L1-Assissted Reciprocal Teaching for ESOL students to improve their Comprehension of English Expository Text Irene Y.Y. Fung, Ian A. G. Wilkinson & Dennis W. Moore The University of Auckland Paper presented at the Joint Conference of Australian Association for Research in Education and New Zealand Association for Research in Education, Melbourne, 1999 Please address all correspondence to the first author: i.fung@auckland.ac.nz
ABSTRACT This study investigated the effects of an L1-assisted reciprocal teaching procedure on ESOL students' comprehension of English expository text. The viability of the procedure was tested in a pilot study using a within-subject ABCD design, and confirmed in a follow-up study using a multiple-baseline design across three schools. In the follow-up study, after a 5-day baseline assessment, 12 Year 7 and Year 8 Taiwanese ESOL students were given 15 to 20 days of L1-assisted reciprocal teaching. This intervention comprised both L1 (Mandarin) reciprocal teaching dialogue while reading Chinese (Mandarin) text and L2 (English) reciprocal teaching dialogue while reading English text. The Mandarin and English dialogues took place on alternate days. On each day, before the reciprocal teaching dialogue, there was a 15-minute session of teacher-directed explicit strategy instruction where students were informed of why a strategy was useful, and how and where to apply it. Results showed that students made gains on both standardised and experimenter-developed tests of reading comprehension. These gains maintained 3 to 4 weeks after the intervention as indicated by results from three follow-up probes. Moreover, students were able to transfer their comprehension fostering and monitoring strategies to novel tasks as indicated by their abilities to recall, and detect logical inconsistencies in, expository text. Treatment effects were also revealed in results from a strategy interview and think-aloud task. The success of this study suggests that students with limited English proficiency can improve their English comprehension through reading strategy instruction that capitalises on students' L1 language proficiency and literacy skills. Introduction To address the learning difficulties of non-english proficient ESOL students has been one of the key issues in many educational settings, this is especially the case in English-only mainstream classrooms in New Zealand. According to the 1990 survey of (IEA) International Association for the Evaluation of Educational Achievement, New Zealand has the largest difference in achievement between native English speaking students and non native English speaking students. This suggests that the learning needs of non-english-speakingbackground students have not been well catered for in this country. One of the important learning needs of ESOL students is to be able to read age-level content area materials for academic learning even though their English language proficiency is limited. However, in most remedial programs for ESOL students, delaying reading comprehension instruction until students are fluent in oral English is a common practice. ESOL programs, by and large, focus on improving English language knowledge and competence. At most, reading and reading instruction in ESOL class would focus on developing students' decoding skills or knowledge of syntax or vocabulary for literal comprehension only. Few would focus on teaching higher level reading comprehension strategies to facilitate students' cognitive/academic learning. Underlying this limited pedagogical practise is the assumption that reading depends primarily on language and, therefore, only when language proficiency is improved will reading comprehension improve. However, findings of bilingual education research suggest that it normally takes two to three years to become proficient in basic communication skills in English and four to ten years to approach grade level competence in English academic skills. New immigrant students who need to continue their secondary education may not have enough time to wait for their
English language maturity before they can read to learn. After all, even students whose first language is English can still have reading comprehension problems if they cannot read strategically. Thus improving English language proficiency alone may not ensure automatic improvement in reading comprehension for academic learning. The purpose of the present study was to investigate the effectiveness of an L1-assisted reciprocal teaching procedure in improving limited English proficient students' comprehension of English expository text. In particular, this study sought to explore the qualitative changes, if any, in students' declarative, procedural and conditional knowledge of reading comprehension following the reciprocal teaching intervention. There is substantial evidence that comprehension strategy instruction can help Englishspeaking poor readers improve their reading comprehension. Research findings show that reciprocal teaching, in particular, provides a feasible teaching method for teaching cognitive and metacogitive strategies for reading comprehension to poor readers even before they are fully able to decode. As limited English proficient ESOL students have been shown to have similar reading difficulties to those experienced by young and slow learners whose L1 is English, it is possible that they may also benefit from reading strategy instruction. Reciprocal teaching, developed by Palincsar and Brown, is an instructional procedure designed to improve reading comprehension in students who are able to decode, but have difficulty in comprehending, age-level reading text. The teaching procedure focuses on the dialogic teaching (i.e. teaching through dialogue between teacher/leader and students) and guided practising of four reading comprehension fostering and monitoring strategies: questioning, summarising, clarifying, and predicting. Working together with a small group of students reading a common text, the teacher thinks aloud and models the use of the four strategies so that the original covert comprehension fostering and monitoring processes are made visible to students. Initially, the teacher takes major responsibility for leading the discussion among the group. Then the students take turns to take over the teacher's role to lead the discussion for a segment of the text. The teacher supports students' participation by prompting, praising, altering the demand on the students, or providing extra scaffolding when necessary so as to ensure students' successful participation. As we can see reciprocal teaching requires students to be able to cope with the concurrent cognitive demands of high-level English language processing and high-level strategic thinking for reading comprehension. Not many ESOL students can handle these competing demands. This may be one explanation for the inconclusive findings from studies using reciprocal teaching with ESOL students. Researchers may here have overlooked the possibility that ESOL students can capitalize on the resources of their first language knowledge and literacy experience during their second language literacy acquisition. The idea of capitalizing on students' L1 language proficiency and experience in the course of second language acquisition has theoretical support. According to Vygotsky the foreign (or second) language acquisition process does not repeat the course of the first language acquisition, but is an analogous system that develops in a reverse direction. Each system complements the other and the two languages interact to the advantage of each. Success in learning a foreign language is contingent on a certain degree of maturity in the native language. The child can transfer to the new language the system of meanings he or she already possesses. The corollary of Vygotsky's argument is that the ability of meaning construction during the reading process is also transferable across languages, and the development of L1 and L2 reading abilities are complementary. Goodman proposes that " the reading process is fundamentally the same in all languages except for minor degrees of differences" (p. 26), and L2 readers compensate for less well
developed L2 skills by means of their L1 reading skills. This theory also suggests that L1 and L2 reading ability complements each other; and that some reading skills and strategies are transferable. Similarly, Cummins proposes that there is a 'common underlying proficiency' across languages, so prior knowledge and experience acquired from L1-literacy activities can lay the foundations for L2-literacy acquisition. Cummins' interdependence hypothesis assumes that, although the surface aspects of different languages are clearly separate, there is common underlying cognitive/academic proficiency that facilitates the transfer of cognitive/academic, literacy-related skills across languages. Although it has been theorized that L1 language proficiency and experience has an important role to play in second language literacy acquisition, it is necessary to identify what specific reading skills and strategies are common across language and hence transferable, as well as how the transfer can be enhanced if an effective reading instruction program is to be designed. Recent developments in reading models suggest that there are four distinctive and interdependent component processes operating in parallel during the reading process. These component processes are: decoding, literal comprehension, inferential comprehension, and comprehension monitoring. Gagne et al. define these four component processes as follows: Decoding involves using the printed word to activate word meanings in memory, either through a direct association of the printed word and its meaning or through the intermediate step of representing letter-sound correspondences. Literal comprehension involves putting activated word meanings together to form propositions. Inferential comprehension involves going beyond the idea explicitly stated to integrate, summarise, and elaborate on these ideas. Comprehension monitoring involves setting a reading goal, checking to see if the goal is being reached, and implementing remedial strategies when one's goal is not being reached. According to Gagne et al (1993) the underlying expertise required for successful comprehension involves three elements: conceptual understanding, automated basic skills, and strategies required for accomplishing each of the four component processes of reading comprehension. Conceptual understanding includes knowledge about letters, phonemes, morphemes, vocabulary, ideas, text schemas, and metacognition of reading processes as well. Automated basic skills include word decoding skills and the ability to construct propositions from strings of words. Strategies include varying one's approach to reading depending upon one's goal and monitoring one's comprehension. The knowledge that forms the basis of conceptual understanding is classified as declarative knowledge, whereas the basic skills and strategies are classified as procedural knowledge (Anderson, 1983). Paris, Lipson, and Wixon (1983) further divide procedural knowledge into two subcategories: cognitive procedural knowledge, and metacognitive procedural knowledge, or conditional knowledge (knowledge of where, when, why to apply the cognitive procedural knowledge). Skilled reading, therefore, requires a highly complex capability involving extensive declarative, procedural, and conditional knowledge to carry out all four component processes for successful reading comprehension. Although problems can occur in any or more of the component processes, Paris et al. (1983) highlight the importance of conditional knowledge in reading comprehension. If ESOL students are skilled L1 readers, theoretically, they can compensate for their less developed L2 reading by means of their L1 language proficiency and reading skills for meaning construction through these four component processes. In decoding, though the activation of meaning in memory can be achieved through a direct association of the printed word and its meaning, it is also possible through access to its L1 translation equivalent by consulting a bilingual dictionary or a bilingual adult. In literal comprehension, though ESOL students may not have the English language proficiency to put activated word meaning
together to form propositions in English, they can form propositions in their L1. In inferential comprehension, ESOL students may not be able to integrate, summarize, or elaborate on ideas in English, but they can do these and construct meaning in their L1. In comprehension monitoring, ESOL students may switch to their L1 or the stronger language in their inner speech as they reflect on their ongoing comprehension fostering and monitoring process. If ESOL students lack, or fail to activate, their metacognitive knowledge and skills to facilitate their reading skills transfer, explicit reading strategy instruction may help. In the present study, it was hypothesized that reciprocal teaching would be feasible with ESOL students with limited English proficiency if it was modified to incorporate both L1- reciprocal teaching and L2-reciprocal teaching. The former would assist students in their first language reading comprehension, and the latter would encourage transfer to promote students' English reading comprehension. Hence, the comprehension fostering and monitoring strategies would be internalized during L1-reciprocal teaching without linguistic barriers, and the cognitive demands during L2-reciprocal teaching would be lessened as students apply the strategies to L2 reading. Method The viability of L1-assisted reciprocal teaching was first tested in a pilot study using a withinsubject ABCD design. The effectiveness of the procedure was more systematically investigated in the main study using a single-subject AB design with a multiple-baseline across three schools. Twelve Taiwanese students from ESL classes in three intermediate schools in Auckland, New Zealand participated in the main study. Following a 5-day baseline assessment, these 12 students received 15 to 20 days of L1-assisted reciprocal teaching intervention. Table 1. The L1-assisted reciprocal teaching intervention program in the main study L1-assisted reciprocal teaching L1-explicit strategy instruction (15') L1-reciprocal dialogue while reading Chinese expository text (20') English reading comprehension test (25') ***************************************************************** L2-explicit strategy instruction (15') L2-reciprocal dialogue while reading English expository text (20') English reading comprehension test (25') The intervention (see Table 1) comprised both Mandarin (students' L1) reciprocal teaching dialogue while reading Chinese expository text and English (students' L2) reciprocal teaching dialogue while reading English expository text. Mandarin and English reciprocal dialogues took place on alternate days. On each day, prior to the reciprocal dialogue, there was a 15-minute session of teacher-directed explicit strategy instruction where students were informed of why a specific strategy was useful, and how and where to apply it. The
medium of instruction was the same as the one used in the reciprocal dialogue on that day. A short-answer English reading comprehension test was then administered immediately after each reciprocal teaching session. In the main study multiple sources of evidence were collected (see Table 2). To assess gains in students' reading comprehension of English expository text, researcher-developed daily comprehension tests were used. In addition, a standardized measure of reading accuracy and comprehension, the Neale Analysis of Reading Ability (Neale, 1988), was given as a pretest and posttest. Maintenance of reading gains was measured by comparing daily comprehension test results across baseline, intervention, and follow-up phases. To ascertain students' ability to transfer their newly acquired reading comprehension fostering and monitoring strategies to a novel task, a transfer test was administered in a preand post-test format. The transfer test required students to give a free recall after reading each of three English expository texts and to detect any logical inconsistencies in the texts. In order to explore any qualitative changes in students' declarative, procedural and conditional knowledge of reading comprehension following the intervention, a strategy interview and a think-aloud task were conducted prior to and following the intervention. The individual strategy interviews were used to gauge students' metacognitive knowledge in terms of person, task, and strategy variables involved in both Chinese reading and English reading. The think-aloud tasks were used to assess if there were any changes in students' procedural and conditional knowledge of reading comprehension within and across reading languages following the intervention. This assessment required students to do a think-aloud task twice, first while reading a Chinese passage, and second while reading an English passage. The strategy interview, the think-aloud task and the transfer test were all conducted in students' first language (even when they were asked to read English text) because the use of students' stronger language to discuss text written in English should reveal a more complete picture of students' reading comprehension. Table 2. Data Collection Instrument Variables Format 1. The Neale Analysis of Reading Ability Reading comprehension of English narrative text Pretest-posttest Researcher developed comprehension tests Reading comprehension of English expository text Daily assessment 3. Strategy Interview Declarative metacognitive knowledge of reading comprehension in terms of person, task, and strategy variables involved in both L1 and L2 reading 4. Think-aloud tasks Procedural and conditional knowledge of reading comprehension involved in both L1 and L2 reading Individual pretest-posttest Individual pretest-posttest
5. Transfer test Ability to transfer reading comprehension monitoring and fostering strategies to a novel task Individual pretest-posttest Results and Discussion The results of this study indicate that using the 'L1-assisted reciprocal teaching' procedure with Mandarin-speaking ESL students to improve their reading competence was successful. There are three sources of evidence to support this indication. First, quantitative data reveal treatment effects on students' L2 reading comprehension. On the daily reading comprehension measure (see Figure 1), all 12 students from three schools demonstrated improvement. Follow-up probes show that the improved level of performance was maintained three to four weeks after the intervention. On the standardized measure, the Neale Analysis of Reading Ability, there was a statistically significant mean increase of 12 months in both the comprehension (see Table 3) and accuracy reading ages (see Table 4) of 11 students (one student left the school after the holidays and did not take the posttest). Figure 1. Mean daily comprehension performance across baseline, intervention, and followup phases: Main Study
Table 3. Standardised comprehension on Neale Analysis of Reading Ability test scores at pretest and posttest: Main Study Pretest Posttest Student Reading Age Raw Scores Reading Age Raw Scores School 1 1 6.0 3 7.7 10 2 6.3 4 6.5 5 3 6.3 4 7.1 8 4 6.3 4 6.2 4 School 2 5 7.10 11 9.7 19 6 8.1 12 9.10 20 7 8.6 14 10.3 22 8 6.3 4 7.1 8 School 3 9 9.5 18 10.9 24 10 7.10 11 8.3 13 11 8.9 15 9.10 20 12 8.1 12 - - Mean 9.09 13.91 SD 5.43 7.29
Table 4. Standardised accuracy test scores on Neale Analysis of Reading Ability at pretest and posttest: Main Study Pretest Posttest Student Reading Age Raw Scores Reading Age Raw Scores School 1 1 7.2 25 8.6 40 2 7.3 26 8.9 43 3 6.5 17 7.0 24 4 7.1 24 8.2 37 School 2 5 8.8 41 10.4 60 6 9.10 54 9.11 56 7 8.10 43 11.0 68 8 7.1 24 7.1 25 School 3 9 10.6 61 11.7 74 10 9.1 46 10.3 59 11 10.8 63 11.0 67 12 8.1 45 - - Mean 38.55 50.27 SD 16.28 17.39
Second, qualitative data indicate treatment effects on students' declarative, procedural and conditional knowledge involved in the reading comprehension process. Data from the strategy interviews show that there was a qualitative change in students' conceptualization of their L1 and L2 reading processes. After intervention, students saw reading as more a meaning-making process; they saw themselves having a more active role to play in the process of reading; and they were more conversant with strategies that might help them read better. Data from the think-aloud task show that there was also a qualitative change in students' procedural and conditional knowledge involved in their L1 and L2 reading comprehension (see Table 5). At posttest, the think-aloud protocols indicated that students' comprehension processes were more successfully accomplished (see Table 6). They demonstrated a wider repertoire of strategies for inferential comprehension (see Table 7) in both their L1 and L2 readings, and they could recall more idea units in the text. Table 5. Mean percent of text segments in which strategies were used in the think-alouds during L1 and L2 reading at pretest and posttest: Main Study L1 Reading (N = 11) L2 Reading (N = 11) Pretest Posttest Pretest Posttest Mean SD Mean SD Mean SD Mean SD Strategies Decoding 0.00 0.00 0.00 0.00 73.55 26.80 64.36 35.30 Literal 98.00 4.45 99.00 3.31 97.82 3.74 96.45 9.42 Inferential 37.18 20.14 59.64 18.32 19.55 10.96 62.90 20.94 Comprehension Monitoring 38.27 24.67 40.36 20.85 81.18 19.82 82.64 16.72 Table 6. Mean percent of idea units recalled from text in the think-alouds at pretest and posttest: Main Study (%) Assessment L1 Reading (N = 11) L2 Reading (N = 11) Pretest Posttest Pretest Posttest Mean SD Mean SD Mean SD Mean SD Prior knowledge 4.45 3.11 1.91 1.92 8.36 4.54 7.91 7.13 Think-aloud 57.29 8.34 56.64 11.96 59.27 24.05 74.73 9.82
Text-retelling 28.55 15.10 37.36 15.57 26.91 11.18 44.73 15.93 Table 7. Frequencies of inferential strategies used in the think-aloud protocols at pretest and posttest: Main Study Strategies for Inferential Comprehension L1 Reading (N = 11 x 9) L2 Reading (N = 11 x 13) Pretest Posttest Pretest Posttest Previewing text & illustrations 0 0 0 1 Drawing inferences 22 33 16 66 Questioning 12 28 6 32 Using structural cues 0 3 0 0 Clarifying 0 4 0 14 Summarizing 0 2 0 4 Predicting 1 4 0 7 Visualizing 2 7 8 11 Total 37 81 30 135 Note: Frequencies are expressed in terms of total number of times a strategy was noted in the think-aloud protocols. There were 9 units of analysis in the L1 text and 13 units of analysis in the L2 text at pretest and posttest. Table 8. Mean percent of idea units recalled on the free recall measure, and mean percent correct on the comment measure of the transfer test at pretest and posttest: Main Study Assessment Pretest (N =11) Posttest (N = 11) Mean (%) SD Mean (%) SD Free Recall 41.18 8.83 60.55 5.99 Comment 53.09 25.21 78.73 16.77
Note. The comment assessment indicated students' ability to detect logical inconsistencies in three English passages. Third, results of the transfer test (see Table 8) indicate that students were able to transfer their newly acquired comprehension-fostering and comprehension-monitoring strategies to novel tasks as demonstrated by their ability to recall more information from, and to detect logical inconsistencies in, expository text. Students' critical comments made at pretest usually were about the choice of words, the punctuation, the length of the text, or the amount of information presented in the text. However, at posttest, students' comments focused more on the logical structure of the text, and the majority of them were able to suggest sensible solutions to the contradictions that existed in the problem passages. Over all, students at posttest demonstrated possession of cognitive and metacognitive knowledge that made possible their successful performance. Conclusion Findings from the present study indicate that L1-assisted reciprocal teaching may be a effective procedure to improve ESL students' English reading comprehension. It was hypothesized that the reading processes in L1 and L2 are similar, so explicit instruction of the same set of comprehension fostering and monitoring strategies for reading comprehension in both L1 and L2 would enhance students' transfer of reading strategies across languages. It was also hypothesized that explicit strategy instruction in the students' stronger language would help students to internalize the reading strategies more efficiently. Results of the present study support these hypotheses. Instruction in higher-level cognitive and metacognitive strategies to ESOL students who have difficulties in reading English expository text appears to be feasible. Students' conceptualization of these strategies is shown to be enhanced by capitalizing on students' L1 language proficiency and their L1-literacy knowledge and experience. Therefore, consistent with the theoretical perspectives held by Vygotsky (1962), Goodman (1973) and Cummins (1979), ESOL students' L1-language proficiency and L1-literacy experience are valuable resource that can be used to assist cognitive/academic learning while they are developing their second language literacy. References Anderson, J. R. (1983). The architecture of cognition. Cambridge, Mass.: Harvard University Press.