The Effects of “L-S Reading” in English Extensive Reading
Courses
Yuichi Amitani
Tokyo University of AgricultureAbstract
Leading theorists of extensive reading (ER) have emphasized the importance of students’ choosing books freely in ER. This method has proved to be effective, but one cannot conduct an ER course without a large collection of books. In this study we tried a different approach suggested by Professor Yasuhiro Shirai. In this method, called “L-S reading,” the students all listen to and then read the same book in the classroom. This paper reports an experiment where 17 university students in Japan took an ER course using this method while the control group was taught by the grammar-translation method. The results show that those in the treatment performed better in reading comprehension and had a sense of improvement, suggesting that L-S reading could be a serious alternative to conventional ER instruction for many teachers, particularly when they find difficulties in conducting a conventional ER course.
Keywords: extensive reading; L-S reading; listening 1. Introduction
Extensive reading (ER) English courses have become increasingly popular at universities in Asia-Pacific countries. ER courses have two distinctive features compared to conventional reading courses: Reading easy materials in large quantities (Day and Bamford 1998). English ER courses are currently offered in many universities in Asian countries, including Japan. Takase (2010) reports that 30 Japanese universities offered English ER courses in 2008 and their number has probably increased.
However, there is one caveat to this story: Many ER courses rely on a university library with a sufficiently large collection of English books. For example, many of the ER courses in Japan described in the Magazine for Extensive Listening and Reading (a special issue published in September 2014) use a collection of easy readers stored in the library. This might give us the impression that we cannot have a good ER course without the help of a library. Indeed, several researchers (Robb and Kano 2013; Jeon and Day 2016) note that book acquisition and management could be a serious problem when introducing ER into a school. In this study we try a different approach to an ER course suggested by Professor Yasuhiro Shirai (Shirai 1986; 1987). In
this method, called “L-S reading” (‘L’ for listening and ‘S’ for scanning), the students all listen to and then read the same book in the classroom instead of going to the library and choosing the books they would like to read.1In the academic year 2016, the present author taught ER in core
English courses for sophomores and compared the effects of this course in their reading comprehension and evaluation of the course with the more conventional reading course. If this method is effective, L-S reading could be an option for many teachers, particularly those who wish to start an ER course with little support from their institution.
2. Background
In this section we provide an overview of L-S reading. We first review the relevant literature.
2.1 L-S reading
Shirai notes that L-S reading comprises six stages (Shirai 1986, 126; pagination from the reprint in Shirai (2012)): a. Pre-Question b. Listening c. Questions d. Silent Reading e. Questions
f. Explaining the Text in Japanese
In this method, students stay in a regular classroom during the class. Instead of choosing books on their own, the students read the same material designated by a teacher. In the class a teacher first gives a couple of questions on the material (a. Pre-Question), so that the students can stay focused while listening. Then the students listen to the material (b. Listening). After listening, true or false questions are given to check students’ understanding (c. Questions). Then the students read the written text (d. Silent Reading). The teacher adds extra T/F question(s) after the students start to read (e. Questions). After confirming that most of the students have finished reading, the teacher explains the story so that the students can understand the story, the meaning of unfamiliar words, grammatical points, and so on (f. Explaining the Text).
The key difference between the two methods lies in the fact that in L-S reading students do not read materials they themselves choose. Thus, students might end up reading materials that are too easy, too hard, or uninteresting to them. However, this method retains the key feature of ER courses: Students would listen to and read much more than in a typical English course focusing on translation. Another difference is that Shirai’s method integrates listening and reading in one class,
while students only read in the conventional ER teaching. This enables the students to experience the materials more deeply. Also, L-S reading gives the students ample opportunities to check how well they understand the story (via questions) and to enhance their understanding (via explanation of the text). Finally, because the students do not have to go to the library for this course, in this method one can offer an ER course even when the instructor does not receive assistance from their institution, although the teacher does need to choose reading materials.2
2.2 Relevant studies
There is one study of the effects of L-S Reading. Amitani (2015) taught ER courses with L-S reading at a university over two semesters (30 weeks in total; nine students participated in the courses and proficiency tests in full). Comparing pre-tests and post-tests on reading comprehension, Amitani found that the students made statistically significant progress in their reading comprehension. The students also expressed positive feelings toward the ER courses in the questionnaires. Out of 28 students taking the course, none of them gave negative answers to the questions “Do you like this course?” and “Do you believe that this course helped you improve your English abilities?” The main limitation of the study is that due to the lack of comparison groups we cannot clearly attribute these changes to the intervention. Also, since the courses in which L-S reading was conducted were not compulsory courses for the students, it might be that only those students who liked and could take advantage of this kind of course took them.
Audio-assisted reading is a method somewhat similar to L-S reading, where learners listen to and read the text simultaneously, not sequentially. In Chang and Millett (2015), 64 tenth-grade Taiwanese secondary students participated in the study. Students in the treatment read and listened to the books simultaneously (only once) while those in the comparison group read the same material silently without audio. Both groups took a pretest and posttest that measured their reading speed and comprehension. It was found that the treatment group made significantly greater progress than the comparison group did in reading speed and comprehension, although the difference between pre- and posttests for the comparison groups was also significant in both measures.
2.3 This study
To overcome the limitations of these earlier studies, we taught different courses to different groups of students taking compulsory English courses at a university. For the treatment group we taught an ER course with L-S reading; for the comparison group we largely adopted the “grammar-translation” method traditionally used to teach English reading in Japanese schools. Our hypothesis in this study is that L-S reading is effective in improving students’ ability of reading comprehension and motivating them. The research questions are:
2. Do the students feel progress in their English abilities?
3. Do the students have a positive attitude toward the L-S reading course?
A note is in order. Since in our discussion thus far we have compared L-S reading and conventional ER, one may wonder if this experiment should have compared them instead of L-S reading and the traditional method. We stress, however, that this is the first comparative study to attempt to show the effectiveness of L-S reading, and in this sense the study is preliminary. We leave it to future studies to examine whether L-S reading is as effective as conventional ER.
3. Method
To assess the effects of L-S reading, we compared an ER course taught by L-S reading and regular reading courses taught by the traditional grammar-translation method in the academic year 2016 at the Faculty X of Y University in Hokkaido, Japan.
3.1 Courses and participants
Faculty X of Y University has four departments (A-D). Students in each department take core (compulsory) English courses in their 1st and 2nd years. In the 2016 academic year, we taught one course per term (two courses per academic year) to the sophomores of each department. ER courses were taught for the students of Department A (treatment, 25 students (Spring Term) and 24 students (Fall Term)) while the standard reading courses were assigned for students of Departments B-D (control, 86 students (Spring Term) and 88 students (Fall Term)). In the spring semester 14 classes were taught (90 minutes per class, 1 class per week) and two exams were conducted (midterm and final exams). In the fall semester we taught 13 ER classes3 and
conducted the two exams. 3.2 Extensive reading classes
The details of the ER courses were as follows. Each class comprised a free-reading section and an L-S reading section, the former of which was also conducted in the comparison group. In the free-reading section, students freely chose books from the collection of leveled readers and read them for about 10 minutes at the beginning of each class (thus we spent approximately 130-140 minutes in this section per semester). The books were relatively short (80–350 words) so that they could be finished within the time limit. Several series of books were available, but the majority of the books were from the Oxford Reading Tree series (Oxford University Press, from Stage 3 to Stage 6). Students documented their choice of books and impressions on the “reading logs” distributed in the class (see Appendix 2). Their activity in this section was evaluated by the number of books they read.
method described by Shirai (1986, 1987), but with slight changes to some stages. This section is divided into three parts. First, after briefly explaining (in English) the background of the story that they would read (this part is not included in Shirai’s original description, but it was needed especially when we read one story in two weeks) and giving two true-false questions on the blackboard, the students listened to the story for approximately 10–15 minutes.
In the second part the students read the text of the story for approximately 10–17 minutes (approximately 900 words; see the next section). The books are from the starter level or Level 1 of the Oxford Bookworms series and Level 1 or 2 of the Macmillan Readers, thus the readability of the books is A1 or A2 at the CEFR level. From the experience of similar courses, we expected that the students would read the books of these levels with sufficient ease and speed. A couple of minutes after they started to read, a sheet of the true-false questions was handed out. On the sheet the first two questions were the same as those on the board, but we added a third question that required a more accurate understanding of the story to answer correctly in order to motivate them to read carefully even though they already had some idea of the story. Some students finished reading the story and answering the questions earlier than others. Those students were allowed to continue the free reading section until it was confirmed that most of the students finished reading. We then explained the story to them for approximately 30 minutes in Japanese. In the explanation we did not give a Japanese translation of the text. Our explanation mainly concerned the meanings of unfamiliar words and sentences important to understanding the outline of the story (i.e., we do not explain all the sentences in detail).4
As can be seen, we skipped Stage (c) of the method originally described by Shirai, as we judged that if we let the students answer the questions at this point it might be somewhat redundant for them. Apart from this, however, we generally followed the method as originally described.
3.3 Regular classes
In the control condition (regular reading classes), students first did the free-reading section for approximately 10 minutes, after which we taught the kind of reading comprehension class that many teachers conduct in Japan: Explaining the meaning of new words, having students translate parts of the text, explaining most sentences of the text, and having students read aloud a couple of sentences of the text.
3.4 Proficiency tests
The participants took three tests of reading comprehension administered at the beginnings of the spring term (PT1; April 2016) and the fall term (PT2; September 2016), and at the end of the fall term (PT3; January 2017). Each test included two problems taken from grades pre-1 and 2 of the STEP EIKEN tests (http://www.eiken.or.jp/eiken/en/eiken-tests/). According to their website, grade pre-1 corresponds to B2 on the CEFR scale and grade 2 to B1. From an earlier
study we decided that the easier problems were not suitable to evaluate their reading ability fully, because too many students would get full marks. The Flesch-Kincaid levels of the passages are from 8.5 to 10.3 for the grade 2 problems and from 12.1 to 14.2 for the grade pre-1 problems. In each problem students read one passage (362.0 words for grade 2 problems, 306.7 words for grade pre-1 problems on average) followed by three to four multiple-choice questions on the content of the passage. They solved two problems in 18 minutes. The number of correct answers in the proficiency tests were analyzed in a 2 (departments) 3 (timing of the proficiency tests) repeated measures ANOVA. As post-hoc analyses, independent sample t-tests and paired sample t-tests were performed to see whether there was a significant difference between the treatment and comparison groups on PT3 and a significant difference between PT1 and PT3 in the treatment group.
3.5 Questionnaires
At the end of each term we asked the students for their impressions of the course in questionnaires. Here we only report the results of the survey for the fall term because the other survey did not have questions for our study. The questionnaires contained 16 questions, 2 of which concerned our study. They are: “Did you like this course?” and “Do you believe that this course helped you improve your English abilities?” Students answered those questions anonymously by responding “I like it very much / Strongly agree,” “I like it / Agree,” “Neutral / Neither agree or disagree,” “I disliked it / Disagree,” and “I disliked it very much / Strongly disagree,” respectively. Students spent about 10 minutes finishing the survey. In Department A 22 students answered the questions while 82 students did so in Departments B-D. The results of the questionnaires were analyzed with chi-square tests of independence.
4. Results 4.1 How much students read
Free-reading section
The results of the free-reading section (conducted in the treatment and control conditions) are as follows (Table 1). In the spring term, 25 students of Department A (treatment) read 35.28 books on average, while 86 students of Departments B-D (control) read 36.52 books on average. In the fall term, 24 students of Department A read 35.04 books on average, while 88 students of Departments B-D read 33.26 books on average.
L-S extensive reading section
The results of the L-S ER section (conducted in the treatment condition) are as follows (Table 2). As noted before, in this section all the students read the same material. Students read 12,653 words altogether in the spring term and 12,107 words in the fall term. Since there were 13
ER classes in the fall term and 14 classes in the spring, the students read slightly more per class in the fall term (931.31 vs. 903.79 words).
Table 1
The Number of Books Read in the Free-Reading Section.
Departments Term Students Mean SD
A S 25 35.28 17.01
F 24 35.04 13.24
B–D S 86 36.52 10.47
F 88 33.26 10.37
Note. Those who did not submit reading logs are excluded. S: spring term, F: fall term.
Table 2
The Number of Words Read and Readability when L-S Reading was Conducted in the Treatment Condition.
Term Words (average per class) Readability
S 12,653 (903.79) A1/A2
F 12,107 (931.31) A1/A2
Note. Readability is measured by the CEFR level. S: spring term, F: fall term.
4.2 Results of reading tests
The results of the reading proficiency tests are as follows. Since each test has two problems (one from grade 2 level questions, the other from grade pre-1 level questions), we report the results for the entire test, the grade pre-1 level questions, and the grade 2 level questions.
Descriptive statistics
A total of 17 (Department A) and 73 students (Departments B-D) took all the proficiency tests and their scores were analyzed. Tables 3–5 show the average number of correct answers and standard deviations for the treatment and control classes for the entire test (Table 3), the grade pre-1 level questions (Table 4), and the grade 2 level questions (Table 5). The Flesch-Kincaid grade levels of the passages are reported. The numbers of students differ from those reported in the previous section because we only included those who took all the tests.
Table 3
Results of Reading Proficiency Tests (the Entire Tests).
Departments Students PT1 PT2 PT3
A 17 3.06 (1.89) 4.23 (1.52) 3.76 (1.52) B–D 73 2.99 (1.3) 3.90 (1.31) 2.88 (1.47)
Note. The average number of correct answers (out of seven questions). SDs are in parentheses.
Table 4
Results of Reading Proficiency Tests (the Grade Pre-1 Level Questions).
Departments Students PT1 [14.2] PT2 [12.1] PT3 [12.7] A 17 1.06 (0.97) 1.00 (0.79) 1.59 (0.94) B–D 73 1.03 (0.80) 1.03 (0.80) 0.81 (0.83)
Note. The average number of correct answers (out of three questions). SDs are in parentheses. The numbers in
brackets indicate the Flesch-Kincaid grade level of the passage. Table 5
Results of Reading Proficiency Tests (the Grade 2 Level Questions).
Departments Students PT1 [8.5] PT2 [10.3] PT3 [9.8] A 17 2.00 (1.41) 3.24 (1.15) 2.18 (1.07) B–D 73 1.96 (1.03) 2.88 (1.03) 2.07 (1.24)
Note. The average number of correct answers (out of four questions). SDs are in parentheses. The numbers in
brackets indicate the Flesch-Kincaid grade level of the passage. Table 6
Results of Repeated Measures ANOVA on the Entire Tests.
Sources Type III SS df MS F p-value Partial Ș2
Between Subject Department 8 1 7.7 2.63 0.11 0.0290
Error 257 88 2.9
Within Subject Time 32 2 17.0 10.25 0.00 *** 0.1043
DepartmentTime 5 2 2.5 1.53 0.22 0.0171
Error 276 167 1.7
***: p< 0.001, **: p<0.01, *: p<0.05, .: p<0.1. Inferential statistics
To determine the homogeneity of the treatment and comparison groups in English reading abilities before taking the courses, the numbers of correct answers on the first proficiency test were analyzed by an independent sample t-test, which found no significant difference between the two groups (t(19.7)=0.15, p=0.88), nor on the grade pre-1 or grade 2 level questions (t(88)=0.14, p=0.89 for the grade pre-1 level questions and t(88)=0.14, p=0.89 for the grade 2 level questions).5
The numbers of correct answers on the three proficiency tests were analyzed with a 2 (departments) × 3 (timing of the proficiency tests) repeated measures ANOVA. The Shapiro-Wilk normality test and Mauchly’s test of sphericity were conducted before applying ANOVA. The Shapiro-Wilk normality test demonstrated that the assumption of normality was violated for the
data of the proficiency tests (PT) 1–3 (W=.9474, p<.01 for PT1 on all tests, W=0.9182, p<.01 for PT2 on all tests, W=0.948, p<.01 for PT3 on all tests, W=0.849, p<.01 for the grade pre-1 problem on PT1, W=0.8434, p<.01 for the grade pre-1 problem on PT2, W=0.8340, p<.01 for the grade pre-1 problem on PT3, W=0.9135, p<.01 for the grade 2 problem on PT1, W=0.8244, p<.01 for the grade 2 problem on PT2, and W=0.911, p<.01 for the grade 2 problem on PT3). However, as Sokal and Rohlf (2009) note that two-way ANOVAs are robust to the violation of the normality assumption, we continued the analyses.
Entire tests
The results of the analyses of the number of correct answers on the entire test (out of 7 questions) are as follows. Mauchly’s Test of Sphericity indicated that the assumption of sphericity ZDVQRWYLRODWHGȤ2(2)=0.944, p=.0821.
A two-way analysis of variance was conducted on the influence of two independent variables (timing of the proficiency tests, department) on the number of correct answers on all tests (Table 6). The timing of the proficiency tests included three levels (PT1, PT2, PT3) and the department consisted of two levels (A, B-D). The main effect for timing yielded an F ratio of F(2,176)=10.25, p<.001, indicating a significant difference between PT1 (M = 3.00, SD = 1.41), PT2 (M = 3.97, SD = 1.35), and PT3 (M = 3.04, SD = 1.51). The main effect for department yielded an F ratio of F(1,88)=2.63, p>.05, indicating the effect for department was not significant: Departments A (M = 3.69, SD = 1.69) and B-D (M = 3.26, SD = 1.43), nor was the interaction effect significant, F(2,176)=1.53, p>.05. After the two-way ANOVA, two t-tests were conducted as follow-up tests. An independent sample t-test found a nearly significant difference on PT3 between the treatment and comparison groups, t(88)=2.227, p=0.0285, d=0.59, suggesting medium practical significance. A paired sample t-test found no significant difference in Department A between PT1 and PT3, t(16)=1.46, p>0.05, d=0.41, suggesting small practical significance.
The grade pre-1 level questions
A two-way analysis of variance was conducted on the influence of two independent variables (timing of the proficiency tests, department) on the number of correct answers in the grade pre-1 level problems (Table 7). Mauchly’s Test of Sphericity indicated that the assumption RIVSKHULFLW\ZDVQRWYLRODWHGȤ2(2)=0.954, p=.129. The timing of the proficiency tests included three levels (PT1, PT2, PT3) and department consisted of two levels (A, B-D). The main effect for timing yielded an F ratio of F(2,176)=0.85, p>.05, indicating the effect for timing was not significant, PT1 (M = 1.03, SD = 0.83), PT2 (M = 1.02, SD = 0.79), and PT3 (M = 0.96, SD = 0.90). The main effect for department yielded an F ratio of F(1,88)=3.68, p=0.058, indicating the effect for department was nearly significant, Department A (M = 1.22, SD = 0.92) and B-D (M = 0.95, SD = 0.81), while the interaction effect was significant, F(2,176)=4.36, p<.05. After the
two-way ANOVA, two t-tests were conducted as follow-up tests. An independent sample t-test found a significant difference on PT3 between the treatment and comparison groups, t(88)=3.412, p<0.001, d=0.92, suggesting high practical significance, while a paired sample t-test found a non-significant difference in Department A between PT1 and PT3, t(16)=1.77, p=0.0952>0.05, d=0.56, suggesting medium practical significance.
Table 7
Results of Repeated Measures ANOVA on the Grade Pre-1 Level Problems.
Sources Type III SS df MS F p-value Partial Ș2
Between Subject Department 3 1 2.8 3.68 0.058 . 0.0402
Error 68 88 0.8
Within Subject Time 1 2 0.5 0.85 0.431 0.0095
DepartmentTime 6 2 2.8 4.36 0.014 * 0.0472
Error 113 176 0.6
***: p< 0.001, **: p<0.01, *: p<0.05, .: p<0.1. Table 8
Results of Repeated Measures ANOVA on the Grade 2 Level Problems.
Sources Type III SS df MS F p-value Partial Ș2
Between Subject Department 1 1 1.2 0.67 0.41 0.0076
Error 155 88 1.8
Within Subject Time 38 2 20.2 18.42 0.00 *** 0.1731
DepartmentTime 1 2 0.4 0.38 0.69 0.0043
Error 180 165 1.1
***: p< 0.001, **: p<0.01, *: p<0.05, .: p<0.1. The grade 2 level questions
A two-way analysis of variance was conducted on the influence of two independent variables (timing of the proficiency tests, department) on the number of correct answers in the grade 2 level problems (Table 8). Mauchly’s Test of Sphericity indicated that the assumption of sphericity was not violated, Ȥ2 (2)=0.93, p=.0435. The timing of the proficiency tests included
three levels (PT1, PT2, PT3), and department consisted of two levels (A, B-D). The main effect for timing yielded an F ratio of F(2,176)=18.42, p<.001, indicating a significant difference between PT1 (M = 1.97, SD = 1.11), PT2 (M = 2.94, SD = 1.05), and PT3 (M = 2.09, SD = 1.21). The main effect for department yielded an F ratio of F(1,88)=0.67, p>.05, indicating the effect was not significant for Department A (M = 2.47, SD = 1.32) or B-D (M = 2.30, SD = 1.17), nor was the interaction effect, F(2,176)=0.38, p>.05. After the two-way ANOVA, two t-tests were conducted as follow-up tests. An independent sample t-test found no significant difference for PT3 between
the treatment and comparison groups, t(88)=0.331, p>0.05, d=0.09, suggesting small practical significance. A paired sample t-test found no significant difference in Department A between PT1 and PT3, t(16)=0.42, p>0.05, d=0.14, suggesting small practical significance.
4.3 Results from questionnaires
As noted before, two questions in the questionnaires concerned our study. The first question was “Do you like this course?” to find the general impression of the ER (for Department A) or regular (for Departments B-D) courses. The second question (“Do you believe that this course helped you improve your English abilities?”) was intended to find whether they believed their English abilities improved during the courses.
The results are as follows (Tables 9 and 10). The ratios of positive answers (answers coded as 1–2 in the Tables) and negative answers (those coded as 4–5 in the Tables) in the populations were compared between Departments A and B-D by Chi-square tests of independence. The ratios of positive answers were not significantly different between Departments A and B-D Ȥ2(1)=0.138,
p>.05, IRU4XHVWLRQ Ȥ2(1)=0.176, p>.05, =0.04 for Question 2), while the ratios of
negative answers were not significantly different between Departments A and B-D for Question 1 Ȥ2 (1)=1.546, p>.05, =0.12). For Question 2, the ratios of negative answers were nearly VLJQLILFDQW Ȥ2(1)=3.684, p=0.0549>.05, =0.19).
Table 9
Responses to Question 1 (“Do you like this course?”)
Departments 1 2 3 4 5
A 1 8 9 4 0
(4.5%) (36.4%) (40.9%) (18.2%) (0.0%)
B–D 8 22 26 19 7
(9.8%) (26.8%) (31.7%) (23.2%) (8.5%)
Note. Response 1 codes “I like it very much,” 2: “I like it,” 3: “Neutral,” 4: “I disliked it,” and 5: “I disliked it
very much.” The numbers in parentheses indicate the ratio of that answer in the total respondents. Table 10
Responses to Question 2 (“Do you believe that this course helped you improve your English abilities?”)
Departments 1 2 3 4 5
A 0 8 10 4 0
(0.0%) (36.4%) (45.5%) (18.2%) (0.0%)
B–D 3 22 24 23 10
(3.7%) (26.8%) (29.3%) (28.0%) (12.2%)
Note. Response 1 means “Strongly agree,” 2: “Agree,” 3: “Neither agree or disagree,” 4: “Disagree,” and 5:
5. Discussion
In this study we attempted to determine the effects of L-S reading on reading comprehension and motivation. From the results of the reading tests, it appears that students in the treatment class improved their reading comprehension skills. In the pretest we did not find any significant difference between the treatment and control groups on reading comprehension. On the other hand, we found a significant interaction effect of the timing of the tests and departments in the grade pre-1 level problems, which suggests that the students in Department A could understand the passages better than the other students. Also, post-hoc analysis revealed a nearly significant difference on the entire test between the two groups. Since only the students in the treatment group were exposed to the L-S reading section, this suggests the possibility that the students in Department A improved their abilities because of L-S reading.
We also found a significant main effect of timing of the tests on overall tests and the grade 2 level problems. This strongly suggests that some conditions were different, particularly in the grade 2 level problem of the second proficiency test, but it is not easy to determine what these differences were. One possibility is that the difficulty of the problem was different from that of the others, but this is counter-indicated because the problem was actually more difficult as measured by the Flesch-Kincaid grade level.
The results of the surveys suggest the possibility that the two groups responded differently to the questions asking their general impressions of their courses and how much confidence they felt in the improvement of their English abilities. Although the ratios of positive and negative answers to those questions did not significantly differ between the treatment and control groups, considerably fewer students in Department A expressed negative feelings toward their course. In addition, an almost significantly fewer number of students in the ER course answered negatively to the possibility that their courses improved their abilities. Given that students in the control condition were also exposed to a component of a typical ER course (the free-reading section) and that students in Departments B-D expressed satisfaction with that part in other questions of the questionnaire, the contrast between the conditions may have been starker without it. Together with the results of the reading proficiency tests, it is likely that the students in the treatment class did improve their abilities and some of them felt they had made progress.
However, there are several limitations in this study. Students made better scores on PT2 (particularly on the grade 2 level problem), but we are not sure of the cause. Since the same instructor taught the ER and regular courses, there is a possibility of implicit biases in favor of ER. There is a reason to believe that not many students took English courses outside the compulsory courses, since no such courses are offered, particularly for sophomores. Nevertheless, one cannot deny the possibility that some students in the treatment group studied by themselves, which helped them develop their abilities. There were no confirmed differences between departments in the likability of their courses. There may be some confounding factors causing a nearly significant
difference in the recognition of their progress between departments, because the homogeneity of the populations is not statistically confirmed in this respect.
6. Conclusion
Results of the reading proficiency tests and the questionnaire suggest that the students in an ER course made significant progress in their reading comprehension and experienced a sense of improvement. This suggests that L-S reading could be an alternative to teaching a conventional ER course. For one thing, as discussed above, L-S reading is relatively easy to implement. In this study only 10–15 graded readers were used for a single course. Therefore, we have shown that it is possible to conduct an ER course at a university without a library having a large collection of English books for ER. Nevertheless, as this study indicates, L-S reading could improve students’ reading comprehension abilities so much that they could recognize this fact compared to the control condition.6
Some possible directions for future research are as follows. First, an obvious limitation of this study is that we compared the effects of L-S reading with those of the traditional grammar-translation method, not conventional ER, as this was a preliminary study. However, L-S reading would only be a true alternative to conventional ER instruction when its effects are compared to those of conventional ER. The second possible direction is to find which component of L-S reading contributes to the results.7L-S reading has several components which could help
the students’ progress in reading comprehension and motivation: quality (difficulty) and quantity of texts, listening, activities focused on the meaning of the texts (e.g., questions) and so on. But we do not know exactly which components are effective in enhancing the students’ reading abilities and motivation. Finally, since we assessed the effect of L-S reading on reading comprehension, we could later assess its effects on other abilities, such as reading speed, vocabulary building, and listening.8
Notes
1. Shirai (2012) incorrectly notes that he called this method “L-R reading” in the past (Shirai, personal communication).
2. Shirai (1987) mentions this feature of his method, although he does not compare it with a conventional ER course.
3. We did not do one class when the students were taken to view an international student conference held at the university.
4. In total we spent approximately 780-840 minutes in the L-S reading section per semester. 5. We cannot give any statistical evidence for the homogeneity of the two groups in motivation for learning. But from our impression, both groups have a relatively similar level of motivation for
learning. For one thing, in Faculty X of Y University the students take different core English courses depending on their proficiency, and the students in our study were in the classes for those with the highest proficiency.
6. One might wonder why the student made progress in understanding materials of higher levels (in the proficiency tests) by reading easier materials in the regular classes (We thank an anonymous reviewer for bringing this to our attention). Day and Bamford (1998) give a possible explanation. According to them, reading materials easier than their ability helps students increase their sight vocabulary, i.e., the words which learners can automatically recognize. This helps the students even when they read more difficult materials, because it allows them to allocate more cognitive resources to unknown vocabulary and the syntactic structure of the texts.
7. We thank an anonymous reviewer for bringing this to our attention.
8. The author would like to thank Professor Harumi Nishida and two anonymous reviewers for their valuable comments and suggestions to improve the manuscript.
References
Amitani, Y. (2015). Tokyo nougyou daigaku Okhotsk campus ni okeru eigo tadoku jugyo no jissen houkoku [English extensive reading courses at Tokyo University of Agriculture, Okhotsk: Prospects and challenges]. Okhotsk Journal of Business Science 24, 47–58.
Chang, A. C. S., & Millett, S. (2015). Improving reading rates and comprehension through audio-assisted extensive reading for beginner learners. System 52, 91–102.
Day, R. R, & Bamford, J. (1998). Extensive reading in the second language classroom. Cambridge University Press.
Magazine for Extensive Listening and Reading Editorial Board, ed. (2014). Tacho tadoku magazine bessatsu: Eigo no tadoku saizensen [Magazine for extensive listening and reading special issue: Cutting-edge approaches to teach extensive reading]. Tokyo: Cosmopier.
Jeon, E. Y., & Day, R. R. (2016). The Effectiveness of ER on reading proficiency: A meta-analysis. Reading in a Foreign Language 28, 246–65.
Robb, T., & Kano, M. (2013). Effective extensive reading outside the classroom: A large scale experiment. Reading in a Foreign Language 25, 234–47.
Shirai, Y. (1986). Krashen riron no eigo kyoiku heno shisa [Implications of Krashen’s hypothesis to English education]. Koko Tsushin 11, 4–6.
Shirai, Y. (1987). Yakudoku hoo o yori ‘communicative’ na mono ni [A suggestion for making the ‘grammar-translation’ method more communicative]. In Shin Eigo Kouza Henshu Iinkai (Ed.), Sin eigo kyooiku kooza (5): Zyugyoo katei to kyoozyuhoo 3 [New course for English Language Teaching, Vol. 5: Classroom processes and teaching methods (3)] (pp. 276–79). Tokyo: Sanyusha.
Acquisition: An Introduction for English Language Teachers]. Tokyo: Taishukan. Sokal, R. R., & Rohlf, J. F. (2009). Introduction to Biostatistics, 2nd Ed. Dover.
Takase, A. (2010). Eigo tadoku tacho shido manual [Teaching Manual for Extensive Reading and Listening]. Tokyo: Taishukan.
Appendices
Appendix 1. Questionnaire items (original items are in Japanese; the other questions are omitted because they are not relevant to this study)
Q1: Do you like this course? Please circle your answer.
I like it very much. I like it. Neutral. I disliked it. I disliked it very much.
Q2: Do you believe that this course helped you improve your English abilities? Please circle your answer.
Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree
Appendix 2. Reading log. In the free-reading section, students document the information and their impression about the books they read in a reading log (see Figure 1). The information includes: date (the 2ndcolumn from the left), the title of the book (the 3rdcolumn), the series and level of the
book (the 4thcolumn), their overall satisfaction of the book (on a scale of 1 to 5; the 5thcolumn),
and their evaluation of readability (on a scale of 1 to 5; the 6thcolumn). The students write their
impressions in the rightest column. Takase (2010) helped us design the structure of the log.
