Copyright
Daniel Starch.

Educational psychology online

. (page 20 of 41)
Online LibraryDaniel StarchEducational psychology → online text (page 20 of 41)
Font size
QR-code for this ebook


Fig. 55. — Progress of pupils in learning Spanish. After Swift ('06).

I

be derived from the previous study of another language^| The ex-
periment was carried out witli two classes, composed of 24 boys
and 24 girls in a St. Louis high school, who were beginning the study
of Spanish. Weekly tests were made to measure their progress and
individual abilities in learning Spanish. The classes were taught in
the usual manner and the pupils knew nothing of the purpose of the
tests. A record of progress was kept for the first 15 wrecks. The
results are shown in the following graphs (Figure 55), which in-
dicate the relative progress of the three groupsTlTSmcl}', those who
had previously studied one year of Latin and one year of German,



TRANSFERENCE OF TRAINING 23 1

those who had only one year of Latin, and those who had never
studied a foreign language before. The "Latin" and the "Latin
and German " groups stood considerably higher at the beginning
than the "Spanish only" group, but the "Spanish only" group
gained gradually so that at the end of the 15 weeks it had made up
about two-thirds of the dilTerence. Swift concludes:

"The number included in these tests was too small to serve as a basis
fi)r aiuihing more than tentative conclusions, but the results certainly
open the question whether the advantage to beginners of a new language,
so generally thought to accrue from the study of Latin, may not be due,
chiclly if not solely, to grammatical information that would be carried
o\'er from one language to another, and which would naturally help
enormously at the start. In acquiring facility in the use of the Spanish
gender, to cite one example, Latin would aid materially, since the ma-
jority of Latin feminincs are feminine in Spanish, and a large part of
Latin masculines and neuters become masculine in Spanish. The de-
clension of Spanish adject i\x"S for gender and number, and their agree-
ment, in these respects, with their nouns, woidd give Latin students a
further advantage. The teacher of the Spanish classes noted that more
frequent and detailed explanations of case were needed by those who
had not studied Latin. . The order of words, also, was more readily
mastered by those familiar with the Latin arrangement. Finally, in
learning the conjugations and in understanding the signilicance of tenses,
the assistance of the information acquired under these topics in Latin
was found to be especially great. The indications, however, are that the
higher records made by the Latin and German pupils were the result of
the substance of language information obtained from these studies rather
than of any so-called 'language' or 'mental discipline.'" (Swift, '06,
pp. 250 a.)

The writer ('15) made a comparison of the scholastic records
of university students who had entered the university with two
to four years of Latin with the records of those who had entered
with two to four years of Germani. The average grade for the
four years of college work of each of the graduates of the Col-
lege of Letters and Science of the year 1910 was computed. The
median mark of the 104 students who had entered the university
with Latin was 85.7 and the median mark of the 45 students who
had entered with German was 84.0. Hence the difference between
the two groups was only 1.7 pointsi

The explanation for this small advantage of Latin over German
may be sought in three directions: First, the disciplinary difference



232 EDUCATIONAL PSYCHOLOGY /

between Latin and German is either zero or very small. Second,
whatever difference they may have produced originally may have
tended to disappear in the four years of college work, owing to the
freedom of electives, pursuit of different courses, disciplinary effect
of other studies, etc. l Third, the small difference in scholastic
records may be due to an original difference in the students them-
selves, owing to the possibiHty that one language may attract a
better class of pupils than another.

To determine what part, if any, the first two factors played, the
average grade of each of the 738 Freshmen of the year 1909-1910
was computed. The median grade of the 416 Freshmen who had
entered with Latin was 82.4 and that of the 322 Freshmen who had
entered with German was 81.0. Hence the difference between the
two groups was only 1.4 points, or approximately the same as that
for the graduates.

The next problem was to compare the grades of these two groups
in specific subjects as follows:

TABLE 59. After Starch ('15)

Median grade in modern languages of 362 Freshmen who had entered

with Latin 84 . 5

Median grade in modern languages of 293 Freshmen who had entered

with German 82.3

Difference in favor of the Latin group 2.2

Median grade in Freshman English of 54 students who had entered with

Latin only 83 . 9

Median grade in Freshman English of 97 students who had entered with

German only 82.7

Difference in favor of the Latin group 1.2

Median grade in first-year French of 27 Freshmen who had entered with

Latin only 81 .5

Median grade in first-year French of 34 Freshmen who had entered with

German only 82.0

Difference in favor of the German group 5

The differences again are very small. The claim of language
teachers, so commonly made, that beginners in French who have
had Latin are much superior to those who have not had Latin, or
that students in English with previous training in Latin are superior
to those without such training is ill founded.

Another tabulation was made to'^show the scholarship records of



I



ij



TRANSFERENCE OF TRAINING



233



Freshmen in relation to the amount of foreign language studied,
irrespective of what the languages were.

TABLE 60. After Starch ('15)

Years of Foreign Number of Median Grade in all

Languages Students Freshman Studies



o 25 81

1-2 224 81

3-4 19s 83

5-6 155 84



The next problem was to measure the extent to which a pupil's
English vocabulary is increased through the study of Latin. The
method employed measured the percentage of words of the entire
English vocabulary, as well as the approximate absolute number of
words, whose meaning a person knows. The test was made with
1 89 university students and with 46 Juniors in the Madison High
School.

TABLE 61. After Starch ('15)

Size of English vocabulary of 139 university students who had studied

Latin 60 . 9

Size of English vocabulary of 50 university students who had not studied

Latin 58 . 2

Size of English vocabulary of 14 high-school Juniors who had studied

Latin. ._ 54.7

Size of English vocabulary of 3 2 high-school Juniors who had not studied

Latin 50.2

The differences between the Latin and the no-Latin groups are
surprisingly small. Nevertheless, the study of Latin does produce
an appreciably larger English vocabulary.' This advantage becomes
less in university students with whom it is partly counterbalanced
by the increase in vocabulary due to wider experience.

Partridge compared the standing in the regents' third year Eng-
lish examination of 7S3 pupils by dividing them into groups ac-
cording to the number of years of Latin they had studied. His
tabulation is as follows:

TABLE 62. After Partridge ('15)

The entire 783 papers divided on a basis of the number of years Latin was

studied

Number years studied o

Number papers written 181

Average standing (percentage) .... 65



I


2


3


123


220


2,S9


65


69


76



234



EDUCATIONAL PSYCHOLOGY



The pupils with no Latin may have had one or more years of
other languages and consequently Partridge presents the following
table for students who had studied Latin only:

TABLE 63. After Partridge ('15)

^' Includes papers of pupils having Latin only and no other language to their
credit. Total number of papers, 167

No. of 3'ears studied o i 2 3

No. of papers written 28 25 42 72

Average standing (per cent) . . 63 61 69 78



TABLE 64. After Partridge ('15)

Includes papers of pupils having German or French onh'^ and no other lan-
guage to their credit. Total number of papers, 1 76

No. of years studied o' i 2 3

No. of papers written 28 41 57 50

Average standing (percentage) 63 61 65 68



1-

J



X Partridge believes that the "superiority of the classical over the
non-classical pupils is due not solely to initial natural ability, but
to the training received in Latin." He has, however, failed to show
the differences in initial ability and consequently any inference of
this sort is doubtful.

Harris made a study of the effect of knowledge of Latin upon
ability to spell English words by submitting a list of 50 words of
Latin origin to 324 freshmen in the University of Illinois. He gives
the following table:

TABLE 65, After Harris ('15)



Vkaus of Latin





1


2


3


4


No. of students


90
82.1


41

82. 4


05

So. 2


54
Si.S


44


Average


90.1





He further submitted to the same group of students 10 words
of Latin origin which were to be defined. This test gave the follow-
ing result:

TABLE 66. After Harris ('15)



^F.ARS OF Latin





1


2


3


4


No. of students


90
30-5


41
44.2


95
45


54
53


44


Average


85.3



' It is obvious that the o columns in these two tables will contain record of exactly
the same pupils.



TRANSFERENCE OF TRAINING



235



He also compared the grades in Rhetoric of students who had had
various amounts of Latin as follows:

TABLE 67. After Harris ('15)



Years of Latin





1


2


3


4


No. of students

Grade


53
77.2


41
79.2


66
79-5


28
80.6


26
81.8



Harris concludes: "From these various results the conclusions
in so far as these students are concerned, are obvious. In all fields
the four-year Latin students showed a marked lead, and in all but
the spelling — which I have considered above — there is a steady re-
trogression although for the practical purposes the one-and-two-
year Latin students might be classed together."

The interpretation of these figures is by no means so obvious.
Harris has made no allowance for the native superiority of the stu-
dents with more years of Latin study. \ In fact, the probability is
that, if we may infer from other studies in which such a deduction
has been made, a large part of the superiority is due to original
nature. Harris's results as they stand prove little or nothing con-
cerning the effect of training in Latin.

F. M. Foster performed a similar experiment at the University of
Iowa; 503 freshmen, about equally divided between the sexes, were
given a spelHng test of forty words of Latin derivation. The results
are given in the following table:



2


3


4


25


24


17


29


28


27



TABLE 68. After Foster ("17)

Number of years of Latin o i

Average % of errors (girls) 23 28

Average % of errors (boys) 39 37



As in the study made by Harris there appears to be a decided re-
lationship between ability to spell words of Latin derivation and the
number of years devoted to Latin. In this case, however, it hap-
pened that Professor Irving King had previously given intelligence
tests to these same students by means of which it is possible to se-
cure a more accurate notion of the forces really producing the
better scores of the Latin students. The following table shows the
relation between the index of intelligence (that is, the percentage
above or below the average adult intelligence), the number of years



236 EDUCATIONAL PSYCHOLOGY

spent in the study of Latin and the abihty to spell words of Latin
derivation for the two extreme Latin groups.

TABLE 69. After Foster ('17)
Best 14 of 4-year Latin girls, mental ability av.



Poorest 13 "




Best 12 "


" boys "


Poorest 12 "


(( ti li


Best 10 of o-year


girls "


Poorest 10 "


" "


Best 10


boys "


Poorest 10 "


" "



25.8,


speUing


av. 5% error


-95,




" 33% "


13-4,




" 7% "


-12.4




" 44% "


19. 1,




" 11% "


-14.2,




" 34% "


50,




" 23% "


-27.2,




" 63% "



This table shows clearly that students who chose to study Latin
had on the average a distinctly better native intelligence than the
non-Latin students and that the ability to spell words of Latin der-
ivation was to a considerable extent due to this superior intelligence
rather than to the study of Latin.

The secretary of the College Entrance Examination Board made
an extensive tabulation of the records of the classical and the non-
classical students who took the examinations in 1914, 1915, and
19 1 6. The classical students are the ones who offered Latin or
Greek, or both; non-classical students are those who offered neither
Latin nor Greek. A total of 21,103 candidates are concerned in the
following table which is based on the marks in all subjects except
Latin and Greek (reported in Value of Classics, 1917, p. 366):

Combined Ratings in All the Non-Classical Subjects
Candidates who obtained a rating of 90 to 100:

2.95% of all the classical candidates.
2.05% of all the non-classical candidates.

The classical students show a superiority of 44%.
Candidates who obtained a rating of 75 to 89:

17-31% of all the classical candidates.
12.31% of all the non-classical candidates.

The classical students show a superiority of about 40%.
Candidates who obtained a rating of 60 to 100:

51 .96% of all the classical candidates.
40.97% of all the non-classical candidates.

The classical students show a superiority of about 27%.



^ I



TRANSFERENCE OF TRj\INING 25



3^



A committee in connection with the Princeton Conference made a
comparison of the honors received by classical and non-classical
students upon graduation from high schools, academies and colleges.
The table is based upon 2,799 classical and 5,606 non-classical
students from 19 high schools and academies, and upon 4,092
classical and 2,003 non-classical students from 17 colleges and
universities:

"The combined data from the nineteen high schools and academies
reporting yield the following results:

"Students receiving High Honors at Graduation were 18% of all the
classical students, but only 7.2% of all the non-classical students.
"That is: the classical students show a superiority of 150%.

"Students receiving Honors at Graduation were 32.1% of all the
classical students, but only 30.8% of all the non-classical students.
"That is: the classical students show a superiority of 36.7%.

"Students receiving Honors or Prizes for Debating, Speaking or
Essay-writing were 8.8% of aU the classical students, but only 3.5% of
all the non-classical students.
"That is: the classical students show a superiority of 150%."

"The combined data from the seventeen colleges and universities
reporting yield the following results:

"Students receiving High Honors at Graduation were 17.3% of all
the classical students, but only 6.6% of all the non-classical students.
"That is: the classical students show a superiority of 162%.

"Students receiving Honors at Graduation were 46.5% of all the
classical students, but only 38.5% of all the non-classical students.
"That is: the classical students show a superiority of 20.7%.

"Students elected to Phi Beta Kappa were 16.8% of all the classical
students, but only 8.9% of all the non-classical students.
"That is: the classical students show a superiority of 88.8%.

"Students winning Prizes or Honors for Scholarship in Other than
Classical Subjects were 13.5% of all the classical students, but only
9.3% of all the non-classical students.
"That is: the classical students show a superiority of 45.2%.

"Students serving on the Editorial Boards of Student Newspapers and
Magazines were 15.1% of all the classical students, but only 9.2% of all
the non-classical students.
"That is: the classical students show a superiority of 64.1%.

" Students acting as Members of Intercollegiate Debating Teams were
5.1% of all the classical students, but only 3.2% of all the non-classical
students.

"That is: the classical students show a superiority of 59.4%." (Reported
in Valtie of Classics, pp. 381-383.)



238



EDUCATIONAL PSYCHOLOGY



These statistics are interesting enough, but they represent a
mingling of^raining and native ability in the superiority shown and
as such do not afford conclusive proof for the efficacy of classical
training.

Wilcox ('17) made an inquiry with the endeavor to ascertain the
amount of difference in original capacities. If the superiority of the
Latin students is due to their study of Latin, we ought to find that
they were not superior, or at least not as much superior, to the
other students before they undertook Latin. If, however, we should
find that the Latin students were as superior before they studied
Latin as afterwards, we may infer that Latin had nothing to do
with their superiority.

Wilcox tabulated the records of pupils in the Iowa City High
School graduating during a period of ten years. He tabulated
separately the grades made by all the students with Latin or Ger-
man, but not with both. These results are shown in the following
table in which the numbers were obtained by transposing the
symbols E, G, M, P, and F into numerical values of 4, 3, 2, i, and
o respectively.

TABLE 70. After Wilcox ('17)

Median grades in English of Towa City High School students who studied

Latin or German





Freshman


Sophomore


Junior


Senior


All Latin (184)


6.21

4-93
1.28


6.29

5-29
1 .00


5.80
4.72
i.oS


6. II


All German (120)

Difference


4.92
1. 19



The comparison is graphically shown in Figure_3^. It will be
noticed that the superiority of the classical group is found in the
freshman year and continues throughout the course.

A comparison was also made of the English grades of students
having four, three, and two years of Latin. This is shown in
the following table:

TABLE 71. After Wilcox ('17)

Median grades in English of Towa City High School students who had 4, 3,

and 2 years of Latin





Freshman


Sophomore


Junior


Senior


4 yrs. Latin (31)

3 yrs. Latin (27)

2 yrs. Latin (1 26)


714
6-33
5-96


7-37
6.56

5-93


6.70
6.30
5-39


7-33
6.45
5 50



TRANSFERENCE OF TRAINING



'■39



A graphical comparison is shown in Figure 57. It is evident
that those who were destined to take four years of Latin were
already in their freshman year clearly superior to those taking
less Latin.

Wilcox made a similar investigation of the graduates of the
high school of Cedar Rapids, Iowa.









T^atin


^.




G^













12 3 4

Years of High School

Fig. 56. — Median grades in Eng-
lish in the four successive high-school
years of students taking Latin or Ger-
man. After Wilcox ('17).




1234
Years of High School

Fig. 57. — Median grade in Eng-
lish in the four successive high-school
years of students who had 4, 3, or 2
years of Latin. After Wilcox ('17).



Comparisons were made of the grades of 150 graduates having
Latin, German or no foreign language. This is shown in Table

72.

TABLE 72. After Wilcox ('17)

Median grades in English of Cedar Rapids High School students who studied
Latin, German or no foreign language





Freshman


Sophomore


Junior


Senior


All Latin (70)


91
86
86


89
86

83


89
85
84


86


All German (60)

No foreign language (30) .


87

85



It will be observed that in the freshman year the classical group
was superior to the other two, but that by the senior year there was
very little difference in the three groups. ■

In Table 73, comparison is made of the English grades of stu-
dents having four, three or two years of Latin.



240



EDUCATIONAL PSYCHOLOGY



TABLE 73. After Wilcox ('17)

Median grades in English of Cedar Rapids High School students with 4, 3,
or 2 years of Latin





Freshman


Sophomore


Junior


SeNI; R


4 years Latin (19)

3 years Latin (9)

2 years Latin (38)


94
90


94

88
89


93

S7
88


92

85
84



Here again the people with four years of Latin maintain their
superiority throughout the course.

" Conclusions: It seems evident, so far as the Iowa City and Cedar
Rapids high schools are concerned, that the frequently demon-
strated superiority of students who have had Latin is not due to the
special discipline or training secured in the study of Latin. It is
probably due to the fact that, as a whole, the students who elect
Latin are somewhat superior to those who refuse to take it."

Perkins ('14) made an investigation to determine the effect of
emphasizing the derivation of English words from Latin words in
the instruction of Latin in the commercial course in the Dorchester,
Massachusetts, high school. This investigation was designed to
eliminate as far as possible the differences in original abilities
between pupils with Latin and without Latin. His report follows:

"Obviously, the first step was to select two sets of pupils of equal
ability, one set in the second year of Latin, and the other in the second
year of a modern language. Accordingly, we chose pupils such that each
group had virtually the same average mark in Latin, on the one hand,
and modern language, on the other, and also in English, with the result,
in actual figures, that the non-Latin group in the two studies averaged
0.5 of 1% the higher. To make doubly sure that the Latin pupils were
not favored, the non-Latin group were taken from the section of Mr. Mur-
dock, a classical scholar, who in his English teaching emphasizes the
Latin element in the language. There were twenty-one pupils in each
set, all in the second year class of the school.

"Five measurements were made, one in spelling, one of the use of
words in sentences, the third in definitions and parts of speech, the
fourth in the meaning of words and spelling, and the fifth in excellence
in vocabulary.

"Miss Humphrey selected the words in Nos. 1-4, and the subject in
No. 5. In Nos. I and 2 the words were taken from the 600 or 800 deriva-
tives in the notebooks of a fourth-year pupil of the class, who was ex-
cluded from the measurements. Moreover, to be fair to the non-Latin



TRANSFERENCE OF TRAINING 24 1

group, care was taken not to select words too difficult. In No. 3 the
words were taken from the 'Tale of Two Cities' which the pupils of both
groups were reading at the time in connection with their work in Eng-
lish II. Of the twenty words in No. 4, ten were taken from the 'Tale of
Two Cities' and ten from other sources. The subject in No. 5 was,
'What I like to do best.' The papers were marked by teachers in the
English department and the results given to me. Altogether, six teachers
of English assisted in the measurements.

"To these five measurements is added a sixth — in my opinion most
impressive of all. This test was made last June, shortly after I had
received Professor Holmes's letter, by Miss Gormley, with her pupils in
English II. As it happened, IMiss Gormley, who was also the 'home-
room' teacher of all the pupils and consequently had access to their
marks, in making up the two groups to be composed of pupils of equal
ability, took into account not only foreign language and English II, as
was the case in measurements 1-5, but also the studies the pupils had
taken during the year. Hence we have even more reason in this case
than in the others to assume that the pupils were of equal ability. In
each set there were seventeen second-year students. The words were
taken entirely from Franklin's Autobiography and Silas Marner which
all were reading at the time. The Latin pupils were selected from the
first class I had had in the subject, just as they were completing the
course at the end of the second year.

"The result of the six measurements were as follows:



January and February, 1914:

1. Spelling

2. Use of words in sentences

3. Definitions and parts of speech .

4. Meaning of words and spelling.

5. Excellence in vocabulary

June, 1913:

6. Meaning of words and spelling .



Averages

32.18
Difference 29 . 1 2%

"In No. I, the spelling measurement, the words were not difficult,
but such as ordinary pupils of sixteen should know something about,
whether they had studied Latin or not — as 'valedictory,' 'competition,'
'occurrence,' 'benevolence,' 'legible.'

"In No. 2, the pupils composed sentences containing the derivatives,
some of which, in this measurement also, ought not to be unfamiliar to



Averages




Latin


Non-Latin


Per Cent


Per Cent


82.5


72.6


57-5


40.6


695


33-3


57-0


27s


36.0


6.8


63-3


12.3


367.8


193 I


61.3


32.18



242 EDUCATIONAL PSYCHOLOGY



Online LibraryDaniel StarchEducational psychology → online text (page 20 of 41)