Tomatis International Convention SEGOVIA 2005 Sunday 8th of May No. 15 Topic
Results of Tomatis Training in a Japanese high school: 2nd year Research into how listening improvement through use of the Electric Ear influences the development of English Ability
Speaker
Yoko Miyoshi : English teacher at Kanto International high school in Tokyo, Japan
Abstract
This paper discusses: 1. the results of a listening test that compares Ear Training Effect between operation term – short term concentration and long term dispersion. 2. comparisons between other schools on the nationwide English test done before and after the listening training. The tests check several items such as vocabulary, grammar, reading , listening and writing ability. 3. the change of speed and pitch and the existence of high frequency waves using a sound analyzer on an ondoku (reading aloud) test.
1. The results of a listening test that compares Ear Training Effect between operation term – short term concentration and long term dispersion. The listening test was given to all the students who took Tomatis training. The research was done on two different groups respectively for two years in a row in order to compare and examine the effect on listening: short term concentration for the first year and long term dispersion for the second year. All students were first year high school students studying English as a second language. The training was tailored for American frequencies. The training time tables are shown in Table 1. T-tests were administered to evaluate meaningfulness of the results. In the t-test, the closer to 0.000 the score is, the more reliable the training effect is. Table 2 shows the short term results and Table 3 shows the long term results. Table 1 Time table Training style Examinees Short term
Number
1st year in 2003 1st year in 2004
Long term
2nd
40
Training hours 1st semester 4.5hs/w × 10wks
Total hours
180
3hs/w × 10wks
1.5hs/w × 10wks
semester 0
45 45
Table 2 The results of listening test: Short term in 2003 ear
125 before
r i g h t
l e f t
after
250 before
after
500 before
after
750 before
after
1000 before
after
1500 before
after
2000 before
after
3000 before
after
4000 before
after
6000 before
after
8000 before
after
japanese
Am.E
range
range
before
after
before
after
average 8.8 9.0 8.6 9.2 9.0 9.0 9.0 9.1 9.0 9.0 8.7 9.0 8.2 8.5 8.0 8.3 8.0 8.5 8.6 8.8 8.9 9.1 8.8 9.1 8.4 8.6
STDEV 1.2 1.0 1.1 0.5 0.8 0.8 0.9 0.9 1.0 0.8 1.1 0.9 1.1 0.9 1.2 0.7 1.2 0.9 1.0 1.0 1.2 1.3 0.5 0.4 0.7 0.6
ttest 0.215
0.002
0.884
0.496
0.798
0.111
0.123
0.115
0.032
0.440
0.463
0.027
0.018
average 9.2 9.3 8.6 8.8 9.1 8.8 8.6 8.6 8.6 8.7 8.3 8.4 7.7 8.7 7.4 8.0 7.4 8.3 8.1 8.5 9.0 9.3 8.7 8.8 7.9 8.6
STDEV 1.2 0.7 1.2 1.0 1.1 0.9 1.3 1.1 1.1 1.0 1.1 1.0 1.3 0.9 1.5 0.7 1.3 0.9 1.1 1.1 0.9 1.0 0.7 0.5 0.7 0.6
ttest 0.482 0.390 0.100 0.874 0.768 colored:items of meaningful difference
0.517
0.000
0.008
0.003
0.050
0.066
0.922
0.000
Table 3 shows that the students’ sensitivity to high frequencies over 1500Hz, which is sustainable to American English frequencies, improved considerably. The T-test scores in colored cells mean there were 1
marked differences between the results of the pre-training and post-training. The last two columns in Table 3 are the scores for Japanese sound frequencies and the American English sound frequencies. All the T-test figures of the areas show the training worked fairly well. It is interesting to note that the low frequencies area, characterized as dominant Japanese sound frequencies, were also enhanced, though the training was supposed to help high frequencies. The reason for the decline in the 6000Hz is not clear, but it may result from excessive use of portable music devices. Table 3 The result of listening test: Long term in 2004 Japanese
r i g h t
l e f t
Am.E
range range 125 250 500 750 1000 1500 2000 3000 4000 6000 8000 before after before after before after before after before after before after before after before after before after before after before after before after before after average 8.4 7.6 9 8.9 9 8.9 9.3 9.1 8.9 9.1 8.4 8.5 8.3 8.6 8 8.2 7.7 8.5 7.6 7.2 9.6 9.4 8.9 8.7 8.5 8.7 STDEV 1.6 1.8 0.8 0.9 0.9 1 0.9 0.8 1 0.8 1.1 1.1 1 1 1 1 1.1 1.3 2.1 2.7 0.9 1.3 1.1 1.2 1.1 1.1 ttest 0.011 0.424 0.512 0.492 0.491 0.711 0.075 0.235 0.004 0.391 0.440 0.047 0.014 average 9.4 8.8 8.4 8.9 8.6 9 8.5 8.9 8.4 8.7 8 8.5 7.8 8.4 7.3 7.9 6.7 7.8 6.4 6.7 9.5 9.3 8.6 8.8 7.9 8.5 STDEV 0.9 1.4 1 0.9 1 1.1 0.9 1 1 1 1 1.3 1.1 1 1.2 0.9 1.2 1.3 1.9 2.6 0.8 1.3 1.1 1.1 1.2 1.2 ttest 0.017 0.009 0.070 0.047 0.193 0.034 0.001 0.005 0.000 0.565 0.359 0.007 0.000
colored:items of meaningful difference 10.0
Chart 1 The dispersion of the result over 2000Hz 9.0 term results Short 8.0
right ear
7.0
le ft e ar
10.6.0 0
1 0 .0 9 .5
9.5.0 5 125
250
500
750
1000 1500 2000 3000 4000 6000 8000
9.0
9 .0
8.5
8 .5
8.0
8 .0
7.5
7 .5
7.0
7 .0
before
be fo re
after
afte r
Long term results Am .E. range:right ear
Am .E. range:left ear
10.0
10.0
9.5
9.5
9.0
9.0
8.5
8.5
8.0
8.0
7.5
7.5
7.0
7.0
6.5
6.5
6.0
before
after
6.0
2
before
after
Chart 2
The balance of listening sensitivity in the right and left ears
Short term results le ft e ar
right ear 10.0
1 0 .0
9.0
9 .0
8.0
8 .0
7.0
7 .0
6.0 125
250
500
750
1000
1500
before
2000
3000
4000
6000
6 .0
8000
1 25
25 0
500
750
10 00
after
150 0
be fo re
2 000
30 00
400 0
6 000
80 00
afte r
Long term results right ear
left ear
10.0
10.0
9.0
9.0
8.0
8.0
7.0
7.0
6.0
6.0
5.0
5.0 125
250
500
750 1000 1500 2000 3000 4000 6000 8000 before
125
after
250
500
750 1000 1500 2000 3000 4000 6000 8000
before
after
A Consonant Test A consonant test was originally made to assess the students’ ability to distinguish between consonants with similar sounds such as “s” and “th” or “l” and “r”. The test consists of 100 minimal pairs. The students are asked to circle the sound he/she hears. The words themselves were pronounced and recorded on tape by an American English speaker. On the students answer sheet, the words themselves were not written, only the phonetic signs. It is difficult to get a score over 90. A score of 90 could be considered the border between those who have so-called English ears and those who don’t. Table 5 shows the results of the consonant listening test in short term and long term one. Histogram of minimal check Table 5 18 Short term results 16 Group A=Experimental Group 14 Group B=Control Group 12 histogram 10 Group A Group B 8 range frequency range frequency 6 70 0 70 0 4 75 0 75 0 80 0 80 4 2 85 3 85 8 0 90 9 90 15 70 75 80 85 90 95 100 95 16 95 3
100
8
100
0
Group A 3
Group B
Long term results
Histogram of minimal check
The number of people
histogram before after range frequency frequency 60-64 0 0 65-69 0 0 70-74 4 0 75-79 15 3 80-84 14 18 85-89 10 17 90-94 2 7 95-99 1 1 100 0 0 The number of people:3→8
20 18 16 14 12 10 8 6 4 2 0
before after
60
65
70
75
80 85 Score
90
95
100
It is clearly illustrated that the students’ listening ability grew as a whole, which means their listening sensitivity to consonant sounds improved. Conclusion All the results shown above suggest that the effectiveness of Ear Training is clear both in the short term and long term with the same amount of total training hours. It might also be induced that Tomatis training could be given in less than 45 hours. This would be very convenient to those who have a hard time finding the specified amount of consecutive hours to conduct the training. It would be especially good news for schools as it is almost impossible to spend hours on listening training by sacrificing other English subjects. Still, it is unclear how long training can be conducted in the long term without a decrease in effectiveness. Next year, our experimental group will have 30-hour Ear Training in the long term, at which time we will research how it will work. Additionally, individual observations of students’ attitudes suggest better results are obtained from highly motivated students who have had more serious training. When applying the training in school, training may be treated as an elective class but not as a mandatory class. 2. Comparisons between other schools on the nationwide English test done before and after the listening training. The ACE test (Assessment of Communicative English) assesses vocabulary, grammar, reading, listening and writing abilities in communicative English. The number of the examinees was about 13500. Table 6 shows that the experimental group improved in Listening by 11.2% for two years, while the improvement of the national average was 2.5% .The students who scored higher (27% of the examinees) show a 3.8% improvement. K school, which was almost the same with the experimental group at the very beginning, showed a 6.9% improvement. Table 7 shows that there was a strong correlation (coefficient=0.734) between Listening and Reading abilities in the experimental group. The control group showed a correlation coefficient 0.670. A Writing test was given to both groups just once. Table 8 shows the correlation coefficient between writing and listening for the 2nd year experimental group was 0.455, while that of the control group is 0.031. The 1st year groups had a coefficient of 0.358. These results betrayed our expectation that listening and writing would correlate in the same way as reading and listening. More tests are needed to confirm this.
Table 6 Improvement in ACE Short term results 4
vocabulary
gram m ar
reading com prehension
total
listening conprehension
A pr.'03 Jan.'04 Jan.'05 A pr.'03 Jan.'04 Jan.'05 A pr.'03 Jan.'04 Jan.'05 A pr.'03 Jan.'04 Jan.'05 第2回 第3回 第5回
G roup A G roup B
0.513 0.57 0.478 0.55 the tops nationw ide 0.537 0.59 nationalaverage 0.462 0.501 K .H igh S . 0.498 0.56
0.612 0.574 0.621 0.515 0.613
0.464 0.562 0.647 0.52 0.443 0.523 0.59 0.516 0.526 0.606 0.681 0.638 0.444 0.5 0.561 0.509 0.478 0.58 0.66 0.536
vocabulary vocabulary
80%
0.556 0.593 0.595 0.563 0.6 0.598 0.648 0.66 0.669 0.507 0.506 0.567 0.577 0.64 0.591
0.572 0.575 0.635 0.556 0.557
0.707 0.674 0.65 0.562 0.633
0.534 0.525 0.613 0.51 0.538
0.638 0.619 0.654 0.535 0.637 reading
gram m ar grammar
80%
0.565 0.558 0.616 0.51 0.568
80%
70% 80%
80% 70%
70%
70% 60%
70% 60%
60%
60% 50%
60% 50%
50%
50% 40%
50%
40%
A pr.'03
Jan.'04
Jan.'05
40% A pr.'03
40%
Jan.'04 G roupApr.'03 A G roup B Jan.'05 The tops nationw ide nationalaverage Group A Group B K.H igh S. The tops nationwide national average K.High S.
Apr.'03 G roup A
Jan.'04
Jan.'04
The tops nationw ide Group A K.H igh S. The tops nationwide K.High S.
reading comprehension
Jan.'05
Jan.'05
G roup B nationalaverage Group B national average
listening comprehension
80%
80%
70%
70%
60%
60%
50%
50%
40%
40% Apr.'03
Jan.'04
Group A The tops nationwide K.High S.
Jan.'05
Group B national average
Apr.'03 Group A The tops nationwide K.High S.
total
80% 70% 60% 50% 40% Apr.'03 Group A The tops nationwide K.High S.
Jan.'04
Jan.'05
Group B national average
Long term results 5
Jan.'04
Jan.'05
Group B national average
40% A pr.'03
G roup A The tops nationw K.H igh S.
vocabulary
gram m ar
reading listening com prehension com prehension
total
A pr.'04 Jan.'05 A pr.'04 Jan.'05 A pr.'04 Jan.'05 A pr.'04 Jan.'05 A pr.'04 Jan.'05
1R1 1R2 1S 1SS the tops nationw ide
nationalaverage
0.516 0.493 0.484 0.522 0.589 0.523
0.497 0.465 0.489 0.528 0.576 0.484
0.484 0.458 0.464 0.48 0.554 0.473
0.493 0.465 0.496 0.565 0.632 0.529
0.467 0.378 0.418 0.46 0.581 0.47
0.448 0.437 0.461 0.523 0.616 0.475
0.564 0.508 0.552 0.574 0.633 0.555
0.581 0.56 0.608 0.636 0.644 0.556
0.51 0.454 0.481 0.511 0.567 0.508
vocabulary
0.508 0.487 0.52 0.568 0.621 0.513
grammar
70%
70%
60%
60%
50%
50%
40%
40%
30%
30% Apr.'04
Jan.'05
R1 S The tops nationwide
Apr.'04
R2 SS national average
R1 S The tops nationwide
R2 SS national average
listening comprehension
reading comprehension 70%
70%
60%
60%
50%
Jan.'05
50%
40%
40%
30% Apr.'04
Jan.'05
R1 S The tops nationwide
R2 SS national average
30% Apr.'04 R1 S The tops nationwide
total
70% 60% 50% 40% 30% Apr.'04 R1 S The tops nationwide
Table 7
Jan.'05 R2 SS national average
The correlation between Listening and Reading 6
Jan.'05 R2 SS national average
Short term resluts GroupA (Apr.'03)
Group B (Apr.'03)
0.8
0.8
r 0.6 e a d 0.4 i n 0.2 g
r 0.6 e a d 0.4 i n 0.2 g
0
0
0
0.2
0.4
0.6
0.8
1
0
0.2
listening
1
Group B (Jan. '05)
Group A (Jan. '05) 1
r e a d i n g
0.8 0.6 0.4 0.2 0
0.8 0.6 0.4 0.2 0
0
0.2
0.4
0.6
0.8
1
0
listening
1 0.8 0.6 0.4 0.2 0 0.2
0.4
0.6
0.4
0.6
Correlation coefficient:0.6695
Long term results All the 1st year students (180)
0
0.2
listening
Correlation coefficient:0.7337
0.8
1
listening
Correlation coefficient:0.572
Table 8
0.8
Correlation coefficient:0.6709521
1
r e a d i n g
0.6
listening
Correlation coefficient:0.5946843
r e a d i n g
0.4
The correlation between Listening and Writing 7
0.8
1
Short term results Group A
Group B
1
w r i t i n g
1
0.8
w r i t i n g
0.6 0.4 0.2 0
0.8 0.6 0.4 0.2 0
0
0.2
0.4
0.6
0.8
1
0
listening
0.2
0.4
0.6
0.8
1
listening
Correlation coefficient:0.4549566
Correlation coefficient:0.0309538
Long term results 1
w r i t i n g
0.8 0.6 0.4 0.2 0 0
0.2
0.4
0.6
0.8
1
listening
Correlation coefficient:0.357938726 General Comment for the results of ACE There are several combined reasons for the experimental group having improved so much in their English ability, especially in listening and writing. It is due to Tomatis Ear Training, NETs’ well-prepared Oral Communication and Writing classes and Dual Teaching of Reading classes by NET and JET. Most likely, it is through much exposure to English that most of the students came to understand English without translating into Japanese. That is why they increased the speed and comprehension in their reading and listening and writing. The effect of Ear Training has been maintained and highly motivated the students. 3. The change of speed and pitch and the existence of high frequency waves using a sound analyzer on an ondoku (reading aloud) test. The results of the reading aloud tests were examined in two ways. One was an objective examination by an authorized acoustic laboratory in Tokyo(2004) and Sophia University acoustic laboratory (2005). The other was a subjective evaluation of the recorded tapes by two NETs, who are both Americans, at our school,. The tape evaluations by the NETs were done anonymously to exclude bias as much as possible. The students were asked to read a certain passage taken from a story which includes several 8
unfamiliar words. There was little concern about students remembering the passage. Therefore we felt it was appropriate to have them read the same passage again after the training for comparison. The following are the results of the tests. Analyses with sound analyzer by experts Illustration 1 The acoustic data below was taken from one student’s recorded tape and processed by Japan Acoustic Analysis Laboratory in 2004. It illustrates that the after-training sounds seem to contain more high-frequency elements. It is regrettable that the recording condition was not good enough to illustrate the results more clearly. Data 1
The results of the sound analyzer (presented by Japan Acoustic Analysis Laboratory, Tokyo)
The following results are by a group of university experts in acoustics. Forty three students were chosen at random for analysis as an experimental group and 7 students who had no training were chosen as a control group. Table 9 illustrates the change of the reading speed of one phrase; “Why I say I never forget a face, I really and truly mean it.� The phrase was also broken into 2 parts for more precise analysis. T-tests here again show the degree of reliability in their improvements.
Table 9
The average change of reading speed for 50 students (Data presented by Sophia University 9
Acoustics Laboratory)
average STDEV ttest
duration (when-it) before after 7.3847 6.2356 2.1330 1.4693 0.0001
duration Duration (when-face) (I-it) before after before After 3.1145 2.7911 3.7392 2.8988 0.8833 0.6739 1.4665 1.4665 0.0121 0.0000
duration (pause) before after 0.6081 0.5457 0.4856 0.3201 0.4616
duration(when-face)
duration(when-it) 10.0
4.5
9.0
4.0
8.0
3.5
7.0
3.0
6.0
2.5
5.0
2.0
4.0
1.5
3.0
1.0
2.0 1.0
0.5
0.0
0.0
before
before
after
duration(pause)
duration(I-it) 6.0
1.2
5.0
1.0
4.0
0.8
3.0
0.6
2.0
0.4
1.0
0.2
0.0
0.0
before
after
before
after
after
Sophia analysts also presented results of the sound analyzer that showed improvements in high frequency areas. It was reported, however, that the pre-training recording condition was not good enough for analysis as there were high noise levels. Therefore the analysis with the sound analyzer focused on rather low frequency areas analyzing the sound of /a/ out of [ai](“I�) exclusively. Hardly any drastic improvements in the particular phrase were noted. It was also reported that the rhythm of English might be improved by Ear Training but that of duration could not always be improved through training. Next time better quality recordings are desired for a reliable analysis and consonants, both in clusters and with vowels, should be examined.
Conclusion The hypothesis that there is an improvement in listening ability through Tomatis Ear Training may not have been clearly proved by the university laboratory this time. Still the instructors’ impressions are 10
quite different. It is true some students who do not have much motivation could see hardly any improvements or even a decline in some areas. Poor recording conditions further impeded verification. Reconsideration of parameters for sound analysis might be necessary. In order to clarify the effectiveness on pronunciation of consonants, which characterizes the Tomatis Ear Training, recordings may have to be limited to cluster consonant words.
11