Repeated reading for developing reading fluency and reading comprehension: The case of EFL learners in Vietnam

Available online at www.sciencedirect.com

System 36 (2008) 253–278

www.elsevier.com/locate/system

SYSTEM

Repeated reading for developing reading fluencyand reading comprehension: The case of

EFL learners in Vietnam

Greta Gorsuch a,*, Etsuo Taguchi b,1

a Texas Tech University, Classical and Modern Languages and Literatures, Box 42071,

Lubbock, TX 79409-2071, USAb Daito Bunka University, Department of Japanese, 1-9-1 Takashimadaira, Itabashi-ku, Tokyo 175-8571, Japan

Received 23 August 2006; received in revised form 14 July 2007; accepted 26 September 2007

Abstract

Reading in a foreign or second language is often a laborious process, often caused by underdevel-oped word recognition skills, among other things, of second and foreign language readers. Develop-ing fluency in L2/FL reading has become an important pedagogical issue in L2 settings and one majorcomponent of reading fluency is fast and accurate word recognition. Repeated reading (RR) wasdevised by Samuels [Samuels, S.J. (1979). The method of repeated readings. The Reading Teacher32, 403–408] to develop reading fluency in English L1 readers, and instantiate Automaticity Theory[LaBerge, D., Samuels, S.J., 1974. Toward a theory of automatic information processing in reading.Cognitive Psychology 6, 293–323] in practice. In RR, readers read a simplified text repeatedly to helpautomatize word recognition, leaving more cognitive resources for higher order comprehension pro-cesses. RR used in FL settings is a more rare practice; studies show RR increases FL learners’ readingfluency but not necessarily their comprehension, possibly due to poor comprehension test instrumen-tation. This report describes an 11-week quasi-experimental RR study carried out with university-level Vietnamese learners of English using improved reading comprehension testing procedures.Results suggest that the experimental group (n = 24) gained in reading fluency, and comprehendedsignificantly more than the control group (n = 26). The results have implications for future uses ofRR in FL contexts, future reading comprehension test design, and the need for measurement of work-ing memory during short- and long-term use of RR. The results also imply a need for further study of

0346-251X/$ - see front matter Published by Elsevier Ltd.

doi:10.1016/j.system.2007.09.009

* Corresponding author. Tel.: +1 806 742 3145x246; fax: +1 806 742 3306.E-mail addresses: [email protected] (G. Gorsuch), [email protected] (E. Taguchi).

1 Tel.: +81 03 5399 7374; fax: +81 03 5399 7375.

mailto:[email protected]

mailto:[email protected]

254 G. Gorsuch, E. Taguchi / System 36 (2008) 253–278

a persistent but unsupported belief in FL settings that simply increasing language proficiency guaran-tees reading fluency and that word recognition and fluency need not be developed as skills.Published by Elsevier Ltd.

Keywords: Fluency; Word recognition; ESL/EFL reading

1. Introduction

It is well documented that reading in a second or foreign language (L2/FL) is a labo-rious process (Anderson, 1999; Jensen, 1986; Segalowitz et al., 1991). One cause of thisslow, effortful reading may lie in the underdeveloped word recognition skills of L2 andFL readers, being the lower-level process of reading in which readers access word pronun-ciation and meaning from texts (see Chard et al., 2006; Harris and Hodges, 1995; Kuhnand Stahl, 2003; Samuels, 2006 for L1 comment, and Grabe, 2004; Grabe and Stoller,2002 for L2/FL comment). It is useful to contrast L1 and L2/FL readers in order to under-stand problems the latter may have with word recognition and automaticity: Compared toL1 readers who have learned the mother tongue orally before learning to read, L2/FLreaders are handicapped in terms of the amount of exposure to the language they arelearning. L2/FL readers’ oral language and reading development may start at the sametime. Participants in this study did not begin learning English orally or through print untiltheir first year of secondary school in Vietnam in once-a-week classes.

Looking at L1 reading in English specifically, readers must learn how the 26 alphabet let-ters and the approximately 44 phonemes are combined to form meaningful words. Examina-tion of word recognition processes in English L1 settings illustrates benefits L1 readers have:Competent L1 readers possess developed knowledge about spelling patterns and sound-spelling relationships in English and this knowledge helps them to identify words as singlewhole units automatically from memory (Ehri, 2005). They also seem able to analyze andmanipulate letter or letter clusters in words (Ehri, 1998), and able to pronounce nonsense let-ter strings (Hogaboam and Perfetti, 1978). Although Vietnamese English L2 learners use aRomanized alphabetical writing system in their L1s, their pronunciation system is different,and they have not accumulated the same repertoire of sound-symbol relationships as EnglishL1 learners. Finally, L1 readers have acquired basic grammatical knowledge, and a reper-toire of 5000–7000 words before learning to read (Grabe and Stoller, 2002). Both L1 readingcommentators (e.g., Reynolds, 2000; Ruddell, 1994) and FL/L2 commentators (e.g., Grabe,2004) point out the importance of word recognition accuracy and automaticity as a basis forprocessing extended text within and across ‘‘meaningful grammatical units” (Samuels, 2004,p. 829) and thus comprehending text (see also Koda, 2005, p. 30 for commentary that ‘‘mosttext words are thoroughly processed during reading”). In sum, L2/FL readers, even if theirL1 uses the same writing system as English, have to go a long way before they develop auto-matic word recognition (Grabe and Stoller, 2002; Koda, 1996, 2005).

L2/FL readers’ slow reading may cause motivational problems in L2/FL contexts. Nut-tall posits a ‘‘vicious circle” to describe readers who cannot develop good reading skills(1996, p. 127). Slow readers do not read much, and if they do not read much, they donot understand. If they do not understand, then they cannot enjoy reading. Koda (2005)describes a similar scenario, adding that inadequate reading practice among poor readersdeprives them of opportunities to develop their conceptual growth (p. 30). Day and

G. Gorsuch, E. Taguchi / System 36 (2008) 253–278 255

Bamford (1998) note that only actual reading experience can help L2/FL learners acquirelinguistic, world, and topical knowledge needed to improve their reading skills (p. 19).Therefore, for theoretical and pedagogical reasons, L2/FL educators are more interestedthan ever in finding methods to help L2/FL learners develop efficient word recognitionskills (Day and Bamford, 1998; Eskey, 1988; Grabe, 1991, 2004; Silberstein, 1994).

2. Why fluency as a skill is needed

Samuels (2006) defined fluency as ‘‘the ability to decode and to comprehend the text atthe same time” (p. 9). Other outward indicators of fluency are accuracy, speed, and appro-priate expression in reading texts silently or orally (see also Grabe, 2004). A theoreticalbridge to the crucial role fluency plays in successful reading is highlighted in AutomaticityTheory (AT) proposed by LaBerge and Samuels (1974) and Verbal Efficiency Theory(VET) proposed by Perfetti (1985, 1988). AT posits that attentional resources the mindcan direct at one time are limited, and that the dual process of recognizing printed wordsand then constructing meaning from them ‘‘is slow, effortful, and hard on memory” (Sam-uels, 2006, p. 9). In performing the complex task of reading, part of the reading processhas to be automatized. That is, the process is performed with a minimal amount of atten-tional resources (Samuels, 2006). Higher-level comprehension processes such as integrat-ing propositions, accessing schemata, and monitoring on-going comprehension require aconsiderable amount of attentional resources (Reynolds, 2000). Therefore, it is lower-level,pre-lexical processes of reading (orthographical and phonological identification of words)that can be automatized with practice. AT has contributed to elevating fluency to the fore-front of reading research in English L1 settings (Kuhn and Stahl, 2003; National ReadingPanel, 2000; Pikulski and Chard, 2005). Verbal Efficiency Theory expands the notion ofautomaticity beyond lower-level decoding processes (Walczyk, 2000), positing thathigher-level post-lexical reading processes, such as using basic cognitive and metacognitivestrategies, and activating background schemas, can also be automatized through practice.

In L2/FL reading research, in contrast, the topic of reading fluency has received scantattention (Grabe, 2004). It seems that the prevailing perception among L2/FL educators isthat fluency will develop as learners’ overall proficiency grows. There is, however, littlesupport for this assumption (Koda, 2005). Empirical evidence has rather supported theopposing position that language proficiency and reading fluency do not necessarilydevelop hand in hand (Favreau and Segalowitz, 1983; Segalowitz, 1986; Segalowitzet al., 1991). In other words, high language proficiency does not necessarily ensure goodreading fluency. Further, many practitioners in L2/FL contexts avoid adequate consider-ation of reading fluency partly because of ‘‘a limited understanding of the role of rapid andautomatic word recognition processes” (Grabe and Stoller, 2002, p. 21) in comprehension.It would be fair to say that L2/FL teachers, learners, and programs do not see reading flu-ency as a specific goal to reach for language learning.

3. Repeated reading as one means of fluency development

3.1. Repeated reading in L1 settings

Repeated reading (RR) was devised by Samuels (1979) to develop reading fluency inEnglish L1 readers. It attempted to translate Automaticity Theory (LaBerge and Samuels,


1974) into practice, and aims to develop fluent word recognition skills. It is theorized thatthe method develops automatic, effortless word recognition, freeing readers’ cognitiveresources and allowing them to direct more attention to higher order comprehension pro-cesses. Samuels’ method has learners re-reading a short passage, three or more times, untilthey are able to read at criteria word per minute (wpm) level. RR programs which do notinvolve having learners read along with an audio model of the text are called unassistedRR, while assisted RR programs use an audio model. In the case of the current FL/L2study in Vietnam, the audio model was an audio tape of a given text or one of the authors(the participants’ regular classroom teacher) reading the text aloud.

RR has been extensively studied in English L1 contexts and overall has been shown tobe effective in developing reading fluency and comprehension of monolingual readers ofEnglish (see the National Reading Panel, 2000; Kuhn and Stahl, 2003 for reviews). Re-reading passages has been found to increase students’ oral reading rates and accuracy(Carver and Hoffman, 1981; Chomsky, 1976; Dahl, 1974; Dowhower, 1987; Herman,1985; Rashotte and Torgesen, 1985; Samuels, 1979; Young et al., 1996). This appears tolead to better comprehension of passages (Dowhower, 1987; Herman, 1985; O’Sheaet al., 1985; Young et al., 1996). In addition, practice effects of re-reading a passage arecarried over to new, unpracticed passages in terms of reading rate and accuracy (Carverand Hoffman, 1981; Dowhower, 1987; Faulkner and Levy, 1994; Herman, 1985; Rashotteand Torgesen, 1985; Samuels, 1979) and comprehension (Dowhower, 1987; Morgan andLyon, 1979; Young et al., 1996). RR has a positive effect on readers’ vocabulary develop-ment (Koskinen and Blum, 1984), and seems to enable readers to read in larger and moresyntactically and phonologically appropriate phrases (Dowhower, 1987). It should benoted, however, that unless the degree of overlapping vocabulary between old and newpassages is high, transfer of gains to the new passage is minimal in terms of reading rate(Rashotte and Torgesen, 1985). We were to see the same short-term effect in the currentstudy when participants started a new story with new vocabulary, suggesting that automa-ticity with word recognition, or lack of it, is a significant factor affecting reading fluency.This is discussed in more detail below.

3.2. Repeated reading in L2/FL settings

Developing fluency in L2/FL reading has become a significant pedagogical issue (Grab-e, 2004; Grabe and Stoller, 2002; Nation, 2001). RR is a rather new approach in L2/FLreading contexts and in this approach, L2 learners read specified passages from relativelyeasy texts repeatedly in order to increase learners’ sight recognition of words and phrases,resulting in increased fluency and comprehension (Blum et al., 1995; Dlugosz, 2000; Tagu-chi, 1997; Taguchi and Gorsuch, 2002; Taguchi et al., 2004). ‘‘Relatively easy texts” in thisstudy was operationalized as a graded reader, written with a limited range of vocabularyand grammatical structures. Specifics on the text will be discussed below in the Section 6.Nonetheless, there have not been many studies on RR in L2/FL settings, including howRR affects comprehension. In Blum et al. (1995) home-based RR with an auditory model(audio cassettes) was investigated to learn whether RR was an effective supplement to anL2 literacy program. They concluded that RR improved readers’ ability to read books ofincreasing difficulty fluently and accurately, and readers’ motivation to read. Taguchi(1997) examined the effects of assisted RR in English on oral and silent reading rates of15 Japanese university students in a 10-week, 28-session RR treatment. In each session


readers read a passage silently seven times; three of those times, readers read while listen-ing to an audio model. Taguchi found that silent reading rates increased significantlywithin the practiced passages, but failed to show transfer of practice effects in terms ofsilent and oral reading rates to new, unpracticed passages.

Taguchi and Gorsuch (2002) focused on RR transfer effects for silent reading rate andcomprehension to new passages. They found that the 10-week RR treatment facilitatedparticipants’ (nine EFL readers) reading rates from a pre-test reading passage to a post-test reading passage (a different passage). However, reading rate gains from the first RRsession passage to the 28th (the last) session passage approached but did not exceed sig-nificance. In addition, control (non-RR) and experimental (RR) group readers showedsimilar and modest but not statistically significant transfer gains for reading comprehen-sion from the pre-test to the post-test passage. Taguchi and Gorsuch speculated that thelack of clear transfer effects for reading rate and comprehension of RR group readers wascaused by the shortness of the treatment. In a follow-up, Taguchi et al. (2004) explored theeffects of assisted RR on silent reading rate and comprehension by comparing two meth-ods thought to develop silent reading fluency and comprehension, repeated reading andextensive reading. Taguchi et al. (2004) extended the RR treatment period from 10 weeksto 17 weeks, and increased the total number of RR sessions from 28 to 42. They found thatassisted RR significantly improved students’ silent reading rate from the first to the lastRR session. As for gains in reading comprehension, however, they were not able to detectappreciable transfer effects from the RR treatment.

4. Rationale and research questions

Despite calls for research on reading fluency development and its impact on readingcomprehension in L2/FL settings, there have been few studies addressing this topic. Tagu-chi and Gorsuch (2002) and Taguchi et al. (2004) are rare attempts. Both studies, however,have shortcomings focusing on reading comprehension measurement. In both studies thepre-test and post-test texts used were found not to be at the same difficulty level. Also, thetest format (10 short-answer comprehension questions) in both studies may not have beensensitive enough to detect comprehension gains. In both studies, the test passages wereread multiple times, echoing the RR treatments, and the test questions were answered afterthe first and final readings. Learners were not allowed to see the test passages as theyanswered the questions. We speculate that the participants in both studies had difficultyprocessing information in texts during their first reading of the test passage, so theymay not have been able to process and retain the information needed to answer the ques-tions without referring to the text.

In planning this study, we surmised that other reading comprehension measures such asfree recall, or longer, well-constructed short answer tests capturing an array of inference,gist and detail related propositions in the text might be more sensitive to gains in compre-hension. Based on this speculation, the present study utilized both recall protocols andnewly designed short answer tests. Our research questions are:

1. How much will the reading fluency of learners in the experimental group increase dur-ing the repeated reading (RR) treatment period?

2. How will the reading fluency of the learners in the experimental group compare to thereading fluency of learners in the control group at the end of the study?


3. How will the overall reading comprehension of learners in the experimental group com-pare to that of the control group at the end of the study on two measures: Short answerquestions and recall?

We also had the rare opportunity to pilot RR with a large group (n = 24) of VietnameseEFL learners. This experiment was a welcome opportunity, particularly as learners’ L2exposure in this isolated region is minimal.

5. Method

5.1. Participants

Experimental group participants were 24 third year Vietnamese provincial universitystudents. There were 6 males and 18 females, and their mean age was 22. In the controlgroup were 26 third year Vietnamese university students. There were 13 males and 13females, and their mean age was 21.78. Both groups were placed in the same level in theirfirst year of study, and were considered to have the highest English proficiency of all thirdyear students. According to a cloze test (see Section 6 below) administered at the beginningof the study, the experimental group had somewhat lower reading ability with a meanscore of 31.583 out of 50. Participants in the control group scored higher with a meanof 36.304. This difference was statistically significant using an unpaired t-test(p = .0350). However, we note the difference was not meaningful, as the eta squared effectsize was only .019, suggesting that less than 2% of all variance present in the test is attrib-utable to group membership (experimental versus control group). The cloze tests are mea-suring a number of constructs and can be seen only as a very general comparativemeasure.

6. Materials

6.1. Cloze test

A cloze test for checking pre-treatment equivalence of the experimental and controlgroups was developed using a 425-word passage from A Scandal in Bohemia (Doyle,1999). The Flesch Kincaid Grade Level estimate was 2.4. All semantically and grammat-ically possible responses were accepted.

6.2. Repeated reading treatment texts and audio tapes

The repeated reading treatment texts were comprised of three short stories from agraded reader in the Penguin Readers series (A Scandal in Bohemia, The Red-Headed Lea-

gue, The Boscombe Lake Mystery, Doyle, 1999), segmented into 16 texts. The text seg-ments were from 274 to 670 words, with a mean word length of 526 words. One textchosen randomly from the 16 showed a readability level at 2.8 on the Flesch–KincaidGrade Level. The graded reader is rated by the publishers as pre-intermediate and iswritten with a vocabulary range of 1200 words. The RR treatment texts were chosen onthe basis of reading passages used in placement tests at the university where the studytook place, which averaged 9.37 on the Flesch–Kincaid Grade Level. We noted that few


students got high scores on items based on the passages. With RR, it is important tochoose texts that are not too difficult in order to maximize fluency building. An accompa-nying audiotape was used in the treatments.

Some notes on vocabulary overlap between the three short stories comprising the 16segments are called for here. This has bearing on whether or not reading rate gain resultscan be interpreted as arising from greater automaticity in word recognition, which is pos-ited by Automaticity Theory. A Scandal in Bomemia and The Red-Headed League (the firstand second stories comprising the first through the sixth, and seventh through the thir-teenth treatment segments, respectively) had 65.3% vocabulary overlap, excluding propernouns. The remaining 34.7% was vocabulary unique to either A Scandal in Bohemia or The

Red-Headed League. The non-overlapping (unique) vocabulary accounted for 14.03% ofall occurrences of all words (6426 word occurrences total) in the two combined texts,excluding proper nouns. The Red-Headed League and The Boscombe Lake Mystery (thesecond and third stories comprising the seventh through the thirteenth treatment seg-ments, and 14–16 treatment segments, respectively) had 52.2% vocabulary overlap, exclud-ing proper nouns. This left 47.8% of non-overlapping vocabulary from either story. Thenon-overlapping vocabulary accounted for 18.7% of all occurrences of all words (4643word occurrences total) in the two combined texts (excluding proper nouns). The vastmajority of non-overlapping vocabulary were content words, such as common nouns,verbs, and adjectives.

As with most graded readers, the RR treatment texts were limited in the number ofgrammatical structures used. In all three short stories, simple past and simple present com-prised the majority of structures used, with infrequent occurrences of past progressive,present perfect, and future constructions.

6.3. Short answer pre- and post-test

One 1143 word text, Two Men Visit (Young, 1971) was segmented into two texts, onewith 578 words and one with 565. The 578-word text was used as the basis for Form A ofthe test. The second half of the story (565 words) was used as the basis for Form B. FormA had a Flesch–Kincaid Grade Level of 2.7 and Form B had a level of 2.9. For each of thetwo test forms, 14 short answer items were written, capturing main ideas, supportingdetails, and details. A final item was written to test inference. The directions accompanyingthe test texts were translated into Vietnamese, as were the items. The two texts (with Eng-lish directions) for Forms A and B are given in Appendix A and the items appear inAppendix B. Due to space restrictions, a full description of test development can berequested of the authors.

The test procedure was the same for both the pre- and post-tests, and mirrored the RRtreatments: Participants in either experimental or control group read the assigned Englishpassage once while timing themselves. The test texts were then taken away while partici-pants answered the 15 short answer items in Vietnamese. Participants then read the testtexts a second and third time while hearing an audio model of the texts. Participants readthe test texts silently a fourth and fifth time while timing themselves, and after the fifthreading, participants were given a fresh sheet with the same short answer items theyhad answered after the first reading and answered in Vietnamese.

For scoring, participants’ written answers were translated into English. Main idea, sup-porting detail, and detail propositions captured by the participants in response to the short


answer items were totaled and then converted into percentages. To check scoring reliabil-ity, 10 pre- or post-tests from experimental and control groups were randomly selectedfrom both Forms A and B and given to an experienced ESL teacher for scoring. Authorsand the teacher were 95% consistent on Form A and 82% consistent on Form B.

6.4. Recall pre- and post-test

In order to provide more detail on potential changes in experimental and controlgroups’ comprehension of English texts two forms of a recall test were developed aspre- and post-tests. One 845-word text, Man With No Name (Davies and Town, 1992)was segmented into two texts, one with 416 words and one with 429. The 416-word text,which comprised the first half of the text (about a mysterious plane crash), was used as thebasis for Form A of the test. The second 429-word text was used as the basis for Form B.Form A had a Flesch–Kincaid Grade Level of 2.2 and Form B had a level of 2.3. Thus therecall test texts were easier than the 2.8 grade level of the repeated reading treatment texts.The directions on the test text sheet and answer sheet were translated into Vietnamese.The test texts for Forms A and B are shown in Appendix C and the answer sheet inAppendix D.

The pre- and post-test procedures for the recall test were the same as the short answertest. The recall answer sheets were translated into English. Scoring was accomplished byall reasonably complete propositional units being assigned a point as a main idea, support-ing detail, or detail. The points were totaled and converted into percentages for total prop-ositional units for each student. Scoring reliability was checked by randomly selecting 10pre- or post-tests from the experimental and control groups and then giving them to anexperienced ESL teacher. Authors and the teacher were 79% consistent on Form A and86% consistent on Form B.

7. Procedure

At the beginning of the study, the experimental group took Form A of the recall testand Form B of the short answer test. The control group took Form B of the recall testand Form A of the short answer test. The experimental group then participated in 16repeated reading sessions over an eleven week period as part of their regular English skillsclass. Each RR session followed the same procedure:

1. Participants read an approximately 500-word segment of a short story once while tim-ing themselves with a stopwatch. They wrote their times on a time log sheet.

2. Participants then read the text a second and then a third time while listening to it on anaudiotape or being read aloud by one of the authors.

3. Participants finally read the text a fourth and fifth time, timing themselves for eachreading and marking each time on their time log sheet.

4. At the end of the session, participants wrote a short report either in English or Vietnam-ese, their choice.

Assigned segments for each RR session were contiguous. In other words, RR session #1utilized the first 500-word segment of Scandal in Bohemia, session #2 the next 500-wordsegment of Scandal in Bohemia, and so on. Then the next story in the book was begun.


The control group met with another teacher and did not have any explicit reading course-work. At the end of the study, the forms used in the post-test were reversed: The experi-mental group took Form B of the recall test and Form A of the short answer test and thecontrol group took From A of the recall test and Form B of the short answer test.

8. Analyses

To answer RQ #1, reading fluency rates (words per minute) were calculated for the firstand fifth readings for each experimental group member in each RR session using students’time logs. To estimate how much fluency increased on average for the study, the experi-mental group’s average wpm for the first reading for the first RR session was comparedto their first reading for the 16th and final RR session using a paired t-test. In addition,the group’s average wpm for the fifth reading for the first RR session was compared tothe fifth reading for the 16th and final RR session, again using a paired t-test. The alphavalue set for significance was p = .025 (.05 divided by 2 for two comparisons). To estimatehow much fluency increased on average within RR sessions, the experimental group’s aver-age first wpm rate for all 16 sessions was compared to their average fifth wpm rate for the16 sessions using a paired t-test with significance set at p = .05.

To answer RQ #2, the wpm rate for the experimental and control groups on both thepre-tests and post-tests were calculated using time logs. To estimate potential differences inreading fluency at the beginning of the study, experimental and control groups’ readingrates of the test texts on the first reading for the short answer pre-test were compared usinga paired t-test. This procedure was repeated for the fifth reading of the pre-test texts.Paired t-tests were again done on the first and fifth readings of test texts on the recallpre-test. Alpha probability was set at p = .0125 (.05 divided by 4 for four comparisons).To compare differences between groups at the end of the study, the same proceduredescribed above was done, using wpm reading rates on the recall pre- and post-tests.Alpha probability was set at p = .0125 (.05 divided by 4 for four comparisons).

To answer RQ #3, experimental and control groups’ mean total scores on the first andfifth readings of the short answer pre-test and recall pre-test were compared in order tofurther investigate whether the groups were equal at the outset of the study. In order tocompare pre-test means, a mixed within and between 2 � 2 repeated measures ANOVAwas done with test type (short answer versus recall) and test occasion (first versus fifthreading) as the within subjects factors, and group (experimental versus control) as thebetween subjects factor. The between subjects factor was found to be statistically signifi-cant at p < .05 (F = 15.828, df = 1), and there was a statistically significant within subjectsinteraction between test type and test occasion (p < .05), suggesting that the two groupswere different on both short answer and recall tests at the outset of the experiment. Indeed,the experimental group got lower scores on first readings for both tests, and on the fifthreading on the recall test (see Table 3 below for details). Therefore, the experimentaland control groups’ pre-test scores on the two test types and occasions were averagedinto a single omnibus pre-test variable and then treated as a covariate for the purposesof further analyses being conducted at the end of the experiment. In this way, pre-test dif-ferences between experimental and control groups were removed as a confounding factorwhen considering possible post-test differences. In other words, the experimental and con-trol groups were put on a statistically level playing field in terms of their pre-testperformances.


In order to compare the experimental and control groups’ post-test performances onthe first and fifth readings of the short answer and recall tests, a mixed within and between2 � 2 repeated measures ANOVA was done with test type (short answer versus recall) andtest occasion (first versus fifth reading) as the within subjects factors and group (experi-mental versus control) as the between subjects factor.

9. Results

Experimental group participants’ reading fluency increased during the 11 week RRtreatment period, confirming findings in Taguchi et al. (2004), see Table 1 below:

Participants’ fluency on average increased approximately 55 words per minute on thefirst reading between the first and last RR session. This increase was statistically significantat p < .025 (df = 21, t = 4.553). g2 effect size was strong at .497, meaning that nearly 50%of the overall variance of the data can be attributed to the variable of occasion (first RRtreatment versus 16th RR treatment). On average, participants’ fluency on the fifth read-ing increased 91 words per minute from the first to the last RR session. This increase wasalso statistically significant at p < .025 (df = 20, t = 2.719). g2 effect size was moderate at.2698, meaning that nearly 27% of variance was due to occasion (the first versus the 16thRR session). The remaining 73% of unexplained variance may be due to individual vari-ations in the group in processing an increasingly familiar text within RR sessions (learnerinterest in the story genre, ability to resolve from context or previous knowledge unknownvocabulary noticed in a first or second reading of the text), as was suggested by partici-pants’ comments in book reports written after each RR session.

Another way to visualize participants’ gains with new, unpracticed passages over thecourse of the 16-session treatment is through Fig. 1 below:

Although there are dips and rises, the overall movement is upwards (higher words-per-minute reading rates). At the first session (see also Table 1 above) participants read thepassage the first time at 163.200 wpm. This rate increases until the seventh session, wherea new story is begun. Note the temporary dip from approximately 200 wpm on the sixthsession down to around 165 wpm, likely due to the 34.7% of vocabulary not shared by thetwo stories (described above in the Section 6). In other words, participants cannot processthe new vocabulary efficiently and with automaticity the first time they encounter thevocabulary in a new story (a supposition confirmed by participant self-report). But bythe time the participants continue on to the eighth session, their reading rate has increasedup to around 200 wpm, suggesting positive effects of the repeated exposures to the vocab-ulary in the text. A similar pattern emerges at the 14 and 15 sessions (the second and third

Table 1Reading rates (words per minute) on first and fifth readings, first and last RR sessions

M SD (wpm) Mean difference (wpm)

First reading

Session #1 163.200 49.093 54.575*

Session #16 217.775 62.890

Fifth reading

Session #1 261.020 98.312 90.705*

Session #16 351.725 201.005

Note: *Statistically significant at p < .025.

050

100

150200250

Session1

4 7 10 13 Session16

Words Per Minute

Fig. 1. Words per minute reading rate of learners for first reading of new, unpracticed passages for a 16 RRtreatment session.


session of yet a new story with 47.8% vocabulary not shared by the previous story – seeSection 6 above) where participants begin the story at around 200 wpm but decline to180 wpm (possibly as they begin to encounter new vocabulary) and then return to anever-higher rate of 217.775 wpm by the 16th session.

Participants’ reading fluency also increased within RR sessions, a finding that confirmsfindings in Taguchi et al. (2004). The average reading rate for all first readings (16 RR ses-sions) was M = 184.88 (SD = 18.46) and for all fifth readings was M = 290.11(SD = 37.16). This was an increase of 105.23 wpm, which was statistically significant(df = 15, t = 17.237, p = <.001). g2 was a strong .95, suggesting that 95% of the variancewas due to learners’ repeated experience with the assigned textswithin individual RR ses-sions. The results suggest that RR is effective in increasing fluency when the same text isread five times, and also in aiding transfer of fluency gains to new, unpracticed passages.

When comparing WPM rates of the experimental and control groups on the shortanswer and recall pre- and post-tests the results were inconclusive. Below in Table 2 arethe pre- and post-test wpm rates for the short answer and recall tests.

On the short answer test, the experimental group read significantly faster(M = 148.55 wpm) on the first reading of the pre-test than the control group(M = 122.21 wpm), but by the fifth reading, the control group (M = 233.79 wpm) readslightly faster than the experimental group (M = 212.81 wpm). On the post-test, the exper-imental group (M = 131.35 wpm) read the test text slightly faster than the control group(M = 123.42 wpm) but by the fifth reading the control group (M = 297.02 wpm) had againoutstripped the experimental group (M = 255.82 wpm), although the difference was notstatistically significant. Fluency gains realized by the experimental group through theRR treatment (M = 261.020 wpm on the first reading of the 16th RR treatment on a textof comparable difficulty, see Table 1 above) were not transferred to the short answer post-test text.

The same finding seemed to hold true for the recall tests (see Table 2 above). On the firstreading of the pre-test, the experimental group read the test text slightly slower(M = 123.88 wpm) than the control group (M = 132.10 wpm). By the fifth reading, theexperimental group (M = 231.59 wpm) was reading only slightly faster than the controlgroup (M = 220.66 wpm). On the first reading of the post-test, the experimental groupread the test text only slightly faster (M = 144.30 wpm) than the control group(M = 129.80 wpm) and by the fifth reading, the control group was reading slightly faster

Table 2WPM on the short answer and recall test texts

Short answer test

Experimental group pre-test (Form B) Control group pre-test (Form A)

M SD M SD Mean differenceFirst reading 148.55 37.66 122.21 33.15 26.34*

Fifth reading 212.81 88.08 233.79 91.47 20.984

Experimental group post-test (Form A) Control group post-test (Form B)

M SD M SD Mean differenceFirst reading 131.35 32.15 123.42 31.18 7.93Fifth reading 255.82 84.54 297.02 139.32 41.20

Recall test

Experimental group pre-test (Form A) Control group pre-test (Form B)

M SD M SD Mean differenceFirst reading 123.88 41.72 132.10 44.62 127.90*

Fifth reading 231.59 135.28 220.66 77.12 10.93

Experimental group post-test (Form B) Control group post-test (Form A)

M SD M SD Mean differenceFirst reading 144.30 35.91 129.80 31.04 14.50Fifth reading 227.30 73.48 262.44 75.70 35.14



(M = 262.44 wpm) than the experimental group (M = 227.30 wpm), although the differ-ence was not statistically significant. Again, fluency gains for the experimental groupdid not seem to transfer (261.020 wpm on the first reading of the 16th RR treatment com-pared to 144.30 on the first reading of the recall post-test text).

The picture for reading comprehension was different. The experimental group compre-hended more of both test texts at the end of the study, even though they read at roughlythe same wpm rates as the control group in both test conditions.

On both pre-tests, the control group outperformed the experimental group. On theshort answer pre-test, the control group used a greater percentage of the total possiblepropositions to answer items (M = 21.5%) on the first reading than the experimentalgroup (M = 14.4%). On the fifth reading, both groups were nearly identical (experimentalgroup M = 38.8%; control group M = 38.3%). However, on the recall pre-test, the controlgroup outperformed the experimental group on both the first reading (M = 21.5% versus8.5%) and the fifth reading (M = 38.5% versus 30.3%).

At the end of the study, the experimental group did better than the control group. Onthe first reading of the short answer post-test, the experimental group used a mean of41.2% of the propositions to answer test items while the control group used 28.7%, andon the fifth reading the experimental group used 58.2% of the propositions while the con-trol group used 49%. On the first reading of the recall post-test, the experimental grouprecalled 21.6% of the propositions while the control group recalled 19.3%, and on the fifthreading, the experimental group recalled 61.5% while the control group recalled 36.1%.

On the ANOVA then, it was not surprising that there was a significant between groupseffect at p < .05 (F = 89.338, df = 1). For within groups, there was a significant main effectfor test occasion (first versus fifth reading) at p < .05 (F = 7.429, df = 1), but not for test


type (short answer versus recall). There were significant interaction effects for test occasionand group (p < .05, F = 10.086); and test type, test occasion, and group (p = <.05,F = 51.774, df = 1). The higher order three-way interaction can be interpreted as partici-pants’ performances on test occasion (first versus fifth reading) as being moderated by arelationship between group and test type (short answer versus recall). In other words,the two groups differed not only how well they did from the first to the fifth reading,but also how well they did on the two different types of tests. When one compares thegroup means (see Table 3 above), it becomes clear the experimental group got 41.2% onthe first reading of the short answer post-test, but only 26.1% on the first reading of therecall post-test, for example. The control group got 49% on the fifth reading of the shortanswer test, but only 36.1% on the fifth reading of the recall test. The recall test task wasapparently more difficult than the short answer test task.

The interaction of reading occasion and group can be interpreted as participants’ per-formances on first and fifth readings of both tests as being mediated by group. This islikely related to significant main effect of reading occasion, meaning that all participantsdo better on first versus fifth readings, despite the fact that the recall test is harder thanthe short answer test. In other words, both experimental and control groups improvedfrom the first to the fifth readings on both the short answer and recall post-tests, some-thing noted in an earlier study, where the RR methodology used to administer the testsbenefited comprehension in both groups in the short term (Taguchi et al., 2004). Certainly,four paired t-tests comparing first and fifth readings within groups for the short answerand recall tests were all statistically significant at p < .0125 (.05 divided by 4 for four com-parisons). But the interaction with group suggests that group membership also mattered inhow well participants did. To investigate this, four paired t-tests were done to compare

Table 3Percent of propositions used to complete short answer and recall tests

Short answer test

Experimental group pre-test (Form B) Control group pre-test (Form A)

M (%) SD (%) M (%) SD (%) Mean difference (%)First reading 14.4 7.8 21.5 7.3 7.1*

Fifth reading 38.8 11.7 38.3 7.6 .5

Experimental group post-test (Form A) Control group post-test (Form B)


Fifth reading 58.2 12.4 49.0 10.7 9.2*

Recall test

Experimental group pre-test (Form A) Control group pre-test (Form B)


Fifth reading 30.3 11.0 38.5 12.8 8.1

Experimental group post-test (Form B) Control group post-test (Form A)

M (%) SD (%) M (%) SD (%) Mean difference (%)First reading 26.1 10.3 19.3 8.5 6.8Fifth reading 61.5 10.9 36.1 13.1 25.4*



between groups on first and fifth readings for the short answer and recall post-tests withalpha set at p = .0125 (.05 divided by 4 for four comparisons). Three out of four compar-isons were statistically significant, meaning that the experimental group did significantlybetter than the control group when comparing first readings of the short answer post-test,fifth readings of the short answer post-test test, and fifth readings of the recall post-test (onthe first reading of the recall test, the higher mean of the experimental group approachedbut did not exceed significance). The findings support the notion that the 16 repeated read-ing treatments benefited the experimental group’s reading comprehension more. Indeed,the partial eta2 effect size for the test occasion and group interaction is.177, meaning thatthe two variables account for 17.7% variance in all variables examined in the ANOVA. Bythe end of the study, the experimental group is very different from the control group and atleast some of the difference can be identified as the RR experiences of the experimentalgroup.

10. Discussion

This study was done to investigate the utility of RR in increasing fluency and readingcomprehension with EFL learners in a new setting (Vietnam), and further, to illuminatethe role of the theories we believe underpin RR (e.g., Automaticity Theory and VerbalEfficiency Theory). Finally, we felt that our earlier findings with RR with Japanese learn-ers of English (i.e., RR increases reading fluency but not comprehension) suggested wefocus on creating comprehension measures that were more sensitive to changes in partic-ipants’ comprehension.

We believe the data suggest a robust utility for RR to increase reading fluency and com-prehension with EFL learners. Confirming previous studies in other English learning set-tings participants in the experimental group significantly increased their reading fluencyfrom the beginning to the end of the 11-week RR treatment. Unfortunately, contrary toour findings in previous studies, increases in the experimental group’s fluency did nottransfer to first readings of their post-test texts, and the wpm reading rates of the exper-imental group did not differ significantly from those of the control group. Even on fifthreadings of the post-test texts, the experimental group’s wpm reading rates, while increas-ing as expected, did not differ significantly from that of the control group’s. However, free-answer comments from experimental group participants on their post-tests provide animportant explanation: On the short answer test, every experimental group participantnoted they paid close attention to detail while reading the test text because they knew theywould answer questions afterwards. After the first reading and administration of the shortanswer post-test, 13 participants noted specific items they had trouble answering andnoted they would read for information needed to answer the items on subsequent readings(which they then found answers for, as evidenced by their increasing comprehensionscores). On the recall test, 12 participants noted they would read closely for sequence inthe test text (which they apparently also found in subsequent readings) so they could retellthe story for the subsequent test administration.

The experimental group comprehended more of both test texts on both first and fifthreadings at the end of the study than the control group. This is remarkable in that exper-imental group participants comprehended significantly less than control group partici-pants on the first readings of both short answer and recall pre-tests (see Table 3).Automaticity Theory posits that as readers increase their word recognition skills, more


attentional resources are made available for higher order comprehension processes. Webelieve our improved comprehension tests were more sensitive to changes in the experi-mental group. Further, we believe our results present causal evidence for the efficacy ofRR in increasing comprehension of unpracticed passages. According to time logs keptduring RR treatments, the average wpm of the experimental group increased significantly(see Table 1). Clearly, participants were processing text faster by the time they took thepost-tests. Even though the experimental group participants’ wpm reading rate on thepost-tests was not significantly different than the control groups’, the experimental groupcomprehended more.

We feel we should underscore once again the role of automaticity of word recognitionin reading fluency and thus comprehension. We believe word recognition is a major com-ponent of reading fluency and offer a point for consideration, based on Fig. 1 above. Wenoted temporary descreases in wpm reading rate at the outsets of new stories which didnot share substantial percentages of vocabulary with a stories previously read. We believeparticipants’ faster ‘‘recovery” of wpm in later parts of the treatment and overall upwardmovement throughout the treatment may be due to greater automaticity in word recogni-tion engendered by RR methodology used (repeated exposures to the text).

That the experimental group read more slowly than they were capable of in order to dowell on the comprehension post-tests also suggests use of metacognitive strategies, suggest-ing support for Verbal Efficiency Theory. Comments written by both groups after pre- andpost-tests confirm use of metacognition. For example, in the short answer pre-test, fourexperimental group members mention ‘‘focusing on relevant information to items” and‘‘paid more attention to details,” but not until the fifth reading. Ten members commenton the first and fifth readings of the short answer post-test (‘‘focused on details not yetclear”). While an in-depth analysis is beyond the scope of this paper, the jump in numberof comments from the experimental group on one of the post-test measures suggests agrowing facility with reading strategy use, which could mean that participants had auto-matized some of their post-lexical reading processes.

11. Limitations and directions for future research

First we are concerned that the shortness and relatively low intensity of the RR treat-ment (11 weeks twice a week with a two week hiatus right in the middle) may have weak-ened expected effects of increased fluency and comprehension on unpracticed passages. Wethink a longer and more intensive RR treatment program would result in faster and moreaccomplished fluency and comprehension on the part of learners in the long term. We notethat Fig. 1 above suggests an ever-increasing wpm reading rate, and wonder what patternsof improvement in fluency and comprehension might have emerged over the longer term.We see no downside of RR. It does not consume large amounts of class time (the treat-ments amounted to 35 minutes twice a week) and learners’ processing of texts using RRcan be built on for other lessons with discussions about the text, learners’ roleplayingand elaboration of dialog in the text, etc.

Second, we are concerned that some items on the short answer test, even though theywere designed to capture main ideas, supporting details, and details, may have seemedexclusively detail oriented, causing participants in both groups to focus only on informa-tion in the text needed to answer the questions. For example, item 5 (Form B), Who did

Mr. Hogan telephone?, captured a main idea: Samuel W. Rogers, 722-6329. A moment


later, he was talking to the surprised Mr. Rogers. Assigning an additional recall test in thestudy to more fully guage participants’ comprehension was probably a reasonable solutionto design flaws in the short answer test. Regardless of test type, we noted in many partic-ipants’ comments that memory played a role in completing test tasks. While answering theshort answers and writing the recalls, participants were not allowed to refer to the testtexts. We wonder whether participants could have demonstrated greater reading fluencyand greater comprehension of the texts without the intervening variable of memory,despite the fact they completed test tasks in their L1s and read the texts five times. Butin mitigation, we add that in most ordinary reading tasks done for the purpose of gettinginformation for text, whether in the L1 or L2, texts are read only once, and assumed to bebasically comprehended.

Finally we found possible problems with test equivalency in a post hoc study done withseven advanced-level ESL students in the US With the short answer test, two out of fourstudents answered items for Form A using more text propositions in just one reading, eventhough one student read Form A first (62% of propositions used), then Form B (37%), andone student read Form B first (37% of propositions used), then Form A (52%). Form Aitems on the short answer test may have been easier to answer, even though the text forthe two forms was equivalent in readability. For the recall test, two out of three studentsused fewer propositions on whatever form they took first. These results may suggest aneffect for task type (recall) with students getting better the second time they write a recall,regardless of the text form they read first.

The issue of memory raised in participants’ comments leads us to potential futureresearch, that of working memory and its role in fluency building in FL reading. Conven-tional wisdom asserts that automatizing foreign language reading processes frees atten-tional resources (and processing components of working memory; see Just andCarpenter, 1992) for comprehending more of a text. We would like to expand our conven-tional measurements of reading fluency and reading comprehension to include workingmemory measurement in learners’ L2s to determine whether their demonstrated workingmemory processing component efficiency changes through RR treatments. Added to evi-dence presented in Fig. 1, we may be able (or not) to argue for a strong role of word rec-ognition in reading fluency, and comment on the extent to which working memory isimplicated.

We are also concerned whether RR is perceived as suitable by FL teachers. This causesus to return to the larger issue of reading fluency being given short shrift in language edu-cation programs, due to prevailing misconceptions about proficiency being sufficient causefor reading fluency, and a lack of understanding of the role of fluency in comprehension.At two schools in Vietnam, local teachers remained sceptical, claiming that students wouldbe bored reading one passage five times and that any rate, by the fifth time, ‘‘they wouldjust be skimming anyway.” No amount of explanation about automaticity and attentionalresource emancipation could persuade teachers that skimming on a fifth reading was good,especially if students were comprehending propositions they had not even noticed during afirst and close reading (reported in participant reports and presented in another study). Inorder for RR to be a viable means of reading fluency development on the global scene,locally valid explanations must be devised. One argument that might be used would beto compare the experimental and control groups in the current study, noting that bothgroups experienced the same grammar-based English instruction in their other classes dur-ing the RR treatment period. As the other instruction was thought to increase learners’


proficiency, why then was the experimental group the only group to improve their readingfluency (Table 1) and comprehension (Table 3)? Gains in proficiency (hypothesized) didnot result in increased reading fluency for the control group–therefore a positive linkbetween language proficiency and reading fluency might not be a supportable idea afterall.

12. Conclusion

In this study with university level English learners in Vietnam, an eleven-week assistedRR treatment was found to be effective in increasing reading fluency and comprehension.Previous studies have failed to report reading comprehension gains from repeated reading(e.g. Taguchi and Gorsuch, 2002; Taguchi et al., 2004) and so we feel our findings aresignificant. Repetition is a major unique component of the RR method, and while theexperimental group failed to read post-test passages faster than the control group, theexperimental group took better advantage of this repetition during the post-tests. Theyread more slowly on the post-tests than they did on their treatment passages at theend of the study but comprehended more than the control group. In English in foreignlanguage settings, where a paucity of acquisition-friendly L2 input is likely to be anissue in the decades to come, RR offers an effective method to help readers becomeindependent.

Acknowledgements

We wish to thank Ngo Phuong of Vinh University, Vietnam, for his support. We alsowish to thank Tran Bo Tien, Lam Van Vien, and Dale T. Griffee for their timely help. Butespecially we wish to thank to the students of Vinh University, Class 43A2, for participat-ing in this study.

Appendix A. Short answer test texts for Two Men Visit, Forms A and B

A.1. Two Men Visit (Form A)

Instructions: Start your stopwatch, and read this story once. Stop your stopwatch assoon as you are done and mark the time on your time sheet. Turn the story over, facedown. Then, answer questions on the test sheet, in Vietnamese. Do not look at the storyas you answer the questions. Your teacher will tell you when you can look at the storyagain.

Background: This story takes place in a town somewhere in the middle of the UnitedStates. This story begins as two men, who are new to town, visit a house.

The night was dark. And the house was dark. Dark and silent. The two men ran towardit quietly. They slipped quickly through the dark bushes which surrounded the house.They reached the porch, ran quickly up the steps, kneeled down, breathing heavily, inthe dark shadows. They waited, listening.Silence. Perfect silence. Then, out of the blackness, a whisper: ‘‘We can’t stay outhere.... Take this suitcase.... Let me try those keys. We’ve got to get in!”


Ten – twenty – thirty seconds. With one of the keys the one man opened the door.Silently, the two men entered the house, closed the door behind them, locked it.Whispering, they discussed the situation. They wondered if they had awakened anyonein the house.‘‘Let’s have a look at this place.” ‘‘Careful, Hasty!” ‘‘Oh, there isn’t anybody awake!”And the soft rays of a flashlight swept the room.It was a large room. A living room. Rugs, carefully rolled, lay piled on one side. Thefurniture, chairs, tables, and couches, was covered by sheets. Dust lay like a light snowover everything.The man who held the flashlight spoke first. ‘‘Well, Blackie,” he said, ‘‘we’re in luck.Looks as if the family’s away.”‘‘Yeah, Gone for the summer, I guess. We better make sure, though.”Together they searched the house. They went on tiptoe through every room. Therecould be no doubt about it. The family was away. Had been away for weeks.Yes, Hasty Hogan and Blackie Burns were in luck. Only once in the past ten days didtheir luck fail them. It was with them when they made their big robbery, their trulymagnificent robbery, in San Francisco. It was with them during their thousand-mile tripeastward, by automobile. It was with them every moment but one.That moment came just one hour before. It came when Blackie, driving the car, ranover a policeman. And Blackie, thinking of the suitcase at Hasty’s feet, had drivenaway. Swiftly.There had been a chase, of course. A wild, crazy chase. And when a bullet had punc-tured the gasoline tank they had had to abandon the car. But luck or no luck, here theywere. Alone, and without a car, in a completely strange town. But safe and sound, withthe suitcase.The suitcase lay in the center of the table, in the center of the room. In the suitcase, neatlittle package on neat little package, lay nearly three hundred thousand dollars!‘‘Listen,” said Mr. Hogan. ‘‘We have to get a car. Quick, too. And we can’t steal oneand use it. It’s too dangerous. We have to buy one. That means that we have to waituntil the stores open. That will be about eight o’clock in this town.”‘‘But what are we going to do with that?” And Mr. Burns pointed to the suitcase.‘‘Hide it right here. Sure! Why not? It’s much safer here than with us, until we get a car.”And so they hid the suitcase. They carried it down to the cellar. Buried it deep in somecoal which lay in a corner of the cellar. After this, just before dawn, they slipped out.‘‘Say, Blackie,” Mr. Hogan remarked as they walked down the street, ‘‘the name of thegentleman we’re visiting is Mr. Samuel W. Rogers.”‘‘How do you know?”‘‘Saw it on some of them books. He’s surely got a wonderful library, hasn’t he?”

A.2. Two Men Visit (Form B)

The automobile salesrooms opened at eight o’clock, as Mr. Hogan thought. Shortlybefore nine, Mr. Hogan and Mr. Burns had a car. A very nice little car. Very quiet,plain, and fast.Three blocks from the house, they stopped. Mr. Hogan got out and walked toward thehouse. He wanted to go around to the rear and slip in.


Fifty yards from the house he stopped and stared. The front door was open. The win-dow shades were up. The family had returned!Well, what bad luck. And what could they do? Could they break into the cellar thatnight, and pick up the suitcase? No, it was too dangerous. Mr. Hogan would have tothink of something.‘‘Leave it to me, kid,” he told Mr. Burns. ‘‘You drive the car. I’ll do the special brainwork. Let’s find a telephone. Quick!”Ten minutes later, Mr. Hogan was looking at a telephone directory. Yes, there it was —Samuel W. Rogers, 722-6329. A moment later he was talking to the surprised Mr.Rogers.‘‘Hello,” he began, ‘‘is this Mr. Samuel Rogers?”

‘‘Yes, this is Mr. Rogers.”Mr. Hogan cleared his throat. ‘‘Mr. Rogers,” he said, and his tone was sharp, official,impressive, ‘‘this is Headquarters, Police Headquarters, talking. I am Simpson. Ser-geant Simpson, of the detective division.”‘‘Yes, yes!” came over the wire.‘‘The Chief —the Chief of Police, you know,” Mr. Hogan lowered his voice a little, ‘‘hasordered me to get in touch with you. He’s sending me out with one of our men to seeyou.”‘‘Am I in trouble of some kind?” asked Mr. Rogers.‘‘No, no, no. Nothing like that. But I have something of great importance to talk to youabout.”‘‘Very well,” came the voice of Mr. Rogers. ‘‘I’ll wait for you.”‘‘And, Mr. Rogers,” Mr. Hogan said, ‘‘please keep quiet about this. Don’t say anythingto anybody. You’ll understand why when I see you.”On the way back to the house Mr. Hogan explained his idea to Mr. Burns.Within ten minutes ‘‘Sergeant Simpson” and ‘‘Detective Johnson” were talking with thesurprised Mr. Rogers. Mr. Rogers was a small man, and plain. He had pale blueeyes. Not much of a chin. A funny little face. He was nervous — a badly frightenedman.Mr. Hogan told the whole story. Somewhat changed. Very much changed. And Mr.Rogers was surprised, but delighted.He accompanied Mr. Hogan to the cellar. And together they dug up the suitcase. Tookit to the living room, opened it, saw that it had not been touched. It really did hold asmall fortune. Bills, bills, bills!Mr. Hogan closed the suitcase.‘‘And now, Mr. Rogers,” he announced, in his best official manner, ‘‘Johnson and Imust run along. The chief wants a report. We have to catch the rest of the robbers.I’ll keep in touch with you.”He picked up the suitcase and rose. Mr. Burns also rose. Mr. Rogers also rose. The triowalked to the door. Mr. Rogers opened it. ‘‘Come on in, boys,” he said pleasantly —and in walked three men. Large men. Strong men. Men in police uniform who, withoutfear, stared at Mr. Hasty Hogan and Mr. Blackie Burns.‘‘What does this mean, Mr. Rogers?” asked Mr. Hogan.‘‘It’s quite simple,” said Mr. Rogers. ‘‘It just happens that I am the Chief ofPolice!”


Appendix B. Short Answer Items for Two Men Visit, Forms A and B

B.1. Two Men Visit (Form A)

B.1.1. Test Sheet

Name: ____________________________________ Date: _____________

1. How long did the men wait outside the house?2. How many men entered the house?3. Is anyone in the house? How do you know?4. What room do the men see first?5. Why don’t the men turn on the lights?6. How many miles did the men drive?7. How many times did their luck fail them?8. When the men had bad luck, why did they keep driving?9. Who chased the men? Why do you think so?10. Why did the men abandon their car?11. What is in the suitcase?12. How did the men get the contents of the suitcase?13. In the house, what did the men do with the suitcase?14. What is the name of the man they are visiting?15. Complete the chart below. Give reasons you remember from the story to argue for both

of the positions:

The house belongs to the men. The house does not belong to the men.

The section below is not part of the test, but it is important that you answer. Pleasewrite anything that came into your mind as you took this test. For example, what was easyor difficult about reading the English passage? Were some questions easier or harder thanothers? Did you notice any changes in your thinking as you read the passage or answeredthe questions?

When you are finished, give test sheet to your teacher and wait for instructions.

B.2. Two Men Visit (Form B)

B.2.1. Test Sheet

Name: ____________________________________ Date: _____________

1. What time did the car dealer open?2. What was the men’s new car like?3. Why did Mr. Hogan get out of the car before getting to the house and then walk the rest

of the way?4. What did Mr. Hogan see when he got close to the house?


5. Who did Mr. Hogan telephone?6. When Mr. Hogan telephoned, who did he say he was?7. Why did the person who answered the telephone agree to wait for the men?8. After the telephone call, where did the men go?9. What did Mr. Rogers look like?10. Where did they find the suitcase?11. What was in the suitcase?12. After Mr. Hogan and Mr. Burns went to the door to leave, what happened?13. Did Mr. Rogers tell anyone about the telephone call?14. Why were Mr. Hogan and Mr. Burns surprised when they found out about Mr.

Roger’s job?15. Complete the chart below. Give reasons you remember from the story to argue for both

of the positions:

Mr. Hogan and Mr. Burns are the police. Mr. Hogan and Mr. Burns are not the police.

The section below is not part of the test, but it is important that you answer. Pleasewrite anything that came into your mind as you took this test. For example, what was easyor difficult about reading the English passage? Were some questions easier or harder thanothers? Did you notice any changes in your thinking as you read the passage or answeredthe questions?

When you are finished, give test sheet to your teacher and wait for instructions.

Appendix C. Recall test texts for Man With No Name, Forms A and B

C.1. Man With No Name (Form A)

Instructions: Start your stopwatch, and read this story once. Stop your stopwatch assoon as you are done and mark the time on your time sheet. Turn the story over, facedown. Then, use the provided piece of paper, and in Vietnamese, write down everythingyou remember about this story. It may help to write in the same order as the story. Donot look at the story as you write. Your teacher will tell you when you can look at thestory again.

Background: This story takes place in a quiet part of northern Great Britain. Therearen’t many people living there, and not many people drive cars. This story begins as awoman begins a long walk to buy some things in town.

Next morning, Mary left the cottage. She was going shopping in the village four milesaway. She carried a rucksack on her shoulder. The rucksack was light, and Marywalked quickly.Mary started to climb up the hill behind her cottage. On top of the hill, there was awood. Through the wood, on the other side of the hill, there was a valley. In the valleythere was a small airfield.


Thirty years ago, the airfield had closed and now grass grew on the runway. Everyweek, Mary walked over the old runway. It was the quickest way to the village andthe shops.It was a bright, clear morning and Mary felt happy. The sun was shining on the grass,and the air was clean and fresh. Mary loved the bright, clear air in the north of Scot-land. She liked living alone in Scotland. It was so quiet, so peaceful, so easy to write herbooks here.Soon Mary reached the wood on the top of the hill. It was darker in the wood, still andsilent. But soon Mary left the wood, and the sun shone again. At the top of the hill,Mary stopped for a moment and took a deep breath. She was ready to climb down intothe valley.First, she looked up at the clear blue sky. Then she started to climb down the hill. ThenMary looked down at the old runway. And suddenly, for a second, her breath stopped.Mary had a shock. She closed her eyes for a second. Then she opened them again. Therewas a jet on the old runway.The huge jet filled the small airfield. The sun shone on the plane’s wings. The long grasscovered the plane’s wheels. And everything was silent and still.Mary stopped and stood still. Then she thought quickly. It must be an accident. Theplane had crashed. She must get help. She must look for people, and get help for thepassengers.She started to run quickly down the hill. Then she suddenly stopped and stood stillagain. Something was wrong. She looked down at the plane again. Everything wassilent. She saw no one and she heard nothing. There was no sound and nothing moved.Were all the people on the plane dead?Mary walked slowly now, down the hill. As she walked, she looked at the aircraft. Shesaw no damage. Nothing was scratched or broken. But there was no one, no sound,nothing.

Appendix D.

D.1. Man With No Name (Form B)

Instructions: Start your stopwatch, and read this story once. Stop your stopwatch assoon as you are done and mark the time on your time sheet. Turn the story over, facedown. Then, use the provided piece of paper, and in Vietnamese, write down everythingyou remember about this story. It may help to write in the same order as the story. Donot look at the story as you write. Your teacher will tell you when you can look at thestory again.

Background: This story takes place in a quiet part of northern Great Britain. Therearen’t many people living there, and not many people drive cars. Mary, a woman who livesin the area, has been taking a long walk from her home to town to do some shopping.On her way to town, she has seen something out of the ordinary and has gone toinvestigate.

At the bottom of the hill, Mary stopped again. She heard something now. It was thesound of music. Mary saw something too. The passenger ramp was down. The steps


led from the open door into the long grass. The sound of music came from the opencabin door.Mary shouted. No one answered. She ran quickly over the grass to the aircraft. At thebottom of the ramp, she stopped again and shouted very loudly. Again, no oneanswered. But the sound of music came from the open cabin door.Mary dropped her rucksack and ran up the ramp and into the plane.The plane was empty – there was no one in it. Mary walked quickly, along the gangway,to the front of the plane. She found no one. No one alive or dead. She walked backslowly and looked at all the passenger seats. The seats were empty. There was noone dead or alive on the plane. Where were the passengers?Mary walked again to the front of the plane. She looked again at the empty seats.Where were the crew? At the front of the plane, Mary put her hand on the crew door.Then, for the second time that morning, her breath stopped. A hand tapped her shoul-der, and a voice said, ‘‘Is this yours?”

It was a man. He was carrying Mary’s rucksack.‘‘My name is Ned Harding,” he said. ‘‘I was walking to the village when I saw theplane. . ..”‘‘There’s no one here,” said Mary. ‘‘There’s nobody on the plane.”Mary and Ned searched the plane together. Together they called, but no one answered.Together they shouted, but no one answered. There were no passengers and nocrew.Ned found food and drink in the galley. Mary found a warm cup of coffee in front ofone of the seats. But they did not find the passenger list, and they did not find any handluggage.There was no damage. Nothing was broken. Everything was tidy.‘‘This plane did not crash,” said Ned. ‘‘It landed.”‘‘Yes,” said Mary, ‘‘but where are the passengers and the crew?”

Mary opened her rucksack and took out a notebook. She sat down near the galley, atthe front of the plane. She began to make some notes. Mary stopped writing and lookedat Ned. He was searching the passenger seats. Suddenly he bent down and picked up apiece of paper. He looked at it quickly and put it in his pocket. Mary watched him. Shegot up and followed Ned down the plane.

D.2. Man With No Name (Form A)

D.2.1. Test sheet

Name: ___________________________________ Date: ___________________

Write your recall of the passage below. Use both sides if necessary. When you are finishedwriting the recall, answer the questions on side two.

The section below is not part of the test, but it is important that you answer. Please writeanything that came into your mind as you took this test. For example, what was easy ordifficult about reading the English passage? Were some things in your recall easier orharder to write than others? Did you notice any changes in your thinking as you read thepassage or wrote your recall?


D.3. Man With No Name (Form B)

D.3.1. Test sheet

Name: ___________________________________ Date: ___________________

Write your recall of the passage below. Use both sides if necessary. When you are finishedwriting the recall, answer the questions on side two.

The section below is not part of the test, but it is important that you answer. Please writeanything that came into your mind as you took this test. For example, what was easy ordifficult about reading the English passage? Were some things in your recall easier orharder to write than others? Did you notice any changes in your thinking as you read thepassage or wrote your recall?

References

Anderson, N., 1999. Exploring second language reading: issues and strategies. Heinle & Heinle, Boston, MA.

Blum, I., Koskinen, P.A., Tennant, N., Parker, E.M., Straub, M., Curry, C., 1995. Using audiotaped books to

extend classroom literacy instruction into the homes of second-language learners. Journal of Reading

Behavior 27, 535–563.

Carver, R.P., Hoffman, J.V., 1981. The effect of practice through repeated reading on gain in reading ability using

a computer-based instructional system. Reading Research Quarterly 16, 374–390.

Chard, D.J., Pikulski, J.J., McDonagh, S., 2006. Fluency: the link between decoding and comprehension for

struggling readings. In: Rasinsky, T., Blachowicz, C., Lems, K. (Eds.), Teaching reading fluency. Guilford,

New York, pp. 39–61.

Chomsky, C., 1976. After decoding: what?. Language Arts 53 288–296.

Dahl, P.J., 1974. An experimental program for teaching high speed word recognition and comprehension skills.

Washington, DC: National Institute of Education, Office of Research (ERIC Document Reproduction

Service No. ED 099812).

Davies, E., Town, P., 1992. The man with no name. Heinemann International, Oxford.

Day, R.R., Bamford, J., 1998. Extensive reading in the second language classroom. Cambridge University Press,

New York.

Dlugosz, D.W., 2000. Rethinking the role of reading in teaching a foreign language to young learners. English

Language Teaching Journal 54, 284–290.

Dowhower, S.L., 1987. Effects of repeated reading on second-grade transitional readers’ fluency and

comprehension. Reading Research Quarterly 22, 389–406.

Doyle, A., 1999. Scandal in Bohemia. Pearson Education Limited, Harlow, England.

Ehri, L.C., 1998. Grapheme-phoneme knowledge is essential for learning to read words in English. In: Metsala,

J.L., Ehri, L.C. (Eds.), Word recognition in beginning literacy. Lawrence Erlbaum, Mahwah, NJ, pp. 3–40.

Ehri, L.C., 2005. Learning to read words: theory, findings, and issues. Scientific Studies of Reading 9, 167–188.

Eskey, D., 1988. Holding in the bottom: an interactive approach to the language problems of second language

readers. In: Carrell, P., Devine, J., Eskey, D. (Eds.), Interactive approaches to second language reading.

Cambridge University Press, New York, pp. 93–100.

Faulkner, H.J., Levy, B.A., 1994. How text difficulty and reader skill interact to produce differential reliance on

word and content overlap in reading transfer. Journal of Experimental Child Psychology 50, 1–24.

Favreau, M., Segalowitz, N.S., 1983. Automatic and controlled processes in reading a second language. Memory

and Cognition 11, 565–574.

Grabe, W., 1991. Current developments in second language reading research. TESOL Quarterly 25 (3), 375–406.

Grabe, W., 2004. Research on teaching reading. Annual Review of Applied Linguistics 24, 44–69.

Grabe, W., Stoller, F.L., 2002. Teaching and researching reading. Pearson Education, Harlow.

Harris, T.L., Hodges, R.E. (Eds.), 1995. The literacy dictionary: The Vocabulary of Reading and Writing.

International Reading Association, Newark, Delaware.


Herman, P.A., 1985. The effect of repeated readings on reading rate, speech pauses, and word recognition

accuracy. Reading Research Quarterly 20, 553–564.

Hogaboam, T.W., Perfetti, C.A., 1978. Reading skill and the role of verbal experience in decoding. Journal of

Educational Psychology 70, 717–729.

Jensen, L., 1986. Advanced reading skills in a comprehensive course. In: Dubin, F., Eskey, D.E., Grabe, W.

(Eds.), Teaching second language reading for academic purposes. Addison-Wesley, Reading, MA, pp. 103–

124.

Just, M., Carpenter, P., 1992. A capacity theory of comprehension: individual differences in working memory.

Psychological Review 99 (1), 122–149.

Koda, K., 1996. L2 word recognition research: a critical review. Modern Language Journal 80, 450–460.

Koda, K., 2005. Insights into second language reading: a cross-linguistic approach. Cambridge University Press,

New York.

Koskinen, P.S., Blum, I.H., 1984. Repeated oral reading and the acquisition of fluency. In: Niles, J., Harris, L.

(Eds.), Changing Perspectives on Research in Reading/Language Processing and Instruction. Thirty-third

Yearbook of the National Reading Conference. National Reading Conference, Rochester, NY, pp. 183–187.

Kuhn, M.R., Stahl, S.A., 2003. Fluency: a review of developmental and remedial practices. Journal of

Educational Psychology 95, 3–21.

LaBerge, D., Samuels, S.J., 1974. Toward a theory of automatic information processing in reading. Cognitive

Psychology 6, 293–323.

Morgan, R., Lyon, E., 1979. Paired reading: a preliminary report on a technique for parental intuition on reading

to retarded children. Journal of Child Psychology and Psychiatry 20, 151–160.

Nation, I.S.P., 2001. Learning vocabulary in another language. Cambridge University Press, New York.

National Reading Panel, 2000. Report of the subgroups: National Reading Panel. Washington, DC: National

Institute of Child Health and Human Development.

Nuttall, C., 1996. Teaching reading skills in a foreign language, 2nd ed. Heinemann, Oxford.

O’Shea, L.J., Sindelar, P.T., O’Shea, D.J., 1985. The effects of repeated readings and attentional cues on reading

fluency and comprehension. Journal of Reading Behavior 17, 129–142.

Perfetti, C., 1985. Reading ability. Oxford University Press, New York.

Perfetti, C.A., 1988. Verbal efficiency in reading ability. In: Daneman, M., Mackinnon, G.E., Waller, T.G. (Eds.),

. In: Reading research: advances in theory and practice, vol. 6. Academic Press, New York, pp. 109–143.

Pikulski, J.J., Chard, D.J., 2005. Fluency: bridge between decoding and reading comprehension. The Reading

Teacher 58, 510–519.

Rashotte, C.A., Torgesen, J.K., 1985. Repeated reading and reading fluency in learning disabled children.

Reading Research Quarterly 20, 180–188.

Reynolds, R.E., 2000. Attentional resource emancipation: toward understanding the interaction or word

identification and comprehension processes in reading. Scientific Studies of Reading 4 (3), 169–195.

Ruddell, R.B., 1994. The development of children’s comprehension and motivation during storybook discussion.

In: Ruddell, R.B., Ruddell, M.R., Singer, H. (Eds.), Theoretical Models and Processes of Reading, 4th ed.

International Reading Association, Newark, DE.

Samuels, S.J., 1979. The method of repeated readings. The Reading Teacher 32, 403–408.

Samuels, S.J., 2004. Toward a theory of automatic information processing in reading revisited. In: Ruddell, R.B.,

Ruddell, M.R., Singer, H. (Eds.), Theoretical models and processes of reading, fifth ed. International Reading

Association, Newark, DE, pp. 1127–1148.

Samuels, S.J., 2006. Reading fluency: its past, present, and future. In: Rasinski, T., Blachowicz, C., Lems, K.

(Eds.), Fluency Instruction: Research-based Best Practices. Guilford Press, New York, pp. 7–20.

Segalowitz, N., 1986. Skilled reading in the second language. In: Vaid, J. (Ed.), Language Processing in

Bilinguals: Psycholinguistic and Neuropsychological Perspectives. Lawrence Erlbaum, Hillsdale, NJ, pp. 3–

19.

Segalowitz, N., Poulsen, C., Komoda, M., 1991. Lower level components of reading skill in higher level

bilinguals: implications for reading instruction. AILA Review 8 (1), 15–30.

Silberstein, S., 1994. Techniques and Resources in Teaching Reading. Oxford University Press, New York.

Taguchi, E., 1997. The effects of repeated readings on the development of lower identification skills of FL readers.

Reading in a Foreign Language 11, 97–119.

Taguchi, E., Gorsuch, G., 2002. Transfer effects of repeated EFL reading on reading new passages: silent reading

rate and comprehension. Reading in a Foreign Language 14(1), 1–18. Available: <http://nflrc.hawaii.edu/rfl/

April2002/taguchi/taguchi.html>.

http://nflrc.hawaii.edu/rfl/April2002/taguchi/taguchi.html

http://nflrc.hawaii.edu/rfl/April2002/taguchi/taguchi.html


Taguchi, E., Takayasu-Maass, M., Gorsuch, G., 2004. Developing reading fluency in EFL: how assisted repeated

reading and extensive reading affect fluency development. Reading in a Foreign Language, 16(2), 1–19 [On-

line]. Available from: <http://nflrc.hawaii.edu/rfl/October2004/taguchi/taguchi.html>.

Walczyk, J.J., 2000. The interplay between automatic and control process in reading. Reading Research

Quarterly 35, 554–566.

Young, J.N., 1971. The wrong house. In: Dixson, R.J. (Ed.), Modern short stories in English. Regents, New

York, pp. 59–64.

Young, A.R., Bowers, P.G., MacKinnon, G.E., 1996. Effects of prosodic modeling and repeated reading on poor

readers fluency and comprehension. Applied Psycholinguistics 17, 59–84.

http://nflrc.hawaii.edu/rfl/October2004/taguchi/taguchi.html

Documents

Repeated reading for developing reading fluency and reading comprehension: The case of EFL learners in Vietnam