Implementing Silent Repeated-Reading: How Much Can Non-Native Readers of English Be Expected to Develop their Reading Fluency?

Transcription

1 103 Implementing Silent Repeated-Reading: How Much Can Non-Native Readers of English Be Expected to Develop their Reading Fluency? Timothy Gutierrez Introduction The pedagogical technique of repeated-reading has a well-developed history of being used to develop reading fluency in native speakers of English. Concurrently developed by both Chomsky (1976) as well as Samuels (1979), having students repeatedly reading a text aloud has been shown to develop reading fluency in native speakers. Typically, a native English speaking developing reader who engages in repeated-reading will read a text which is at the reader s grade level repeatedly to a specified number of repetitions or until the reader can read the text to a criterion of a desired reading speed. Research into repeated-reading with native speaking children has shown that repeatedly reading a text aloud can improve native speaking children s reading fluency and reading comprehension (Schwanenflugel & Knapp, 2016, p ). Adapted to developing reading fluency in non-native readers of English, an English learner will read the same text, typically four times, but potentially up to seven times, in a single session. Nation (2009) describes that repeated-reading fits the requirements for a fluency development activity if (a) the students are reading for the meaning contained in the text, (b) the material the students are reading is within the range of control of the students, (c) the student is prompted to go faster while reading, and (d) there is a sufficiently large amount of reading. Repeated-reading fits into Nations (2007) four strands framework for designing a language curriculum for non-native speakers which recommends that about one-fourth of instructional time in a second language curriculum be devoted to fluency development activity. For example, in a reading class, students should be engaged in activities where reading fluency is developed about one-fourth of the time. In order for a teacher or an administrator to make the decision to include repeated-reading in an English as a foreign or second-language program, it is necessary to know what the likely outcomes of engaging in silent repeatedreading will be. The purpose of this paper is to inform such teachers and administrators as to the potential benefits of implementing silent repeated-reading with their students of English as a foreign language.

2 104 Implementing Silent Repeated-Reading To fulfill the purposes of the paper, I review the literature investigating the effects of the pedagogical technique of silent repeated-reading on non-native readers of English. I discuss the findings of studies which investigated the effects of silent repeated-reading on non-native readers of English. The research has looked at several primary research questions. The first question asks how much faster can a native speaker read a text when reading the text repeatedly. The second question asks how much faster do readers read a new passage as they progress along a treatment period of repeated-reading. The third question asks can improvements in reading rate be transferred to texts which are different from the texts which are being used to read in the repeated-reading sessions. The fourth question asks if readers can improve their reading comprehension through engaging in a treatment of repeatedreading. The fifth question asks how repeated-reading improves students reading fluency. Methods I searched for studies to review in the following manner. I used the search term repeated reading on both Google Scholar and the library search engine from Temple University to find studies on using repeated-reading in the English language classroom. The library search engine amalgamated a search of 34 academic databases of research into education and psychology. From the studies I found, I searched the previously mentioned databases for studies from the authors who had authored papers investigating repeated-reading. Following these searches, I manually checked the references list of the papers I found for further studies. The purpose of this study was to review the effects of silent repeated-reading on reading fluency. As such, studies which investigated reading aloud were not included. In addition, this study focused on the results of English language learners. As such, studies which investigated improvements in languages other than English were not included. Seven studies were found which fit the criteria. While most studies of repeated-reading on non-native learners of English focus on multiple aspects of reading ability, for example the effect of engaging in repeated-reading on reading rate and reading comprehension, along with the learners perspectives on the repeated-reading activity. In this review, I analyze each aspect investigated under separate research questions. To facilitate comparison of the effects of repeated reading, each study is reviewed under multiple categories if it investigated multiple aspects of reading fluency. The purpose of this analysis strategy is to highlight just how much and by what qualities an aspect of reading fluency is developed through repeated reading.

3 Implementing Silent Repeated-Reading 105 Results RQ1: Can students read a text faster upon each repetition? The first research question investigated how much students reading changed during the procedure of repeated-reading itself. In this section, I review the four studies that explicitly investigated this phenomenon and reanalyze the data reported from one study to determine if students read faster during the repeated-reading treatment. In all cases, students read faster a text faster on their fifth repetition than they read on the first repetition. Where it was investigated, students also read the text faster on the eight repetition when compared to the first as well as the fifth repetition. The 16 Japanese university students investigated by Taguchi (1997) statistically significantly increased their silent reading rates on the first reading passage from the first repetition ( words per minute) to the fifth repetition ( words per minute), as well as from the fifth repetition to the seventh repetition ( words per minute). On the final session, the 28th session, the students also statistically significantly increased their silent reading rate from the first repetition ( words per minute) to the fifth repetition ( words per minute), and from the fifth repetition to the seventh repetition ( words per minute). While Taguchi and Gorsuch (2002) did not specifically analyze the average increases in the speed by which the 9 Japanese students in their study read the treatment passages as an average, one can compute from the data given that the students read the 28 passages an average of words per minute on the initial reading of a treatment text, read the same passage at words per minute on the fifth repetition, and at words per minute on the seventh repetition. 10 Japanese students with slow initial reading rates statistically significant improved their silent reading rate as an average of their reading rates across all the treatment sessions (Taguchi, Takayasu-Maass & Gorsuch, 2004). Students read the treatment texts on average words per minute on the first reading and read the treatment texts on average on the fifth repetition of reading the text. The average increase of words per minute was significant statistically. A single Japanese advanced proficiency student of English increased her reading speed words per minute on the fifth repetition of the text with the initial average reading being at words per minute and the average fifth repetition being at words per minute (Taguchi, Gorsuch, Takayasu-Maass & Snipp, 2012). Likewise, her average sixth repetition was read at words per minute which was 9.22 words per minute faster than the fifth repetition and words per minute faster than the average initial reading.

4 106 Implementing Silent Repeated-Reading 24 Vietnamese learners of English statistically significantly improved their silent reading rate within the treatment sessions with the students reading on the texts initially at an average rate of words per minute and on the fifth repetition of the text were reading at words per minute (Gorsuch & Taguchi, 2008). This statistically significant gain of words per minute was larger than the gains made by the Japanese students investigated previously. In most cases, students who engage in silent repeated-reading increase their reading speed by approximately 40 words per minute on the fifth time they read a text as compared with the first time they read the same text. Vietnamese students (Gorsuch & Taguchi, 2008) who initially read the texts faster than the Japanese students studies (Taguchi, 1997; Taguchi & Gorsuch, 2002; Taguchi, Gorsuch, Takayasu-Maass & Snipp, 2012; Taguchi, Takayasu-Maass & Gorsuch, 2004), made much larger reading rate gains in reading a text on the fifth repetition than the Japanese students studied. RQ2: Can students read the treatment text faster after engaging in repeated-reading? A second finding from research into the effects of repeated-reading with non-native readers is that students while reading a text during a sustained repeated-reading program can read the treatment texts progressively faster as the repeated-reading program progresses. In this section, I review the six studies which investigated how much faster a non-native reader of English reads subsequent texts while engaging in a program of repeated-reading. As with the finding that students can read a text faster the more times they read the text, the finding that students read treatment texts faster at the end of a period of engaging in repeated-reading is replicated throughout the studies. A number of investigations which have shown that repeated-reading silently can produce benefits in improving silent reading rate. Taguchi (1997) found that 16 Japanese university students could read the passage used in the 28 session faster than students could read the passage used in the first session. When comparing the initial reading of the passage, students read the final passage at words per minute which was a gain of words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. Similarly, when comparing the fifth repetition of reading the passages, students read the final passage at words per minute which was a gain of 29.8 words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. Likewise, when comparing the seventh repetition of reading the passages, students read the final passage at words per minute which was a gain of words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. Even though the gains made by the students in silent

5 Implementing Silent Repeated-Reading 107 reading rate were not statistically significant, this study showed that assisting student s reading by concurrently listening to an oral reading of the passage could increase student s silent reading rate across sessions. The results finding that Japanese students could read English texts faster after engaging in a treatment of repeated-reading was first replicated with 8 Japanese university students by Taguchi and Gorsuch (2002). The students made similar increases in reading rate to Taguchi s (1997) previous study. When comparing the initial reading of the passage, students read the final passage at words per minute which was a gain of words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. Similarly, when comparing the fifth repetition of reading the passages, students read the final passage at words per minute which was a gain of words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. Likewise, when comparing the seventh repetition of reading the passages, students read the final passage at words per minute which was a gain of words per minute from their reading rate of words per minute when reading the passage used in the first treatment session. The authors (Taguchi, Takayasu-Maass, & Gorsuch, 2004) replicated the improvements in silent reading rate with 10 lower proficiency Japanese university students who engaged in repeated-reading for 42 sessions. These lower proficiency students statistically significantly improved their silent reading rate when reading the treatment passages. The silent reading rate, word per minute, from the final four sessions averaged together (sessions 39, 40, 41, and 42) showed that the students could read words per minute faster at the end of the repeated reading treatment period than they could at the beginning as computed by averaging the students reading rate on the first four sessions (83.97 words per minute). Gorsuch and Taguchi (2008) replicated their previous findings of the effectiveness of engaging in repeated-reading in increasing rates of reading English in Japanese students with 24 3rd year Vietnamese university students. The Vietnamese participants statistically significantly increased their reading rate both within treatment sessions as well as when comparing their speed of reading the final text of the treatment period with the speed of reading the first text of the treatment period. The students read the 16th treatment passage initially at words per minute which was words per minute faster than the words per minute by which they read the first treatment passage. Likewise, for the fifth repetition, which the students read the 16th treatment passage at words per minute which was words per minute faster than the words per minute by which they read the first treatment passage.

6 108 Implementing Silent Repeated-Reading A Japanese advanced proficiency non-native reader of English also increased her reading rate through 14 weeks (i.e., 69 sessions) of self-administered repeated reading (Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012). When comparing her initial reading of the passage, she read the final passages (i.e. the average of the final five sessions) at words per minute which was a gain of words per minute from her average reading rate of words per minute when reading the passages used in the first five treatment sessions. 26 Taiwanese college students increased their reading speed through 26 sessions (over 13 weeks) of repeated-reading (Chang & Millett, 2013). When comparing the initial reading of the passage, students read the final two passage at an average of 150 words per minute which was a gain of 47 words per minute from their average reading rate of 103 words per minute when reading the passages used in the second and third treatment sessions. Improvements in reading rate were made by students in all studies on the initial reading of the final treatment text. Two of the studies (Chang & Millett, 2013; Gorsuch & Taguchi, 2008) showed that students who engaging in repeated-reading improved their reading rate on the treatment texts to a statistically significant degree as compared to a comparison group, although students in one study (Taguchi & Gorsuch, 2002) improved their reading rates without statistically significantly outperforming the comparison group. The other studies either did not show statistically significant gains (Taguchi, 1997; Taguchi, Takayasu-Maass & Gorsuch, 2004) or the reading rate could not be compared to others as it was a diary study (Taguchi, Gorsuch, Takayasu-Maass & Snipp, 2012). RQ3: Can students read a new text faster after the repeated reading treatment? A third finding from research into the effects of repeated-reading with non-native readers is that students while reading a text during a sustained repeated-reading program have seldom been able to read a text different from the treatment text at an increased reading rate which the students have been able to read the treatment text. In this section, I review the five studies investigating how much faster a non-native reader of English is likely to be able to read a text which is different from the texts which were used while engaging in a program of repeatedreading. In contrast to the previous two findings, the finding that students can read a text different from the texts used to implement the repeated-reading treatments at a faster rate than a comparison group is only demonstrated in one study. An initial investigation into reading rate changes on different texts was conducted by Taguchi and Gorsuch (2002). Before and after a 28-session treatment period of repeated reading, the 9 first-year Japanese university students in the study were timed on the initial reading of a standardized reading passage. These students who engaged in repeated-reading during the

7 Implementing Silent Repeated-Reading 109 treatment sessions statistically significantly increased their reading rate to words per minute on the posttest which was an increase of words per minute when compared to words per minute, which was the speed by which they initially read an equivalent passage as a pretest. However, these gains were not significantly different from reading rate gains (10.49 words per minute) made by a control group of 9 equivalent proficiency first-year Japanese university students doing intensive reading for the 28 sessions (pretest = words per minute; posttest = words per minute). Taguchi, Takayasu-Maass, and Gorsuch (2004) again investigated reading rate changes on a new text. Before and after a 42-session treatment period of repeated reading, the 10 first-year Japanese university students in the study were timed on the initial reading of a standardized reading passages designed for native speaking fourth grade students. These students who engaged in repeated-reading during the treatment sessions read the posttest at the slightly slower rate of words per minute than the rate, words per minute by which they read the pretest passage. These rates showed no statistically significant difference from each other, nor was there any statistically significant difference from the rates by which 10 students of equivalent proficiency who engaged in extensive reading during the treatment period read the testing passages (pretest = words per minute; posttest = words per minute). Although, the slower rate by which all the students read the posttest was most likely a related to an inequivalancy between the test form for non-native readers of English. The silent reading speed of 24 3rd year Vietnamese university students of Gorsuch and Taguchi s study (2008) was measured in two passages developed to test reading comprehension changes through the 16 sessions (11 weeks) of repeated-reading. The authors investigated the students silent reading rate changes on two sets of texts which were designed to test students changes in reading comprehension. These texts also served as tests of reading rate changes on different texts due to the 16 sessions of repeated-reading. The first type of non-treatment text for which students were tested on was derived from an essay written for young native speakers of English. On this text, the students who engaged in repeated-reading during the treatment sessions decreased their initial reading rate to words per minute on the posttest designed as a short-answer test which was a decrease of 17.2 words per minute when compared to words per minute, which was the speed by which they initially read an equivalent passage as a pretest. Although the rate by which the experimental group read the pretest passage initially was statistically significantly faster than the rate by which the control group read the passage initially (pretest = words per minute), there was no statistical difference between the rate by which the control group read the posttest passage initially (i.e words per minute).

8 110 Implementing Silent Repeated-Reading The second type of non-treatment text for which students were tested on was adapted from a graded reader and was used to test the ability of the students to recall the contents of the story. These students who engaged in repeated-reading during the treatment sessions increased their initial reading rate to words per minute on the posttest designed as a recall test which was an increase of words per minute when compared to words per minute, which was the speed by which they initially read an equivalent passage as a pretest. These results were also not statistically different from the rates by which a control group read the same passages (posttest = words per minute; pretest = words per minute). The advanced proficiency learner in Taguchi, Gorsuch, Takayasu-Maass, and Snipp s (2012) study increased her posttest reading rate after 14 weeks (i.e., 69 sessions) of self-administered repeated reading on a short story divided into two sections used as a pretest and a posttest which was primarily used to test reading comprehension. She read the posttest passage initially at words per minute which was a gain of words per minute when compared to the words per minute rate at which she read the pretest passage initially. Her reading performance on the test passages were close, but less than her reading performance on the first and last five passages averaged together (posttest = words per minute; pretest = words per minute; gain = words per minute). Despite the inconclusive results of the previous studies, one study showed positive results of students being able to read a passage different from the treatment passage at an increased rate comparable to the rate increases on the treatment passages used in repeated-reading. The 26 Taiwanese college students investigated by Chang and Millett (2013) improved their reading rate on narrative passages after 26 sessions (over 13-weeks) of silent repeated-reading of expository texts. They read the posttest passage at 145 words per minute which was a gain of 45 words per minute when compared to the 100 words per minute rate at which they read the pretest passage. These gains were comparable to the gain of 47 words per minute by which the students were able to improve their reading of the expository treatment passages (pretest = 150 words per minute; pretest = 103 words per minute). Clear improvements in reading rate on new texts above a comparison group were only made by students in one study (Chang & Millett, 2013). Students in two studies (Taguchi & Gorsuch, 2002 and the recall tests of Gorsuch & Taguchi, 2008) improved their reading rates without statistically significantly outperforming the comparison group. The other studies showed either non-significant changes (the short answer tests of Gorsuch Taguchi, 2008) or decreases in reading rate on new texts (Taguchi, Takayasu-Maass & Gorsuch, 2004). It should be noted that students read the new texts used as a posttest at a slower rate than the treatment texts used as posttests in all cases.

9 Implementing Silent Repeated-Reading 111 RQ4: How does reading comprehension improve due to silent repeated-reading? A fourth finding from research into the effects of repeated-reading with non-native readers is that reading comprehension can be improved through engaging in silent repeated-reading for a treatment period. In this section, I review the results from five studies which investigated changes in reading comprehension made by students of English. As with the previous three findings, the findings that students can improve their comprehension through engaging in repeated reading is replicated through the studies. However, gains in reading comprehension made through engaging in repeated-reading are comparable to gains in reading comprehension from students who engage in another form of reading instruction. As measured in correct or partially correct responses to close-ended questions designed to test if a reader had understood the passage, the 28 students who engaged in 28-sessions of silent repeated reading scored progressively higher on a new text designed to investigate students silent reading fluency (Taguchi & Gorsuch, 2002). These students scored 9.33 out of 20 after reading the posttest passage once, after reading the passage three times, and after reading the passage seven times. The students score on the comprehension test after reading the posttest passage once was statistically significantly higher than the students comprehension score after reading the pretest passage once. Also, the students scores on the comprehension tests after reading the pretest and posttest passages seven times were statistically significantly higher than the students score after reading the passage only once. However, these scores where not statistically significantly different from the scores of the comparison group which engaged in intensive reading for the treatment period. Similar reading comprehension results on 8 open-ended comprehension questions for a total of 16 points were obtained by 10 students low proficiency Japanese students of English who engaged in 42 sessions of silent repeated-reading (Taguchi, Takayasu-Maass, & Gorsuch, 2004). These students scored 3.90 after reading the posttest passage once, 8.40 after reading the passage three times, and 8.80 after reading the passage five times. As with Taguchi and Gorsuch (2002), both groups statistically significantly increased their reading comprehension from the first reading to the fifth reading in each test. However, there were no statistically significant gains in the posttest comprehension test made by the students who engaged in repeated-reading when their test results were compared to the results of the pretest comprehension test. In addition, the group who engaged in repeated reading did not differ in reading comprehension from the 10 comparable Japanese students who engaged in extensive reading for the same amount of sessions. More promising results were obtained in the following studies. 24 3rd year Vietnamese university students who engaged in 16 sessions (11 weeks) of silent repeated-reading

10 112 Implementing Silent Repeated-Reading demonstrated higher levels of reading comprehension on a test measuring short answer recall and a test measuring proposition recall (Gorsuch & Taguchi, 2008). To measure the students ability to answer short questions about a reading passage, the authors asked the students to answer eight short answer questions after reading a passage that was similar to the repeated reading treatment passages. The students who engaged in repeated reading answered 41.2% of the propositions correctly after reading the posttest passage once and answered 58.2% of the propositions correctly after reading the posttest passage five times. These results where statistically significantly higher than the students results on the pretest. The students also answered statistically significantly more propositions on the short answer test than the students who participated in the control group. The authors also investigated recall of story propositions as reading comprehension. The students who engaged in repeated-reading recalled 26.1% of the story s propositions after reading the posttest passage once and recalled 61.5% of the story s propositions after reading the posttest passage five times. As with the short answer test, the students who engaged in repeated-reading recalled statistically significantly more of the posttest passage after the repeated reading treatment than they recalled on the pretest passage. Likewise, they recalled statistically significantly more propositions than the control group recalled on the posttest passage. The advanced English proficiency Japanese adult woman who engaged in 14 weeks (i.e., 69 sessions) of self-administered repeated reading also successfully increased her reading comprehension (Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012). Similar to the reading rate results reported earlier, the participant s reading comprehension increased across testing sessions (i.e. from pretest to posttest). By answering 5 open-ended questions written to test the participants understanding of the events of the story, the participant increased her reading comprehension score from 4 after the first reading of the pretest passage to 8 after the first reading on the posttest passage. Contrasting to expectations, she received a full comprehension score of 15 after reading the passage twice on the pretest, but she needed to read the posttest passage three times before receiving an equivalent score on the posttest. 26 Taiwanese college students after a 13-week (i.e. 26 sessions) period of engaging in silent repeated reading treatment period clearly maintained their reading comprehension as measured by the two types of texts used to test the students silent reading speed. (Chang & Millett, 2013). The authors investigated changes in comprehension of two types of reading passages. The first type of reading passages were expository passages which were comparable to the treatment passages. After the treatment, students who had engaged in repeated reading answered 70% of the multiple-choice questions about two posttest passages which was

11 Implementing Silent Repeated-Reading 113 statistically significantly higher than their scores on comparable pretest passages. The second type of reading passages used in the study were narrative passages developed from a graded reader. The students also answered 66% of the multiple-choice questions about a story with similar vocabulary composition to the treatment passage, which was statistically significantly higher than their score on the questions after reading the same story before the treatment sessions. The comprehension scores of the treatment group on both posttests were statistically significantly higher than the scores of the group which engaged in a treatment consisting of timed, single readings of the treatment passages. The results of these studies support the notion that students who engage in silent repeatedreading increase their ability to comprehend similar texts after the period of engaging in repeated-reading. The early studies (Taguchi & Gorsuch, 2002; Taguchi, Takayasu-Maass, & Gorsuch, 2004) showed that students comprehended only a small portion of the text on the initial reading of the text. The studies with carefully controlled tests of comprehension (Chang & Millett, 2013; Gorsuch & Taguchi, 2008) showed students can improve their comprehension more than a comparison group. RQ5: How does Repeated-Reading develop reading fluency? A fifth finding from research into the effects of repeated-reading with non-native readers is that non-native readers of English who have engaged in silent repeated-reading have perceived that repeated-reading has improved their reading fluency. In this section, I review the results from three studies which investigated how the learners who engaged in repeated-reading perceived the repeated-reading treatment. Consist through these studies are student perceptions that repetitions assist with fluent reading of the treatment texts, audio renditions of the texts assist students in integrate verbal language with text, and the combined processes of repetition along with the audio assisted vocabulary recognition and boost reading comprehension. Analysis of a post treatment questionnaire answered by Japanese students who engaged in 42-sessions of repeated-reading (Taguchi, Takayasu-Maass, & Gorsuch, 2004) found that students perceived that the repetition in repeated reading had a positive role in developing their reading fluency and comprehension. Students also perceived that the concurrent audio version of the text assisted the development of their listening skills. In addition, repeated reading helped the students develop a tolerance for reading longer texts and cope more efficiently with unknown words. Likewise, the perceptions of an advanced English proficiency Japanese adult woman who engaged in 69 sessions of self-administered repeated-reading were analyzed through diary

12 114 Implementing Silent Repeated-Reading entries following each repeated-reading session (Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012) supported the assertion that engaging in repeated-reading improves a non-native speaker s English reading fluency. The participant reported that repeated reading assisted her comprehension of the text at both the discourse level and the sentence level as well as assisted her comprehension of both grammar and vocabulary used in the passages. However, the participant reported the repetition of reading could be overdone and lead to decreased motivation and could not overcome the hindering effect of a critical mass of unknown words in the text. The participant reported that the audio model helped her develop fluency by making her keep up the pace of reading the passage and assisted her comprehension of the text by helping her comprehend dialogs and recognize the pronunciation of words. However, the audio did not help her overcome unknown features of the text. A descriptive and exploratory analysis of the end of session reports written by 30 Vietnamese college students who engaged in 16-sessions of repeated-reading (Gorsuch & Taguchi, 2010) found five reading behaviors which engaging in repeated-reading helped to develop. Firstly, the students reported a wide use of reading strategies while engaging in repeated reading with topdown and meta-cognitive reading strategies being reported widely in the beginning sessions and top-down strategies being mentioned more widely during the later sessions. Secondly, the repeated reading method motivated the students to read more intently and widely as well as apply the repeated reading method outside of class. Thirdly, students mentioned consistently, but not widely, perceived improvements in reading speed and comprehension, with some mention of the repeated reading treatment as causing the improvement. Fourthly, the participants perceived that the repeated-reading treatment assisted the participants with a variety of language skills other than reading, especially the audio model to which the students listened to while reading. Finally, the participants mentioned to a limited degree that word recognition became easier and the students could focus on higher level features of the texts as the treatment period progressed. Three studies found the perceptions of students who engaged in repeated-reading to support the idea that silent repeated-reading with an audio model increased the students ability to read faster and comprehend more easily. Learner perceptions of engaging in repeated-reading supports the assertion that repeated-reading improves a learners reading fluency, reading comprehension, and ability to implement reading strategies. There is also support that learners move away from lexical and grammatical concerns to a focus on sentence and discourse level meaning, as predicted by Automaticity Theory (LaBerge & Samuels, 1974).

13 Implementing Silent Repeated-Reading 115 Discussion The primary goal of this study was to specify how silent repeated-reading could improve nonnative students English reading fluency by reviewing the literature which investigated reading fluency gains from English language learners who engaged in repeated reading. The studies investigated five aspects of reading improvements through repeated reading. Researchers are interested (a) in how much students read faster while repeatedly reading a text; (b) in how much faster students can read treatment texts at the end of a period of repeated-reading; (c) in how much faster students can read new texts at the end of a period of repeated-reading; (d) in how much readers can improve their reading comprehension at the end of a period of repeatedreading; and (e) the processes by which repeated-reading improves readers reading fluency. RQ1: Can students read a text faster upon each repetition? The first research question asked if non-native speaking students of English read a single text faster as they repeatedly read a text designed for repeated-reading. A clear finding from research into the effects of repeated-reading with non-native speakers is that students can read a text faster, the more times they read the text. This finding is clearly replicated in a series of studies by Etsuo Taguchi and his colleagues. For the most part, Japanese students improve their reading by approximately 40 words per minute on a fifth repetition of a text studies (Taguchi, 1997; Taguchi & Gorsuch, 2002; Taguchi, Gorsuch, Takayasu-Maass & Snipp, 2012; Taguchi, Takayasu-Maass & Gorsuch, 2004), but Vietnamese students (Gorsuch & Taguchi, 2008) were able to improve their reading by words per minute. The results discussed above should take into consideration some benchmarks for a good silent reading speed. A good careful silent reading speed is around 250 words per minute (Nation, 2009, p ). As a comparison, native speaking readers read for enjoyment (i.e. Rauding) at 300 words per minute, read for learning at 200 words for minute, and read for memorizing at 138 words per minute (Carver, 1997). As a further comparison, third-year Chinese university students of English read for enjoyment at 185 words per minute, read for learning at 140 words per minute, and read for memorization at 97 words per minute (Fraser, 2007). The obvious standout for increasing their reading rate in the process of repeated-reading are the Vietnamese learners of Gorsuch and Taguchi (2008). By the fifth repetition, they were reading near the rate of native speakers reading for enjoyment. While the Japanese students investigated in the other studies made more modest increases in reading rate on repeated reading of the texts, they were still able to increase their reading rates to a higher rate on

14 116 Implementing Silent Repeated-Reading repeated readings of a text. This finding was also replicated in reading Japanese with beginning learners of Japanese. These results suggest that by engaging in silent repeated reading, nonnative readers can experience reading for enjoyment on subsequent readings of a text, even though they might read at a rate for learning or memorization on an initial reading of a text. As observed by Taguchi (1997), the apparent ability for non-native readers to increase their rate of reading a text (even up to increases in the eight repetition) supports the assertion that repeated-reading can assist non-native readers in reading English more fluently. RQ2: Can students read the treatment text faster after engaging in repeated-reading? The second research question asking if students can read a treatment text faster at the end of a period of engaging in repeated-reading can be answered that students do read the treatment text faster at the end of the treatment period when compared to the speed by which the students read treatment texts at the beginning of the treatment period. When considering the speed by which students read a passage for the first time in programs where students made substantial improvements in reading speed, students made improvements of 55 words per minute (Gorsuch & Taguchi, 2008), 47 words per minute (Chang & Millett, 2013), and 38 words per minute (Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012). These positive results can generally be attributed to advanced learners and texts carefully controlled for vocabulary and grammar. When considering gains made in programs where students made less impressive gains on the initial reading of treatment passages, students made improvements of 23 words per minute (Taguchi, Takayasu-Maass, & Gorsuch, 2004), 21 words per minute (Taguchi & Gorsuch, 2002), and 20 words per minute (Taguchi, 1997). For initial reading of a treatment passage, more advanced students can be expected to increase their reading rate by about 35 to 50 words per minute through a program of repeated reading. Less proficient students can be can be expected to increase their reading rate by around 20 words per minute through a program of repeated-reading. As a benchmark, the difference between Fraser s (1998) students speed when reading to memorize (i.e. 97 words per minute) and their speed when reading to learn (i.e. 140 words per minute) was 43 words per minute. Also, the difference between these students speed when reading to learn and their speed when reading for enjoyment (i.e. 185 words per minute) was 45 words per minute. Improving students reading rate by 40 words per minute, which is obtainable in a school semester, would give students a push toward being able to read with enjoyment through higher fluency. However, this is not nearly enough to push low or intermediate readers to the 250 words per minute reading rate suggested by Nation (2009). If reading rates consistently increased through a continued treatment of repeated-reading, then two years of undergoing

15 Implementing Silent Repeated-Reading 117 repeated-reading could possibly support students to achieving a good reading speed in English. RQ3: Can students read a new text faster after the repeated reading treatment? The third research question asking if students can read a new text faster at the end of a period of engaging in repeated-reading can be answered that students can read a text different from the treatment text passages faster at the end of the treatment period when compared to the speed by which the students read a comparable text at the beginning of the treatment period. However, the support for this assentation is quite weak. The gain of 45 words per minute on narrative texts by Taiwanese learners (Chang & Millett, 2013) and gain of 40 words per minute by Japanese learners (Taguchi & Gorsuch, 2002) were notable gains which compared favorably to the gains these learners made when reading the treatment texts. However, only gains made by Chang and Millett learners (2013) were statistically significantly different from a comparison group. The single adult proficiency learner investigated by Taguchi, Gorsuch, Takayasu-Maass, and Snipp (2012) was also able to read 24 words per minute faster after the treatment period. The other studies showed either a decrease in reading speed (Gorsuch & Taguchi, 2008; Taguchi, Takayasu-Maass, & Gorsuch, 2004) or an increase which was not statistically different from a control group (Gorsuch & Taguchi, 2008). There were 4 unsuccessful attempts to show that students could read a text different from those used to practice repeated-reading at a rate comparable to the rates which the students could read the treatment passages before a single study showed success. Chang and Millett (2013) posit that the biggest factor was the careful control of vocabulary and grammar in the passages used in the study. The testing passages contained a comparable level of vocabulary words and were less grammatically complex than the treatment passages. It is possible that the testing passages that were used in the studies by Taguchi and his coauthors has been more carefully controlled for vocabulary and grammar, the students in those studies would have shown reading rate gains. More research is needed, but it can be tentatively stated that engaging in repeated-reading assists a retained, greater level of reading rate. RQ4: How does reading comprehension improve due to silent repeated-reading? The fourth research question asking if students can increase their reading comprehension after a period of engaging in repeated-reading can be answered that students can comprehend a text different from the treatment text passages more comprehensively at the end of the treatment period when compared to the amount by which the students comprehended a comparable text at the beginning of the treatment period. According to Nation (2009, p. 141), non-native readers should be able to score 7 or 8 out of

16 118 Implementing Silent Repeated-Reading 10 on a comprehension test when reading with higher scores representing that a reader is going too slow and lower scores indicating that a reader is going too fast. Studies where the proficiency of the students under investigation were clearly higher than the reading materials used for testing reading comprehension (Chang & Millett, 2013; Gorsuch & Taguchi, 2008; Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012) demonstrated that students could read the text at an appropriate rate and improve their reading comprehension. However, only the students in Chang and Millett (2013) scored near the benchmark of 70% to 80% of questions answered correctly proposed by Nation (2009). The other studies which measured reading comprehension as recall of propositions required the students to read the passage five times to approach 50% recall of the text. Mostly, comprehension was operationalized as recall of propositions in the text (Chang & Millett, 2013; Gorsuch & Taguchi, 2008; Taguchi, 1997; Taguchi & Gorsuch, 2002; Taguchi, Gorsuch, Takayasu-Maass, & Snipp, 2012; Taguchi, Takayasu-Maass, & Gorsuch, 2004). Recall is a more difficult test of comprehension than identification (Anderson & Krathwohl, 2001, p.66-70). Even after the repeated-reading treatment, the percentage of propositions recalled is quite low. These studies show that keeping to Nation s (2009) 70% criterion on a first reading is quite difficult, if reading comprehension is measured through what students can recall. Students reading comprehension does improve through repeated-reading, however, the results obtained does not show repeated-reading improves reading comprehension more than another type of reading instruction, such as extensive reading or intensive reading. As with improvements in silent reading rate, the groups where the tests of reading comprehension where the vocabulary contained in the texts and the grammatical composition of the text were clearly within the students knowledge showed the clearest improvements in reading comprehension. This follows from Nation s (2009) criterion for a fluency development activity where students should be in control of the reading material that they are reading. Improvements in reading comprehension through repeated-reading are clearly shown in studies with contrasts between the students who engaged in repeated-reading and students who do not engage in a reading treatment (Gorsuch & Taguchi, 2008) and students who used a comparison group which spent less time reading (i.e. read the same text only once) than the treatment group (Chang & Millett, 2013). RQ5: How does Repeated-Reading develop reading fluency? The fifth research question asking how students can develop their reading fluency through silent repeated reading can be answered that students improve their lexical and grammatical processing so that they can devote more cognitive resources to semantic processing which is a

17 Implementing Silent Repeated-Reading 119 primary tenant of Automaticity Theory (LaBerge & Samuels, 1974). Automaticity Theory posits that readers naturally focus on higher levels of meaning (i.e. episodes contained in the text) and only focus on the textual features if there is less than automatic recognition of the words or phrases being read. Automaticity Theory also posits that language processing is naturally an auditory process for which the visual recognition of text is only an intermediary step in phonologic processing. As such, non-native readers need to automatize the conversion of textual features into auditory language to become more fluent readers. The learners perceptions of engaging in repeated-reading provide some support for this process of developing their reading fluency. Taguchi, Gorsuch, Lems, and Rosszell (2016) synthesized their research into how engaging in repeated-reading develops reading fluency in non-native speakers. They concluded that the research into developing reading fluency has shown that repeated-reading does positively affect the reading ability of non-native readers of English in three primary ways. Firstly, the authors argued that repetition and auditory models assist English language learners by scaffolding the reading process. In this context, this means that audio helps the learners to internally convert the text on the page to sound in their head. Thus, the audio model of the text assists readers in utilizing inner speech which occurs in reading (Pollatsek, 2015) while silent reading. The authors explained their assertion through the implicit prosody hypothesis (Fodor, 2002) which posits that prosody allows readers to hold language that they are reading in working memory which assists in the process of parsing and semantic analysis. The authors supported their view by reviewing the qualitative data from studies into the effects of repeated reading which showed that learners who engaged in repeated-reading perceived that the audio version of the text that they listened to while reading both (a) pushed the speed by which English language learners read the text and (b) assisted English language learners in comprehending story dialog. The studies consistently found that repeated-reading helped learners develop vocabulary recognition skills which allowed readers to focus on higher level features of reading comprehension (i.e. parsing the meaning of sentences and synthesizing the text into a coherent discourse) as opposed to lower level features such as vocabulary identification and grammatical parsing. One of the studies also identified that it is imperative that texts which are used in interventions designed for fluency development are controlled to contain only vocabulary which are within the control of the reader as neither an audio model nor repetition could provide the meaning for words which are completely unknown to, or unable to be guessed through context by, the reader. Secondly, Taguchi, Gorsuch, Lems, and Rosszell (2016) argued that engaging in repeated-