Google
This is a digital copy of a book that was preserved for general ions on library shelves before il was carefully scanned by Google as part of a project
to make the world's books discoverable online.
Il has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books
are our gateways to the past, representing a wealth of history, culture and knowledge that's often diflicult to discover.
Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the
publisher to a library and finally to you.
Usage guidelines
Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to
prevent abuse by commercial parlies, including placing technical restrictions on automated querying.
We also ask that you:
+ Make non-commercial use of the plus We designed Google Book Search for use by individuals, and we request that you use these files for
personal, non-commercial purposes.
+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the
use of public domain materials for these purposes and may be able to help.
+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find
additional materials through Google Book Search. Please do not remove it.
+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just
because we believe a b<x>k is in the public domain for users in the United States, that the work is also in the public domain for users in other
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means il can be used in any manner
anywhere in the world. Copyright infringement liability can be quite severe.
About Google Book Search
Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers
discover the world's hooks while helping authors ami publishers reach new audiences. You can search through I lie lull text of this book on I lie web
at |http : //books . qooqle . com/|
STMFORD-WWEESITY-PBIARY
IHII 3EBDIilIll.-TllIF
SCHOOL OF EDUCATION
LIBRARY
Chinese Students in America:
Qualities Associated with Their Success
By
JENNINGS PINKWEI CHU, Ph.D.
Teachers College, Columbia University
Contributions to Education, No. 127
• ••
• • • •
r • • • • •
• • • • •
• • •■*
• * «■ - „ ■» »
Published by
GTeaefjera College, Columbia Untberrftp
New York City
1922
•*
.**>
■, ■ . c
Copyright, iQ22> by Jennings Pinkwei Chu
291319
c
• <
• • •
• •
• • •
• •
• • •
• •
ACKNOWLEDGMENTS
The author wishes to acknowledge his thanks to many Chinese
students in America for their willing co-operation in supplying a
large part of the necessary material for this study. To the authori-
ties of Tsing Hua College, Peking, and to the Directors of the Chi-
nese Educational Mission and of the Educational Bureau of the
Chinese Ministry of Education, Washington, D. C, he feels greatly
indebted for their permission to make use of many otherwise inacces-
sible records. He is grateful to Professor Edward L. Thorndike and
Professor Truman L. Kelley for much guidance in devising statistical
methods; to Dr. M. R. Trabue and Dr. I. L. Kandel for valuable
suggestions; and to Mr. J. S. Orleans and Miss Elva Wald, both of
the Institute of Educational Research of Teachers College, for tech-
nical assistance. But especially to Professor Edward L. Thorndike,
under whom the investigation was undertaken and whose stimu-
lating criticisms and timely encouragement have been both a guid-
ance and an inspiration, the author wishes to express his deep grati-
tude and indebtedness.
Jennings P. Chu
New York City,
April, 1922
CONTENTS
CHAPTER PAGE
I. Statement of the Problem i
II. Experimental Material 4
1 . The Subjects
2. The Data
The Judgment of Associates
Individual Records
Test Scores in English
High School Marks in China
College Marks in America
III. The Judgment op Associates 11
1 . Transmutation of Rankings into Sigma Positions
2. Comparison of the Judgments in English with Test Scores in
English: the Validity of the Judgments in the Knowledge of
English
3. The Reliability of the Judgment: Self-correlation
4. Raw Correlations
5. Correlations Corrected for Attenuation
6. Final Composite Correlations
7. Interpretation of Results
IV. Comparison op the Judgments with High School
Marks 34
V. Comparison op the Judgments with College Marks 38
1. The Judgments in Scholarship and General Averages in
Colleges
2. High School General Averages Compared with College General
Averages
VI. Comparison of the Judgments with Individual
Records 40
1. A Scale for Judging the Activities of Chinese Students in
America
2. A Discussion on the Validity of the Judgment of Associates
VII. Years Spent in America as a Factor 47
VIII. Conclusions and Suggestions 51
Appendix 54
Chinese Students in America:
Qualities Associated with Their Success
.% •
CHAPTER I
STATEMENT OF THE PROBLEM
For the past fifty years, China has made a practice bf sending
students to the United States to be educated in Americaa-colleges
and universities. Many psychological problems have arisfeji ^con-
nection with this practice, which, fundamental and impqrtgftt'.as
they are, have never been solved or even explicitly formulated. ,
A solution of them, however, would greatly aid in the improvement/
of methods for the preparation and selection of these students. ■'*;
With the steady growth of the number of Chinese students in
America and with the increasing emphasis placed upon the value of
their American education in recent years, these psychological prob-
lems have come more and more to the fore. The need of a logical
statement of some of them and of a scientific attempt towards their
solution is manifest and pressing. Fortunately, the science of
mental measurement, which has recently made great progress in
America, will directly lend itself to the solution of these problems
and indirectly open a new field in China for educational experimenta-
tion, of which the present investigation may be a beginning.
The problem in connection with the education of Chinese students
in America emanates from a consideration of their success as a result
of that education. Briefly stated, it comprises the following:
1 . What should be considered as success of a Chinese student in
America?
2. What are some of the qualities associated with it?
3. How are they related to it?
4. How are they related to one another?
5. How can the optimum amount of success be secured in the
light of our knowledge of its relationship to the qualities associated
with it?
6. What should be the proper method for the preparation and
selection of the prospective Chinese student in America?
CHAPTER II
EXPERIMENTAL MATERIAL
I. THE SUBJECTS
The subjects for this investigation were 664 Chinese students in
America distributed amdng the following college and university
centres: the number in parentheses after each college or group of
colleges is the number of subjects from that college or group of col-
leges studied in this investigation: Baldwin Wallace College (9);
University of Chicago (35); University of Cincinnati (6); Coe Col-
lege, Iowa (8); Cornell University (46); University of Detroit (11);
George Washington University (14); Grinnell College (4); Harvard
University (37) ; Johns Hopkins University and University of Mary-
land (10); Leland Stanford Jr. University (11); Lehigh University
(9); Marquette University, Milwaukee (14); Massachusetts Insti-
tute of Technology (26); Michigan University (45); University of
Minnesota (12); New York University (20); Northwestern Uni-
versity (7); Oberlin College (20); Ohio State University (17);
Oregon Agriculture College (7); University of Pennsylvania (22);
Pomona College (6); Purdue University (l8)j University of Roch-
ester (5); Schenectady Chinese Students' Club {3); Syracuse Uni-
versity (18); University of Washington, Seattle (14); University of
Wisconsin (6); Yale University (15) and a group of girl students
from Vassar College, Wellesley College, Bryn Mawr College and
Mount Holyoke College amounting to {9). In the accompanying
table is shown the distribution in these American colleges and uni-
versities of the Chinese students embodied in this study:
NUMBER OF STUDENTS NUMBER OF COLLEGES
over 50
40-49
30-39
20-29
10-19
4- 9
Experimental Material 5
It must be remembered that the numbers of students in the fore-
going, table are the ones that were embodied in this study, while
the actual numbers,in these colleges are in many cases a little larger
and in a few cases twice as large.
Among these students are -included both the undergraduate and
graduate students. High school students and those engaged in full
time teaching are excluded from this study. They all represent
students from China, of whom two-fifths depend, for their mainte-
nance, upon either government or special scholarships and the other
three- fifths upon their parents or relatives or upon themselves. In
the analysis of facts, no distinction is made as to the course of study
they take, or age and sex, or maintenance.
2. THE DATA
In order to make, a study of these subjects in respect to their
scholarship, leadership, knowledge of English and knowledge of
Chinese, it was found necessary to gather all the relevant data then
available. Inasmuch as the crux of the problem consists in the
finding of the proper relationships of these four qualities rather than
any objective measurement of them according to certain standard-
■ ized criteria, the effort was focused upon a procedure whereby such
relationships as should exist between them could be best secured.
Therefore, the judgment of associates in the form of rankings was
used, with all possible protection against certain defects inherent
in the method, and validated by such other possible measures then
available as high school marks, college marks, tests, and personal
records. These sources of data were secured in the following
manner.
The Judgment of Associates. The fact that the Chinese students
in America generally group themselves in different college centres
in small numbers, all associated closely with each other, and there-
fore thoroughly acquainted with one another, offers an opportunity
seldom found elsewhere for ranking. For this study, in the case of
those centres where the number of students exceeded twenty-four,
the total number was divided into several convenient groups on the
basis of mutual acquaintance or other affiliations. For the purpose
of ranking, a number larger than twenty-four would be too unwieldy
for any judge to make the ranking accurate, and a group smaller
than four could hardly make Lhe ranking significant. Consequently,
6 Chinese Students in America
the size of groups in this study varies from four to twenty-four. The
accompanying table will show the distribution of the number of
groups of these 664 Chinese Students in America.
SIZE OF RANKING GROUP
NO. OF
(i.e. No of Students Rated)
20-24
8
15-19
12
IO-14
14
5- 9
26
Less than 5
4
64
The form of the original directions for the ranking sheet is repro-
duced here.
Directions for the Ranking Sheet
On the next page you will find a list of names of Chinese students in your
locality or school, whom I presume you know well. Look the list through. If
there are persons whom you do not know well, omit their names by marking a
cross (X) after each. Omit also your own name by marking a cross (X) after it.
These marked names will not be considered. Having done so, please observe the
following procedure:
1. Look through the rest of the names. Decide which one in your judgment
would be the best in Scholarship (See Definitions below), and put a figure "1"
after his or her name in the Scholarship column.
2. Next decide which one would be the second best in Scholarship and put a
figure "2" after the name in the same column. Next decide which one would be
the third best in Scholarship and put a figure "3". Continue the same process
until the Scholarship column has in it figures 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and
so on, representing the ranks of these students from best to poorest in Scholarship.
3. Proceed likewise with each of the remaining columns, considering the rank
each person would have in Leadership, Knowledge of English, and Knowledge of
Chinese. (See Definitions below.)
Definitions
Scholarship: means ability in studies, as shown by school standing, degrees,
honors, and other signs or evidences.
Leadership: means ability to take initiative, willingness to assume responsibility
and ability to inspire confidence in others.
Knowledge of English: means knowledge to understand printed and oral English
and to express well orally and in writing.
Knowledge of Chinese: means knowledge to understand printed Chinese and to
express well in writing.
Experimental Material 7
RANKING SHEET
The ranking sheet is divided into five vertical columns, headed
with the following captions: Name, Scholarship, Leadership,
Knowledge of English, and Knowledge of Chinese, as shown on
page 14, in which it has been filled with rankings and their transmu-
tations.
In April, 1920, these ranking sheets were sent with return enve-
lopes to 926 Chinese students grouped under 62 college or university
centres. Out of these, 465 or a little over 50 per cent returns were
received. This means that on the average each person is ranked or
judged by half of the people in the group in which he is one. This
is evidenced by the later finding that the largest number of judg-
ments given to a person is 12 which is one half of the total number of
24 on the list.
The individual who is judging may not know well enough all the
qualities in the individuals judged. This results in a different num-
ber of people within a group being ranked by the same judge for
the four qualities. Thus judge X in group Q judges 22 people of his
group for scholarship, 23 for leadership, 23 for knowledge of Eng-
lish, and 14 for knowledge of Chinese.
Thus in all we find there are 656 people ranked in scholarship;
631 in leadership; 664 in knowledge of English; and 590 in knowl-
edge of Chinese. This would mean that the Chinese students
know most about the knowledge of English of their associates and
least about their knowledge of Chinese.
Individual Records. In connection with the ranking sheet, a ques-
tionnaire was sent out for the purpose of gathering individual records
which would later on lend themselves to the validation of the rank-
ings. The original sheet is here reproduced.
Questionnaire
1 . Name and address.-
2. School attending.^ Major course.-
3. Degrees obtained (where and when)
4. How many years of English did you have in China?
5. School in China where you got most of your English
6. When did you come to the U. S. A.? Month Year
7. Since coming here, if you have been president, vice-president, secretary,
general secretary, treasurer, or manager, please fill out the following:
8 Chinese Students in America
Name of Organization Name of Office Held Year to Year
8. If you have been editor, contributor, or reporter of any magazines, Chinese
or English, please fill out the following:
Name of Magazine Name of Position Held Year to Year
9. If you have written books either in Chinese or in English, please fill out the
following:
Title of Book Pages Publishing Co, Date of Publication
10. If you have been awarded any scholarship, fellowship, assistantship, or
instructorship, please fill out the following:
Position School Subject Year to Year
II. If you belong to any honorary American fraternities, please fill out the
following (e. g., Phi Beta Kappa, Sigma Xi, etc.):
Name of For Excellence in
Fraternity School When Elected What Subject
12. Mention any other honors which have come to you (e. g., graduated with
honor, etc.)
A brief inspection of the questions will show that each one has its
specific purpose and most of them are related to the four qualities
in our study. The evaluation of these questions is found in Chap-
ter VI.
Test Scores in English. Knowledge of English was the only
quality, of which an objective measure could be secured by means of
tests. Four sets of English tests were given to 42 of the subjects
from eight different college centres, namely, Harvard, Yale, Colum-
bia, New York University, University of Pennsylvania, Massachu-
setts Institute of Technology, Johns Hopkins University, and Uni-
versity of Maryland. These four sets are (1) Thorndike Alpha 2,
Part II, (2) Reading Test I 1 , (3) Reading Test II 1 , and (4) Thorndike
1 Reading tests I and II are of the type used in parts II and III of the Thorndike Intelligence
Examination for High School graduates, Series 1919-1923.
Experimental Material 9
Vocabulary Test. Alpha 2, Part II , is known to all who are familiar
with tests. Reading Test I and II are each made of four parts.
The vocabulary test is made of five parts. In order to show the
degree of difficulty of the reading and vocabulary tests, a part of
each is shown below.
Sample of Reading Test
Continuous improvements have enabled engineers to reproduce machinery of
less and less weight for the same power, and at the same time to reduce the spaces
required for its accommodation, the vibrations due to the working of the engines,
and the consumption of fuel per horse power. For engines of high power, quadruple
expansion has sometimes been adopted, scientific methods of balancing have been
employed, improved qualities of steel and bronze have been introduced, the rate
of revolution has been increased, and forced lubricating fitted. In the boilers
higher steam pressures have been used, superheating in some cases being resorted
to; the rate of combustion has been accelerated by supplying air under pressure
in the stockhold or in the furnaces, and sometimes by placing fans in the exhaust
to draw the air and products of combustion more rapidly through the fires; the
former being known as forced draught and the latter as induced draught. In the
navy, with the view of saving weight, water-tube boilers have been adopted, but
boilers of this type have not yet been generally fitted in the mercantile marine.
Steam pressures now in common use vary from 100 to 180 lb. per sq. in. in cargo
ships; from 140 to 220 lb. in passenger ships, including the large Atlantic liners;
from 210 to 300 lb. in large warships where water tube boilers are used; while in
the destroyers and other classes of warships in which small water-tube boilers are
used it varies from 180 to 230 lb. per sq. in.
1 . Of the improvements noted, which do you think is the main cause of the reduc-
tion in vibration?
2. What do you think is superheated?
3. What feature of an ordinary stove would correspond to the exhaust?
4. Give two words which mean the same as "fitted" as used in line 7.
5. Name one product of combustion which would not be drawn through the fire
by an induced draught.
6. Which of the reductions noted do you attribute most particularly to the
increased rate of revolution?
Sample of Vocabulary Test
Ai
Write Ha before every word that means something done with hands.
Write F before every word that means something done with feet.
Write E before every word that means something done with eyes.
Write M before every word that means something done with mouth.
Write T before every word that means a tool.
Write B before every word that means a part of the body.
Write S before every word that means a part of the house.
Write H before every word that means about ships, the sea, and sailing.
10
Chinese Students in America
pulverize
jamb
regurgitate
ratchet
espy
scourge
scour
spinnaker
clamor
gape
slander
gimlet
adze
perambulate
glower
laud
quaver
polka
lugger
gable
mandrel
spokeshave
lathe
pancreas
converge
prate
newel
cajole
alcove
transcribe
wainscot
expostulate
fibula
pirouette
ulna
roam
brig
jettison
calipers
larynx
calumnate
upbraid
purses
wherry
cochlea
limn
dermis
dhow
ilium
swage
High School Marks in China. High school marks in China were
secured for 84 of the subjects. These marks, covering four years
in high school, comprise (1) general averages, (2) marks in English,
and (3) marks in Chinese. They were used to check up the va-
lidity of the judgment in the corresponding qualities and for the
prediction of academic success in America on the basis of that in
China. The detailed nature of these marks will be described in
Chapter IV.
College Marks in America. Since not many students take any
course in English and certainly none would take any in Chinese in
America, the college marks in America are confined only to general
averages as indications of scholarship. Marks of 122 students were
secured from the Chinese Educational Mission and the Educa-
tional Bureau of the Chinese Ministry of Education, Washington,
D. C. They were marks of different subjects and covered differ-
ent numbers of years according to individual cases.
\
CHAPTER III
THE JUDGMENT OF ASSOCIATES
The judgment of associates in respect to a number of human
traits, to result in high validity and reliability, must needs meet
several rigorous conditions. First, there must be a thorough
acquaintance among those who judge one another; second, the
judgments must be comparable; third, the final judgment of a per-
son in a particular trait must be the average of as many judgments
as can be secured; fourth, there must be a common understanding of
the trait to be judged. In this study, the first condition is well met
by the unique situation of the Chinese students, in which they culti-
vate their familiarity to a high degree through meetings, discussions,
conferences and all kinds of publications both in English and Chi-
nese. The second condition has to be met by adequate statistical
treatment as will be explained later. The third and fourth condi-
tions are both rigorously conformed with.
I. TRANSMUTATION OP RANKINGS INTO SIGMA POSITIONS
When the rankings were returned, it was found that they were
extremely unwieldy and irregular. For instance, in the same qual-
ity, say scholarship, a person might be ranked as one among 8
people by one associate, and then as one among n people by
another. Supposing that he is ranked by both as the third, evi-
dently the third among 8 people is not as high as the third among 1 1 ,
for there are more cases in the latter. Again, it was also found that,
on the same sheet ranked by any judge, under scholarship 18 people
might be ranked; then under leadership 15 people might be ranked;
under knowledge of English perhaps 19 people; under knowledge
of Chinese most likely only 14 people. The difficulties with the way
of obtaining correlations by means of Spearman's formula for rank-
ing, are apparent. In order to surmount these two difficulties
pointed out above, it was imperative to transmute these ranking
values of 1, 2, 3, 4, 5, etc., into some other terms so that the third
among 8 will have a different value from a third among 11 and that
correlations could be computed not from the data of every sheet by
12 Chinese Students in America
Spearman formula but from the data of all sheets on the basis of the
number of judgments and by the Pearson product-moment formula.
The method for the transmutation of rankings into sigma posi-
tions can be found in Thorndike's Mental and Social Measurements,
pages 116 and 221 ff. For the convenience of this study, which
involves more than 2,000 such transmutations, a special table was
made which shows the different comparable values of each number
according to its place in groups of different sizes. If we look at
number "1" in Table I, we will see that it has values, 1. 10, 1.27,
1.40, 149, 1.59, etc., as it happens to occur in groups of 3, 4, 5,6, 7,
etc. On page 14 is shown the original ranking sheet filled out by the
judge with figures which, in turn, are transmuted into sigma posi-
tions according to Table I . The number to the left in each column
is the original rank. The signed number is the transmuted sigma
value. The value of this procedure can be seen when we begin to
add all the rankings of a person in a particular quality in terms of
the plus and minus sigmas, which, as we have just observed, are of
comparable values. Thus a person's final score in a quality is the
algebraic sum of the sigma values divided by the number of judg-
ments. To illustrate, if A's rankings in scholarship as given by five
judges are "5" among 10, "5" among 12, "6" among 10, "4" among
13 and "7" among n , then his average score in it in terms of sigma
positions will be the sum of .13, .32, —.13, .61, —.23 divided by 5,
or .14.
All the later correlations of judgments are made of these average
sigma values thus computed.
Judgment of Associates
13
0> M
g
C/5
O rt
•#-» 09
PO
«o ci
CI M
PO ^ O
ci 00
M M
M
1^
00
10
1-
10
PO
PO
PO
M
PO PO •*> »* 00 M
m c« po ^t "J e*
° 1 1 1 1 1 1
<o
00
1
O
M
1
PO
CO
M
1
CI
M CI
1 1
ci
CI
O 00
CI M
O © CI
M O 00
M M
1-
10
M
Ok 00
CI M
OO00 Ok M ^M1
OOMd^wtOOO
1 1 1 1 1 1 1
M
1
O
M
1
00
M
1
s
CI
1
M
CI
21?
CI M
00 00 Ok po
H ftMO
M
00
PO
CI
*
O
Ok ^f O 00 PO Ok 00
O W PO ▼ O t* Ok
1 1 1 1 1 1 1
00
M
M
1
M
1
O
M
CI
1
s
O >Q
CI M
M Ok 1^
M
ss
CI
PO
Ok
1
OkN lQO« ^>0
1 1 1 1 1 1 7
M
1
O
M
CI
1
M
2.
CI M
M O M
M ft (^
M
vt
M
CI
PO
M
PO
M
1
|»» M l/> M O M M
w ^r w» t* Ok m ^r
M M
1 1 I 1 1 I 1
O
M
1
00
M
ft 00
ft PO
M M
00 00
00
10
PO
CI
00
00
1
O
1
k» O 00 O CO N Ok
PO "> O 00 O PO Ok
M M M
1 1 1 1 1 1 1
1-
M
Ok m
ft ro
M M
^f M M
O 00 C
10
PO
M
O
M
1
M
PO
1
l/> M M ^f «J ft
^ 00 po ft
M M M
1 1 1 1 1 1
•
M
ft PO
M M
r* r*
O r* »o
M
Ok
PO
po 00
CI
00
O
1
PO
1
Ok
PO
1
t^ t^ O « ft
U) h O (0 ft
M M M
1 1 1 1 1
M
CI ft 00 CI M
Ot N Ok l^ m
M M
PO
PO M
O
M
1
PO
PO
1
M
10
1
« 00 ft w
t^ ft w ft
M M
1 1 1 1
2
w
si
S
CI <0
Ok *
M M
^r ft i>
0. <© *
ci
Ok
Ok
1
ci
1
1
*
1
ft w ft
M M
1 1 1
PO
M
00 CI
M M
00 M ft
00 ^0 po
ft
M
Ok
M
1
Ok M
PO
1 1
00
00
1
N 00
M M
1 1
04
03
25
M
00 M
M M
W W> PI
00 10 po
M
M
1
ci
PO
1
1
M
00
1
M
M
1
00
M
1
M
M
M M
00 M
M M
00 PO
t^ ^ ci
PO
CI
1
00
1
1
M
M
M
1
M
00
M
1
O
M
<0 t 00 ft PO
t^ O <0 <*> •*
M M
PO
M
1
Ok 00
PO
1 1
f.
M M
1 1
Ok
M 00 O Ok
i> Ok <o «
• • • •
M O
Ok
CI
•
1
•
1
00
Ok
•
1
M
•
M
1
00
PO Ok 00 •/» «/>
»0 00 * M M
M
1
00 Ok
t 00
1 1
s?
M
1
1-
Ok
10 00
• •
PO PO
• •
1
00
•
1
Ok
l/>
•
M
1
Ok 00
M
ci ci 00
ci «
1 1
3
M
1
«>
Q PO
•
PO
IS) ^
• •
O M
1 1
*•
CI PO
• •
M
ci f.
PO CI
• •
M
1 1
ro
M
M O
M
•
1
21
M CI
CO t M)«
t*
00
Ok
O
M
« po ^ «o <0 i> 00
ft
CI
M
CI
CI PO
CI CI
H
Chinese Students in America
Ranking Sheet
Read Directions and Definitions very carefully before you proceed. Use pencil.
NAME
SCHOLARSHIP
LEADERSHIP
KNOWLEDGE
OF ENGLISH
KNOWLEDGE OF
CHINESE
A
X
9
+ .35
X
X
B
X
IO
+ .23 ^
12
— .06
X
C
X
22
-1.52 '
20
— 1.20
X
D
II
- .13
13
- .13
14
- .29
16
~ .76
E
6
+ -55
6
+ .71
8
+ .41
15
- .60
F
I
+2.10
4
+ 1.04
3
+ 1.20
14
- .45
G
X
16
~ 47
13
- .18
I
H-2.IO
H
4
+ .90
5
- .86
. 9
+ .29
2
+1.45
I
16
- .90
20
— 1.04
21
-1.48
II
— .06
J
2
+1.41
2
+ 1.52
1
+2.21
8
+ .32
K
3
+I.II
15
- 35
15
- .41
12
- .19
L
15
- .71
14
- .23
17
- .67
6
+ .60
M
8
+ .27
18
- 71
7
+ .54
20
—2.IO
N
19
—2.10
21
-1.23
19
— 1. 00
19
-1.45
IO
0.00
7
+ .58
10
+ .18
9
+ .19
P
18
-1. 41
17
- .58
18
- .82
18
-1.15
Q
9
+ .13
19
- .86
16
- -54
17
- .94
R
13
- .41
11
+ .13
5
+ .82
3
+1.15
S
X
X
X
X
T
7
+ .41
1
+2.16
2
+ 1.48
5
+ .76
U
14
- .55
12
0.00
6
+ .67
10
+ .06
V
17
— I. II
23
—2.16
22
—2.10
13
- .32
W
5
+ -7i
3
+ 1.23
4
-fi.oo
4
+ .96
X
2
+1.41
8
+ 47
11
+ .06
7
+ .45
No. ranked
19
23
22
20
Your name Date,
COMPARISON OF THE JUDGMENTS IN ENGLISH WITH TEST SCORES
IN ENGLISH: THE VALIDITY OF THE JUDGMENTS IN THE
KNOWLEDGE OF ENGLISH
Before any correlations could be made with the average sigma
values given to each subject in the four qualities, it was first neces-
sary (1) to compare the rankings or judgments of these four qualities
with some other independent criteria of the same qualities, and (2)
to find out how well the judgments of each quality of a certain sub-
Judgment of Associates
15
ject agree with one another. The former is called the validity of the
judgment and the latter the reliability.
We will take knowledge of English for our first consideration.
As described in the last section the four sets of Thorndike English
tests were given. The scores in these English tests were then corre-
lated with the judgments, and the coefficients are shown in Table II .
TABLE II
Showing the Intercorrelations between the Judgments in English
and the Test Scores in English. N=42
THORNDIKE
THORNDIKE
READING I
READING II
ALPHA 2,
VOCABU-
•
II
LARY
J udgments in
English
.4 1 ±.084
.37 ±.088
.32 ±.09 1
.61 ±.064
.56 ±.069
.58 ±.067
49±.078
Reading II
.63^.063
.21 ±.094
Thorndike Alpha
2, II
.34 ±.089
Thorndike Vocabu-
lary
A brief inspection of Table II will at once show that the correla-
tions between the judgments and different test scores are approxi-
mately as high as the correlations among test scores themselves.
The average coefficient of the correlations between the judgments
and test scores is .43, and that of the correlations among test scores
themselves is .47. This means that the judgments in the knowledge
of English are highly valid in so far as they are compared with the
test scores in English. This would warrant our relying on the judg-
ments in English so far as their validity is concerned. But since
the cases here considered are rather few, and since it would be better
if we could secure other independent criteria to check up the judg-
ments, high school marks, college marks, and individual records were
also employed for our purpose, in which more cases were involved,
as will be described in later chapters.
As regards the validity of judgments in scholarship, leadership,
and knowledge of Chinese, since no objective tests were available,
academic and individual records were secured as criteria. In regard
16
Chinese Students in America
to leadership, it has so far remained such an elusive human trait
that at present the only good measure of its quantity and quality
that can still be resorted to is "a consensus of opinion" rendered by
the largest number of best qualified judges.
3. THE RELIABILITY OF THE JUDGMENT: SELF-CORRELATION
The validity of the judgments having been established, the next
step was to ascertain the reliability, or how well the different judg-
ments of a particular quality agreed among themselves. It is to be
TABLE III
Showing the Reliability Coefficients or Coefficients of Self-correlation
of 2 with 2 Judgments, 3 with 3, 4 with 4, 5 with 5
in the Four Qualities
SCHOLARSHIP
LEADERSHIP
KNOWLEDGE OF
ENGLISH
KNOWLEDGE OF
CHINESE
NO. OF
NO. OF
NO. OF
NO. OF
•
CASES
r
CASES
r
CASES
r
CASES
r
2 with 2
Judgments
118
.63 ±.037
121
.72 ±.029
1 10
.60 ±.042
Il8
.7 1 ±.032
3 with 3
Judgments
77
•73 ±.037
79
.66 ±.044
83
.88 ±.017
76
.79 ±.029
4 with 4
Judgments
53
.84 ±.027
44
.70 ±.044
52
.80 ±.034
27
.83 ±.041
5 with 5
Judgments
15
.85 ±.048
5
.82 ±.094
12
.92 ±.030
8
.85 ±.066
noted that the number of judgments on a certain quality varies
widely with different people; it ranges from 3 to 12 judgments. It
would be extremely laborious, if not impossible, to find out the corre-
lation of one judgment with every other for more than 600 people
in four qualities. As statistically justified, it was only necessary to
split the judgments into halves and correlate them. Thus all the
judgments on a quality of a subject were massed in a random fashion;
then the total judgments of 4, 6, 8, 10, 12 in number were split
into 2 with 2, 3 with 3, 4 with 4, 5 with 5, 6 with 6; these halves
were added, due regard being given to the plus and minus signs and
correlations made out of these halves. To illustrate, suppose a cer-
Judgment of Associates 17
tain subject receives the following judgments in scholarship: 1.32,
.23, 1.29, 1.32, 1.26, .68 as they come in random sampling. It is
only necessary to add the first three and next three and call the pair
2.84 and 3.26, the same procedure of thus pairing the judgments
being observed throughout. All the coefficients are contained in
Table III.
The significance of these coefficients can not escape our attention.
We notice two things: first, that the coefficients are substantially
high, showing the high reliability of judgments; and second, that
on the whole, as the number of judgments increases there is a tend-
ency of corresponding increase in the coefficients, which means a
corresponding increase in the reliability of judgments.
From the coefficients in Table III we may estimate the self-corre-
lations for an estimate from any given number of judges with
another estimate from that same number of different judges, the
two groups of judges being drawn at random. The formula to be
used is the usual Spearman prophecy formula or the formula for
self-correlation, as shown in the following:
r PM
P q+(p-q)r<i
in which
gis the number of judgmentswhose correlation coefficient is known ,
p is any number of judgments whose correlation coefficient is to
be found out,
U is the correlation coefficient of q judgments with q judgments and
tv which is unknown, is the correlation coefficient of p judgments
with p judgments.
To illustrate, let us take the facts from Table III. We find there
that the self -correlation coefficient of 2 with 2 judgments in scholar-
ship is .63. Based upon this fact what should be the self-correlation
coefficients when we have 3 with 3 judgments, 4 with 4 judgments,
5 with 5, 6 with 6 and so on? If we substitute the formula we have
the following:
, "2+(3-2).6 3 " 72
We can find out a series of values of r P as p becomes 4, 5, 6, 7, 8,
9, 10, 11, 12, etc., on the basis of q = 2 and r q = .63. This series of
values appears in Column II of Table IV-4 .
Chinese Students in America
TABLE IV
Theoretical Selk-Correlation for Scholarship According to the Number
of Judgments, Estimated from Empirical r's for' 2 with 2 (.63),
3 with 3 (.73), 4 with 4 (.84) and 5 with 5 (.85).
I
II
III
IV
V
VI
VII (VI +4)
VIII
OF JUDG-
9=
q>
3<
g,
S's
CAL VALUE
theoreti-
cal VALUE
3
4
5
6
7
8
9
10
630
720
773
810
836
856
872
884
895
9<>3
911
643
12°.
7«3
819
8+4
S64
878
890
899
908
915
724
798
840
868
887
9'3
922
929
939
■940
S94
773
819
850
872
888
901
gn
919
93a
3
3
3
3
3
3
3
3
3
3
691
315
347
439
5">
564
607
673
698
673
755
804
837
860
878
891
902
911
918
9=5
668
750 '•;
799 '
830
856
871
885
898
909
913
918
1
II
III
IV
V
VI
VII (VI+4)
VIII
NUMBER
JIDGMENTS
2'
Si
$
5'
OF
fl'S
CAL VALUE
THEORETI-
CAL VALUE
3
4
5
6
7
8
9
II
73o
808
844
877
890
904
915
924
931
937
942
366
721
764
795
820
838
858
866
877
886
538
636
744
778
803
824
840
854
865
875
642
733
2 8 5
B20
845
864
879
891
jog
3
3
3
3
3
3
3
3
3
480
836
°5°
305
308
39t
45«
513
553
588
630
709
763
SOI
827
848
864
878
888
897
905
63 3
714 .■>.
768
803
828
849
864
876
889
896
900
Judgment of Associates
TABLE IV
I
II
III
IV
V
VI
VII (VI +4)
VIII
NUMBER
or
TUDGMENTS
3i
9>
$
q>
SUM
OF
GENERAL
THEORETI-
FITTED
3
4
5
6
7
8
9
13
6oo
693
730
790
818
8+n
«57
871
S93
S99
831
880
907
934
936
945
951
957
961
964
967
667
750
800
833
HS7
*rs
689
8Q9
909
917
924
831
874
902
942
949
954
958
965
3
3
3
3
3
3
3
3
3
3
919
197
359
467
543
603
646
681
710
735
755
730
799
840
867
886
901
912
m
934
939
'783
.83O
859
■879
■893
.901
!qi8
■935
Theoretical Self-Correlation for Knowledge of Chinese,
Obtained as in Table IV A
I
II
III
IV
V
VI
VII (VI +4)
VIII
NUMBER
JUDGMENTS
g.
Q>
3*
?•
SUM
OF
g's
GENERAL
THEORETI-
CAL VALUE
FITTED
THEORETI-
CAL VALUE
3
3
4
5
6
7
8
9
13
710
786
831
860
SSo
896
907
917
925
931
937
7M
834
861
882
898
909
9'5
933
938
709
•785
.830
859
.879
-895
■907
.916
.924
-930
■936
694
773-
819
850
872
888
900
911
919
93i
2.837
3134
3.314
3-431
3-313
3-577
3.635
3-659
3-694
3-7 '9
3-742
707
784
839
855
878
894
906
915
924
93°
936
.786
.831
.860
.880
.896
.907
.917
.935
-93 1
-937
20 Chinese Students in America
In a similar manner we find values of r» as p varies on the bases
of the obtained correlations (when q is 3, r t is .73; when q is 4,
r t is .84; when q is 5, u is .85) in the case of scholarship. These
four sets of values are shown in Table XV A , in which, for instance
Column II reads: If the observed coefficient of ?i is .63, then by
applying the Spearman prophecy formula the coefficients, r%, r t , ri,
etc., will be .720, .773, .810, etc., as p becomes 3, 4, 5, etc.
By using the data of Column I , Table \VA as abscissa and the
data in the Columns II, III, IV, V as ordinates, four curves can be
plotted, each of which will show the theoretical rise in self-correla-
tion in the judgment of scholarship. The most probable curve will
be the one plotted by combining the four altogether and fitting to
the empirical .63, .73, .84, and .85.
In Figs. 1, 2, 3 and 4, the solid lines show the newly fitted theo-
retical curves and the dotted lines the observed correlations. The
values for these newly fitted theoretical curves are called "fitted
theoretical values," as shown in Column VIII, Table IV. The
advantage of these fitted theoretical values is that they give the
m
„
■■
-£
■sr
s
M,
'/
3 .
rmfjfrm***
MA.T
JMHTO.
•nan
P „
r8XSSf»1H r SHSSmullS ""'"" ""
:
■
1
Fig. I. Curves Showing the Fitted Theoretical and Observed Self-correlation
Values for Judgments in Scholarship.
Judgment of Associates 21
coefficients from judgments of 2 in each half to 12 in each half, thus
aiding us greatly in computing correlations corrected for attenuation.
The data in Table IV and in Figs. 1, 2, 3 and 4 show that the
obtained values in Chinese agree best with the "fitted theoretical
values" in it, while the obtained values in other three qualities do
not agree so well with their corresponding fitted theoretical values.
This would mean that the judgments agree best on the knowledge
of Chinese, the reason being most probably that Chinese, the
mother tongue, stands out as a prominent quality, in respect to
which Chinese students come to know one another moat readily and
a competent judgment of which can be rendered in a most valid and
reliable way.
It may be contended that, as stated previously, there was in
total a less number of judgments in knowledge of Chinese, showing
that the quality was not so widely known to judges as others; but
this does not change the fact that the judges were better able to rate
knowledge of Chinese, so that when the quality is judged, it is
judged in a more thorough manner. The obtained values in scholar-
„
Kf"*
•sr-
«~
is-
m
vja
«,
M,
<
„„„
Fig. 2. Curves Showing the Fitted Theoretical and Observed Self-correlation
Values for Judgments in Leadership.
22 Chinese Students in America
ship stand second in conformity with the fitted theoretical values
therein. This would seem to indicate that scholarship, as easily
evidenced by school marks, degrees, and rewards, with which the
students are fully acquainted., does not seem to be vague to them.
The obtained and fitted theoretical values in English stand third
in agreement. We are inclined to believe that they should agree
even better, inasmuch as English is a daily spoken language for
these students. It is to be noted, however, that the correlation
coefficient of 5 with 5 judgments in the knowledge of English runs
as high as .92; it is a clear indication that when there are 10 judg-
ments in the knowledge of English the number is highly sufficient.
The obtained and fitted theoretical values in leadership agree
poorest, which, however, was not unexpected. We are all aware
of the vagueness of the meaning of leadership and the lack of stand-
ards to judge by. Still, however, there is a tendency of the obtained
values to rise as the number of judgments increases.
Taken as a whole, the reliability of judgments on these four quali-
ties, as described above, has led to two conclusions: first, the relia-
^
u.
,„
7
N?
/
*T~
-Jri^
/
s
/
E
8
f^n™7r,™vi™"°™
1
,
1 —
■ —
!
FlG. 3. Curves Showing the Fitted Theoretical and Observed Self-correUtio
Value for Judgments in Knowledge of English.
Judgment of Associates
23
bility is very high, and second, on the average, it has accorded with
our general statistical principle that the increase in judgments is
accompanied by a corresponding increase in correlation coefficients.
That is, the judgments are both sufficiently valid and reliable to
justify our reliance upon them.
The validity and reliability having both been established, the
next step is to compute the raw correlations among these four
qualities.
4. RAW CORRELATIONS
When we come to compute the raw correlations among these four
qualities we are confronted with the fact that the judgments given
to different people in each quality vary as much as from 3 to 12.
An average of 3 judgments naturally cannot have as high a relia-
bility as that of 5 judgments; and that of 7 judgments likewise
cannot be compared with that of 9, etc. It would be statistically
erroneous to include under one study judgments by varying num-
bers of judges. It was found proper, consequently, to divide the
JS.
U
>
^S
*r~
' er "
/
1
s
!
i
Fig. 4. Curves Showing the Fitted Theoretical and Observed Self-correlatio
Values for Judgments in Knowledge of Chinese.
24
Chinese Students in America
subjects, for purposes of correlation, into four large groups in each
of the four qualities. They are (i) those judged by either 3 or 4
judges; (2) those judged by either 5 or 6 judges; (3) those judged by
either 7 or 8 judges; and (4) those judged by 9, 10, 11 or 12 judges.
All correlations were computed by Pearson's product-moment
method. The coefficients are presented in Table V.
TABLE V
Showing the Raw Intercorrelations of Scholarship, Leadership,
Knowledge of English and Knowledge of Chinese
CORRELATION
3 or 4
judgments
5 OR 6
JUDGMENTS
7 OR 8
JUDGMENTS
9, 10, II OR 12
JUDGMENTS *
t
OP
NO. OP
CASES
r
NO. OP
CASES
r
NO. OP
CASES
r
NO. OP
CASES
r
Scholarship with
232
.578 ±.030
I8S
.536 ±.036
130
.331 ±.053
62
.416 ±.072
Scholarship with
Knowledge of
English
200
.533 ±.031
180
.609 ±.032
129
.506 ±.044
59
.695 ±.047
Scholarship with
Knowledge of
Chinese
212
.334 ±.043
180
•4S3±.0?9
128
.307 ±.054
58
.139 ±.088
Leadership with
Knowledge of
English
242
.525 ±.041
197
.583 ±.033
129
.526 ±.043
4«
.705 ±.040
Leadership with
Knowledge of
178
.229 ±.049
198
.347 ±.043
120
.ISO ±.060
51
.31 1 ±.085
Knowledge of Eng-
lish with Knowl-
edge of Chinese..
177
.027 ±.051
200
.014^.048
134
.078 ±.059
63
.01 1 ±.085
A discussion of the significance of these coefficients, as shown in
Table V, will be presented in Section 7 of this chapter.
5. CORRELATIONS CORRECTED FOR ATTENUATION
All correlation coefficients in Table V, it should be noticed, are
subject to errors due to paucity of measures. What we are con-
cerned with in this study is what would be the true correlations
among these four qualities, were it possible to secure all the possible
competent judgments in each quality. In practice, such correla-
Judgment of Associates 25
tions cannot be secured. Mathematically, however, we can com-
pute such true relationships between these qualities by means of
Spearman's formula for the correction of the correlation coefficients:
y/fx K y x . Tx& t . YxiVt . ?x&i
Yxy—
VTxft . Ty x y t
1
Another form of this formula is:
yTx x y t . Yx t y x
Txy —
ytx x X t . Ty x y t
Xi and X2 are the two independent series of measures of quality A;
yi and y% are the two independent series of measures of quality B.
To illustrate the application of this formula to our problem, it is
necessary to refer to Tables IV and V. Let us work out, for example,
the true correlation between scholarship and leadership of those sub-
jects uniformly judged by 3 or 4 judges. Then r% x % t will be .775, the
average of .750 and .799, as contained in Column VIII, Table IV4
the fitted theoretical values for 3 and 4 judgments in scholarship;
ry x y t will be .741, the average of .714 and .768, as shown in Column
VIII, Table IVB, the fitted theoretical values for 3 and 4 judgments
in leadership.
In this particular case, the values rx x y t and r%# x were not obtained
since they are by the nature of the case approximately the same and
equal to the raw correlation txy. We therefore write
Vr Vl * t . rx# x = V(.578) 2 =.578
By substituting these values, the Spearman formula for the cor-
rection of correlation coefficients is greatly simplified,
r -578
which is the correlation corrected for attenuation between scholar-
ship and leadership in the case of those judged by either 3 or 4
judges. The rest of the correlations are likewise corrected, as shown
in Table VI.
26 Chinese Students in America
TABLE VI
Showing the Correlations Corrected for Attenuation between
Scholarship, Leadership, Knowledge of English, and
Knowledge of Chinese
3 or 4
5 or 6
7 or 8
9, IO, II OR 12
JUDGMENTS
JUDGMENTS
JUDGMENTS
JUDGMENTS
correlation of
f
f
f
f
Scholarship with Leader-
ship
.764
.648
.382
.464
Scholarship with Knowl-
edge of English
.684
.713
.572
.762
Scholarship with Knowl-
423
.530
.338
.151
Leadership with Knowl-
.679
.693
.600
.78I
Leadership with Knowl-
.296
.413
.171
.343
Knowledge of English
with Knowledge of Chi-
ese
-O33
.016
-.087
.012
A brief inspection of these corrected correlations as compared
with the observed correlations (Table V) , shows that on the whole
the values of the former are only slightly higher than those of the
latter, which would indicate that the obtained measures attain a
high degree of approximity to the true measures.
6. final composite correlations
All the correlation coefficients in Table VI , though they have been
corrected for attenuation, represent various values of single true
correlations, due to the difference in measures. It is desirable to
give one single expression of the correlation in each case. This
single expression — the final composite correlation — we compute
as the average of the four separate correlations of Table VI each as
the square root of the number of cases.
The final composite correlations thus obtained are shown in
Table VII.
Judgment of Associates
27
TABLE VII
Showing the Final Composite Correlations between Scholarship, Lead-
ership, Knowledge of English, and Knowledge of Chinese.
correlation of
NO. OF CASES
r
Scholarship with Leadership
609
628
578
616
547
574
.3Q2±.OSO
Scholarship with Knowledge of English
.7IQ±.027
Scholarship with Knowledge of Chinese
.389±.047
.679 ±.022
Leadership with Knowledge of English
Leadership with Knowledge of Chinese
.309 db. 030
— .024 ±.014
Knowledge of English with Knowledge of Chinese. .
The deviation of each correlation coefficient in Table VI from the
corresponding final composite correlation in Table VII was obtained.
The sigma of these deviations was computed. The probable error is
therefore this sigma divided by Vw (in this case 4) and multi-
plied by .6745. Thus, for the correlation between scholarship and
leadership, the calculation is as shown in the accompanying table:
Corrected
Deviations of
Correlation
Column (i) from
<P
Coefficients
.592 of Table VII
.764
+ .172
.029584
.648
+ 056
.OO3136
.382
— .210
.O44IOO
.464
-.128
.OI6384
Sum d* = .093204
V .093204
4
P. E. = .6745— =^- - .674 -f - .050
7. interpretation of results
The correlation coefficients, as shown in Table VII, are significant
and certain interpretations of them with respect to each one may be
made here.
a. The correlation between scholarship and leadership is rather
high (r= .592 ±.05). This high correlation coefficient clearly indi-
28 Chinese Students in America
cates that academic work has a significant bearing upon non-aca-
demic work. But it has an even more significant meaning:
it indicates that the non-academic work of Chinese students in
America is of a highly academic type instead of a social, recrea-
tional , and athletic type , as is the case with American students . The
non-academic activities in which Chinese students in America
generally participate and develop their capacities for leadership
include, almost exclusively, academic discussions, speeches, debates,
writings, etc., to which scholarship evidently contributes a large
part. On the other hand, non-academic activities such as athletics,
social affairs in various forms, in which leadership would require
qualities different from those essential to scholarship, are rarely
participated in by Chinese students in America.
The writer is not aware of any experiment that has been conducted
to show the relationship between scholarship and school leader-
ship. Studies have been made between scholarship and leadership
after the individuals have graduated from college through Phi Beta
Kappa and Who's Who records; they, however, are somewhat dif-
ferent from our study in that the leadership element considered
in them is of the after-school period.
Another significant fact that this rather high correlation brings
out is a confirmation of the traditional practice in China of predict-
ing non-academic leadership almost exclusively upon the basis of
academic attainment. For centuries past, the selection of adminis-
trative officers in China had been purely determined by accomplish-
ments along academic lines. The old examination system was
typical. In the Western countries, leadership in all walks of life is
recognized irrespective of academic attainment though the latter
has its important part to play. In China, or in the minds of the
Chinese, it is hardly conceivable except in rare cases that recognition
be given to non-academic leadership without first ascertaining the
academic background. Just as the Greeks believed that the philoso-
phers should rule, so do the Chinese hold that the intellectuals or
literati should lead. It is within the limit of reason to expect that
if a scientific determination is to be made in China among Chinese
students there to find out the relationship between scholarship and
leadership, the correlation coefficient will be still higher than .592.
b. The correlation between scholarship and knowledge of English
is high (r= .7 1 9 =b .027) and the correlation between scholarship and
Judgment of Associates 29
knowledge of Chinese is low (r= .389 =b .047), the former being the
highest of all the correlations obtained. These two correlation
coefficients tend to settle the long-standing dispute between two
theories: one maintains that knowledge of Chinese which carries
with it knowledge of Chinese history, Chinese philosophy, Chinese
attitudes, interests, habits, forms a valuable asset in the acquisition
of Western knowledge, and that knowledge of English is not essen-
tial; the other maintains just the opposite, that is, knowledge of
English is essential. These two correlations found in this study are
undoubtedly in support of the second theory. Their significance lies
in the fact that while knowledge of English is more causal to the
degree of scholarship, knowledge of Chinese is not a good measure of
the scholarship of a Chinese student in America, since in all
probability the correlation found (.389) is due in large part to
the influence of intelligence, which is undoubtedly highly correlated
with both scholarship and knowledge of Chinese. Thus, if it were
possible to "partial out" the influence of intelligence, the correlation
between scholarship in America and knowledge of Chinese would be
much nearer zero. The explanation for the existence of the high
correlation between scholarship and knowledge of English is simple.
Scholastic attainment in America presupposes, other things being
equal, an efficient instrument to work with and a good method to
work by. The former consists of a knowledge of English as a lan-
guage and the latter consists of proper attitude, interests, habits,
and ideals acquired through a knowledge of English. The extent
to which a Chinese student acquires this method and instrument
conditions the extent to which he will succeed in understanding,
appreciating, and mastering readings and lectures in English and in
approaching and solving problems that are Western in nature. It is
to be noted that while these two correlations obtain among Chinese
students in America they may not hold true among Chinese students
in China. While no scientific determination has been made, it is
very likely that among students in China knowledge of Chinese is
more closely associated with scholarship than knowledge of English.
c. The correlation between leadership and knowledge of English
is high (r= .679 ±.022) and the correlation between leadership and
knowledge of Chinese is low (r = .3091L .030) . The situation here is
somewhat similar to that described in the preceding paragraph.
Other things being equal, leadership in America demands an instru-
30 Chinese Students in America
ment and a method, or a language and an attitude. The language is,
of course, English which Chinese students in America generally
speak, and the attitude implies all those modes of behaviour, con-
duct, and bearing that are Western . This statement is evidenced by
the use of English in most, if not all, of the discussions, writings and
speeches among Chinese students here and by the adoption of ways
of appealing, convincing, and influencing people, which the students
have to acquire anew in this country.
A correlation of .309 between leadership and knowledge of Chinese
shows that if intelligence, which is decidedly correlated highly with
both leadership and knowledge of Chinese, is "partialed out,"
knowledge of Chinese will be correlated extremely low with leader-
ship in America. On the other hand, if a scientific determination is
to be made among Chinese students in China on the relationship of
leadership with knowledge of English and with knowledge of
Chinese, the coefficients may be just the reverse; that is, leadership
is correlated very low with knowledge of English and rather high
with knowledge of Chinese. This statement, however, still awaits
further verification.
d. The correlation between knowledge of English and knowledge
of Chinese is virtually zero (r= — .024=11.014). The existence of
this slightly negative or substantially zero correlation raises two
fundamental questions in respect to the teaching of English and of
Chinese: First, is there any fundamental difference between the
English language and the Chinese language? Second, is there any
great divergence of emphasis upon and time in teaching these two
languages?
In the attempt to answer the first question, it may be said on the
positive side that English and Chinese are two totally different
languages. In writing, English goes horizontally from left to right,
while Chinese goes vertically from right to left. English is pho-
netic, while Chinese is ideographic. English grammar has been quite
fully developed; Chinese grammar has not had concise formation
and has not been taught as such. In English, the written and spoken
languages are almost, if not perfectly, identical; in Chinese the
gulf between the two has been as wide as that between the language
of Paradise Lost and modern American journalistic language.
The English words are mostly polysyllabic, but withoutexception the
Chinese characters are monosyllabic. By virtue of these differences,
Judgment of A ssociates 3 1
it may be inferred that neither identity of procedure nor identity
of elements would or could exist in the process of learning these two
languages. Thus the lack of transfer tends to reduce the correla-
tion between these two languages to zero.
The second interpretation of the lack of correlation between
English and Chinese is found in the existing fact that the emphasis
upon and the time in teaching these two languages in different
schools in China have been so varied. Essentially, there are three
types of schools in China which, for the sake of argument, let us call
types "A," "B," and "C." A short description as to how English
and Chinese are taught in these schools may be of service here. In
type "A" school, the students commence to learn Chinese very
early in their school life and master it thoroughly. It is not until
very late, just prior to their coming to America, that they begin to
learn English. The preponderance of emphasis upon and time
devoted to the study of Chinese produces the type of Chinese stu-
dents in America that generally excels in the knowledge of Chinese
and is handicapped in the knowledge of English. To type "A"
belong the old-fashioned Chinese schools, most of which are gov-
ernment schools. In type "B" school, students generally have
acquired a good knowledge of Chinese, but they also begin to learn
English fairly early with due emphasis upon it. As a result, these
students possibly possess a corresponding attainment in their knowl-
edge of English and of Chinese. To type "B" belong most of the
modern schools in China, both government and private. In type
"C" school, knowledge of English stands out as a paramount feature
in the curriculum, overshadowing and crowding out the rest of
school studies, among which is included the study of Chinese.
The product of this type of school is a group of students well versed
in English, especially in spoken English, but poorly prepared in the
knowledge of Chinese. To type "C" belong most of the missionary
schools in China. Now, if we make a study of the correlation be-
tween Chinese and English among students from the type "A"
schools, it will be highly negative; the same relationship will exist
among students from the type "C" schools. But if the correlation
between English and Chinese is to be worked out among students
from the type "B" schools, it will in all probability be highly posi-
tive. The question now is how the students from the three types of
schools are distributed in America. It is recalled that question 5
32 Chinese Students in America
in the questionnaire (page 7) attempts to find out the school in
China where each student acquired most of his knowledge of Eng-
lish, which will also indirectly indicate the particular type of student
he is. A distribution of these three types of students in America
according to the types of schools they come from in China is made
as shown below:
1. Students from type "A" schools in China — 20 per cent
2. Students from type "B" schools in China — 52 per cent
3. Students from type "C" schools in China — 28 per cent
The number of Chinese students who were born and brought up
in America are so few that they are negligible for our purpose . Now ,
if we add the students from schools of types "A" and "C," they
comprise 48 per cent of the total number of students included in this
study or practically one half; the students from schools of type "B M
are roughly another half of the total. The correlation between
English and Chinese among "A" and "C" students will be negative
while that among "B" students will be positive. A correlation of
something around zero is therefore highly probable when these three
types of students are assembled in one study.
These two interpretations evidently represent two theories: one
attributing the zero correlation between knowledge of English and
knowledge of Chinese to the intrinsic and fundamental differences
between these two languages; the other to external and artificial
causes. If the former is true, then the correlation, as it exists
now, will persist permanently so long as English and Chinese
remain fundamentally different. If the latter is true, the zero
correlation can be increased from a low positive value to a high
positive value in proportion as human effort is brought to change
the environmental influences. The writer is inclined to believe
that the latter interpretation is more likely to be true, though the
former is not to be ignored. The correlation probably can be
raised considerably higher on the positive side, but it is unlikely
that it could be raised as high as that between, say, English and
French or even between Russian and Spanish.
e. One of the interesting points of this general problem is the
correlation of leadership in student days in America with leader-
ship in future years in China. This particular phase of the
problem cannot be solved now, for the records of the older gen-
Judgment of Associates 33
erations of Chinese students in America both for the student
period and for the period after their return to China are incom-
plete. The writer hopes, however, to take up this question by
preserving the records of the six hundred and sixty-four subjects
of this study and by following their careers upon their return to
China. In the meantime, it may be said that in all probability
the dependence of leadership upon general intelligence will make
this correlation high and that the dependence of leadership upon
the formation of habitual responses to enviromental factors will
will tend to make the correlation low. The latter part of the
statement will become clearer when it is remembered that the
American environment is at many points different from the Chinese
environment, and that consequently responses which are effective
in either one of the two environments may be ineffective or even
handicapping in the other.
CHAPTER IV
COMPARISON OF THE JUDGMENTS WITH
HIGH SCHOOL MARKS
In Chapter III we showed the validity of the judgments of asso-
ciates in English by comparing them with test scores in English.
The purpose of this and the following chapters is to prove the valid-
ity of the judgments of associates in the other three qualities,
namely: scholarship, leadership, and knowledge of Chinese.
The validity of the judgments in the qualities of scholarship,
knowledge of English and knowledge of Chinese will be proved by
comparing them with the high school marks from China of 78 of
the subjects who studied in the same school. The students from
this school are fairly representative of the Chinese students in
America. These marks cover a period of four years in the high
school, comprising the following: (1) general average in High School
I, II, III, IV; (2) marks in English in High School I, II, III, IV;
and (3) marks in Chinese in High School I , II , III , IV. The average
mark of any student in any subject, say, English, is the average of
the marks of English for four years and this average is taken as an
index of the knowledge of English of that student. The following
tables show the averages, standard deviations, and corrections for
(A) general average, (B) English, and (C) Chinese.
A brief inspection of the foregoing tables will show that in each
case the averages are approximately the same; so are the standard
deviations. The correction in each case is computed by comparing
the average of each year with the general average of the four years
and finding the deviations therefrom with either a plus or minus
sign, as the case may be. In order to make the work less laborious
and much simpler, these corrections can be used for computing the
final average in each subject for each person. Thus, if A has in four
years in the high school the following marks in English: 88.5, 86.3,
80, 79. 5, then these values, when corrected according to the cor-
rections in Table VI I LB, will become 88.5, 70, 84.3, 83, 80.5, which
are averaged again to be 85 as a single index of A's mark in English
in the four years in the high school.
Comparison with High School Marks
35
TABLE VIII
Showing the Averages, Standard Deviations, and Corrections of High
School Marks for Four Years in (A) General Average,
(B) English, and (C) Chinese
A
GENERAL average
H. S. I
H. S. II
H. S. Ill
H. S. IV
No. of Cases...
130
146
109
102
81.83
83.53
79.48
82.68
Standard
Deviation . . .
44
4-5
5.1
4.3
Correction
—2
+2
— 1
B
ENGLISH
H. S. I
H. S. II
H. S. Ill
H. S. IV
No. of Cases. . .
131
I89
154
149
Average
80.85
82.74
77.88
79.85
Standard
Deviation . . .
51
6.2
6.2
50
Correction
— 2
+3
-hi
C
CHINESE
H. S. I
H. S. II
H. S. Ill
H. S. IV
No. of Cases. . .
119
179
172
137
80.46
8I.4I
82.01
78.45
Standard
Deviation . . .
74
7.1
72
71
-hi
— 1
4-3
The marks for the general average and for Chinese are computed
in a similar manner.
It must be noted here that in comparing the high school marks
with the judgment, in order to find out the validity of the latter,
the general average mark in the high school is taken to be indicative
36
Chinese Students in America
of scholarship, the mark in English to be indicative of knowledge
of English, and the mark in Chinese to be indicative of knowledge
of Chinese. The correlations thus computed are shown in Table IX.
TABLE IX
Showing the Correlation Between the High School Marks and the
Judgment of Associates in Scholarship, Knowledge of
English, and Knowledge of Chinese
correlation of
no. of cases
r
H. S. General Averages with the Judgment in
Scholarship
68
76
68
.42 1 db .068
.307^.070
.374 ±.07 1
H. S. Marks in English with the Judgment in
Knowledge of English
H. S. Marks in Chinese with the Judgment in
Knowledge of Chinese
The coefficient of .307 in Table IX shows that the judgment in
English does not correlate so highly with high school marks in Eng-
lish as it does with test scores in English, for, if we refer back to
Table I, the average correlation coefficient of the judgment with
the four English test scores is .43. But still the difference is not so
great as to make us doubt the validity of the judgment as com-
pared with high school marks.
The three correlation coefficients of .421, .307, and .374 are
enough to warrant us to rely upon the judgment in scholarship and
knowledge of Chinese.
By using the high school marks, inter-correlations between the
general averages, marks in English, and marks in Chinese were also
computed, as shown in Table X.
TABLE X
Showing the Correlations Between the High School General Averages,
Marks in English and Marks in Chinese
correlation of
NO. OF CASES
r
H. S. General Averages with Marks in English. . . .
H. S. General Averages with Marks in Chinese ....
H. S. Marks in English with Marks in Chinese ....
•
84
72
77
.683 ±.039
.190=*= .077
.191 ±.074
Comparison with High School Marks 37
The rather high correlation coefficient of .683 between general
averages and marks in English as compared with that of .719 as
secured on the basis of the judgment is interesting. It shows very
conclusively that the judgment of associates and high school marks
agree very well in their estimation of scholarship and knowledge of
English. And this agreement adds, of course, more weight to the
validity of the judgment.
The low correlation coefficient (r = .190) between general averages
and marks in Chinese gives weight to the fact that Chinese has a
very small part to play in the estimation of scholarship, or better
still, in determining the general averages in that particular high
school. And such is the actual case. For, to the knowledge of the
writer, in that high school, the department of Chinese is a separate
and distinct unit which has nothing to do with the Western depart-
ment, so called, which includes all modern subjects, taught in Eng-
lish.
CHAPTER V
COMPARISON OF THE JUDGMENTS WITH COLLEGE
MARKS
I. THE JUDGMENTS IN SCHOLARSHIP AND GENERAL
AVERAGES IN COLLEGES
If it is desirable to compare the judgment of associates with high
school marks in China, it will be still more desirable to compare it
with college marks in America. This was done in a crude way, for
any fine statistical manipulation did not seem possible, though
profitable, due to the erratic nature of the material in hand . College
marks of 122 of the students in our study were secured. Since some
had more marks and others less, due to the difference of the number
of years which had been spent here aortd to other causes, such as that
some colleges made reports of marks while others did not, it was
decided to include only those students who had a total of at least
7 marks in different subjects. The method for weighting these
marks from different colleges was a simple one. Either by corre-
spondence or by studying the catalogues, the different marking sys-
tems of more than 50 colleges were collected and compared. It was
found that, numerous as they were, these systems did not differ
very much. Inasmuch as many systems preferred letters or abbre-
viations to designate marks such as "A" standing for "excellent"
which is equal to "90-100" and "B" for "very good" which is equal
to "80-90," it was imperative, for the purpose of transmutation of
capital letters into numbers, to devise a definite method by means
of which to transmute. This was done by taking the mid-value of
a range of values that was given to a certain designation. To illus-
trate: the marking system of a certain college is somewhat like the
following: H.H.C. =95-100; H.C. = 90-94; C. = 85-89; etc. If
we want to assign numerical values to H.H.C, H.C, and C, etc.,
we simply take the mid-values of their corresponding numbers,
which are 97, 92, 87, etc. Thus, if a student receives a mark of
H.H.C. in a certain subject, the most convenient way to express it
numerically is to call it 97, which would probably be the average of
all the H.H.C marks received.
Comparison with College Marks 39
After all the transmutations were made, the marks of every stu-
dent were averaged as an index of his scholarship. The average
of these 122 marks was found to be 81.2 with a standard deviation
of 7.1. These two values, however, are not as significant as they
might appear to be, since they were computed from marks from
different systems.
The correlation of college marks with judgments in scholarship
was worked out and the coefficient was 4141b .052. This coefficient
warrants us again to trust the judgment in scholarship as reasonably
valid.
2. HIGH SCHOOL GENERAL AVERAGES COMPARED WITH COLLEGE
GENERAL AVERAGES
Another correlation that is of interest and might be mentioned
here is that between high school averages and college averages.
The correlation coefficient was found to be .41 2 ±.074. This must
be taken to mean the correlation between the entire high school
work in China and the entire college work in America. Here again
the coefficient is not as significant as it might be, due to the difference
in the marking systems. Still it indicates how accomplishment in
high school in China will predict accomplishment in college work in
America, and to that extent it is significant and of practical use.
CHAPTER VI
COMPARISON OF THE JUDGMENTS WITH INDIVIDUAL
RECORDS
I. A SCALE FOR JUDGING THE ACTIVITIES OF CHINESE
STUDENTS IN AMERICA
When the ranking sheets were sent out, it is to be noted that the
judges were not asked to rank themselves among those whom they
ranked, the reasons being that, first, the extreme modesty which is
characteristic of Chinese might result in ranking oneself too low,
thus rendering the rankings inaccurate and, second, the embarrass-
ment put to the judge might even hinder his otherwise willing coop-
eration in this work. Consequently, a questionnaire was sent with
the ranking sheet to each one, asking him or her to fill it out for
himself or herself as the case may be, concerning the four qualities
in our study. This questionnaire is shown on page 7. The facts
stated in these questionnaire sheets, however, were found to be so
unwieldy as to make a direct statistical treatment of them seem
impossible. To evaluate these facts by a number of judges without
at first some objective criteria to fall back upon would be to expose
oneself to gross error. Consequently, a method was devised
whereby all the facts were classified under four headings: scholar-
ship, leadership, knowledge of English, and knowledge of Chinese.
Then the procedure was greatly simplified by evaluating each posi-
tion or achievement once and for all by twelve judges. This saved
the judges the trouble of going over all the original papers, which
would mean a tremendous amount of time and labor. To illustrate:
If ten students happened to have Phi Beta Kappa keys from ten
different colleges, it was only necessary to take this particular fra-
ternity key as an evidence of scholarship common to all, and ask
the judge to give a numerical value to it. It might be contended
that Phi Beta Kappa keys from different colleges might carry differ-
ent values and they again might differ with different individuals.
True, but for our purpose it was not essential; in fact, not desirable
to make too fine a distinction. It was sufficient to mark off a man
Comparison with Individual Records
41
with such an honor key from one who had a different one or from
one who did not have any.
This being the general principle, all those common activities in
terms of positions held, degrees gained, articles written, or other
works accomplished, with the exception of a few activities peculiar
to certain individuals, were listed under the four headings to which
they belonged and submitted to twelve judges who were asked to
give them numerical values. The judgments were made on a scale
of 10, i. e., the judgments varied from o to 10. "o" is assigned to
those activities that deserve no credit; "1" to those that deserve
the least amount of credit; "10" to those that deserve the highest
amount of credit. The rest of credits, 2, 3, 4, etc., are assigned to
positions that fall in between. The reason for selecting just twelve
judges instead of more or less is that given in Chapter III, where it
was pointed out that this seems to be the optimum number because
of its high self-correlation coefficient. These twelve judges were
fully acquainted with the activities of Chinese students in America,
mature in scholarship, and sound in judgment. Thus each activity
received twelve values or judgments. These judgments were split
into halves and correlations were computed to show their reliability,
as shown in Table XI .
TABLE XI
Showing the Reliability of 12 Judgments on Individual Records
CORRELATION OF
NO. OF CASES
r
6 with 6 Judgments in Scholarship
17
78
40
42
.636 ± .096
.855 ±.020
.892 ±.023
.865 db. 026
6 with 6 Judgments in Knowledge of English
6 with 6 Judgments in Knowledge of Chinese
These high self -correlation coefficients warrant us to make use
of the judgments of these twelve judges for the treatment of indi-
vidual records. The next step is to give a single score to each activ-
ity by averaging the twelve judgments.
With the omission of a few individual activities which partake of
personal character, the scale thus made out is produced in Table
XII. These omitted activities were such as to reveal the identity
of the individuals holding the positions involved.
42
Chinese Students in America
TABLE XII
A Scale for Judging the Activities of Chinese Students in America
Position
President
Vice President
Secretary
Treasurer
Councilman
Representative
Auditor
Election Officer
Leadership
Organization
ines
« Students'
Alliance
in A
^mei
a
a
«
u
«
a
a
«
u
u
u
u
u
tt
u
«
u
«
tt
tt
u
a
«
u
u
a
a
tt
«
u
a
u
u
«
tt
Points
9.2
57
6.8
6.3
5.8
4.8
• 44
4-4
Chairman
Vice Chairman
Secretary-
Treasurer
Section of Chinese Students' Alliance in America
u u u
u u u
tt tt a
«
tt
« u a
U tt u
« tt tt
8.0
5.0
5.8
5.4
President
Vice President
Secretary
Treasurer
Manager
Auditor
Committeeman
Local Chinese Students' Club in America
tt
tt
u
tt
tt
tt
u
u
u
u
u
tt
tt
tt
«
tt
tt
tt
«
«
tt
tt
tt
tt
tt
u
u
u
tt
tt
u
u
tt
tt
6.7
4-3
4.8
4-5
43
31
3-8
«
«
General Manager Chinese Students' Monthly
Advertising Manager
Circulation Manager
Associate Manager
u
tt
7.3
6.0
5-5
4-4
General Manager Chinese Students' Quarterly
Advertising Manager
Circulation Manager
Associate Manager
«
tt
u
«
u
u
6.4
5.1
4.8
3.9
Chairman
Secretary
Treasurer
Committeeman
Annual Students' Conference
tt
a
tt
u
u
u
a
«
tt
79
6.1
57
57
Position
President
Manager
Secretary
Treasurer
Chairman
Secretary
Treasurer
Comparison with Individual Records
Organization
College Cosmopolitan Club
a u a
u
Chinese Prohibition League
u u a
u u u
43
Points
6.3
4-9
4.5
4.3
5.3
4.4
43
President
Vice President
Secretary
Treasurer
Committeeman
Auditor
Chinese Students 1 Professional Societies or Clubs,
Such as Banking Club, Educational Club, Engi-
neering Club, etc.
u
u
u
a
u
u
u
u
u
u
a
a
a
a
u
u
a
a
u
a
6.2
4.3
47
4.3
3.5
2.8
President
Vice President
Secretary
Treasurer
Committeeman
Auditor
Alumni Association, Such as Nankai Alumni Asso-
ciation, Soochow Alumni Association, Tsing Hua
Alumni Association, etc.
u
u
u
u
u
a
u
u
u .
u
a
a
u
a
u
u
u
a
u
a
u
a
u
a
a
6.3
4-3
4.8
4.1
36
2.8
President
Secretary
Treasurer
Chinese Patriotic Committee
a
a
«
u
u
6.9
5.6
5.1
President
Vice President
General Secretary
Chinese Students' Christian Association
a
u
u
u
u
u
u
«
7.3
4.8
7.3
Chairman
Secretary
Treasurer
Local College Y. M. C. A.
u u u
4.6
3.6
3.1
Cabinet Member Local College Y. M. C. A.
Member Volunteer Band « « «
Chairman College Students' Bible Class
Leader College Students' Discussion Group
2.7
1.8
3.0
33
44
Chinese Students in America
Scholarship
Position or Achievement
Institution
Fellowship
Scholarship
Graduation with Honor
College or
a a
a «
University
«
Ph.D.
a a
«
Instructor
a m
<<
Lecturer
a m
m
Assistant
u u
a
Member
Member
A.B. with
A.B. with
Phi Beta Kappa Society
Sigma Xi "
Magna Cum Laude
Summa Cum Laude
Points
7-7
5.8
7.0
8.9
7.0
6.8
5.8
74
7.5
6.5
6.4
Posiition or Achievement
Editor-in-Chief
Associate Editor
Contributor
Reporter
Editor-in-Chief
Associate Editor
Contributor
Reporter
Editor-in-Chief
Associate Editor
Contributor
Reporter
First Prize Essay
Second Prize Essay
First Prize
Second Prize
Member
Knowledge of English
Organization
Chinese Students' Monthly
« « a
u « «
<(
Christian China"
u
tt
u
American College Paper
« « «
u
u
u
Chinese Students' Monthly
tt it «
College English Oratorical Contest
« a u «
College English Debating Team
Points
8.9
6.0
4.9
34
7*
5.3
43
34
8.9
6.5
5.1
37
6.9
5*
6.9
56
6.6
Position
Editor-in-Chief
Associate Editor
Contributor
Reporter
Knowledge of Chinese
Organization
Chinese Students' Quarterly
it
tt
it
«
m
Points
8.3
6.1
49
38
Comparison with Individual Records
45
Position
Organization
Poin
Editor
"Science" in Chinese
7.8
Contributor
u u u
5-8
Editor-in-Chief
"Political Science Review"
or "Cheng Hsueh
Ts'ung KW
7.1
Editor
u u u
u u
5-9
Contributor
u a u
u u
5.1
Contributor
Shanghai Shun Pao
SO
Contributor
"The Renaissance"
59
Contributor
"Eastern Miscellanies"
57
With the help of this scale, the individual records were turned
into numerical values. If a student had several records in any one
quality, then all the transmuted values were combined representing
his total credit therein. These individual records, then, in terms of
numerical values were correlated with the judgments of associates
and the coefficients were represented in Table XIII.
TABLE XIII
Showing the Correlations of Judgments with Individual Records
CORRELATION OF
NO. OF CASES
r
The Judgment of Associates with Individual Rec-
ords in Scholarship
58
145
44
47
.706 ±.044
The Judgment of Associates with Individual Rec-
ords in Leadership
• 1 ^~ • ^^
.623 ±.034
The Judgment of Associates with Individual Rec-
ords in Knowledge of English
.703 ±.038
The Judgment of Associates with Individual Rec-
"1 JV ~~~ V
.543 ±. 070
From these correlation coefficients, it is evident that the judgment
of associates agrees better with individual records than with other
data heretofore given. This would indicate that the judgment was
largely based upon a knowledge of the accomplishments of the person
judged, and further that the acquaintance among any group of
Chinese students in this study had reached the stage where thorough
knowledge of one another's accomplishment was hardly to be
doubted.
2. A DISCUSSION ON THE VALIDITY OP THE JUDGMENT OF ASSOCIATES
In order to find out which source of data agrees best with the judg-
4 6
Chinese Students in America
ment, a resume of all the correlation coefficients showing the validity
of the judgment from the previous chapters is shown in Table XIV.
TABLE XIV
Showing the Correlation of the Judgment with Test Scores, High School
and College Marks, and Individual Records
correlation of
SCHOLAR-
SHIP
LEADER-
SHIP
KNOWLEDGE
OF ENGLISH
KNOWLEDGE
OF CHINESE
The Judgment with Test
Scores
.421
414
.706
.623
.430
.307
•793
The Judgment with High
School Marks
374
The Judgment with College
Marks
The Judgment with Indi-
vidual Records
•543
This table shows that the judgment in scholarship is checked up
three times; in English three times; in Chinese twice; and in
leadership only once. It is also seen that judgments agree best
with individual records and practically equally well with test scores,
high school marks, and college marks. As has been hinted, this
substantial correlation between the judgment and individual records
shows conclusively the reliance of the former upon the latter. This
reliance rids the judgment of personal prejudices and whims and
adds a greater weight to it.
The fact that the judgment in leadership is only checked up once
raises the question as to whether its validity has been sufficiently
proved. The answer is that any human trait like leadership par-
takes of such elusive character that, for the present at least, the
only way to judge or measure it is to resort to subjective opinions,
directly or indirectly, of the largest possible number of competent
judges. The procedure, in the light of the present investigation,
must needs meet the following few requirements: (1) a thorough
acquaintance of the judge with his subjects; (2) a sufficient number of
judges (this investigation shows that a total of twelve judges will
yield a self-correlation coefficient of about .90) ; (3) a fairly common
understanding of the qualities to be judged; and (4) the use of
rankings.
CHAPTER VII
YEARS SPENT IN AMERICA AS A FACTOR
In the preceding chapters it has been shown that knowledge of
English rather than knowledge of Chinese is more closely associated
with success in America, academic and non-academic. The next
question that comes consequent upon this conclusion is: How can
knowledge of English be better secured? To put it differently, is it
more profitable to spend more initial years studying English in
China, or is it more profitable to spend more additional years in
America? How is the number of years spent in America associated
with scholarship, leadership, knowledge of English, and knowledge
of Chinese? Questions 4 and 6 in the questionnaire sheet tell us the
number of years a student spent in China in studying English and
the time of his arrival in America. With the data supplied there-
from, it is possible to work out the correlation coefficients as shown
in Table XV.
TABLE XV
Showing the Correlation of the Number of Years Spent in America with
Scholarship, Leadership, Knowledge of English, and Knowledge
of Chinese
correlation of
NO. OF CASES
r
Years Spent in America with Scholarship
276
279
282
263
.119 ±.040
.2i8=fc.038
.205 ±.038
— .090=41.042
Years Spent in America with Leadership
Years Spent in America with Knowledge of English
Years Spent in America with Knowledge of Chinese
It must also be stated here for later comparisons that the correla-
tion between the initial number of years spent in studying English
in China and the knowledge of English in America is .212 with 277
cases, while the correlation between the additional number of years
spent in America and the knowledge of English is .205 with 282
cases, as shown in Table XV. The correlation between the initial
number of years spent in China in studying English and the years
48 Chinese Students in America
spent in America is — .522 with 287 cases. If we designate the
knowledge of English in America as 1 , the initial years in China in
studying English as 2, and the years spent in America as 3, then
by using the usual partial correlation formula,
V(i-r a I3 )(i-r% 3 )
the following two partial correlation coefficients are yielded:
fi*.8 = .382 (1)
ris.a=.378 (2)
Equation (1) reads that the correlation between the knowledge of
English in America and the initial years spent in China in studying
English, with the influence of the disturbing factor of years spent
in America eliminated or equalized, is .382. Equation (2) reads that
the correlation between the knowledge of English and the additional
years spent in America, with the influence of the disturbing factor
of the initial years spent in China in studying English eliminated or
equalized, is .378.
Interpretation of the correlation is important. In the first place,
it should be noticed that neither the initial years spent in studying
English in China nor the additional years spent in America (it must
be noted that these additional years are not spent primarily in
studying the English langauge, but in acquiring it as a matter of
course) has any very high correlation with the knowledge of English.
This would point to the fact that knowledge of English does not
depend, to any great extent, upon the number of years spent in
studying it before coming to this country, other factors such as
individual difference and environmental effect being, perhaps, of
much greater influence. This is what we expect.
In the second place, but in a more striking manner, it should be
noticed that the two partial correlation coefficients of .382 and .378
would show that the initial years spent in studying English in China
and the additional years spent in America have practically an equal
influence upon one's knowledge of English. This does not argue,
however, that the initial years spent in America in studying English
would have the same effect upon one's knowledge of English as the
initial years spent in China in studying it. In all probability, the
former would exert a much greater influence. The significance of
Years Spent in America as a Factor 49
these two coefficients lies in the fact that after one has spent a
reasonable number of years in studying English in China, so as to
acquire a knowledge of speaking and reading, then if he comes to
America the number of years he stays here will not mean more to
his further acquisition of English than his initial years in China
would mean to his initial improvement in it.
The low correlation between the number of years spent in America
with both scholarship and leadership (r= .119; r= .218) bring us to
a discussion of the problem of preparing students in China that are
to be sent to America. We have seen that scholarship and leadership
in America depend largely upon knowledge of English; that knowl-
edge of English does not depend so much (as we presumed) upon
the years of stay in America provided a reasonable number of years
is spent in China in studying it. Thus, scholarship, leadership,
and knowledge of English, which are more or less substantially
correlated with one another, all have an equally very low correlation
with the years of stay in America. Substantially, the years spent
in America as a factor have very little significance in respect to the
qualities in our study.
The correlation of —.552 between the years spent in China in
studying English and the number of years spent in this country is
merely a result of chance. It can be explained by the fact that when
the study was made more than one- third of the students in this study
had been in this country for a short time, about 6 months, and had,
nevertheless, spent about five or six years in China in studying Eng-
lish; at the same time there were those older students who had been
in America long and who had spent generally fewer years in China
in studying English. This phenomenon indicates clearly that stu-
dents of more recent years had a more thorough preparation in
English in China than those of the older days.
The very slightly negative correlation (r= — .090) between the
years spent in America and knowledge of Chinese simply means
that knowledge of Chinese, once acquired, whatever its amount
may be, does not decrease with the lack of practice resulting from
staying a few years longer in America. The correlation —.090 is
so small that it amounts practically to zero. Its significance should
rid us of the current fear that a few years' stay in America tends to
result in the retrogression of knowledge of Chinese. What actually
happens is in all probability that, through lack of practice, the
50 Chinese Students in America
ability of reading and writing Chinese temporarily lies in a state of
dormancy which to some people may seem to be in the state of
retrogression but which, upon a renewal of study and practice in
China, will regain its normal condition.
CHAPTER VIII
CONCLUSIONS AND SUGGESTIONS
Some of the larger conclusions of the study may be restated as
follows:
1 . Scholarship is correlated rather high with leadership (r » .592) .
2. Knowledge of English is correlated rather high with both
scholarship (r = .719) and leadership (r = .679) .
3. Knowledge of Chinese is cprrelated low with both scholarship
(r = .389) and leadership (r = .309) .
4. Knowledge of English and knowledge of Chinese have a very
slightly negative correlation or are practically uncorrelated, or
independent of each other (r= — .024).
Other findings that are of import are (1) scholarship, leadership,
and knowledge of English are all correlated low with the length of
time spent in America (r=.,H9; r=.2i8; r=.205 respectively);
(2) knowledge of Chinese is practically uncorrelated with the length
of time spent in America (r = — .090); and (3) the initial years spent
in studying English in China and the additional years spent in
America both have equally low correlations with knowledge of
English (r= .382 and r=* .378 respectively).
In the light of the foregoing conclusions, a few suggestions may
be here presented which may be of use to a consideration of the
policies governing the process of preparation and selection of Chinese
students to be sent to America.
In the first place, since knowledge of English is correlated rather
high with both scholarship and leadership and since knowledge of
Chinese is correlated rather low with both of them, it is evident
that in the preparation of Chinese students to be sent to America
more emphasis should be placed upon knowledge of English and
less upon the knowledge of Chinese if these students are expected
to do well in scholarship and leadership in America. It is to be
regretted that it has not been possible to find out the intelligence
of these students in this study, but it is very likely and safe to say
that intelligence has a high correlation with both knowledge of
English and of Chinese and that if intelligence is "partialed out,"
52 Chinese Students in America
there will still be a fair correlation of knowledge of English with
both scholarship and leadership; but there will be perhaps a zero
correlation of knowledge of Chinese with both scholarship and
leadership. Thus the scholarship and leadership of a Chinese stu-
dent in America is determined by intelligence and knowledge of
English, while knowledge of Chinese is almost indifferent to the
attainment of both. While this is true, it by no means proposes to
neglect the study of Chinese. Knowledge of Chinese, while it is
impotent in the career of a Chinese student in America, will certainly
be important upon his return to China. For this reason, it is suggested
that in the preparation of Chinese students to be sent to America
attention should be directed to an equal attainment of knowledge
of both English and Chinese so that these students not only will
best succeed during their period of education in America by means
of their knowledge of English but also will do well during their entire
life of service in China by means of their knowledge of Chinese.
In the second place, since knowledge of English and knowledge of
Chinese are uncorrelated or independent of each other, it will be
in the future worth our greatest effort to teach these two languages
in such wise that improvement in one will be associated, and if pos-
sible, closely associated with, improvement in the other. Besides
the schools giving equal emphasis upon both English and Chinese,
there are many other schools in China that are unfortunately divided
into two extremes: one places the emphasis on the study of Chinese
at the expense of the study of English and the other does just the
reverse. The suggestion is to build up a reasonably secure founda-
tion in Chinese at an early period in a student's school life and then,
but not until then, to take up the study of English in a rather sys-
tematic fashion until a good speaking and reading knowledge of
English is adequately attained. When such is the case, it is highly
probable that the correlation between knowledge of both English
and Chinese will be positive.
In the third place, since scholarship, leadership, and knowledge
of English all have low correlations with the length of time spent
in America, it is evident that these traits are affected only to a small
extent by a longer stay in America, and that they are more or less
constant, just as intelligence with which they must be highly asso-
ciated . One of the unsolved problems in connection with the sending
of Chinese students to America is the number of years they should
Conclusions and Suggestions 53
stay here. Approximately it varies from one to ten years. Inas-
much as a longer stay will bring with it only slight improvement in
scholarship, leadership, and knowledge of English, ajid inasmuch as
it will involve greater financial expenses as well as prolong the
detachment of these students from their own country, it is here
suggested that it would be desirable to shorten the period of their
sojourn in America. But the shortening of the period necessitates
maturity of both knowledge and age. If a student is sent to America
to begin with the Freshman year in a college, it will require on the
average seven years to finish the graduate work, which is entirely
too long. It is therefore suggested that in order to acquire advanced
knowledge in a short period, say, three or four years, it is almost
imperative to send out only those students who have received the
A.B. degree in China or who have had a training which will qualify
them to enter the graduate schools in America. This plan of send-
ing out advanced and mature students for a- short period of three
or four years of stay in America will bring with it at once many
advantages: (1) it will save the educational cost; (2) it will avoid
the danger of many students becoming unfamiliar with their own
country, less appreciative of (heir own culture and civilization, and
therefore less ready to rejnder their service to it; (3) it will take in
only students who will better judge Western civilization and
can thus make a better selection of what is desirable; and (4) it will
include only those students who, because of their maturity, have
already, before their coming to America, gained a considerable
respect of their own people and who, therefore, upon their return,
will make their influence more readily felt and their work more
productive. While this suggestion is borne out by the results of
this study, it happily meets with the rising consciousness among
educators in China which is gradually expressing itself in the process
of reshaping educational policies respecting the education of Chinese
students in America.
APPENDIX
The following are samples of tables containing the original data
used in this study, which are now kept in the Library of Teachers
College, Columbia University.
TABLE I A
The Judgment in Sigma Values and the Number op Judges in Each Case
for 664 Chinese Students in America in the Qualities op (I)
Scholarship, (II) Leadership, (III) Knowledge of English, and
(IV) Knowledge of Chinese
The first column under each quality is the number of judges; the second column
under each quality is the average value of their judgments transmuted in terms of
sigma. The table should read "Individual i was rated by 3 judges in Scholarship,
the average of the judgments being .18 sigma; he was not judged at all in
Leadership; he was judged by 3 judges in Knowledge of English, the average of
the judgments being .73 sigma; he was not rated at all in Knowledge of Chinese."
Ind.
Ind.
No. I
II
III
IV
No. I
II
III
IV
1 3 -18
3 .73
34 8 .39
8 i.io
7 48
7-34
2 4 25
3-62
4 49
4 32
35 8 -.20
7-.09
7 -.33
8 1.46
3 3-49
3 .20
3-44
4 .36
36 8 .17
7 .28
7 .15
6 .34
4 3-.6I
4 .70
3-.5I
3-44
37 6 -.07
5 -.25
6 .01
5-122
5
3 -.89
4-75
38 7 -.14
7 .05
7 1.85
5-1.02
6 3-.85
3-1.25
3 -.68
3-112
39 8 - .07
8 49
8 .55
7 .06
7 3-.3»
3 43
3 .21
40 7 .28
5-42
6 -.55
6-. 13
8
3 47
3 1-51
3-74
41 7-78
7 27
7 .08
6 — .41
9 3 135
3 .37
3-05
3 1-37
42 7 — .20
6 — .69
7 .T 2
7 -.24
TABLE II A
The Scores in English Tests for 42 Chinese Students in America
I. Thorndike Alpha 2,
Part II. III. Reading Test II.
II. Reading Test I.
IV. Thorndike Vocabulary
Test
Ind.
Ind.
No. I
II
III
IV
No. I
II
III
IV
16 8.6
70
37.5
50
301 8.6
74
12
128
17 77
80.5
50
163
305 79
67 5
34-5
125
18 7.1
45
25
50
306 8.6
78
64.5
41
106 7.0
515
15
65
312 8.4
66.5
23
104
no 8.1
77
158
385 77
57
133
60
Appendix
55
TABLE III A
Marks in High School in China for 78
I. General Average for Four Years.
III. Marks in Chinese for Four Years.
Ind.
No.
12
16
19
64
67
I
84
75
76
85
82
II
79
75
81
87
84
III
84
86
84
75
81
Ind.
No.
169
170
174
177
188
I
83
84
83
74
82
II
82
89
81
89
79
Chinese Students in America
II. Marks in English for
Four Years.
Ind.
I
III
87
82
84
89
82
No.
390 85
403 86
406
414
78
87
416 88
II
83
86
77
82
93
III
89
94
78
76
94
TABLE IV A
College Marks in American Colleges for 122 Chinese Students in America
Ind.
No.
12
16
18
19
46
General
Average
88
70
75
77
68
Ind.
No.
133
134
137
154
165
General
Average
85
82
79
86
85
Ind.
No.
284
289
296
299
301
General
Average
89
9i
89
92
79
TABLE V A
Individual Records of 156 Chinese Students in America on a Point Scale
for the Qualities of (I) Scholarship; (II) Leadership;
(III) Knowledge of English; and (IV) Knowledge
of Chinese
Ind.
Ind.
No.
I
II
1
5.2
20.4
7
7.0
19.0
8
5.2
17.8
9
4.8
10
14.8
III
IV
No.
I
II
III
IV
133
1 1. 6
10.5
24.0
134
19.0
137
13.5
18.3
159
95
160
22.5
TABLE VI A
Length of Time in Terms of Months that 286 Chinese Students in America
(I) Studied English before Coming to America and (II) Have
Been in America
Ind.
Ind.
Ind.
No.
I
II
No.
I
II
No.
I
II
1
84
19
90
60
44
179
96
19
7
48
61
92
36
59
183
94
20
8
120
19
93
84
8
184
72
43
9
60
19
94
36
125
186
24
31
10
48
55
95
151
187
84
19
■or
"""'*«■•»» In Awhor '
3 8105 032 558M6