Reading in a Foreign Language, Vol. 5, No. 2 (1989)

Extensive Reading vs Skills Building in an EFL Context

Thomas N. Robb and Bernard Susser

Kyoto Sangyo University and Doshisha Women's Junior College, Japan

Reading has been taught by the translation procedure in EFL situations such as Japan. But today there is a trend towards the use of ESL-type "skills building" text books and procedures and, to a much lesser extent, towards extensive reading. There is a considerable difference between these two procedures, not only for teachers and learners. but aiso for the allocation of institutional resources. Despite this. there is little useful research comparing them. This paper examines previous research on extensive reading. and then describes an experiment comparing the improvement of reading comprehension by Japanese college freshmen taught by either a skills-based or extensive reading procedure. The results suggest that extensive reading may be at least as effective as skills-building. with the important advantage that it is more interesting: for the learners.


In many parts of the world, reading has traditionally been the skill most emphasized in FL and EFL situations, and this is still true today in spite of the current emphasis on oral/aural approaches. Japan may serve as a good example: although Japanese society needs more people who can communicate in English and other foreign languages, and although most students want to improve their speaking and listening skills (Koike 1985: 158), English instruction at the secondary and university levels is devoted largely to reading taught by the translation procedure.

This procedure has not been particularly successful, considering the time and energy devoted to it (Kobayashi 1975; Hino 1988; Reischauer 1978:397ff; Miller 1982:22lff, 275ff; Tanaka 1985:4-8). Most undergraduates who study English, including many of those who major in it, cannot use English with any facility; they cannot read well enough to read English books for pleasure. being abide to decode only at the sentence level.

One procedure that may solve this problem is extensive reading. Referring to FL teaching in general, Nuttall says, "an extensive reading programme . . . is the single most effective way of improving' both vocabulary and reading skills in general." (1982:65) This is Smith's "learn to read by reading" (1985:S8), the belief that exposure to large quantities of written material combined with the effort made in reading helps to develop reading ability. Extensive reading is part of the ESL/FL curricula in many places, but has attracted comparatively little attention in Japan [-239-] (for a history. see Matsumura 1987; 165ff; for recent work, see Kitao and Shimatani 1988, Kraemer 1982, Boys 1987, and Lupardus 1987).

Previous Research On L2 Extensive Reading

Most general works on FL reading that discuss extensive reading do so in terms of book selection and course administration; some even have detailed drawings showing how to display books (Nuttall 1982: 175ff; see also Bright & McGregor 1970:65-80). In addition. some articles (e.g. Tangitau 1973, Kalb 1986, and Marbe 1979) have appeared discussing various aspects of teaching FL reading by extensive reading. Unfortunately, little data-based research has been done on extensive reading as an L2 pedagogic procedure. The following are the only research-based studies on FL/EFL extensive reading that we could find.
  1. Saragi et al. (1978) did an interesting study on how learners pick up vocabulary; their study "shows that extensive reading results in a substantial amount of vocabulary learning" (p.78). But they do not study extensive reading itself.
  2. Hamp-Lyons (1985) describes the development of an extensive reading course, from concentration on macro-level mechanical reading skills to handle sheer volume of reading, to emphasis on cognitive aspects of the reading process such as scanning and skimming; in this paper she gives no details on teaching methods or results.
  3. Laufer-Dvorkin (1981) compared "intensive" to "extensive" reading, but for her. "intensive" means 1 to 3 pages and "extensive" 7 to 10 pages. In the former, the emphasis is on vocabulary, syntax and discourse; in the latter it is on distinguishing main and peripheral ideas, explicit and implicit information, etc. From our point of view, there was no substantial difference between her two experimental conditions; we do not consider that she was studying extensive reading at all.
  4. Elley and Mangubhai (1983) tested the effect of giving Fijian school children a large number of high-interest story books in English. In a carefully conducted experiment over two years they found that the children in the experimental group improved their "general reading comprehension skill at over twice the normal rate" (61). Unfortunately, the authors do not provide data showing how much each individual pupil read, so we cannot know how "extensive" their reading was.
  5. Mason (1987; 1988; forthcoming) established an extensive reading program using graded readers mostly at the 1000 to 1600-word level. Some of the students read more than 2000 pages, but the results measured by a cloze and the TOEIC, were disappointing. Mason wonders if even 2000 pages is enough for an effect to show, and also questions if her testing devices were appropriate. [-240-]
  6. Hafiz and Tudor (1989) reported the results of an extra-curricular extensive reading program. finding "statistically significant levels of improvement" for their experimental group, particularly in writing. The "experiment" cannot be taken seriously, however, because their control group did no extra reading at all. While the authors have demonstrated that more reading results in higher reading ability, their research does not prove that extensive reading is superior to any other form of additional practice.

Summarising the problems with these studies, we can say first that there is no agreement on the meaning of "extensive reading"; how much reading must be done before It can be called "extensive"? An experiment on extensive reading must indicate clearly how much the students read. Second, there is a tendency to confuse "extensive reading" with the so-called "cognitive reading skills" of skimming and scanning; an experiment on extensive reading as a language teaching procedure should contrast this procedure with a skills-based or translation-based procedure. Third, insufficient attention has been paid to what is being read; "extensive reading" as a language teaching procedure may require or at least favor certain kinds of materials over others.

The experiment described below was designed to see if extensive reading without any overt instruction can improve students' reading ability. In another paper (Susser and Robb, forthcoming) we discuss related theoretical issues such as the definition of extensive reading; graded readers, readability formulas, and text quality; and transfer of L1 reading ability.

The Experiment

This experiment was designed to shed light on the following questions:

Can extensive reading alone improve students' reading ability?

Are skills better learned when specifically taught?

The experimenters here define one group as EXTENSIVE readers based on the relatively large volume of reading required during the experiment. We chose 50 pages as a minimum; this is roughly five times what first-year college students in Japan are required to read in their translation-type reading courses. In fact. the average number of pages read by students in the extensive group was 641; fully one-third of the students read more than 700 pages.

The SKILLS group is likewise defined as the group that read little and approached their studies from a skills-building standpoint. The textbook used by the SKILLS group was 269 pages. of which approximately one-third was texts for reading.


1. That the EXTENSIVE treatment would remain equal or be superior to the [-241-] SKILLS group in all areas where skills were specifically taught to the SKILLS group:
   Getting the main idea                  E >= S 
   Understanding the important facts      E >= S 
   Guessing vocabulary from context       E >= S 
   Making inferences                      E >= S
2. That the EXTENSIVE group would become faster readers:
   Reading speed                          E > S 

Experimental Design

Four intact groups of freshman English majors (N = 125) at Kyoto Sangyo University were randomly assigned to two instructors (the authors), with each instructor teaching one class of each treatment. As a pre-test, Form X of the Multiple Skills Series Midway Placement Test (Boning 1977) was administered during the freshman orientation and the first two weeks of class. Results are shown in Table 1 below. There were no significant differences between the groups at the outset of the experiment.

Table 1: Pre-Tests (N=125) Multiple Skills Series Midway Placement Tests (Form X) (Boning 1977)
                                       EXTEN      SKILLS     F VAL

General Skills 

  Best title (10 items)                 6.54       6.42      0.201 ns
  Getting the facts (20 items)         15.14      14.45      2.217 ns 
  Making inferences (10 items)          4.96       4.34      3.730 ns
  Vocabulary skills (10 items)          5.67       5.32      1.204 ns

Specific Skills "Making Inferences" (S's had to label statements drawn from
reading as "True," "False" or "Inference") 

  Total                                26.78      26.10      1.022 ns
  Inference items only                  8.96       9.2l      0.316 ns 

Reading Speed 

   Words in first minute               79.31      78.50       0.017 ns 
   Seconds for total passage          238.55     366.17       2.517 ns



In Class: Students read modules from the SRA Reading Laboratory Kits 2c and 3a. Students read at their own pace, all starting from the Lime level (grade level 2.5). [Davies and Widdowson point out that the SRA materials have "been used with [-242-] good results in both the first and the second language situation" (1974: 1 86); see also Boey (1976) and Eskey (1973:177), and Lynn (1972:61) for a negative opinion.] Students were permitted to go to a higher level after achieving a score of 809{:o or greater on the comprehension check of three readings of a particular level. Students spent most of each class period reading from the SRA Reading Lab, answering questions on each "Power Builder" and "Rate Builder" card, checking their answers with the answer cards, and recording their scores on their individual SRA progress sheets. While the SRA reading laboratory may be considered a "skills-based approach" to reading, we have in this experiment applied it to the EXTENSIVE group for the following reasons:
  1. We felt that it would not be acceptable to the students of our colleagues for the experimenters to let students merely read in class. We therefore needed something more goal-orientated and controlled for in class use.
  2. Although the students were required to answer the first set of comprehension questions that accompany each reading, they were not taught any skills overtly.
  3. The SRA materials as we used them suited the principles of the extensive reading classes: students could choose what they wanted to read from a reasonably wide selection; they worked individually, at their own pace; and they were engaged in silent reading for much of the class time. Motivation was high; the average number of SRA cards read during the course of the experiment was 36.2, with the range from 13 to 52.

Homework: The students were required to read a minimum of 500 pages at home during the year. A library of readers designed for American teenagers (Perfection Form Company's readers, including their "Bestsellers", "Windswept Mystery Romances," "Judy Blume," etc.) was made available, although students were permitted to select other books so long as no Japanese translation was available. The students were required to write short summaries in a notebook so that the instructors could check that the reading was actually being done.


In Class: Interactions II. A Reading Skills Book (Kirn and Hartmann 1985) was the primary text for the course. This textbook has 269 pages; each of its 12 chapters contains two reading sections (2 to 4 pages in total) and one item of "realia" per chapter. Class work consisted of going through one or two sections during the class period (90 min.) with the students individually reading the passage and doing the exercises designed to teach the skills of effrcient reading. The text was appropriate for these students in interest and difficulty. Defects of the book inciude many poorly designed questions (a common problem with such texts) and a big jump in [-243-] difficulty level near the end.

Homework: An additional section of the text was assigned for homework. A short 2-item quiz was administered at the beginning of each period to make sure that the students were doing the required reading at home. At the end of the experimental period. Form Y of the Multiple Series Midway Placement Test was administered as a posttest. Scores were analysed using Analysis of Co-variance (ANCOVA).


Results of thc post-test are summarised in Table 2.

Table 2: Multiple Skills Series Midway Placement Tests (Form Y) (Boning 1977)

                                        EXTEN      SKILLS     F VAL 

 a. Best title (10 items)                9.32       9.14      1.305    ns
 b. Getting the facts (20 items)        19.45      18.84      8.678 <.01
    Adjusted means from ANCOVA          19.41      18.88      6.68  <.02
 c, Making inferences (10 items)         8.50       8.37      3.730   ns
    Adjusted means from ANCOVA           8.45       8.42      0.019   ns
 d. Vocabulary skills (10 items)         8.88       8.62      1.538   ns
 Specific Skills "Making Inferences"		
 e. Total                               32.27      32.04      0.099   ns
 f. Inference items only                 7.59       7.39      0.223   ns 
 Specific Skills "Best Title"
 g. Total Reading Speed                  8.00       7.58      2.431   ns
 h. Words in first minute               86.55      76.75      4.580  <.05
 i. Seconds for total passage          336.39     411.90     2S.921  <.001
 SRA Reading 3c Gold No. 15 ("Sweet William") 
 j, Comprehension                        7.311       6.667    7.897  <.05
 k. Vocabulary Skills                    6.721       5.778    7.396  <.05

The pre- and post-tests are not strictly equivalent. Thus what is being examined is not improvement, but differential effects of the treatment.

All the hypotheses were confirmed. With respect to 'Getting the main idea' and 'Making inferences', there were no significant differences between the groups. The Experimental group scored significantly higher on 'Understanding the important facts' ard 'Guessing vocabulary from context'. This group's reading speed was aiso significantly faster. These results are summarised in Table 3. [-244-]

Table 3: Summary of Results


                                        Hypothesis      Result
  Getting the main idea                  E>=S           E =  S   a,g
  Understanding the important facts      E>=S           E >  S   b
  Guessing vocabulary from context       E>=S           E >= S   d,k
  Making inferences                      E>=S           E =  S   c,e,f
  Reading speed                          E> S           E >  S   h,i 


The results appear to show that extensive reading was superior to a skills approach, but we must interpret these results in light of the population under study and the experimental conditions. There are many factors that make our results considerably more tentative.

  1. Not An Optimal Skills Procedure. The skills textbook and the teaching procedure merely led the students through the exercises, assuming the prior existence of the skills/strategies rather than teaching them and developing a metacognitive awareness of their use. The students could work through the exercises mechanically without deepening their understanding of how to use these skills or transferring this ability to other reading. There was no metacognitive training in the use of reading skills (e.g. Jolley 1985); in Haas and Flower's terms, the emphasis was on content strategies with some work in function/feature strategies, but with an almost total, absence of rhetorical strategies, their most important category (1988:175-76).
  2. Contamination. The students were taking a total of 6 English courses concurrently. In addition to our "Extensive Reading" course, there was also an "Intensive Reading" course, two sections of which (one of each treatment), were reading Hawthorne and James, and two sections, D.H. Lawrence. Very little material, however, was covered during the one year term. Other classes focused on grammar, writing, speaking and listening.
  3. Student Readiness. The students came into the course having had practice with passages of only a few hundred words of length, because this is what is taught in the high schools and what they had been taught to decode for the entrance examinations. It could be that an extensive approach is more beneficial at this stage in their L2 development because it forces them to adopt new strategies to cope with the volume of material that they are forced to read. A skills approach, on the other hand. might prove more beneficial in the second and later years. This important issue must be explored further.
  4. Differences In Study Time. As revealed in the following questionnaire, the EXTENSIVE group spent twice as much time studying at home. [-245-]

Differences could have emerged purely because the EXTENSIVE group studied more.. Had we been able to make the SKILLS group spend an equal amount of time studying (whether this is possible given the nature of the materials is another question). perhaps the results would have been radically different. It may be just that the EXTENSIVE readers are able to spend more time reading thanks to the nature of the material. If "extensive" reading is, at least to some extent, "pleasure reading," students may spend many hours at home reading, but they might soon rebel if they are forced to do an equivalent number of hours of selected readings with skills-building exercises at home. In other words, it could be that if an equivalent number of hours were spent with either procedure, the results would be similar, but only in the extensive approach will students study for such an extended period of time.

Attitude Questionnaire

The questionnaire in Table 4 was administered in Japanese to the students in the last regular class. Our hypotheses and the resuits are shown to the right. The first 10 questions were answered on a scale of 0 to 5 (Disagree - Agree).

Table 4: Results of Questionnaire
Sense of Improvement
1. My reading ability improved E=S 3.113.06E=S ns
2. Thanks to this course, I can read faster and more accurately.E>S3.343.30E=S ns
Attitude Toward Method Used
3. I would like to be taught by the same method next year. E=S 3.49 3.27E=S ns
4. The work we did in class was useful for me. E=S 3.653.67E=S ns
5. The practice in class was interesting.E>S3.343.55E=S ns
6. The amount of reading homework was suitable for me.E3.003.17E=S ns
7. There was too much homework.E>S 2.932.56 E=S ns
8. We were required to do too much writing at home.E>S 2.482.46 E=S ns
Intrinsic Interest of Homework
9. The homework was interestingE>S 3.272.90E>S<.05 F= 4.457
Sense of Improvement
10. It was useful to have to do various kinds of writing at home.E>S3.783.40E>S<.05 F= 4.457
Time Spent
1l. About how many minutes did you spend reading at home each week?98.8234.81E>S<.001 F= 50.34
12. About how many minutes did you spend writing at home each week62.6743.25E>S<.01 F= 9.49

Discussion Of Questionnaire

Despite the greater amount of work required of the students in the EXTENSIVE group, as clearly shown by their self-reported homework hours (items 11 & 12), there was no difference in their attitudes concerning their improvement during the year, the classwork, or the amount of homework. The EXTENSIVE group, however, felt that the homework was more interesting and that they received considerable benefit from the summary writing that they were required to do at home. The former result, that the EXTENSIVE group would find the homework more interesting, is a logical consequence of the fact that the students could choose material interesting to them, while the SKILLS class had no choice at all. The average was only 3.2 on a scale from 0 to 5, which indicates that even the EXTENSIVE group did not find their material exceedingly interesting - only more interesting than the SKILLS group did.

It came as a surprise to the authors that the most significant difference perceived between the groups pertained to writing practice, because the course was ostensibly to teach reading! It may be that the required summary writing was perceived to be a useful exercise to the students - a type of writing practice that they could do at home and on which they got regular feedback from their instructors - both on the content of the work, because it was reviewed to check if they were writing actual summaries or merely copying from the reader, and also on the quantity of writing. This feedback, free of criticism of grammar or usage errors, might have given them a keen sense of accomplishment, particularly when they looked over their thick notebooks at the end of the term.

Further, Smith (1988:277) notes that there is evidence that "writing is one way of promoting engagement with a text. which leads to better comprehension." It may be that the students realised that their writing was a useful means to better understanding of their reading. Hafiz and Tudor (1989) also noted that there was a [-247-] very marked improvement" in their experimental group's writing tests where "a more pronounced enhancement of the subjects' test performance was observed than on the reading tests" despite the fact that the focus of the experiment was extensive reading.


The teaching of reading dominates the EFL situation in Japan, and the research reported here suggests that the extensive reading procedure is an effective and pleasurabie way for students to learn to read English as a foreign language as an alternative to translation or skills courses in which students are not free to choose reading material that interests them, and which are frequently boring for both students and teachers. If the extensive reading procedure is as effective as the skills procedure in terms of test scores, the implications for the teaching of FL/EFL reading are profound. By reading what they choose and (more or less) enjoying their homework. students' motivation to learn will increase, which will in turn benefit their eventual acquisition of the target language. Teachers will have to conduct their classes differently: they will be less dependent on inappropriate textbooks and they will have to spend more more time in individual conferences with students.

Our resuits are tentative. because many important theoreticai issues in reading are not yet resolved, and because of the limitations of our experimental conditions and design. We are now in the process of analysing the results of a repetition of the experiment in the 1988 academic year with an improved design, and hope to report the results in the near future. [1]

We would like to thank the following persons for their heipfui comments on earlier drafts of this paper: Andrew Cohen, S. Kathleen Kitao. David Peaty. Steven Ross and Vlvian Zamel. This research was presented at TESOL '88 (Chicago. March 7-12).


Boey. Lim Kiat (1976) Facilitating reading comprehension among undergraduates from Malay-medium schools. In Edward M. Anthony and Jack C. Richards (Eds) Reading, Insights and Approaches. Singapore: SEAMEO Regional Language Centre/Singapore University Press, 1-18.

Boning, R.A. (1977) Multiple Skills Series Midway Placement Test. Baldwin, New York: Barnell and Loft, Ltd.

Boys, T. (1987) Reading English without "translation." Paper presented at JALT '87. Tokyo, Japan. November 21-23, 1987.

Bright. J.A. and G.P. McGregor. (1970) Teaching English as a Second Language. London: Longman Group Ltd. [-248-]

Davies, Alan and H.G. Widdowson. (1974) Reading and writing. In J.P.B. Allen and S. Pit Corder (Eds.) The Edinburgh Course in Applied Linguistics, 3, Techniques in Applied Linguistics. London: Oxford University Press, 155-201 .

Elley, Warwick B. and Francis Mangubhai. (1983) The impact of reading on second language learning. Reading Research Quarterly, 19(1), 53~7.

Eskey, David E. (1973) A model program for teaching advanced reading to students of English as a foreign language. Language Learning, 23(2), 169-184.

Hafiz, P.M. and I. Tudor. (1989) Extensive reading and the development of language skills. ELT Journal, 43(1), 4-13.

Hamp-Lyons, Elizabeth. 1985) Two approaches to teaching reading: A classroom-based study. Reading in a Foreign Language, 3(1), 363-73.

Haas, Christina and Linda Flower. (1988) Rhetorical reading strategies and the construction of meaning. College Composition and Communication, 39(2), l67-83 .

Hino, Nobuyuki. (1988) "Yakudoku": Japan's dominant tradition in foreign language leaming. JALT Journal, (10), 45-53.

Jolley, Jeanne Swafford. (1985) Metacognition and reading, theoretical background and implementation strategies for c!assroom teachers. ED 259 301 .

Kalb, Gertrud. (1986) Die Dadaktik des extensiven Lesens im Englischunterrichtder gymnasialen Oberstufe [Teaching of Extensive Reading in English Instruction at the Senior Gymnasium Level]. Die Neueren Sprachen, 85(4), 420-430.

Kirn, Elaine and Pamela Hartmann. ( 1985) Interactions II, a Reading Skills Book. New York: Random House.

Kitao, Kenji and Hiroshi Shimatani. ( 1988) Jishu-teki na eigo no dokusho shido [Pleasure reading: Setting up a special English section in the library]. The Language Teacher, 12(2), 47-49.

Koike, Ikuo (Ed.) (1985) Daigaku Eigo Kyoiku ni kansuru Jittai to Shoraizo no Sogoteki Kenkyu (II) Gakusei no Tachiba [General survey of English language teaching at colleges and universities in Japan - students' view]. Tokyo: Daigaku ippan eigo kyoiku jittai chosa kenkyu-kai.

Kobayashi, Yuko. (1975) A new look at reading in the college program. Engiish Teaching Forum, 1 3(34): 188-195. [-249-]

Kraemer, Helen. (1982) Developing extensive reading. JALT Newsletter, 6(l), 1,4-5.

Laufer-Dvorkin, Batia. (198 1) "Intensive" versus "extensive"reading for improving university students' comprehension in English as a foreign language. Journal of Reading, 25(l), 4043.

Lupardus, Karen. (1987) Intensive/extensive reading at the advanced level. Paper presented at JALT '87, Tokyo, Japan, November 21-23, 1987.

Lynn, Robert W. (1972) Tue preparation of a reading programme. RELC Journal, 3(1-2), 60-69.

Marbe, Alan. (1979) The fourth book is quite easy. (Strategies and Techniques for Extensive Reading). Alon Lemorim leanglit/English Teachers' Journal (Israel), 2 1 : 3842.

Mason, Beniko. (1987) Effects of a self-selecting reading program on the results of the TOEIC. Paper presented at the 26th annual JACET Convention, Kyoto, Japan, September 10, 1987.

Mason, Beniko. (1988) Poster session at JALT Osaka meeting, January 17, 1988.

Mason, Beniko. Forthcoming. The effects of a self-selecting reading program on the results of the TOEIC.

Matsumura, Mikio (Ed.) (1987) Eigo no Riidingu [English Reading]. Tokyo: Taishukan Shoten, 2nd edition.

Miller, Roy Andrew. (1982) Japan's Modern Myth. Tokyo: Weatherhill.

Nuttall, Christine. (1982) Teaching Reading Skills in a Foreign Language. London: Heinemann Educational.

Parker, Don H. (1978 and 85) Mark II Reading Laboratory 2c, 3a. Science Research Associates. Chicago:

Richards, Jack C. and Ted Rodgers. (1982) Method: Approach, design, and procedure. TESOL Quarterly, 16(2), 1 53-168.

Reischauer, Edwin O. (1978) The Japanese. Tokyo: Charles E. Tuttle Company.

Saragi, T., I.S.P. Nation and G.F. Meister. (1978) Vocabulary learning and reading. System, 6(2), 72-78.

Smith, Carl B. (1988) Does it help to write about your reading? Journal of Reading, 32(3) 276-277.

Smith. Frank. (1985) Reading without Nonsense. Second edition. New York: Teachers College Press. [-250-]

Tanaka, Chie. (1985) A Study of the Effectiveness of Reading Instruction at l College lgvel in Japan Based on Psycholinguistic Theory. Unpublish Ph.D. dissertation, University of Kansas, DAI 8608453.

Tangitau, Vuki. (1973) Intensive and extensive reading. TESL Reporter, 7(1), 7-9

[This originally placed at the foot of the first page, p. 239]

Thomas N. ROBB. an associate professor at Kyoto Sangyo University, is a former president of the Japan Association of Language Teachers and is presently its executive secretary, His publications and presentations have been on composition, pronunciation, reading, and classroom speaking tasks.

Bernard SUSSER is a professor at Doshisha Women's Junior College. His publications and presentations have been on business communications, CALL, composition, reading, and communicative activities.

The authors may be contacted via Professor Robb, Department of Foreign Languages, Kyoto Sangyo University, Kamigamo Motoyama, Kita-ku, Kyoto 603, Japan.

[1]Note: The second experiment was conducted with a tighter design but never written up. The results again showed no signficant differences between the two groups.