Reference data:

The impact of a Nishizawa, H., Yoshioka, T., & Fukada, M. (2010). The impact of a 4-year extensive reading program. In A.
M. Stoke (Ed.), JALT2009 Conference Proceedings. Tokyo: JALT.

4-year extensive This study reports the impact of a long-term extensive reading (ER) program on reluctant EFL learners.
The ER program consisted of sustained silent reading (SSR) classes, 45 minutes a week, for 120 weeks
over 4 consecutive academic years. Thirty-seven students, ranging in age from 20 to 22 years old, fin-
ished the program. A comparison between three groups of students in the ER program showed a strong

reading program correlation between their TOEIC scores and the amount of reading. The most critical factor for success
was reading at least 300,000 words, which was found to be the enabling threshold for the subjects to
feel at ease while reading English texts. To ensure the students were reading this amount, the program
needed scheduled reading time and easy English texts, especially in the first year. The readability levels of
the English texts for our students were far easier than the ones recommended by the Edinburgh Project
on Extensive Reading.

Hitoshi Nishizawa 英語に苦手意識を持つ学習者に対する長期多読授業の成果を報告する。授業は毎週１回45分の授業時間内多読で、年間３

０週を４年間継続させた。全期間受講したのは、２０〜２２歳の学生３７名である。受講学生を読書量で三分割して比較したと
ころ、群別読書量とTOEIC得点に明確な関係が認められた。多読授業の正否を分ける要因は、最低30万語以上の累積読書量
Toyota National College of であり、定期的な読書時間の確保と、特に初年度の図書のやさしさが大切である。本校学生に適した本は、エジンバラ・多読プ
ロジェクトで推薦されるものよりも、かなりやさしいものであった。
Technology (Toyota Kosen)
Takayoshi Yoshioka Background
Reading amount and readability of reading materials
Toyota National College of Day and Bamford (1998) described ten characteristics of successful extensive reading pro-
Technology (Toyota Kosen) grams, two of which were “students read as much as possible” and “reading materials are well
within the linguistic competence of the students in terms of vocabulary and grammar” (pp.
Momoyo Fukada 7-8). However, we felt that both the expected reading amount and the optimum readability of
the materials were not stated clearly enough to design an effective ER program for Japanese
Toyota National College of EFL learners. In the late 1990s, Sakai (2002) proposed 1,000,000 words as a milestone for Japa-
nese EFL learners to feel they can read independently. He also suggested using much easier
Technology (Toyota Kosen) books than those that were being used in contemporary ER programs. These proposals were
supported and promoted by the Starting with Simple Stories (SSS) English Study Group and
the Japan Extensive Reading Association (JERA).

JALT2009 Conference Proceedings

ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 632
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

These proposals led to numerous extensive reading studies age high school students of the same age (Kameyama, 2010).
within the Japanese context. Sakai and Kanda (2005) proposed a The kosen students in the fourth or higher grades however,
detailed methodology for conducting ER programs based on the have lower proficiency in English than university students of
proposals, and Kanda (2009) suggested focusing on the joy of the same age, even though they are promising engineering
reading, based on a case study of a university student who had students (Kameyama & Ozawa, 2002). For example, the national
taken a 3-year ER program. Takase (2008) confirmed the positive average TOEIC score of seventh grade kosen students was 373
effect of reading very easy-to-read books, and pointed to it as in 2007, which was 40 points lower than the national average of
one of the two most critical features of a successful ER program. fourth year university students majoring in engineering, science,
Furukawa et al. (2010) compiled a list of 13,000 easy-to-read and agriculture, in the same year (ETS, 2008, p. 9, 11).
reading texts, and evaluated their readability as an index called The poor performance of kosen students in English is
the Yomiyasusa Levels (YL), specifically designed for Japanese sometimes explained by them not needing to pass severe
EFL learners. Nishizawa, Yoshioka, and Itoh (2006) compared entrance examinations at the age of 18. Another reason is that
the readability levels proposed by the Edinburgh Project on they have fewer English lessons compared to high schools and
Extensive Reading (EPER) (Hill, 1997, cited in Day & Bamford, 4-year universities (Kameyama, 2010). However, graduating
1998, pp. 173-194) with the YLs. They found that the easiest students were dissatisfied with their low performance even
books recommended by Furukawa et al. were not in the EPER after some kosen employed native English-speaking teachers,
list, and the books in the EPER starter, beginner and elementary or were equipped with language laboratories, or even had set
levels were very difficult for low-level Japanese EFL learners to up computer-assisted language laboratories with cutting edge
read without translating the English texts into Japanese. They technology (Nishizawa et al., 2004).
recommended that Japanese EFL learners should read far easier
Since 1995, the Department of Electrical and Electronic Engi-
books than the EPER recommended.
neering at Toyota Kosen has made several attempts to improve
The ER program described in this paper used the YL read- the situation for their engineering students. Both vocabulary
ability scale as its guide, and the subjects started their ER with building exercises and ondoku-hisha, a technique aimed at in-
books far easier-to-read than those indicated by the EPER scale. creasing the volume of reading and listening, were unsuccessful.
The ER program described in this paper was the Department’s
third attempt to improve the English ability of the students.
English education at Toyota Kosen
Toyota Kosen (a college of technology), where the ER program
described in this paper was conducted, is a small but specialized Research questions
institution for engineering education. In Japan, kosen typically
The research questions in this study were:
accept graduates from junior high schools, and educate them
in a 5-year foundation course and a 2-year advanced course. 1. How much reading do typical engineering students need
Usually 20% of the 5-year course students go on to the advanced to do to increase their TOEIC scores from 370 to 500, which
course. First-year (age 16) kosen students are usually excellent is the national average of fourth grade university students
in mathematics and science, and are as good in English as aver- in all majors? And how much ER should be included in our

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 633
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

curriculum to achieve this amount of reading? genres. The logbooks were collected one or two days before the
2. What are the other critical factors apart from the volume of lessons, and then returned at the beginning of the lesson with
reading? advice from the teacher added to them. The logbook, designed
by the SSS English Study Group (2005), has 320 slots for read-
3. What are the possible obstacles for implementing and run-
ing records, each of which stores the date of reading, serial
ning an ER program?
number of the record, title and series-name of the book, its YL,
the length of the text, the cumulative amount of reading, their
The ER program personal evaluation of the story, and the reader’s comments on
each book. The data from the YL and accumulated words read,
The ER program started in April 2004 in six classes in the De-
showed if the students had read easier or more difficult books
partment of Electrical and Electronic Engineering. The depart-
relative to their peers. The evaluation and comments section
ment has one class in each grade, and all the classes from second
allowed the students to tell us their favorite genre or type of
to seventh grades joined the program at the same time. The
books. Because the main role of the teachers was to introduce
students also attended compulsory English lessons during the
level and age appropriate books and genres, the teachers were
same years. In the 2007 academic year, 37 students completed
expected to read the books themselves, and to know them well.
their fourth year of the program.
Most of the lessons took place in the college library, which in
March 2008 had about 15,000 easy-to-read English books and Reading histories and examinations
2,000 audio CDs. There were three kinds of books: picture books After 3 years of the ER program, 50% of the students had read
for young English-speaking children; graded readers, which are 450,000 words, and 75% of them had read more than 300,000
reading materials for EFL learners; and story books for English- words. At this point, according to Nishizawa et al. (2006), stu-
speaking children and young adults. dents typically were comfortable reading easy English texts, had
In the lessons, students selected their reading materials, with increased their reading speed, and reported that they mainly
guidance from a teacher if needed. The teacher’s main role was used English instead of Japanese to comprehend the stories.
to help students select appropriate books and they avoided After 4 years, 50% of the students had read 690,000 or more
teaching about English so as to ensure as much reading time as words, and the percentage of the students who had read more
possible during the ER lessons. Students were strongly recom- than 300,000 words had increased to 84%.
mended to do out-of-class reading as well. The students read at From their reading logs, we analyzed the students’ average
their own pace, without referring to dictionaries, and recorded YL by their reading amount, and confirmed that the students
their reading histories in logbooks. Some of them listened to gradually increased the YL they were reading at as the amount
the stories on audio CDs while reading because listening-while- of reading increased. Representative books they read by year are
reading helped them to avoid translating into Japanese (Sakai & shown in Table 1.
Kanda, 2005). The logbooks were a valuable source of informa-
tion for the teachers to know which books the students had
already read and helped them to learn the student’s favorite

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 634
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

Table 1. Typical books students read during the program Increased TOEIC scores
We evaluated the effect of the ER program on their TOEIC
Time in the Reading YL Typical book series
scores because TOEIC measures the ability to comprehend
program amount
spoken English, the ability to read quickly and the ability to
The beginning of 0.3 Oxford Reading concentrate in English and they all depend on an ability to proc-
first year Tree Stage 3 ess large amounts of language quickly and accurately (Graham-
Marr, 2010).
The early months 250,000 1.4 Cambridge English
of second year words Readers Level 1

Third year 400,000 2.0 Oxford Bookworms

words Level 1

Fourth or fifth 800,000 2.8 Macmillan Readers

year words Level 3

As we had been expecting an increase in their reading levels

over the long term due to the ER program, we set and modified
the expected readability level of term-end examinations in light
of this. For example, the target was YL 1.2 for the first year stu-
dents, YL 1.8 for the second years, YL 2.2 for the third years, and
YL 2.4 for the fourth year students, in the 2009 academic year. In
the examinations, students were required to read an unfamiliar
English text in a limited time. The time allowed was calculated
assuming a reading speed of 100 words per minute. Students
were not allowed to take any notes during the reading time. Af-
ter the reading time, they had to answer ten questions about the Figure 1. TOEIC score distributions of groups who
story. They could easily answer 6 of the 10 questions if they read read different amounts
the whole story in the set time and were able to grasp the basic
meaning of the story, but they needed to understand critical
In Figure 1, we compared the TOEIC score distributions of
events and some details of the plot in order to answer the other
the four groups of students listed in Table 2. The students in the
four questions. In the first year of the program, they were re-
three groups (A, B, C) belonged to the ER program and were
quired to read a 3,000-word YL 1.2 text in 30 minutes. From our
doing their fourth year of the program in 2007. Seven students,
past experience, we believe this would have been challenging
who had stayed in English-speaking countries for more than
for an average graduate from a Japanese engineering college.

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 635
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

3 months during the 4-year duration of the reading program, grade kosen students (373) (ETS, 2008, p. 9, 11). The mean TOE-
were excluded from the groups. The remaining 30 students were IC score of group B (498) was as high as the national average of
divided into three groups by their cumulative reading amount. fourth grade university students in all majors in 2007 (499). The
Group A consisted of nine students whose median reading TOEIC scores of group C (604) and D (605) were the same, and
amount was 310,000 words, group B consisted of 13 students the score distributions of both groups were quite similar too.
whose median reading amount was 660,000 words, and group Even though the sample size is small, we can roughly estimate
C consisted of eight students whose median reading amount the increase in TOEIC scores from the amount read by compar-
was 1,800,000 words. The 69 students in group D had not par- ing the average TOEIC scores and the median reading amount
ticipated in the full ER program. They were third grade kosen of three groups. These data show a gain of about 18 points per
students in the academic years 2006 to 2008 who had studied in 100,000 words between groups A and B: (498-435) / (660,000-
English-speaking countries for 10 months. Some of them had ER 310,000), and 9 points per 100,000 words between groups B and
lessons, some before and some after their studies abroad, but the C: (604-498) / (1,800,000-660,000).
duration was less than one and a half years. The data show the potential of a 4-year ER program. Even
Group A, who read only in class, showed significantly higher
average TOEIC scores than their fellow kosen students. The
Table 2. The four groups of students shown in Figure 1
mid-range readers (group B), who read about the same amount
Group N ER Students’ grade Reading amount in and out of classes, achieved TOEIC scores equivalent to the
pro- and academic (Median) Min.-Max. national average of university students from all majors, includ-
gram year ing English. The top readers (group C), who read large amounts
(310,000) 280,000- outside class on their own, reached TOEIC scores as high as stu-
A 9 dents who had had their engineering education at kosen for two
390,000
and half years, and then studied in English-speaking countries
5th to 7th grade (660,000) 490,000-
B 13 Yes for 10 months. This means it is possible for engineering stu-
in 2007 820,000
dents, who are generally reluctant to learn English, to improve
(1,800,000) 1,100,000- their proficiency in English by participating in a long-term ER
C 8
12,000,000 program without sacrificing their engineering education.
Students who studied
To examine the effect of reading more than Sakai’s (2002) sug-
3rd grade in in English-speaking
D 69 No gested 1,000,000 words, we selected 19 students who read more
2006-2008 countries for 10
than 1,000,000 words and who had taken several TOEIC tests in
months
the academic years 2005 to 2008, and examined the relationship
between their TOEIC scores and the cumulative reading amount
The mean TOEIC score of group A (435), was 15 points higher (Figure 2). Students who had stayed in English-speaking coun-
than the national average of fourth grade university students tries for more than 3 months were excluded from the analysis.
majoring in engineering, science, and agriculture in 2007 (420) There were eight students whose TOEIC scores increased rap-
and was 62 points higher than the national average of seventh

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 636
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

idly at around 1,000,000 words, although such rapid increases increased their TOEIC scores by 4.1 points per 100,000 words.
did not continue in the long term. The other students were They needed to read 2,000,000 words to score 500. Remarkably,
divided into three groups, based on their TOEIC scores when the long-term gain rates for TOEIC scores were very close among
they reached 1,000,000 words. the three groups, although the students’ competency at the start
differed widely. This shows the applicability of ER at various
proficiency levels, including elementary levels, in EFL settings.

Critical factors for success

The results show that the most critical factor for success in an
ER program is the amount of reading. We could not find signifi-
cant differences in mean TOEIC scores of classes when the stu-
dents’ reading amount was around, or less than, 100,000 words.
A reading amount of 300,000 words was the threshold where
many students started to show significant increases in TOEIC
scores. To achieve this volume of reading, we need to have at
least 2-3 consecutive years of SSR and a guaranteed scheduled
reading time. Students will read from 100,000 to 200,000 words
in a year if the length and frequency of SSR lessons are the same
as our program.
The second factor that led to the success of the reading pro-
Figure 2. TOEIC score and reading amount of selected gram was to start with simple stories: SSS. When students read
students (see also larger version on final page) very easy-to-read books, they can avoid concurrent translation
from English to Japanese and comprehend directly from English
texts. If they avoid translation, they can relax and enjoy the
The most competent three students (1-3) scored 590 on stories without loss of comprehension. If they have to translate,
average when they had read 1,000,000 words and increased their their reading changes into decoding. When they decode English
TOEIC scores by 4.6 points per 100,000 words. Two of them (1, text, students think hard in Japanese but they do not use Eng-
2) further increased their scores at the same rate until 800, by lish (Nishizawa et al., 2006). There is no joy of reading either.
when they had read 6,000,000 words. The least competent three Because the joy of reading is the main motivating factor for the
students (4-6) scored 375 at 1,000,000 words and had an increase students to continue ER for a long period, it is important for
in TOEIC score of about 4.0 points per 100,000 words. From the them to avoid concurrent translation in the ER program.
data, we estimate the least competent students need to read
Displaying the appropriate readability levels of books is
4,000,000 words to score 500. The other five students (7-11), who
another important factor, and our ER program uses far easier-
were of middling competency, scored 460 at 1,000,000 words and

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 637
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

to-read books than the ones recommended by programs such Possible obstacles
as EPER. For example, we found in the first year of our trial
Short-term ER programs do not seem to demonstrate the ef-
ER program that Penguin Readers Easystarts (YL 0.8) were not
fectiveness of ER to students. Even if students feel the joy of
easy enough for some of our fifth grade students. In our ER
comprehending easy-to-read English texts at an early stage, they
program, we suggest our students read about 1,000,000 words,
tend to lose confidence when they don’t make significant gains
or more, before they start to read Macmillan Readers Level 3 (YL
in their English test score. According to our experience, typical
2.8), which is set at EPER level D (lower intermediate) and is
Japanese teenage students need to read at least 300,000 words
recommended for learners with a TOEIC score of 300. Our expe-
of comprehensible English texts before they reach the threshold
rience suggests that Japanese EFL learners with a TOEIC score
when the average TOEIC score increases significantly. The fre-
of 300 are not able to read a book at this level without concur-
quency of reading in our program suggests it will take from 2 to
rent translation. The easy-to-read books for these students are
3 years of regular SSR lessons to do this. An ER program of one
the first seven stages of Oxford Reading Tree (YL 0.0-0.7), or the
year or less may not overcome this threshold and may become
first three levels of Cengage’s Foundations Reading Library (YL
an obstacle to both perceived and actual success.
0.6-0.8).
Another possible obstacle is the lack of people who are expe-
We found that if we could satisfy the above two critical fac-
rienced with extensive reading in foreign languages in Japan.
tors: SSR and SSS, motivating reluctant students to read English
Very few Japanese adults, including English teachers, have had
books was not so difficult because the fun of reading motivated
the experience of reading English texts fluently without English-
the students. Some teachers may think that easy-to-read books
to-Japanese translation, and few native English teachers read
are too childish for college students, but the reading histories of
non-English texts fluently. As a result, most students have no ER
our students contradict that claim. If they can avoid concurrent
role model around them. It is therefore natural that curriculum
translation, even male students who play rugby are delighted
designers hesitate to include ER in their curriculums when they
to read the Rainbow Magic series: fairy stories that are popular
have no experience of ER themselves. Promoting ER among
among first grade school girls in the United Kingdom.
adult EFL learners and trainee educators might be a necessary
The third factor is the reading experience of the teachers. preparation for starting an ER program.
To aid the success of an ER program, we believe that teachers
themselves have to read at least 1,000,000 words of the books
that students are reading. Without this reading experience, the Conclusion
teachers could not share the joy of reading with the students, This study reports the impact of a long-term ER program on re-
nor have the knowledge to recommend books at the right level luctant engineering students learning English. The program was
and genre for each student, and recognize the improvement of 45 minutes a week of sustained silent reading (SSR) in which
their students. The teachers must have flexible minds to enjoy students started reading simple stories (SSS). The students read
children’s picture books and language learner literature. a median 690,000 words of easy-to-read books and increased
their average TOEIC score to 507 by their fourth year. These
data suggest it may be wise to include 4 credits of ER lessons

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 638
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

into the English curriculum of any educational institution so as References

to improve the students’ reading/listening fluency. We conclude
Day, R. R., & Bamford, J. (1998) Extensive reading in the second language
that a successful ER program for Japanese EFL learners needs classroom. Cambridge: Cambridge University Press.
long-term SSR and SSS, built on the rich experiences in ER of
ETS (2008). TOEIC test data & analysis 2007.
their teachers and supporters.
Furukawa, A., Kanda, M., Mayuzumi, M., Sato, M., Nishizawa, H.,
Hatanaka, T., & Miyashita, I. (2010). Eigo tadoku kanzen book guide
Acknowledgement [Complete book guide for extensive reading in English] (3rd ed.).
Tokyo: CosmoPier.
This study was supported in part by Japan Society for the
Graham-Marr, A. (2010, April 22). Praise, criticism of TOEIC not always
Promotion of Science, Grant-in-Aid for Scientific Research (C),
called for. The Daily Yomiuri, p. 15.
19520540.
Kameyama, T., & Ozawa, S. (2002). Koutou senmon gakkou ni okeru
eigo kyouiku no genjou to kadai [English education of colleges of
Bio data technology and its issues]. Research report for Grant-in-Aid for Scientific
Research (C) 13898006, Japan Society for the Promotion of Science.
Hitoshi Nishizawa is a Professor of the Department of Electrical
Kameyama, T. (2010). Kosen ni okeru eigo kyouiku no genjou to kadai
and Electronic Engineering at Toyota National College of Tech-
[The situation and perspective of English language education at col-
nology. His research interests include computer-assisted learn-
leges of technology]. Journal of Japanese Society for Engineering Educa-
ing in mathematics and extensive reading/listening in English. tion, 58(3), 28-31.
He is the project manager of extensive reading/listening in the
Kanda, M. (2009). The pleasures and pains of extensive reading. In A. M.
college. <nisizawa@toyota-ct.ac.jp>
Stoke (Ed.), JALT2008 Conference Proceedings (pp. 1201-1211). Tokyo:
JALT.
Takayoshi Yoshioka is an Associate Professor of the Depart- Nishizawa, H., Noda, K., Yamashita, S., Yamada, K., Hamachiyo, I., Shi-
ment of Electrical and Electronic Engineering at Toyota National mizu, T., & Andoh, H. (2004). Sotsugyo-sei ankehto ni yoru kyouiku
College of Technology. His research interests include computer- hyouka to kyouiku kaizen heno katsuyo [Graduates’ assessment of
aided educational technology and extensive reading/listening educational outcomes and improvement]. Journal of Education in the
College of Technology, 27, 555-560.
in English. He started the ER program in 2002 at the college.
<yoshioka@toyota-ct.ac.jp> Nishizawa, H., Yoshioka, T., & Itoh, K. (2006). Eigo tadoku ni yoru
kougaku-kei gakusei no eigo unyou-nouryoku kaizen [Improvement
of engineering students’ communication skills in English through
Momoyo Fukada was a Professor of the Department of General extensive reading]. Transactions of the Institute of Electrical Engineers of
Education at Toyota National College of Technology until March Japan, 126A-7, 556-562.
2010. She completed her MA TEFL at the University of Reading. Sakai, K. (2002). Kaidoku hyakumango [Toward one million words and
Her research interests include Communicative Language Teach- beyond]. Tokyo: Chikuma shobo.
ing and extensive reading/listening in English. <m7fukada@ Sakai, K., & Kanda, M. (2005). Kyoshitsu de yomu eigo 100 mango [Reading
na.commufa.jp> one million words in the English classroom]. Tokyo: Taishukan.

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 639
 Nishizawa, Yoshioka, & Fukuda • The impact of a 4-year extensive reading program

SSS English Study Group (2005). Mezase 100-mango eigo tadoku kirokute- Takase, A. (2008). The two most critical tips for a successful extensive
cho [Toward one million words reading record book] (2nd ed.). Tokyo: reading program. Kinki University English Journal, 1, 119-136.
CosmoPier.

Figure 2. TOEIC score and reading amount of selected students

JALT2009 CONFERENCE
PROCEEDINGS ONLINE  PREVIOUS PAGE NEXT PAGE  FULL SCREEN 640