«Article Revelations from Three Consecutive Studies on Extensive Reading Sy-ying Lee National Taipei University, Taiwan, ROC lwen ...»
Revelations from Three Consecutive
Studies on Extensive Reading
National Taipei University, Taiwan, ROC
■ This paper presents three consecutive studies on the effect of extensive
reading on the development of reading and vocabulary for Taiwanese university nonEnglish majors. Each study used a different approach, with subsequent studies
adjusting the methodology in response to the results of the previous year. These
results conﬁrm other ﬁndings, using different subjects in other countries, that (1) extensive reading can be integrated into an EFL curriculum, termed in-class sustained silent reading, at the university level; (2) extensive reading is at least as effective and efﬁcient as traditional instruction in acquiring English as a foreign language and is more effective than traditional instruction when the treatment duration is longer; (3) book access and self-selection of reading are two keys to the success of a sustained silent reading program.
Keywords ■ [please supply 6-8 keywords].
Introduction Extensive Reading (ER) is not a new idea in English-as-native-language contexts. It has been well supported by a number of different researchers who have related their results to well-established theoretical frameworks.
The consensus among researchers in this area is that we acquire language by understanding messages, by being exposed to a large quantity of print that is comprehensible and compelling, by joining a literacy club in which all the members are learning on a ‘social and collaborative basis’, with no risk of being evaluated and excluded (Krashen 1982, 1985a, 1985b, 2004;
Smith 1983, 1988; Vygotsky 1978).
For ﬁrst language development and the development of literacy, Smith hypothesizes that ‘Learning about language is not the primary aim, but Vol 38(2) 152-172 | DOI: 10.1177/0033688206■■■■■■ © 2007 SAGE Publications (Los Angeles, London, New Delhi and Singapore) http://RELC.sagepub.com Revelations from Three Consecutive Studies on Extensive Reading rather the by-product of some other activities… Language is learned for its uses at the time’ (Smith 1988: 7). There is a great deal of evidence supporting this view: studies consistently show that children who grow up in a print-rich environment display superior competence in several different aspects of language and literacy, including syntax, vocabulary, spelling, and knowledge of history, culture, literature, and practical information (Cipielewski and Stanovich 1990, 1992; Chomsky 1972;
Goodman and Goodman 1982; Nagy, Herman, and Anderson 1985; Nagy, Anderson, and Herman 1987; Shu, Anderson, and Zhang 1995; Stanovich and Cunningham 1992; West, Stanovich and Mitchell 1993).
The impact of reading exposure on one’s ﬁrst language development has also been found in Mandarin using a survey study. In Lee (1995, 1996), 200 subjects from three senior high schools representing three different levels of academic achievement took part in the survey, including students from the best senior high school to those from a mediocre high school in Taipei. In this study, home environment (indicated by parental education, parents’ reading behavior, parents’ view toward reading, and number of books owned in the home) signiﬁcantly predicted the subjects’ free reading behavior, which in turn was the only signiﬁcant variable among others (leisure writing and writing apprehension) that predicted the subjects’ Chinese writing performance for a nationwide entrance examination.
Clearly, we cannot assume that what works for the child in the ﬁrst language situation will also work for the second or foreign language acquirer, but so far there is good evidence that second language acquirers also proﬁt from free voluntary reading. Those who report doing more reading in their second language outside of school do better in writing (Janopolous 1986; Lee 2005a), have greater grammatical competence (Lee, Krashen, and Gribbons 1996), and on the TOFEL examination (Constantino, Lee, Cho and Krashen 1997; Gradman and Hanania 1991).
There is also consistence evidence that pedagogical approaches based on the comprehension of messages through interesting reading materials, adapted to the ESL and EFL context, are successful. Research done in ESL/EFL situation consistently shows that a curriculum incorporating extensive reading is a teaching practice at least as effective as a curriculum based on skill-building. More often than not, the ER curriculum works better. Extensive reading, with no direct instruction on formal aspects of language, has been shown to be very effective for children, teenagers, and college students in acquiring a second or foreign language (Cho and Regional Language Centre Journal 38.2 Krashen 1994; Cho 1995; Cho and Kim 2004; Elley and Mangubhai 1983;
Elley 1980, 1989, 1991; Haﬁz and Tudor 1990; Lai 1993; Lao and Krashen 2000; Mason and Krashen 1997; Mason 2003; McQuillan 1994;
Tsang 1996; Tudor and Haﬁz 1989; Young 2001). More important, extensive reading is very pleasant to do and is a lifelong approach for language acquisition and intellectual growth (Krashen 2004).
Taiwan has been a productive laboratory for the study of extensive reading in school, or ‘sustained silent reading’ (SSR), investigating the potential of extensive reading as a tool for language acquisition and literacy development. Results thus far have shown that students participating in sustained silent reading make gains that are equivalent to or better than gains made by comparison students in classes not including SSR, in reading and vocabulary (Cheng 2003; Hsu and Lee 2005; Lee 2005b, 2005c, 2005d; Liu 2005; Sheu 2004; Sims 1996; Yuan and Nash 1992), writing (Lee and Hsu 2005), grammar (Sheu 2004), and attitudes toward reading (Lee 1998; Sheu 2004). In addition to the positive results favoring readers, researchers also suggest that an extensive reading class can be ‘less labor intensive for the teacher’ (Yuan and Nash 1992). It allows teachers also to enjoy some good quality reading as a member in the same literacy club they are inviting their students to join.
The duration of the studies listed above was either one semester or one academic year. Studies with the same duration, however, sometimes produce different results. Hence the inquiry: What makes an SSR program better than another? What are the conditions underlying successful SSR?
Our ultimate goal is to make it possible for foreign language students to have at least some of the advantages second language students have, to be able to proﬁt from extensive reading.
In order to determine why some studies produced better results than others, the author conducted three studies in three consecutive years, a one-semester study and two one-year studies, involving three different approaches to extensive reading, with each approach an improvement over the one preceding it.
The three studies were done at the same university. Experimental classes were taught by the same teacher (the researcher), and compared with control groups following the regular ﬁrst-year university English curriculum. Students participating in these studies were not taking other classes using English as a medium of instruction and had little exposure to English outside of school. Moreover, because classes are taught by different teachers with different styles and approaches, more than one Revelations from Three Consecutive Studies on Extensive Reading comparison group was included in the studies. This was done to increase ecological validity, to make it more likely that the ﬁndings can be generalized beyond the conﬁnes of the studies reported here.
Pure SSR in One Semester (12 Weeks, Spring 2001) This study examined the impact of extensive reading under less-thanoptimal conditions: Students read for only 12 weeks, had access to a limited amount of reading (215 graded readers), were asked to write summaries of what they read, and their in-class reading took place only once a week. In addition, it is likely that the students were not serious about English class. The study took place in the second half of a year-long course; the ﬁrst semester was devoted to viewing ﬁlms with Chinese subtitles. The results of the pretests shown in Table 1 revealed not only their signiﬁcantly lower proﬁciency level before treatment, they might also reﬂect their low motivation in learning English. For obvious reasons, a new instructor was brought in for the second semester, this researcher.
These conditions were not set up on purpose: they were a result of practical constraints. Nevertheless, the situation offered an opportunity to see how robust extensive reading is, and to determine if it is worthwhile to utilize an extensive reading approach when the situation is not optimal.
Consistent with common practice in sustained silent reading, students were not tested on the content of what they read. The ER class also included class discussion on language acquisition theory, including a presentation of the research evidence showing the efﬁcacy of reading as a means of developing competence in a second or foreign language. This was done to give students an orientation to the sustained silent reading approach, and to give them conﬁdence that self-selected reading would indeed positively impact their language acquisition.
Two comparison groups were used. Comparison group 1 used a textbook and did traditional reading comprehension and writing exercises.
In comparison group 2, outside reading was encouraged, but no record of the reading was kept. In addition, the instructor of group 2 devoted about 70 to 80% of class time to explaining vocabulary students encountered in the assigned text as well as related words. Neither comparison class did grammatical analysis or form-focused exercises. Both included culture, role-plays, discussions, presentations, and direct teaching of reading strategies and vocabulary.
Regional Language Centre Journal 38.2 The Nation Vocabulary Test (Nation 1990), and a cloze test constructed by Mason (Mason and Krashen 1997), consisting of approximately 1600 words with every tenth word deleted, were used as pre- and posttests.
Because of pretest differences, posttest scores were adjusted by ANCOVA. The reading group did slightly better than the ﬁrst comparison group on the vocabulary test, but the difference was not signiﬁcant (p =.32). Comparison 2 did better than the reading group, and the difference fell just short of statistical signiﬁcance (p =.07). Inspection of the raw data (Table 1) shows that the comparison 2 group’s advantage was due to performance on the part of the test that contained less frequent words, those at the 5000 word level. This may be due to the fact that subjects in the experimental group read mostly graded readers written at the 2000 and 3000 word level. It can be postulated that given more time for the reading group, allowing the subjects to move on to higher level materials, they would very possibly show improvement on the less frequency words.
Note. The vocabulary level tests were taken from Nation (1990) [Typesetter: suggest all cols of numbers are decimal point aligned] On the cloze test, the experimental and comparison 2 groups made clear gains, but comparison group 1 did not, as shown in Table 2. The reading group made larger gains than comparison group 1 (p.05).
Comparison 2 did better than the readers, but the difference was not signiﬁcant.
The results of this study are consistent with previous reports of the efﬁcacy of using graded readers (Mason and Krashen 1997), and with the desirability of sharing language acquisition and reading theory with students (Lee 1998). It was also shown that students of English as a foreign language can improve without producing language, without formfocused activities, and without being tested on what they read.
By Figure 1, which provides a visual representation of all the numerical data, it can be observed that the Experimental group made most gain on the 2000 word level test and the cloze test. Control Group 2 made most gains on the 5000 word level test as well as the cloze test; while Control Group 1 made gains that were more reserved.
The limitations of Study I were obvious: the experimental group was signiﬁcantly inferior to their two comparison groups on all measures in the pretests; and the two comparison groups might have reached the ‘ceiling effect’ on the 2000 VLT, for which the full score was 54. These ﬂaws were repaired in the study in the following academic year by lengthening the treatment duration, providing more books, and having groups with equal performance on all pretest measures.
Spring 2003, with the Second Semester Conducted as an Assigned Reading Literature Class) In this study, some of the ﬂaws of Study I were repaired. All groups had equivalent levels of English competence at the start of the treatment and the treatment lasted one year. In this study, three comparison classes (Com 1 = 40; Com 2 = 45; Com 3 = 54), randomly selected from the 26 Freshman English classes at the same university and one experimental class (N = 67) taught by the researcher were involved.
In this study, the vocabulary measure developed by Schmitt (2000) replaced the Nation measure used in Study I. Schmitt’s measure covers 2000, 3000, 5000, 10,000 and academic level words. According to Nation (2001), Schmitt’s measure was ‘a major improvement’ over the original measure because lower frequency words and academic level words are also assessed.
The results of a Multivariate Analysis of Variance indicated that the three comparison groups were not signiﬁcantly different on the pretests, so they were combined into one group. Table 3 presents the descriptive pretest data for the experimental group and the three comparison groups.
An Independent t test showed that the experimental group (n = 67) and the combined comparison group (n = 139) were not signiﬁcantly different on all pretests (for total vocabulary, F = 1.84, p =.145; for the cloze, F = 1.65, p =.18).
Table 3. A Mean Comparison Process Comparing the Pretest Scores of Each Group