting the ex-
periment was simple. The subject was seated comfortably
at a table, on which the blanks were presented one at a time,
after it had been ascertained that the subjects thoroughly
understood the instructions. In the number checking, form
substitution and hard directions tests replies were filled in in
pencil. In the remainder of the tests the replies were spoken.
In the tests where the replies were spoken the experimenter,
who sat next to the subject, had a copy of each test and if
an inaccurate reply was made the subject was required to
correct it before giving a reply to the next stimulus word,
the subjects being told before hand that they would have to
do this. In the number checking test each group was checked
which contained any combination of the digits eight and nine,
b. Instructions to the subjects.

In the instructions to the subjects an effort was made to
combine the proper comprehension of the experiment by the
subject with an ideal uniformity of instruction. In the in-
dividual tests instruction was given by description, illustra-
tion, and execution. The subject was first clearly told the
meaning of the test, then the experimenter performed a
small sample of it, and finally the subject herself performed
a small sample of the work. These samples were presented
in typewritten form and the same samples used with all
subjects. Six illustrations of each test were used with all
subjects. In the class tests the subjects were simply told
what they were to do and the experimenter performed a
small sample of the work on the blackboard.



SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION 23

An attempt was made to impress upon the subjects the
fact that speed was the main object of the tests. They were
given a short talk on the applications of psychological tests
and were told that their mental reactions were going to be
timed and that their individual results were going to be
compared with the results of the rest of the class. They
were urged to go as fast as possible but were warned that
they would either have to correct their mistakes as they went
along or be penalized later for their mistakes (according
to whether a class or individual test) . It seemed that all the
subjects tried to make all the speed possible.

c. Records.

In all the tests in groups 2 and 3 and in all the tests but
the two class tests in group 1 the time was taken by a split
second watch reading to fifths of a second. The watch was
started when the experimenter judged that the subjects' eyes
met the first figure or word on the test sheet, which was un-
covered by the experimenter. In the class tests instructions
were given that as soon as anyone had finished she was to
raise her hand. When the first hand was raised the experi-
menter called upon the class to stop, having instructed them
beforehand that he would do this. Thus the time for all was
the same and the results were worked up for the amount
accomplished.



24 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION



4. Results and how obtained.
A. Scores in the tests.

The following table gives the scores made in the tests for
groups 1, 2 and 3.



Opposites
Verb-object
Agent-action
Action-agent
Mixed relations
Color naming
Hard directions
No. checking
Form subs.

No. c. Errors
Form s. '
Hard d. "

No. of subjects



Group 1
Av. S.D

28.56 8.26

31.43

31.92

33.06

76.03

53.81
110.01
151.52
124.40



7.91
10.09

6.75
28.65

8.99
28.21
28.27
13.11



4.70

.48

1.64

45



3.38
.67
.45



Group 2.

, Av. S.D.

39.00 12.25

38.59 10.26

40.61 16.21

46.33 11.83

99.47 34.30

62.09 8.17

186.25 60.00

157.03 34.72

150.92 21.70



8.20 7.62 .74 1.25

2.63 4.48 4.48 3.85

4.33 2.48 1.69 4.51

38



2.
X Z

.36 .48
.23 .29
.27 .60
.39 .75
.31 .20
.15 +.09
.69 1.13
.04 .23
.21 .65


Group
Av. S.D.
43.37 16.12
37.23 10.39
39.81 14.28
50.19 17.35
101.00 34.71
62.07 9.80
188.72 53.91
176.03 33.52
152.58 30.46


3.
X

.52
.18
.25
.51
.33
.15
.71
.16
.23


Z

.95
.31
.41
1.57
.21
.09
.91
.19
1.32



7.95 7.83 .68 1.32

3.17 3.22 5.46 3.81

4.00 2.96 1.44 5.58

65



Av. average score of the group in seconds and hun-
dredths of a second, in the tests from opposites down through
form substitution.* In the last three tests not the time but
the average number of errors and hundredths of an error
is given.

X percentage that the average of the group is above
or below the average of group 1. A + quantity indicates
that the average is below the average of group 1, or a better
performance, and no sign before the figures indicates that
the average is above the average of group 1, or a poorer
performance.

Z percentage that the standard deviation of the group is
above or below the standard deviation of group 1.

* Since in group 1 the grades in the number checking and form sub-
stitution tests are measures in terms of the amount accomplished (av-
erage in number checking 117.31, average in form substitution 76.31)
the figures, in the table under group 1, for these two tests, are roughly
derived, for means of comparison in this table only, by dividing the
time (constant for the group in each case) by the amount done and mul-
tiplying by the number of groups of figures in the number checking test,
and the number of figures in the form substitution test.

time
amt. done no ' of cases '

Groups 2 and 3 showed themselves to be greatly inferior
to group 1 in performing the tests. Groups 2 and 3 were made
up of girls who were working long hours for very low wages



SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION 25

and who came from families in poor circumstances. Twenty-
four per cent of the subjects of group 2 and twelve per cent
of the subjects of group 3 came from homes where a foreign
language was spoken. The average foreign-born perform-
ance, in each test, in each group, was compared with the
average performance of the rest of its group and was found
to be higher by no more than six per cent in any one test.
The inferiority in performing the tests in comparison with
group 1 could, therefore, not be traced to difficulty in using the
English language due to speaking a foreign language in the
home.

B. Treatment of errors.

At the outset of the experiment it was planned to treat
separately the time required to perform the test and the
number of errors made and to correlate the number of errors
made in the number checking, form substitution, and hard
directions tests with the grades in stenography, grammar,
and typewriting. The standard deviations of the number of
errors made in each of the tests being in every case almost
as high if not higher, than the average number of errors made,
no attempt to correlate the errors with the work has been
made since the number of errors made is so unreliable as a
measure.

C. Methods of combining the measures.

When several tests have been made of an individual's
abilities it is often desirable to show the success of the in-
dividual in the series of tests taken as a whole. A good way
of doing this is by the order of merit method. The measures
of the individuals are so arranged that it can be stated that
an individual stood thirteenth from the top in one test,
twenty-third in another test, eighteeth in still another test
and that her average rank was eighteenth. This method
was used in correlating performances in the tests with ability
in stenography and grammar.

Although the order of merit method is rather a rough
method it has worked well with certain kinds of material
in the past. But to transmute a series of quantitative
measures into an order of merit series is to throw away
a great deal of information contained in the series. Another
method consists in taking the average of the group as
zero and expressing the individual's standing as a devia-



26 SOME EMPIRICAL TESTS IN VOCATIONAL SELECTION

tion above or below this average. The measure of deviation,
usually the standard deviation, is taken as the unit and
all deviations are expressed as multiples of this unit.
What this method does is to assign to each individual a
quantitative position in the distribution of the group. This
method was used in getting reduced measures in the tests
which were correlated with ability in typewriting in groups
2 and 3 and with grades in typewriting in group 1.

D. Correlations.

a. Stenography and grammar.

I. Method used.

The rank differences formula was used in correlating the
tests with abilities in stenography and grammar,

6Sd 2
n(n 2 -!)

II. Correlations.

The following table gives the correlations with the tests
and abilities in stenography and grammar.

Steno. Grammar

Verb-object 36 .37

Number checking 07 .22

Color naming 34 .38

Action-agent 23 .35

Agent-action 19 .37

Form substitution 40 .16

Hard directions 46 .54

Mixed relations 31 .43

Opposites ... .45 .40

No. of subjects 45 45

b. Typewriting.
I. Method used.

In correlating the tests with abilities in typewriting, in


2 4

Online LibraryHerbert Wesley RogersSome empirical tests in vocational selection → online text (page 2 of 4)