Improving students' fluency through voice recording

Improving students’ fluency through voice recording* Aldy Rizky Nesapiradana & Rahmila Murtiana LBPP LIA Banjarmasin Abstract Fluency as an important goal in learning English is the ability to produce the language smoothly and naturally. The problem is, learners often fail to achieve the expected fluency despite the time spent in learning English. In an attempt to help learners improve their fluency, we conducted this classroom action research. Two classes of Intermediate 4 at LIA Banjarmasin were selected to be the object of this research. The students in both classes were asked to read a monologue and their voice was recorded. The result was then played, and the teacher pointed out which parts needed improvement. At this stage, the teacher modeled the correct chunks, intonation, and stress. The next step, the students were asked to re-read the monologue, re-record their voice, and play the result to see the improvement. The result is, through this procedure most of the students were able to improve their reading fluency, as evidenced by their correct use of chunking, phrasing, intonation, and word stress. Background One important goal in learning English is fluency, that is, the ability to produce the language smoothly and naturally. Fluency involves a degree of automaticity and the ability to retrieve ready-made chunks of language (McCarthy, 2010). The term fluency is more often associated with speaking skills, while actually it can also be applied to reading skills. In reading, fluency is defined by the National Assessment of Educational Progress (NAEP) as "the ease or 'naturalness' of reading, including how a reader groups or phrases words as revealed through intonation, stress, and pauses, and expresses oneself in feeling, anticipation, and characterization during oral reading” (http://www.learningrx.com/reading-fluency.htm). Reading fluency is a challenge for students who learn English as a foreign language, not only they have to deal with the pronunciation but also with the appropriate chunk and word stress to make the text meaningful. Ford (2012) stated that fluent readers are able to read words accurately and automatically with expression, and they can focus their attention on the text's meaning because they do not have to slow down in order to concentrate on decoding the individual words in a text. In this way, fluency acts as a bridge between word recognition and comprehension. Unfortunately, teachers often overlook the power of reading aloud, and emphasize more on silent reading activity. There has been a belief that reading aloud is an inauthentic language activity, while as a matter a fact, reading aloud can contribute to spoken fluency (Thornbury, 2005; Ford, 2012). Among the advantages of reading * Paper presented at the 6th LIA Research Colloquium, 22 October 2013. Published in LIA CAR Journal vol.6. aloud is it provides learners a framework to focus on lower level features of talk such as pronunciation, word stress, and intonation. It is also a form of scaffolding for learners before they are assigned to do a spoken activity. For example, prior to acting out a dialog, a role play, or giving an oral presentation, learners can rehearse the script first by reading it aloud (Thornbury, 2005, p. 70). In LIA promotion test, reading aloud is even a part of oral test which assesses the student spoken ability. In real life, reading fluency is needed in certain occasion such as giving an announcement, giving a presentation, or reporting news. Based on the consideration that improving learners’ fluency in reading is equally important as fluency in speaking, we decided to conduct this classroom action research. Our main purpose is to tap the students’ awareness of the importance of fluency and help them improve their ability in producing smooth and natural language. The technique we used in improving students’ fluency is through voice recording. Voice recording is believed to be an effective way because when the voice is recorded, it can be replayed and students can listen again to it and notice the mistakes or which parts need improvement. The questions of this research are: 1. How will the use of voice recording improve the students’ fluency? 2. Which component of fluency is most improved and which component is least improved? Theoretical Framework More fluent vs. less fluent reader Several sources acknowledge that fluent readers can read aloud effortlessly and with expression, because they recognize the word automatically, and their reading sounds natural as if they are speaking. Hook and Jones (2004), for example, wrote that “Good readers read fluently with adequate speed and when they read aloud, they use appropriate phrasing, intonation, and their oral reading mirrors their spoken language” (p.18). In contrast, less fluent readers are marked by their halting and inconsistent rate, poor phrasing, and inadequate intonation pattern; they read slowly, word by word and their oral reading is often choppy, and without natural expression (sources:http://www.reading-skills-pyramid.org/reading-fluency.htm; http://www.learningrx.com/reading-fluency.htm). A study on reading fluency conducted by the NAEP revealed that there is a significant correlation between reading fluency and reading comprehension. Another study conducted by Researchers at the National Institute for Literacy investigated two major instructional approaches related to reading fluency. The first is repeated and monitored oral reading, where students read passages aloud several times and receive guidance and feedback from the instructor. The second is independent silent reading, where students are encouraged to read extensively on their own. The key finding was "repeated and monitored oral reading improves reading fluency and overall reading achievement" (National Institute for Literacy website, 2006). In the case of reading fluency of L2 learners, Robertson (2009) argued that “hearing the text out loud may not necessarily provide a step towards comprehension as it is likely to do for native speakers”. Many L2 learners can read fast and accurate because they are good readers in their native language and have strong decoding skills but it does not mean they automatically understand the text (Robertson, 2009). A more important factor to be fluent and at the same time understand the text is by reading in meaningful chunks. In a study which investigated the relationship between chunking and L2 reading fluency, Yamashita and Ichikawa (2010) found that the difficulty of chunking a text negatively affected comprehension and smoothness for the intermediate learners. Therefore, they suggested that reading in meaningful chunks or grouping words into grammatical and meaningful units is one of the fundamental processes necessary for fluent reading (Yamashita & Ichikawa, 2010). Students often fail to put chunks when reading because the prosodic features are not present explicitly in a text. In a spoken language, the features are easily recognized by gestures, facial expression, intonation, and stress. In a text, prosodic features are sometimes signaled through punctuation, but it is often not enough to guide the students to read in a meaningful chunks. Therefore, the students should be trained to map or to connect the prosodic features of spoken language onto a text (Hook & Jones, 2004), in order to enhance their fluency. Instruction technique to improve fluency To help students become more fluent readers, teachers can provide models of fluent reading, then have students reread the text on their own. By listening to good models of fluent reading, students learn how a reader's voice can help written text make sense. Usually, having students repeatedly read a text four times is sufficient to improve fluency (source: Reading Pyramid). Teacher can have the student immediately read it back to the teacher, or allow students to practice reading along with the taped text, which is an excellent way for them to learn appropriate pronunciation and phrasing (Robertson, 2009). Several other ways of reading activities to improve fluency can also be done such as: 1). choral reading, that is, a group of students read the same material aloud; 2). paired reading, where students work with one competent reader and one that is not as strong in reading; 3). recorded reading , that is, students follow the reading by listening to the book from a tape or CD; 4). echo reading , that is, when the teacher reads a sentence or phrase and the student echoes back; 5). buddy reading, that is, students at about the same reading level are paired up and read together (source: http://www.learningrx.com/reading-fluency.htm) To develop student awareness of the prosodic features such as intonation, stress, and phrasing, Thornbury (2005) suggested that before reading aloud activity, students are asked to mark on the stressed words and divided parts of the text into meaningful chunks. A more structured and systematic instruction to facilitate students to recognize prosodic features in a text was offered by Hook & Jones (2004). The first focus is by practicing on intonation. Teachers can introduce a short sentence and model how to read it with different word stress. When the students can practice it until they are fluent, teachers can modify and expand the sentence and vary the rate, intensity, and pitch. Teachers can also vary the sentence by alternating punctuation marks at the end of the sentence, so the students can see the different intonation and meaning between sentence with full stop, with question mark, or with exclamation point. The second focus is practicing phrasing. Teachers can assign the students to recite a group of alphabet in chunks, then when they understand the basic concept of chunking, teachers can move on the next step, that is, assign the students to chunk a text into syntactic or meaningful units such as noun phrase, verb phrase, or prepositional phrase. Teachers can give example by marking the phrases with slashes, then ask the students to continue putting the slashes. An alternative way is by asking the student to read a text while scooping under the phrases with their finger or a pencil. Below is the example of slashing and scooping: a) slashes: Meg told Jim / her kite was stuck / in a tree b) scooping: Meg told Jim her kite was stuck in a tree Measuring students’ fluency Ford (2012) proposed three components of measuring reading fluency, that is, accuracy, automaticity, and prosody. Accuracy refers to the percentage of words a reader can read correctly in a given text. Reading accuracy can be divided into three levels: independent level, instructional level, and frustrational level. A student is at independent reading level if he/she can read with 98-100% accuracy, or with no more than approximately 1 in 20 words difficult for him/her. The accuracy range for instructional-level text is 90-97%, or with no more than approximately 1 in 10 words difficult for the student. Whereas frustrational reading level is if the student can read with less than 90% accuracy, or with more than 1 in 10 words difficult for the student. At this level, the student would have difficulty reading and comprehending the text, even with the support of the teacher (Ford, 2012). The second component, automaticity is usually measured as reading rate, or the number of words a student reads per minute (WPM). Teacher can use a stopwatch to count the minute, or in the voice recording, it is automatically shown on the screen. To count the reading rate, multiply the total number of words in the passage by 60, and divide by the student's reading time in seconds. For example, if a student read a 200-word passage in 2 minutes and 32 seconds, which is equal to 152 seconds, the reading rate is (200 x 60) ÷ 152 seconds = 79 words per minute (wpm). The third component, prosody refers to the appropriate use of intonation and phrasing in reading, which involves paying attention to punctuation signs, assigning appropriate stress to individual words within a sentence, and raising or lowering voice intonation to match the meaning of the text. Chunking is included here. In assessing prosody, teacher can use rubrics to assess whether students are reading with appropriate pitch variation, intonation, phrasing, and expression. One such rubric is the Oral Reading Fluency Scale created for the National Assessment of Educational Progress (NAEP). To use this rubric, simply have a student read a short passage of text, and assign a score of 1-4 (see appendix 2). As a comparison to this scale, Fountas and Pinnell (2001) also developed a fluency scale which consists of 6 components: rate, phrasing, intonation, pausing, stress, and integration. The first component, rate refers to pace in reading, which is not too fast or not too slow; in other words, the reader should moves along the reading text appropriately with few slow downs, stops or pauses. The second component, phrasing refers to the way the reader puts words together in group to represent the meaningful units of language, to make it sound like an oral language. The third component, intonation is sometimes called expression, or the variation of the voice in tone, pitch, and volume to reflect the meaning of the text. The fourth component, pausing refers to the way the reader pauses at punctuation such as commas, dashes or full stops. The fifth component, stress, refers to the emphasis put on particular word, which is marked with a louder tone. These four components: intonation, phrasing, pausing, and stress are part of prosody. Finally, integration as the sixth component is to assess the overall fluency which involves the consistent use of the five aforementioned dimensions (Fountas & Pinnell, 2001). The score for each component ranges from 1 to 4 (see appendix 3 for more details). Data Analysis In conducting this research, we followed an action research cycle: plan, act, observe, and reflect, as suggested by Burns (2010). 1. Plan: It is the first stage in which we started to identify the problem, develop research questions, and consult relevant literature. The identified problem was students of Intermediate levels cannot yet be categorized as fluent readers. Most of them still struggled to read smoothly, naturally, accurately, and automatically. Thus, we developed an action plan, that is, by using voice recording, and analyzed whether it could help improve student fluency. In addition, we consulted literature related to fluency including the theories and previous research on reading fluency 2. Act: It is the second stage in which we started to collect data by recording the students’ voice. The data were collected during term 2/2013. Two classes of Intermediate 4 at LBPP LIA Banjarmasin were selected as the object of this research. We chose Intermediate 4 students because they have been learning about language chunk and word stress. The first recording was conducted after covering lesson 3, where at the end of the unit there is a listening practice on language chunk. The students read a monologue (see appendix 1), and their voice was recorded using a cellular phone. In the next meeting, the result was played, and the teacher pointed out which parts needed improvement. At this stage, the teacher modeled the correct chunks, intonation, and stress. Then, the students were asked to re-read the monologue, and re-record their voice. The total number of the students who were involved in this research was 16, consisting of 5 males and 11 females. Actually there were more students in the class but we could not involve all of them because a few students did not come either during the first recording or the second recording, so we decided to eliminate them from the data presentation and analysis. 3. Observe: It is the third stage in which the results of the plan are analyzed. As Burns (2010) points out, analyzing data in an action research cycle is a continuing process of reducing information to find explanations and patterns, which involve some steps such as: assembling the data, coding the data, comparing the data, and building meanings and interpretation. To implement the data analysis, we played both recordings and compared the results to see whether there was improvement in student fluency. This stage also included scoring the student fluency using a fluency scale rubric. In assessing the students’ fluency, we decided to use Fountas and Pinnell Fluency Scale because it includes the components of automaticity (reading rate), and prosody (phrasing, chunking, intonation, and word stress). However, we deliberately eliminated the sixth component (integration), as the sixth component is meant for overall performance, while we intended to focus on each component and see which component improved most and which one improved least. For accuracy, we assumed that the students were already between independent and instructional reading level. 4. Reflect: The last stage is making reflection and planning for further action. Reflecting includes reflecting on practice, reflecting on the research process, reflecting on beliefs and values, and reflecting on feelings and experiences (Burns, 2010). It is the point where we as the researchers question our action and the results whether it has made significant improvement to the students or not, and whether there is other alternative plan or further action which is more effective. Result and Discussion In analyzing the result we used both quantitative and qualitative analysis. Quantitative analysis was used to measure students’ fluency in terms of nominal scale from 1 to 4, whereas qualitative analysis is used in describing the fluency components produced by the students while reading the text. As previously explained, the first recording was conducted after the student finished learning Lesson 3 of Intermediate 4 book. At the end of the lesson there was a listening exercise where students had to identify the correct chunks in a monologue they heard by putting slashes on the text. As a follow up of this activity, the students were asked to read the text and their voice was recorded. Although the students have heard the text being read to them from the cassette, on the first recording, some of the students were unable to reproduce the text as smooth as the example. The result is presented in the following table: Table 1. Result of the first recording of Mon-Wed class No 1 2 3 4 5 6 7 Name Iyet Josh Michael Nabila Nazmi Ridho Sean Rate 3 3 3 3 3 2 3 Phrasing 2 3 3 2 2 2 2 Intonation 2 2 3 2 3 2 3 Pausing 2 3 3 3 3 2 2 Stress 2 3 3 3 3 2 3 Generally, students of Monday - Wednesday class were able to read the monologue at appropriate rate, so the score for the first component (rate) was quite good; although it still needs some improvement to make them achieve a higher score. The students were able to start reading the text quite smoothly, however, they stumbled over the words in the middle of the text, especially in the part : “We didn’t check everything twice the way we normally do and instead decided to start immediately. It turned out that the wire at the other end was not properly secured and it came loose.” The sentence was quite long, so students had to figure out themselves where to pause or group phrases. It was evident that most of the students still had problem in placing the appropriate phrasing and chunking in the text. For example, Iyet read: “twice/ the /the /way/ we normally” and “at/ the/ the/ other end /was/ not properly/ se/cured” instead of putting them into one phrasing. Beside repeating the word when she stumbled over some words, she also missed some words such as “that” after “it turned out” , and she misread some words such as “was” became “ways”, “the other” become “the others”, and “this business” became “the business”. Thus, Iyet needed to improve her reading ability not only in terms of fluency but also in terms of accuracy. The other student, Josh was quite fluent but he read too fast, so he slipped in words such as “decided” became “deciding”. His intonation was also rather flat, little variation in tone and voice. Similarly, Ridho had a flat intonation; he seemed to read in a hurry that he mispronounced some words such as: “secured” became “square”, “the way we normally do” became “the way our normal do”, “came” became “come”, and “loose” became “lost”. Other examples of inappropriate phrasing and chunking are by Michael in reading “the way we / normally / do”, Nazmi in reading “the other end wasn’t/ properly secured”, or Sean in reading “everything/ twice the way/ we normally do”, and “the wire at the other/ end was not properly”. Sean, Nazmi, and Nabila had been trying to put stressing on certain words although it still did not sound natural. Particularly for Nabila, not only that she used wrong phrasing, but she also had some problems with pronunciation, so she needed more practice to improve her fluency. Meanwhile, the result of the first recording of Tuesday- Thursday class is presented in the following table. Table 2. Result of the first recording of Tue-Thu class No 1 2 3 4 5 6 7 8 Name Ferina Naufal Rezka Rima Rina Rizwan Roshi Yulita Rate 3 3 4 3 2 3 1 1 Phrasing Intonation Pausing 2 3 2 3 2 3 3 3 4 3 3 3 3 2 3 3 2 3 1 1 1 1 1 2 Stress 3 2 3 3 2 2 1 1 9 Windy 3 3 3 3 3 As shown in the table, the fluency of students of Tuesday-Thursday class was quite varied. Rezka was the best in terms of rate, as well as other components compared to her classmates. She was able to read smoothly with appropriate pace and pausing. The only part she hesitated was when she read: “at the other /end”. Rima and Windy were also quite fluent, but Rima used inappropriate phrasing in reading “check everthing/ twice/ the way we normally”, whereas Windy made some pronunciation errors when she read immediately, lesson and she stumbled on reading ”concentration” Other students mostly have problems in phrasing, stress, and intonation. For example, Ferina made inappropriate phrasing in reading “It wasn’t/ because I lost my concentration”, “We didn’t check/ everything/ twice the way/ we normally/ (do) and instead decided”, “the wire at the other end/ wasn’t/ properly /secured”. Rizwan stumbled over the part “check everything twice the way/ we normally do/ and instead decided to/ start immediately”. Similarly, Naufal and Rina used inappropriate phrasing in reading “everything twice/ the way we normally do”. In terms of intonation and stress, Naufal was less expressive, while in terms of rate, Rina read rather fast so she made some pronunciation errors in reading immediately, twice, came, and lesson. The worst fluency was shown by Roshi and Yulita. Referring to three components of fluency: accuracy, automaticity, and prosody as suggested by Ford (2012), these two students were the least fluent. It actually surprised us that students at Intermediate 4 level could be at frustrational reading level (less than 90% accuracy). Roshi read poorly and awkwardly as she tended to read word per word with lots of hesitation, and her reading rate was 58 seconds, much longer compared to her peers who took only between 30 – 35 seconds to read the monologue. She misread words such as it wasn’t became I wasn’t, and she also had difficulty in pronouncing some words such as concentration, immediately, wire, secured, came, business, disastrous, crucial. All this affected other component of fluency, that is, prosody, which includes phrasing, intonation, pausing, and stress. Similarly, Yulita still lacked in accuracy and automaticity. She read slowly with lots of hesitation, repeated some words when she felt unsure about the way to say it, misread some words such as everything became anything, and paused too long at certain parts such as a... mistake, rush…ing. She used inappropriate chunks such as “Yes, we made a/mistake that day”, “we normally / do and/ instead/ decided”, “It turned /out that / the wire /at the other / end was not properly/ secured”. As for reading rate, it took her 59 seconds to finish the monologue. It means she had problem in recognizing and decoding words automatically, not to mention reading them smoothly with appropriate phrasing and intonation. Finding out from the first recording that the students had not yet been able to read fluently, we decided to re-teach the students about the importance of phrasing/chunking as part of reading fluency, and modeled how to read the text with appropriate chunking, intonation, and stressing. Then we asked the students to reread the monologue, and their voice was again recorded. The results of the second recording are presented in table 3 and 4. Table 3. Result of the second recording of Mon-Wed class No 1 2 3 4 5 6 7 Name Iyet Josh Michael Nabila Nazmi Ridho Sean Rate 4 4 4 4 4 3 4 Phrasing Intonation Pausing 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 4 4 4 4 Stress 4 4 4 4 4 4 4 Table 4. Result of the second recording of Tue-Thu class No 1 2 3 4 5 6 7 8 9 Name Ferina Naufal Rezka Rima Rina Rizwan Roshi Yulita Windy Rate 4 4 4 4 4 4 1 4 4 Phrasing Intonation 4 4 3 4 4 4 4 4 4 4 4 3 2 2 3 4 4 3 Pausing 4 4 4 4 4 4 3 4 4 Stress 4 4 4 4 3 4 2 4 3 These tables show that most of the students from the two classes have been able to make much improvement in all components of fluency. Rate improved as this time most of the students no longer read with hesitation, too slow, or too fast. They were able to read with intonation and stress, and make the text more meaningful. What is more important was in terms of phrasing and pausing. The correct chunking made the students sound more fluent and the monologue more meaningful. Iyet, for example, who previously had problem in pausing at certain part, finally were able to improve, not only in terms of chunking but also in terms of stress and intonation. Ridho, although still sounded in a hurry at the beginning part of the text, also could improve in terms of pausing and stress. Rina also improved a lot, despite her struggles with correct pronunciation. To some students, pronunciation problem had made them hesitant in saying certain words and pause at inappropriate place, thus hampered their fluency. The most striking change was shown by Yulita. On the first recording she read poorly, but on the second recording she made significant improvement in almost all components. Her automaticity improved as it took her 33 seconds to finish the monologue, and she was able to recognize prosodic features of fluency by using appropriate phrasing, pausing and intonation. She only stumbled over “we were/late”. The only student who seemed difficult to improve was Roshi. She struggled to use correct phrasing, and in terms of automaticity she still read too slowly (57 seconds), and pronounced most of the words with difficulty. The teacher (Aldy) acknowledged that among her classmates Roshi is a weak student, but she is eager to learn. Her difficulty in producing an automatic and smooth language might be due to some psychomotor factors. Conclusion and Suggestion Based on the findings above, it can be concluded that teachers could not just take for granted that students at intermediate levels are already able to read fluently, as there are various components of fluency that should be paid attention to and be practiced on. We as teachers also need to ensure that students are aware of the importance of producing the language smoothly and naturally, and should facilitate them to achieve the expected reading fluency because it will lead to spoken fluency. In this classroom action research, students are able to improve their fluency in reading through voice recording, as from their recorded voice students could identify the components of fluency which they were still weak at. After re-reading the text and re-recording the voice, it was found that the component of fluency that most improved was rate, phrasing, pausing, and stress. It was partly because the students had been shown the way to use the appropriate chunking and phrasing by marking parts in the text with slashes and then read them accordingly. The voice recording also helped them learn where to put the appropriate stress on certain words. While automaticity and prosody improved, it was found that the least improved component was intonation, or where to raise or lower voice to emphasize on meaning. It was found as well that pronunciation is another area that least improved. Pronunciation error made the students fail to read the text smoothly, and this hampered their fluency. Voice recording and chunking practice conducted in this action research is only one of many ways to help students produce the language smoothly and naturally. As our reflection we acknowledged that there was a limitation of this research as we did not do further action such as checking students’ comprehension to prove whether their fluency significantly related to their comprehension of the text, neither did we conduct a follow-up activity such as speaking tasks to prove that reading fluency contributes to spoken ability. In regards to students’ weakness in pronunciation, students would need a lot of repeated exposure on correct pronunciation as well as repetition drills, and this was not done in our action research. Therefore, further research on these aspects is strongly encouraged. References Burns, Anne. Doing Action Research in English Language Teaching. New York: Routledge, 2010. Explore 4 Student Book. Jakarta: Pusat Penerbitan Lia, 2006. Fluency Instruction (article online), accessed on 22 July 2013, available from http://www.reading-skills-pyramid.org/reading-fluency.htm Ford, Karen. ELLs and Reading Fluency in English. 2012, accessed 28 July 2013, available from http://www.colorincolorado.org/article/54200/ Fountas, Irene C., and Gay Su Pinnell. Teaching for Comprehending and Fluency. Heinemann, 2001, accessed 27 July 2013, available from http://books.heinemann.com/comprehending/formsResources.html Hook, Pamela E., and Sandra D. Jones "The Importance of Automaticity and Fluency for Efficient reading Comprehension." Perspective (Spring 2004): 1621. McCarthy, Michael. "Spoken Fluency Revisited." English Profile Journal 1, No. 1.(2010): 1-15. Reading Fluency (article online), accessed on 21 July 2013, available from http://www.learningrx.com/reading-fluency.htm Robertson, Kristina. Reading 101 for English Language Learners. 2009, accessed 22 July 2013, available from http://www.colorincolorado.org/article/33830/ Thornbury, Scott. How to Teach Speaking. Pearson Longman, 2005 Yamashita, Junko, and Shingo Ichikawa. "Examining reading fluency in a foreign language: Effects of text segmentation on L2 readers." Reading in a Foreign Language 22, No. 2 (October 2010): 263-283. Appendix 1: Monologue to be read by students Yes, we made a mistake that day. It wasn’t because I lost my concentration or anything like that. On that day we were late and people had been waiting to see the show. We didn’t check everything twice the way we normally do and instead decided to start immediately. It turned out that the wire at the other end was not properly secured and it came loose. But we learned our lesson. In this business rushing things can be disastrous; being careful is crucial. (84 words) (Source: Explore 4 Lesson 3 p. 19) Appendix 2: NAEP Oral Reading Fluency Scale NAEP Oral Reading Fluency Scale Fluent Reads primarily in larger, meaningful phrase groups. Although some regressions, repetitions, and deviations from text may be Level present, these do not appear to detract from the overall structure 4 of the story. Preservation of the author's syntax is consistent. Some or most of the story is read with expressive interpretation. Level 3 Reads primarily in three- or four-word phrase groups. Some small groupings may be present. However, the majority of phrasing seems appropriate and preserves the syntax of the author. Little or no expressive interpretation is present. Level 2 Reads primarily in two-word phrases with some three- or fourword groupings. Some word-by-word reading may be present. Word groupings may seem awkward and unrelated to larger context of sentence or passage. Level 1 Reads primarily word-by-word. Occasional two-word or threeword phrases may occur – but these are infrequent and/or they do not preserve meaningful syntax. Nonfluent SOURCE: U.D. Department of Education, Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress (NAEP), 2002 Oral Reading Study Appendix 3: Scale for assessing fluency by Fountas & Pinnel