UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Making meaning : using syntax as a tool for reading comprehension Chretien, Dianne Dorothy 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0206.pdf [ 3.2MB ]
Metadata
JSON: 831-1.0078231.json
JSON-LD: 831-1.0078231-ld.json
RDF/XML (Pretty): 831-1.0078231-rdf.xml
RDF/JSON: 831-1.0078231-rdf.json
Turtle: 831-1.0078231-turtle.txt
N-Triples: 831-1.0078231-rdf-ntriples.txt
Original Record: 831-1.0078231-source.json
Full Text
831-1.0078231-fulltext.txt
Citation
831-1.0078231.ris

Full Text

MAKING MEANING: USING SYNTAX AS A TOOL FOR READING COMPREHENSION by DIANNE DOROTHY CHRETIEN Bachelor of  Arts, University of  British Columbia 1987 Bachelor of  Education, University of  British Columbia 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES LANGUAGE AND LITERACY EDUCATION DEPARTMENT THE FACULTY OF EDUCATION in The University of  British Columbia 2003 We accept this thesis as conforming to the required standard THE MlVERSITY OF BRITISH COLUMBIA April, 2003 © Dianne Dorothy Chretien Wednesday, April 16, 2003 UBC Rare Books and Special Collections - Thesis Authorisation Form In presenting this thesis in partial fulfilment of -the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date http://www.library.ubc.ca/spcoll/thesauth.html ABSTRACT In summary, this study was designed to investigate whether teaching students to approach scientific  texts using their knowledge of  language (specifically  syntax) would help them better understand the passages they are reading. It has long been taken for  granted that the study of  grammar is unhelpful  in teaching students to write. Much of  the research in the past decades has been used to provide evidence that the study of  the structure of  our language restrains the generative, creative process. The reading process, however, must be admitted to be a more analytic process. Constructive, yes, in that the reader does more than merely receive meaning from  the text; rather, he or she constructs meaning from  it. But the text is a whole which must be first broken down into parts in order for  the reader to reconstruct  the frame.  Simultaneously, the reader is adding to the text frame  from  his or her own store of  knowledge, and the final product of  the act of  reading will be the construction of  something new and unique to that reader. The theoretical framework  upon which the study is based is van Dijk's and Kintsch's concept of  discourse structure which represents reading as a process involving the building of both a text base, the aforesaid  "frame"  and a situation model, the new and unique product of the reader. William Adler's theoretical framework  upon which he based his book How  to Read  a Book  also serves as a point of  reference.  Adler describes reading as an active process that requires taking apart an argument proposition by proposition (a process which he says requires a certain amount of  grammatical knowledge) in order to fully  comprehend it and reconstruct it. Such a complex analytical approach is usually unnecessary when reading fictional narrative, but the reading of  content area text presents different  and greater challenges to the young student, and, of  all content areas, science can present the most complex concepts. Processing science concepts becomes even more difficult  when the concepts are counterintuitive, as in the study of  forces  in physics. The questions addressed in this study were the following: 1. Is direct teaching of  syntax effective  in raising syntactic awareness in young students? 2. Can young students be taught to apply syntactic knowledge during the reading process? 3. Will an increased awareness of  syntax and application of  syntactic knowledge be accompanied by an increase in reading comprehension of  science text? In the course of  the study the experimental group received a pretest and a posttest testing syntactic knowledge and reading comprehension and a seven week intervention which included whole class lessons teaching students to recognize sentence structure, chunk sentences and use connectives as cues for  relationships between propositions. The intervention also included guided reading sessions applying the syntactic concepts to reading science text. A control group received only the pretest and posttest. Results indicate that the experimental group improved in syntactic knowledge and maintained their scores comprehension, while the control improved slightly in syntactic knowledge and, on the whole, did not improve in comprehension. Furthermore, observation notes from  the intervention indicate that the intervention allowed students to gain a sense not only that they needed to adjust their pace of  reading while reading dense, science text, but a notion of  how and where to do so, using commas, connectives and their knowledge of phrases and clauses as points at which to stop and think. What follows  in this chapter is an interpretation of  the findings  and the conclusions that can be drawn in light of  the research questions. The quantitative data composed primarily of  the pretest and posttest results and the qualitative data in the form  of  the observation notes on the intervention will be discussed separately. CONTENTS ABSTRACT ii LIST OF TABLES ACKNOWLEDGMENTS  vnx CHAPTER I: INTRODUCTION 1 The Question 1 Overview of  Preliminary Study: 6 Measurements 7 Results 8 Meta-cognition 9 Discussion of  Grammar (10 questions) . 10 Sentence Boundaries 10 Subject and Predicate 11 Chunking 12 Independent and Dependent Clauses 13 A Comparison of  All Variables . . . 15 Analysis of  Results : 15 Discussion 20 Implications for  the Thesis Study 21 CHAPTER II: LITERATURE REVIEW 22 Grammar Instruction 22 Going Beyond the Sentence 23 Discourse Theory: The "structure" of  discourse 24 Discourse Theory: The process of  comprehending 25 Establishing Relations between Chunks 28 Understanding relations—inferencing:  text base or situation model? 33 The Discourse Structure, the Domain of  Science and the Young Reader 38 CHAPTER III: METHODS AND PROCEEDURES . 41 The Study 41 Participants 41 Baseline Data 42 Pretests and Posttests 44 The Intervention 45 Procedure 46 CHAPTER IV: RESULTS 48 Baseline Data—Low Group 48 Baseline Data—Characteristics of  Medium and High Groups 50 Pretest and Posttest Results: Comprehension 53 Syntactic Knowledge 53 Chunking 57 Reasons for  Chunking 59 Connectives within Sentences 60 The Experimental Medium Group 63 Results of  the Low Group • 65 CHAPTER V: THE INTERVENTION 69 Qualitative Data Gathered During the Intervention 72 CHAPTER VI: DISCUSSION AND CONCLUSIONS 80 Question One 82 Chunking (Refer  to table 15) 82 Reasoning (Refer  to table 16) 83 Connectives (Refer  to table 17) 88 Total Syntax (Refer  to table 18) 90 Low Group Data 91 Question Two 92 Question Three 93 Conclusions 95 Limitations 96 Recommendations • 96 APPENDIX A 99 Reading Test One 99 Reading Test Two 101 APPENDIX B: THESIS STUDY: PRETEST 104 Reading Passage ' 104 THESIS STUDY: POSTTEST 107 Reading Passage 107 Worksheets—in order of  presentation to students Worksheet One—Chunking Sentences • • • 109 Worksheet Two—Chunking Sentences in a Paragraph 110 Worksheet Three—Chunking Sentences in a Paragraph 110 REFERENCES = U2 LIST OF TABLES Table 1: Reading  Comprehension  8 Table 2: Meta-cognition  9 Table 3: Sentence  Boundaries  10 Table 4: Subject  /  Predicate  11 Table 5: Chunking  12 Table 6: Reasoning re Chunking  • 12 Table 7: Dependent  and.  Independent  Clauses  13 Table 8: Reasoning re Clauses  13 Table 9: Total  points for  Grammar Discussion 14 Table 10: Comparison Chart  . 15 Table 11: Low group baseline data  48 Table 12: Medium  & High  baseline data  50 Table 13: CELF  and  Stanford  scores 51 Table 14: Comprehension—Medium  & High  Groups 54 Table 15: Chunking  for  meaning 58 Table 16: Reasons for  Chunking  59 Table 17: Connectives  within Sentences  61 Table 18: Total  Syntax  62 Table 19: Summary  63 Table 20: Chunking  Sentences—Low  Group 65 Table 21: Reasons for  Chunking—Low  Group 65 Table 22: Connectives  within Sentences—Low  Group 66 Table 23: Total  Syntax—Low  Group 66 Table 24: Comprehension—Low Group 67 Table 25: Summary—Low  Group 67 Table 26: Intervention  schedule  69 ACKNOWLEDGMENTS There are many people who have supported me in this project and without whose help I would not have been able to complete it. I am grateful  to: Dr. Elizabeth Lee, my advisor whose encouragement spurred me on and whose insight inspired me, Dr. Lee Gunderson who served on my committee, Dr. James Anderson who served on my committee, Ms. Wendy MacKinlay, my administrator, who believes thinking and researching about teaching is important and who encouraged me to take the time to do so, My husband who "always knew" I would get my master's, My sons who were quiet (sometimes) when Mommy had to study, And my class who cheerfully  subjected themselves to testing and instruction and who were a real pleasure to be with. CHAPTER I INTRODUCTION The Question What role does grammatical knowledge play in reading comprehension? The following  preliminary study and subsequent thesis study grew out of  this question. The question arose during a graduate seminar discussion of  Rumelhart's interactive model of reading. At the heart of  his model lies a three dimensional message center which processes text by simultaneously testing a number of  different  hypotheses at various levels: the feature level, the letter level, the letter-cluster level, the lexical level, the syntactic level and the semantic level. Once a certain competency is reached, these hypotheses arise automatically as the reader's eyes fall  upon the features,  letters and words within the context of  the text. Reading instruction targets some of  these levels directly. Children receive direct instruction in how to work with the feature  level, the letter level, and the letter cluster level. Much effort  is expended to ensure that children learn to recognize and reproduce the letters and their features.  They are taught to recognize common letter clusters encountered in our language. Direct teaching on the semantic level is a priority. They learn that words have meaning and that reading is all about making meaning. But educators often  do very little, direct teaching of  syntax, particularly in the context of  reading instruction. This level, though foundational  to the construction of  the text, is given cursory treatment. Children are taught that sentences are representations of  whole thoughts and begin with capitals and end with periods. They are told that when there is a comma, readers are to take a breath. Beyond this the subject of  syntax is rarely broached even as the sentences students are required to read become more and more syntactically complex. Children embark on their school careers already possessing a good deal of  linguistic knowledge (Karmiloff-Smith,  1986). This knowledge has been acquired as they have learned how to speak. Children bring their knowledge of  their spoken language to their interaction with written language. Phonemic awareness is one type of  linguistic knowledge children possess: Children have a highly developed knowledge of  phonemes long before learning to read; if  they did not, they could neither produce nor understand oral language. But, again, this is working knowledge, not conscious knowledge. It is deeply embedded in the subattentional switches, sensors, and gears of  their oral language machinery. (Adams, 1990, p. 303) Building on phonemic knowledge, educators teach children to focus  their attention on words in an effort  to "break the code". They teach children to chunk words, separating them into onset and rime. They teach children to find  bits of  sound they can recognize and chunks of meaning such as "script" in the words "description", "postscript" and "scripture". They help students understand the roles of  prefixes  and suffixes. What is true of  phonemic knowledge is also true of  syntactic knowledge. It is deeply embedded. It does not follow  that because children possess such working knowledge they will be able to apply it to the task of  reading. And, just as words follow  a code that must be "broken", syntax is the code of  the sentence, a code fundamentally  similar to that governing the composition of  words. The engineering trick behind human language—its being a discrete combinatorial system—is used in at least two different  places: sentences and phrases are built out of  words by the rules of  syntax, and the words themselves are built out of  smaller bits by another set of  rules, the rules of "morphology". (Pinker, 1996, p. 127) Yet inexplicably, though direct instruction on the word level is considered helpful,  direct instruction on the syntactical level is not. The term "breaking the code" is fitting  because reading is an analytic process and the first  step of  analysis is breaking the whole down into parts. Adler (1972) assigns grammatical, or syntactical knowledge a principal role in this analysis: You must discover the proposition or propositions that each of  these sentences contains. This is just another way of  saying you must know what the sentence means. You discover terms by discovering what a word means in a given usage. You discover propositions similarly by interpreting all the words that make up the sentence, and especially its principal words. Once more, you cannot do this very well unless you know a little grammar. You must know the role that adjectives and adverbs play, how verbs function  in relation to nouns, how modifying  words and clauses restrict or amplify  the meaning of  the words they modify,  and so forth.  Ideally, you should be able to dissect a sentence according to the rules of  syntax, although you do not necessarily have to do it in a formal  way. (Adler, 1972, p. 125) Adler's book, How  to Read  a Book,  presents the concept of  active reading, taking an argument apart and then putting it back together as a means of  comprehending it. This process has been described as constructing meaning, a current popular metaphor for  the reading process. For construction, it is necessary both to be familiar  with one's tools and one's building materials, or the structure of  language and the "rules" which govern it, and to know how one might best use those tools. Surely in the teaching of  reading, it would be helpful  to make children's intrinsic knowledge of  syntax extrinsic and thus applicable to the decoding of  complex sentences and larger text. Adams (1990) suggests: ...development of  syntactic competence may be far  more important than is generally recognized in reading instruction. Without the independent ability to recognize the within-sentence phrasal and clausal boundaries that permit interpretive recoding, the young reader has no rational option but to try to conquer the whole sentence at a time. (p. 415) The problem so many students encounter is getting lost in text. This is particularly true of  young children who have only just mastered decoding and will often  forget  what the beginning of  the sentence says by the time they reach its end. It is also true of  more experienced readers who may be able to stick with a simple sentence to the end, but when asked to read a complex sentence composed of  several clauses and phrases, they lose the thread of  the discourse. But if  students were taught how these sentences were put together— what the parts were and how to consider each one singularly and then in relation to the others—how could this not be of  aid to them? Considering the parallels between the word level and the syntax level, the initial question takes a more definite  form:  Following the pattern of  chunking words for  sound and meaning, can chunking sentences for  sound and meaning enhance reading comprehension? The necessity of  this ability to chunk sentences is most evident in the content areas, where students not only have to deal with complex sentences, but also the complex concepts described within these sentences. Science text can be particularly problematic. As a young English major in college, attempting to absorb the concepts of  various biological functions,  I fell  back on my knowledge of  the language to trace the arguments presented in the ponderous text. I wrestled with the chapter until I had a sufficient  outline of the text into which I might integrate the new knowledge which would be presented to me in lectures and upon which I would have to act in the laboratory that week. With my knowledge of  language, I had the means of  constructing a logical framework  upon which I might hang the concepts contained in vocabulary yet to be learned. Unfortunately,  many post secondary students are not well schooled in the structure of their language. They have not had the benefit  of  much grammar instruction, nor have they had extensive reading instruction past the first  few  years of  school. In the intermediate grades and in high school, students may have encountered science classes in which they were required to read and learn from  text, but they have seldom been taught how to do this. What is needed is a bridge between the closely guided direct instruction primary students receive to allow them to break the code of  words and the type of  instruction that presumes a student can read and understand a science text independently. This has become a focus  both in educational research and in classroom instruction. The proposed bridge in the thesis is instruction in chunking sentences, but in order to examine the efficacy  of  chunking sentences to enhance reading comprehension of  science text in young students—those in elementary school as opposed to post-secondary—it is necessary to back up a few  steps. There are questions to be considered within a preliminary study: 1. What do young readers (in this case, grade three students) need to know in order to chunk sentences? 2. What do young readers already know about chunking sentences? 3. Is there a relationship between reading comprehension and syntactic awareness? Overview of  Preliminary Study: This is an experimental study using both qualitative and quantitative measures. Due to the limitations of  a small sample size, descriptive statistics in the form  of  percentages were used. Participants were selected from  a grade three class of  24 students in a K-7 school of approximately 500 students in a middle class suburb. The class was composed of  nine girls and eleven boys and included two high incidence SLD (severely learning disabled) students (one of  whom was also labeled BD, or behaviour disordered). (N.B. a student is considered "high incidence" if  there is a relatively high percentage of  children suffering  from  a similar level of  disability within a population and "low incidence" if  there is a relatively low percentage of  children suffering  a similar level of  disability within a population. For example, a child with Down's syndrome would be considered low incidence.) The participants included three boys and three girls divided into three ability groups (high, medium and low), one of  each gender in each group: Child # 1, Child #2 - high ability Child #3, Child #4 - medium ability Child #5, Child #6 - low ability The level for  each student was determined by the Brigance reading test given in September 2001. Students read both vocabulary lists and grade leveled reading passages to determine their approximate grade level reading ability. Students were placed in guided reading groups and had been functioning  within these groups throughout the school year. The study was undertaken in March and April, 2002. Measurements Students were given two reading passages to read aloud and were asked to respond orally to questions based on the passages. (See Appendix A for  the passages and protocols.) Both the reading and the responses were taped and transcribed. Both tests were designed by the researcher. Two passages were taken from  the "Let's Read and Find Out" series : You Can't  Make  a Move  Without  Your  Muscles  by Paul Showers , readability 2.9 (Flesch-Kincaid) and High  Sounds,  Low Sounds  by Franklin M. Branley, readability: 2.6 (Flesch-Kincaid). All the participants were reading at the readability level of  the text or above. To assess basic comprehension, students were asked: "What is the passage about?" and then were asked to explain the processes of  how the muscles move in the first  test and of  how the ear hears in the second test. The participant's level of  meta-cognition was assessed through two questions: "Where was it hard to understand?" and "How did you figure  it out?" Students' ability to chunk sentences and their syntactical awareness was assessed through the following activities: identifying  beginning and ending of  sentences, identifying  subject and predicate, chunking complex and simple sentences and distinguishing between dependent and independent clauses. Several test items in both protocols were excluded from  the analysis. The protocols were designed to include a great variety of  question types as part of  the purpose of  the study was to determine what manner of  questions were comprehensible to young students and elicited useful  information  regarding syntactic awareness. Those which did not fit  the criteria were omitted. Omissions are noted on the protocols (see appendix A). Results For the comprehension questions (questions #1 and #2 on the protocols) the total number of  points possible was ten; thus for  the two tests together, the total number of  points per child is twenty. The scores for  the pairs of  children in each level group were collapsed into one score out of  fourty. Table  1: Reading  Comprehension Groups Reading Comprehension /40 Reading Comprehension % High 26 65 Medium 16 40 Low 17 43 Most participants (8 out of  12) responded to the first  question "What is the passage about?" with a single word or phrase, leading the researcher to the conclusion that the stated question would need to be revised in future  studies to a request for  a summary of  what they had read. Participants were all then asked, "Can you tell me more?" and proceeded to do so. Points were awarded for  each major idea recalled, for  relationships noted between ideas and for  clarity. The second comprehension question (questions #2 on the protocols) requested students to explain a process. In the first  test, they were asked to explain the process by which a muscle bends and, in the second test, the process by which the ear hears. One point was awarded for  each step mentioned. Students found  the first  process much more difficult  to explain. In the second test, one of  the high level group gave a response that was uncodable. The child was trying a high level strategy, giving an analogy, but failed:  "Uhm well, there— it's kind of  like a piece of  metal, uhm, that has been connected together with nails and screws and it moves and when you move the... metal and it squeaks. Or it doesn't squeak when you move your muscles, but it squeaks when you move the piece of  metal." Generally, the high group members were able to describe what they read more fully and clearly and scored higher than the other two groups, achieving 65%. The medium and low groups scored approximately the same, 40% and 43% respectively, with the low group slightly outperforming  the medium. Meta-cognition Table  2: Meta-cognition Groups Meta-cognition /16 Meta-cognit ion % High 14 88 Medium 10 63 Low 8 50 For both meta-cognition questions (question #3 and question #4), one point was awarded for  giving one word or phrase in response, two points given for  a fuller  response. For each passage, the total number of  points possible was four;  thus for  the two tests together, the total number of  points per child is eight. The scores for  the pairs of  children in each level group were collapsed into one score out of  sixteen. In response to the question, "Where was it hard to understand?" ten of  the twelve responses identified  vocabulary words they did not know. Two responses identified  an area in text where understanding broke down, and both these responses were given by high level readers. The question, "What did you do to figure  it out?" elicited word attack skills such as sounding out and chunking. Higher level respondents explained in more detail, achieving 88% and consistently outperforming  the two other groups. The medium group outperformed the low group, achieving 63% to the low group's 50%. Discussion of  Grammar (10 questions) This section included identifying  sentence boundaries, identifying  subject and predicates, chunking sentences and giving reasons for  the divisions and determining dependent and independent clauses. Sentence  Boundaries Table  3: Sentence  Boundaries Groups Sent Bound/8 percent High 8 100 Medium 8 100 Low 7 88 In establishing sentence boundaries, students were directed: "Show me the beginning of  sentence" and "Show me the end of  the sentence"(questions #7a and b in test one and questions #8a and b in test two). One point was awarded for  the correct response. Total possible points for  each child was two on each test, four  for  both. The scores were collapsed by group for  a possible total of  eight. All were able to identify  both the beginning and the end of  the sentence in question for  both tests except for  one of  the subjects in the low group who identified  the beginning of  the line for  the beginning of  the sentence in test two. There was very little variance between ability levels regarding sentence boundaries. Subject  and  Predicate Table  4: Subject  /  Predicate Groups Subject / Predicate /16 % High 9 56 Medium 10 63 Low 10 63 For the Subject / Predicate test items, three sentences were isolated from  the passages, two in test one (questions #14 a and b) and one in test two (question #15). In each sentence, students were asked to identify  the subject. Two points were awarded if  the participant correctly identified  the subject of  the sentence and one point was given if  the participant interpreted "subject" in the general use of  the term, that is as the topic of  the sentence, and correctly identified  that. This led to six total possible points per child, two more possible points were awarded for  identifying  predicates, for  a possible total of  eight per child, collapsed to sixteen per group. The sentences were progressively more difficult.  The last one was a long sentence containing many nouns from  which to choose. In one sentence in test two, they were asked to identify  the predicate also (question #16). This sentence contained two parallel predicates. Two points were awarded for  identifying  both, one was awarded for  identifying  one. Only child #1 identified  both. This exercise proved difficult  across the ability levels with very little difference  in achievement between the groups. Chunking The participants were shown four  sentences within the body of  the text and were asked to chunk them (test one #8 a and #11 a and test two #10 a and #14 a). They were then asked to give their reasons for  choosing the spots they chose (test one #8 b and #11 b and test two #10 band #14 b). Table  5: Chunking Groups Chunking /10 % reasons /28 % High 9 90 26 93 Medium 7 70 8 29 Low 9 90 14 50 The points for  reasoning were assigned on the basis of  the answers received, the best answers serving as the measuring stick. This led to a variable number of  possible points for  the reasons given for  the choice of  chunking spots for  each sentence. Because of  this, a more detailed table has been included. Table  6: Reasoning re Chunking Child #8b/6 #11b/4 #12/8 #10b/6 #14b/4 High 5 4 7 6 4 Medium 1 2 0 2 3 Low 4 1 3 4 2 Basic determination of  points included indication of  an awareness of  clauses, acknowledgment of  commas, and statements indicating that "it made sense" or "sounded right". Question #12 (test one) asked about relative difficulty  of  the sentences for  chunking; all agreed the complex sentence with the comma was easier to chunk than one without, but the high group members were much more capable of  explaining why, achieving 7 out of  8, or 88%, compared to the medium group's 0% and the low group's 38%. Overall, the high group proved very strong in reasoning, achieving 26/28 or 93%, clearly outperforming  the other two groups. The low group outperformed  the medium group, achieving 50% compared to the medium group's 29%. Independent  and  Dependent  Clauses Students were shown sentences in the text and asked whether certain clauses within the sentences could stand on their own. They were shown a dependent clause first,  then an independent. All identified  the first  dependent clause as not being able to stand on its own. (test one questions #9 a and b and test two questions #11 a and b.) One point was awarded for a correct answer, yes or no, with a possible total of  four  per child, eight per group. Table  7: Dependent  and  Independent  Clauses Groups dep/indep /8 % High 6 75 Medium 5 63 Low 6 75 Here the high and low groups both achieved 75%, while the medium group achieved 63%. Table  8: Reasoning re Clauses Groups reasons /20 % High 18 90' Medium 11 55 Low 5 25 There was a possible two points for  reasons given for  the independent clauses and three for  the dependent clauses yielding a possible score of  five  per sentence, ten per child, twenty for  each group of  two children. In discussing the independent clauses, two points were awarded for  the observation that it needed more information,  one was awarded for  the observation that it sounded right. The discussion of  the dependent clauses were less uniform, and thus were rated in relation to one another. The second dependent clause presented more difficulty,  largely because of  the "they". Students felt  that since the clause didn't disclose who or what the "they" was, it was not able to stand on its own. Overall, the high group substantially outperformed  both the medium and low groups in recognizing dependent and independent clauses and giving reasons for  their designations, achieving 90% to the medium group's 55% and the low group's 25%. Table  9: Total  points for  Grammar Discussion Groups Grammar Discussion /90 Grammar Discussion % High 77 86 Medium 49 54 Low 50 56 When scores for  the four  types of  questioning were totaled, it is clear that the high group outperformed  the medium and low. This becomes even clearer when one focuses  on the questions requiring reasoning. The medium and low groups are very close in scores, with the low leading by one point or two percent. There is no clear reason for  the low group's outperforming  the medium group. It may be accounted for  by one member of  the medium group proving to be the most reticent to provide any kind of  answer at all for  the questions. This student is quite timid and quiet by nature. Also, in terms of  daily classroom discussion in guided reading groups, the low and medium groups were consistently quite similar. The members of  the high group were students of  a very high calibre in reading and in reasoning. A Comparison of  All  Variables Table  10: Comparison Chart Groups Reading Comprehension /40 Reading Comprehension % Meta-cognition /16 Meta-cognition % Grammar Discussion /90 Grammar Discussion % High 26 65 14 88 77 86 Medium 16 40 10 63 49 54 Low 17 43 8 50 50 56 In summary, as shown in table 10, the high group is consistently higher than all the other groups in all three areas, reading comprehension, meta-cognition and the ability to identify  and discuss points of  grammar. There is little variance between the other two groups in terms of  reading comprehension (40% for  the medium group and 43% for  the low group) and grammar (49% for  the medium group and 50% for  the low group), but the lowgroup does slightly outperform  the medium group. In meta-cognition, the medium group achieves a score which is 13% higher than that of  the low group. Analysis of  Results In answering the comprehension questions, there was a tendency to give one word answers or phrases for  what the passage is about. The students found  it easier to give a clear explanation of  how the ear hears than of  how muscles move. This may be partly due to readability of  passage, (The ear, 2.6; The muscle text 2.9), and the fact  that hearing was a topic of  study for  science (though up to that point they had not studied the role of  the ear in receiving sound). Also, there was a much clearer diagram for  the ear passage. The students may also have found  it easier to trace the movement of  sound waves through the ear canal than to explain the adversative relationship between muscles. The generally low scores on comprehension may be partly due to the perception, expressed by one low group member, that science texts are incomprehensible: Uhm, I don't know—it says something—I forget.  Well, my dad and I were doing this science project. My dad told me to read it, and I read it all. It was like three pages. And when I was done, my dad said, "Okay, what do you do first?"  And I'm like... I read it, but I don't get anything it said. (Child #5, Test two) In considering the responses to the meta-cognition questions, it became clear that the students viewed their own understanding as based primarily on whether they were familiar with the words, for  instance "cochlea". Even when they understood what the cochlea was, the sound of  the word was difficult  and unfamiliar  and continued to throw them off.  When encountering difficulty  understanding, sounding out is still the first  acknowledged strategy, though some were beginning to apply chunking of  words. The more advanced students recognized heavier explanation as a flag  for  slowing down and attending more closely to ensure comprehension. "Where was it hard to understand what you were reading?" "I think like right here. It started like more explaining stuff  and then they get really hard words like cochlea and stuff."  (Child #2, Test Two) The students were able to establish sentence boundaries. This indicates this is explicit knowledge by grade three. They had difficulty  distinguishing between "subject" as a grammatical concept and "topic", but some of  the oral responses indicate they were beginning to understand and identify  subject and predicate. What is the subject of  this sentence? "Your biceps and triceps work together." The subject is about like your biceps and triceps work together. So your biceps and triceps, that'd be the subject and "work together", that would be the predi - uh... Predicate? Predicate. (Child # 3, Test one) The children were beginning to get comfortable  with the idea of  chunking sentences. You'd chunk it where you see the comma. That's where I'd chunk it. "Every time you move any part of  you (pause) you use your muscles. It's kind of  like a period. You stop, but it doesn't really stop the whole entire sentence, it keeps going. (Child #1) (chunked at comma).. .because it almost like ends, but not quite. Those kind of  match (the two clauses), but you just kind of  need that to make it a sentence. (Child #2) Generally, they recognized the significance  of  the comma. Some still saw it as a pause for  breathing, some were beginning to see it's role in separating parts of  a sentence (i.e. clauses). This indicates an explicit knowledge of  the use of  the comma as a signal. They obviously did not possess a full  understanding of  the role of  a comma, but this understanding is in the developing stage. Students had much more difficulty  chunking sentences without commas, for  example "You even use your muscles when you read". Most were still able to do it, but had a harder time explaining why: Uhm, you should chunk it... you should say, "You even use muscles" and then you'd put a comma there if  there was one there and then you say "when you read". (Why?) Uhm, I think because of...,  if  you cut it and you said, "You eve use" then that wouldn't make sense, but "muscles when you read", that doesn't make sense either. So, if  you just cut it to the next—you just add another letter to "you even use" it'll be "you even use muscles". (Child #1) With the sentence with 2 predicates, "The outer part of  the ear catches the vibrations and carries them into the opening of  your ear", one student recognized you could make two sentences out of  it and gave elaborate instructions on how to do so: I'd chunk it right there: "The outer part of  the ear catches the vibrations", that could be one sentence, but "and carries them into the opening of  your ear", I don't think that could be a real sentence. (What does it need?) It needs, uh... you should take away the "and" and put "they" and take away the "s" and add "y", so it would say "They carry them into the opening of  the ear". (Child #1) The students generally recognized dependent clauses as dependent—but had more difficulty  when the clause began with a "wh" question word. This may be due to their tendency to answer or hear questions answered with a dependent clause. For example "When are we going?" will often  receive the response, "When we are ready". The most prevalent reason given for  designating a clause dependent or independent was based on sufficiency  of information.  These grade three students had come to understand that a sentence is a complete thought. In examining the sentence "When vibrations hit your eardrum, they make it vibrate", the following  discussion ensued: Does it make sense if  we only read the first  half? No, because, like, it says "when vibrations hit your eardrum"... Well, it doesn't tell you what it DOES at the end. And, like, you have to know once you read it, you have to know what it does, but it doesn't tell you. Does it make sense if  we only read the second half? Yes, because it tells you what it does, like "they make it vibrate" and stuff... (Child #6) The reasons given for  designating an independent clause as being able to stand on its own were more difficult  to assess than those given for  dependent clauses. Reasoning regarding the syntax was much more complex and varied with independent clauses partly because of  the presence of  a pronoun in the sentence above (students were looking for  the antecedent). No [it can't stand on it's own] because it's not telling you a subject that will make, like "they'll make it vibrate". You don't know what'll make it vibrate. But if  you read that, it'll tell you like what there is."(Child #3). Perhaps students also had an easier time saying why a clause doesn't have enough information  rather than why it has all it needs. This may well be due to a lack of  extrinsic knowledge of  what a sentence is, a lack of  the concepts of  subject and predicate. Some believed the length of  a clause determined its ability to stand on its own as a sentence. Most were able to isolate a modifying  clause not separated by a comma. Discussion The preliminary study was designed to answer the following  questions: What do readers need to know in order to chunk sentences? What do young readers (i.e. grade three students) already know about chunking sentences? Is there is a relationship between reading comprehension and the ability to chunk sentences? The children who participated in the study had an extrinsic knowledge of  where sentences begin and end through the signposts of  capitals and periods. They had learned to pay attention to internal punctuation, such as commas, as some sort of  signpost—often  for breathing i.e., chunking for  sound. Some were also beginning to develop a sense of  chunking for  meaning, separating two parts of  a sentence or two related ideas. These children had a • sense that a clause can be a sentence when it gives you enough information.  Some were also developing a sense of  what is necessary to constitute enough information:  knowing "who", or having a subject, and knowing what is happening, or having a predicate. (Note that they had received instruction regarding this in class.) Those children who had a high reading ability also had a relatively high syntactic awareness and an ability to express it. The ability to explain and give reasons for  syntactic designations indicates a high level of  understanding which means that those concepts are more accessible for  extension and application. That is, the high performing  students had syntactic information  accessible to apply to the reading task. There was little difference  between the low and medium groups regarding either syntactic awareness or comprehension. Implications for  the Thesis Study The comma was clearly a sound starting point for  teaching chunking. It would be reasonable to assume that older students would be familiar  with this marker as a place to pause. A decision was made to focus  on complex sentences including these explicit markers and to develop the understanding of  the composition of  clauses and the use of  connectives. The thesis study was also designed to examine more closely whether or not understanding relationships between clauses can enhance reading comprehension of  science text. The intervention includes some exercises which would explore whether children are better able to determine whether clauses or phrases can stand alone if  they are presented in isolation. The thesis addresses the following  questions: 1. Is direct teaching of  syntax effective  in raising syntactic awareness in intermediate students? 2. Can intermediate students be taught to apply syntactic knowledge during the reading process? 3. Will an increased awareness of  syntax and application of  syntactic knowledge be accompanied by an increase in reading comprehension of  science text? CHAPTER II LITERATURE REVIEW In an effort  to explore the efficacy  of  making children's implicit knowledge of  syntax explicit to provide them with an analytical tool for  the reading of  science text, this study lies at the point of  intersection between several fields  of  inquiry: discourse theory, meta-cognition, reading comprehension, science instruction. It is an attempt to link theory and practice within the context of  an elementary classroom. Grammar Instruction It has already been noted that when the structure of  language is a focus  in current reading instruction, the focus  is on the word level, teaching children to chunk words for sound and meaning. Instruction at both word level and sentence level fall  under the umbrella of  'grammar', "the general term for  the study of  language" (Fowler, 1984, p. 220). Educators tend to consider study at the word level a separate area, encompassing such topics as "spelling" and "phonics", when it is really a subset of  grammar and takes many forms: phonology, morphology, accidence, orthoepy, and orthography. Study at the sentence level is a study of  syntax, or "how words are arranged in sentences" (Fowler, 1984, p. 220). Given that words and sentences are under similar structural constraints, both functioning  as discrete combinatorial systems (Pinker, 1994), this study will explore the possibility of  adopting the strategy of  chunking used in decoding words and applying it to decoding sentences. Establishing an understanding of  the linguist's definition  of  a sentence and of  how the study of  syntax has evolved over the last few  decades will provide a foundation  for  the application of  this strategy. Going Beyond the Sentence Noam Chomsky and many of  his contemporaries considered the sentence to be the longest formal  syntactic structure which could be studied by linguists (Grimes, 1995). They took sentences and broke them up into their elements, drawing phrase structure trees. They showed how sentences could be transformed  by the shifting  and embedding of  these phrases. But they did nothing to show how a sentence functions  within a text to contribute to the overall meaning of  the text, nor how a reader processes a sentence in the course of  processing an entire text. Their emphasis was on syntax at the expense of  semantics in the representation of  sentence structure. This same delineation between structure and semantics has marked the traditional grammar teaching in schools. It has been, and often  still is, limited to studying parts of  speech or establishing a cursory understanding of  the sentence as needing a subject and a predicate without providing a bridge between these concepts and their relation to the meaning communicated through their forms.  But unlike educational practice, linguistic theory has progressed beyond these limitations: Criticisms of  these two aspects of  Chomsky's theory, the emphasis on syntax and the adoption of  the sentence as the unit of  analysis, have led to a new emphasis on semantics and on discourse as the unit of  analysis. (Frederiksen, 1975, p. 373) An examination of  discourse theory will determine how it might serve as a new basis for grammar instruction, particularly grammar instruction as applied in the teaching of  reading. Discourse Theory: The "structure" of  discourse Discourse theory unites structure and semantics. The theory maintains the sentence as the largest structural unit, but situates the sentence within the semantic unit of  the text.  The text is a semantic unit, not a structural unit, which can be said to be realized by, but not consist of,  sentences. The text is the meaning behind a group of  sentences, it is a discourse. A seeming oxymoron, the term 'discourse structure' is employed by linguists to describe the structure of  a nonstructural unit. However, discourse structure can be seen as a type of structure in the sense that a discourse takes a contained form—a  series of  sentences organized in a series of  paragraphs. It is not defined  as a formal  structural unit because its parts (propositions) are not tied to one another by syntactic relations but by semantic relations. ... a text typically extends beyond the range of  structural relations, as these are normally conceived of.  But texts cohere; so cohesion within a text— texture—depends on something other than structure. There are certain specifically  text-forming  relations which cannot be accounted for  in terms of constituent structure; they are properties of  the text as such, and not of  any structural unit such as a clause or sentence.... They are ...semantic relations, and the text is a semantic unit. (Halliday and Hasan, p. 7, 1976) Thus we can speak of  two levels of  a passage: the syntactic level composed of structural units such as phrases, clauses and sentences united by syntactic relations; and the semantic level, composed of  the propositions which are united by semantic relations to form the discourse. These two levels are not distinct from  one another. Semantic relations between propositions are cued by syntactic markers—punctuation and connectives. The term "proposition" has been employed in many different  ways within linguistics and certainly has a different  meaning in philosophy. For the purposes of  this study, we will use van Dijk's and s, Kintsch's definition  which unites the syntactic structure of  the clause with the semantic unit, the proposition: Propositions, then, are constructed in our model on the basis of  word meanings, activated from  semantic memory, and syntactic structures of clauses. In principle, we will assume that there is a one-to-one relationship between propositions and clauses: One clause expresses one proposition, (van Dijk and Kintsch, 1983, p. 14) Discourse Theory: The process of  comprehending In understanding discourse, one needs to recognize and understand both the propositions and the relations between them. In their model of  discourse comprehension, van Dijk and Kintsch (1983) present two levels of  comprehension: the text base and the situation model. Definitions  given by McNamara, Kintsch, Songer, and Kintsch (1996) will serve for this study: The text base contains the information  that is directly expressed in the text, organized and structured in the way that the author had organized the material. It has a local structure (the microstructure) as well as a global structure (the macrostructure). The construction of  the text base involves the extraction of semantic information  from  a text.... this takes the form  of  an interrelated networks of  propositions.... On the basis of  the text base, readers can verify statements they have read, they can answer questions about the text, they can recall the text, or they can summarize it. (p. 3) The situation model integrates the information  provided by the text with prior knowledge, often  reorganizing and restructuring it in terms of  the reader's understanding of  the knowledge domain as a whole rather than the particular text just read.... The formation  of  such situation models requires two things as a basis: active inferencing  and adequate prior knowledge, (p. 5) The first  step in comprehending text is building a text base. This involves first isolating then understanding propositions. According to van Dijk and Kintsch, "...the evidence for  the psychological reality of  propositional chunks is overwhelming" (1983, p. 41). In their studies, they found  that chunking was a natural process: "readers and listeners are sensitive to the phrase structure of  sentence and chunk sentences accordingly" (1983, p. 28). van Dijk and Kintsch were referring  to adults, experienced language users. Adams (1990) notes the same tendency in children. She refers  to Karpova's 1955 seminal study of word awareness. Karpova found  that when asked to count the words within a spoken sentence, children first  break up sentences by propositions or idea units, then into subject and predicate and, later in their development, learn to isolate each word. This sensitivity to syntactic boundaries is also natural in reading, and it is also developmental. The development works in a reverse order—from  words to propositions. Once a child is comfortable  with decoding words and is able to put words together, she begins to encode sentences, storing the meaning of  one word in her short term memory as she decodes the next word. At the beginning of  the reader's career, sentences are short and simple and generally do not contain numerous propositions. But very soon the child encounters longer, more complex sentences. This strains the limits of  a child's short term memory. It is these very limits of  the short term memory that necessitate chunking. Kintsch and van Dijk (1983) noted that "one cannot retain everything one reads or hears for  very long, and thus it becomes crucial to know when to discard what.... Jarvella (1971) and others have shown that readers hold the final  phrase in short-term memory, dumping it when they arrive at a clause boundary" pp. 28-29). The sentence is chunked to allow it to be preserved within the short term memory. But without the ability to find  natural breaks, or to chunk the sentences, the child quickly becomes lost in verbiage. And though chunking is a natural process, literate language is,not natural but formal.  Thus a child unaccustomed to it may fail to comprehend it, for  when a child is labouring to decode word by word, word groupings are not necessarily obvious to her. Adams (1990) notes that "because readers must interpretively collapse text at and only at syntactic boundaries, such sensitivity stands as a strong, if indirect, determinant of  comprehension" (p. 153). van Dijk and Kintsch (1983) suggest that chunking is not necessarily automatic for adults either, but rather strategic: Strategic behaviour is neither necessarily controlled nor necessarily automatic... Like other behaviors, it varies from  controlled (i.e., slow, sequential, resource demanding) to automatic (fast,  parallel, effortless)  as a function  of  stage of  practice. The child, in the process of  acquiring a strategy, uses it quite differently  than does the mature speaker of  the language. For the latter, what were once demanding tasks have become fully  automated with very low resource demands—unless, of  course, the smooth operation of strategies is blocked (as in the garden path sentences), and attention-demanding, controlled, repair processes are required, (p. 31) Strategies can be taught. Chunking is strategic; therefore,  it can be taught. Establishing Relations between Chunks The second process in building a text base is establishing relations between chunks. This requires the recognition of  explicit and implicit ties. This step is not entirely separate from  the first;  in fact,  the mature reader seeks a tie even before  the entire proposition is read. The assumption underlying the presence of  local strategies is that it is most unlikely that language users will process full  sentence pairs before  establishing some coherence relation. Rather, they will try to relate fragments  of  a new proposition to the proposition already processed, (van Diijk and Kintsch, 1983, p. 154) The immature reader will be slower in recognizing ties, and may in fact  miss them altogether (Zinar, 1990). This is partly due to the fact  that the immature reader may be working more from  the bottom-up than from  the top-down, decoding word by word and encoding phrase by phrase, sometimes seeing no relation at all between the sentence being read and the one just previous to it. Unless the reader is aware of  the existence of  a macrostructure and is endeavoring to establish global coherence while establishing local coherence, the proposition may be processed in isolation, and if  its interpretation is heavily dependent upon the relation it bears to another proposition, it may be unintelligible. It must be noted that, though the stages described here are represented as moving from  the less complex processes to the more complex, the cognitive processes at each level interact with one another, and the building of  the situation model does not happen after  the construction of  the text base is complete, but simultaneous to it. But neither the text base nor the situation model will be constructed if  the reader is unable to establish relationships between propositions; that is, if  they are unable to make the text cohere. Halliday and Hasan describe cohesion simply as "the set of  semantic resources for linking a sentence with what has gone before"  (1976, p. 10). This link is formed  through reference  to some other item within the discourse. Such a reference  may be explicit, such as when a pronoun refers  back to a proper noun (one case of  anaphora) or when connectives such as "and", "but", and "although" are used to link a clause to one which either precedes or follows;  or it may be implicit, in which case the reader will be required to make an inference. These connectives are another structural unit that served as the joining hardware which link prepositional building blocks. They are function  words—words that cannot be defined  apart from  the function  they perform  as compared to form  words, such as nouns, verbs and adjectives which carry meaning within themselves. Recognizing and understanding explicit referential  ties can also require inferencing.  It involves several processes which, due to their interactive nature, may or may not be engaged in the following  sequence: noticing the reference,  understanding what it means (e.g. she = female  person or animal, although = adversative tie), tracing back or forward  to find  the referent  and finally  comprehending how the referent  and the element to which it has been linked are related. Various factors  such as the degree of  ambiguity of  the reference  or the proximity to the referent  determine how much work the reader is required to do and how much background knowledge she must supply, i.e., how much inferring  she must do. This brings us to the second great demand on the young reader, the making of  inferences. Making inferences  has long been acknowledged as one of  the things a skilled reader does. Inferencing  places further  demands on the already taxed short term memory. "... when skilled readers are in the course of  reading a clause, their ability to recall its precise wording is extremely rapid and accurate; in contrast, this fine,  verbatim memory for  the clause is all but lost just as soon as they start reading the next. As another outward sign of  this recoding, skilled readers characteristically pause at the end of  major syntactic units. Appropriately, the duration of  this 'wrap-up' time is significantly  increased if  the interpretation of  the just-read clause requires inference  or subtle resolution of  pronouns. (Adams, 1990, p. 186) What is true for  the skilled reader is doubly true for  the less-skilled. Stoodt (1970) found  that young readers had trouble keeping the individual propositions encoded in a sentence in the mind while establishing the relationships between them. It was particularly difficult  for young readers to maintain and comprehend differences  between ideas (or adversative relations) as opposed to similarities (causal or descriptive relations). Cote and Goldman (1999) also found  evidence that failure  in comprehension of  the overall message of  the text is caused by an inability to establish connections due to the limitations of  short term memory: It is well established that readers often  do not construct coherent propositional or situational representations of  text information.  One reason for  failures  to detect or generate the connections necessary for  coherent discourse representations is the unavailability of  relevant information  in working memory. In the process of  reading a long text, keeping relevant information active in working memory often  requires monitoring the coherence of  the evolving representations and strategically using discourse and topic knowledge to identify  relevant information  in order to reinstate selectively previous text information,  retrieve or reinstate information  from  long-term memory or both (e.g., Ericsson & W. Kintsch, 1995; Fletcher, 1986; W. Kintsch, 1988; W. Kintsch & van Kijk, 1978). (p. 170) Young readers have not only had less time to develop efficiency  in storing and accessing items in their short term memories, they have had less time to acquire background knowledge. Zinar (1990) found  evidence to suggest less-skilled readers ignored connectives if  they lacked the background knowledge required to make the connection between ideas. "The particular linguistic form  in which a particular relationship is expressed may be of  little moment if  the reader does not have sufficient  understanding of  real-world phenomena which underlie the relationship" (p. 197). This lack of  background knowledge can be noted at various levels—at a specific  factual  knowledge, e.g., the child has little or no knowledge of how a volcano erupts, at a vocabulary level, e.g., the child is unfamiliar  with.specific  terms such as magma, at a textual level, e.g., the child is unfamiliar  with cause and effect  text patterns or at a domain level, e.g., the child is unfamiliar  with the domain of  geological science. Young readers often  suffer  from  a lack of  background knowledge at any or all of these levels when reading science text. Making inferences  within a narrative can be relatively easy and more opportunity is granted a child in learning to do so. Making predictions, a form  of  forward  inferencing,  is encouraged, even guided, by various narrative conventions. Cote and Goldman note that "... the dominant mode of  explanation appropriate to understanding stories is causal inferencing". They contrast the comparatively straightforward  inferencing  to that encountered in instructional texts, where "prior knowledge limitations and a greater variety of  expository content structures constrain the success of  efforts  to understand not only underlying causal mechanisms but also other types of  relationships among concepts... " (1999, p. 187). In summary, in attempting to chunk propositions as well as establish relations between propositions, the young reader must contend with the limitations of  short-term memory and the lack of  background knowledge. Readers may overcome these limitations naturally as they develop cognitively or as they gain experience. Zinar (1990) found  evidence to support Johnston's and Pearson's (1982) three stages of  development in the understanding and use of  connectives as explicit tools for  establishing coherence: Younger, as well as less-skilled intermediate readers, process text in word or proposition—length units and do not attend to connectives. More skilled intermediate readers, as well as less skilled adult readers, are aware of connectives and use them to integrate what they read, but they are still dependent upon the presence of  explicit cues. Fluent adult readers are independent of  cueing and actively integrate what they read even in the absence of  explicit connectives, (p. 197) Instruction should be directed at what Vygotsky called the learner's zone of  proximal development, that zone between a child's actual developmental level, working independently and their potential development, working under guidance (Wertsch & Tulviste, 1992). We need to pave the way for  students to progress from  one developmental stage into the next. As Johnston and Pearson have indicated, it is also true that students who struggle with decoding do not progress in their reading as they should. Cain and Oakhill (1999) suggest that less skilled readers are less able to make inferences  because they are focusing  more on word reading accuracy than on monitoring their comprehension. It is their entire approach to reading that causes the comprehension break down. Because they do not see the text as a whole made up of  parts which must cohere, they do not strive to make it cohere. When readers cannot actively integrate text, they may be able to provide answers from  the text verbatim, lifting  phrases and propositions from  the text, but they cannot provide a summary of  the text—they cannot construct a text base. And, though it does not follow  that if  the reader is able to build a text base he will be able to create a situation model, it does follow  that if  the reader is unable to build a text base he will be unable to create a situation model. Understanding relations—inferencing:  text base or situation model? The role of  inferencing  within the two levels of  comprehension is a contentious issue. McNamara et al would place inferencing  with the situation model, seeing the results of "active inferencing"  as "the formation  of  more links between the incoming information  and information  in the personal—knowledge base (Mannes & Kintsch, 1987)" (p. 3). Frederiksen (1977) sees inferencing  as a process of  taking the text and combining it with old knowledge to create something new: ... there are two levels of  processing in discourse comprehension that involve propositional data structures. The first  processing level, the interpretive level, involves those processes by which propositional knowledge is "recovered" from  text; the second level, the inferential  level, involves the generation of new propositional knowledge from  a network of  propositions which is given—from  prior discourse, from  context, or from  stored knowledge of  the world, (p. 68) But inferencing  must be present at the interpretive level, during the processing of  the text. Whenever a proposition is not expressly stated and must be supplied by a reader, an inference  is made. This can be true even when a proposition contains a referent  that directly ties it to another, as in the case of  "Because it was cold, Mary didn't want to swim." To tie the two together, here the reader must infer  an additional proposition to the effect  that "Mary doesn't like to swim in the cold". Irwin (1986) cites several "recent theories" (including Frederkisen's) that place the role of  inference  in establishing cohesion—noting and processing the relations within a text. She draws attention to Crothers (1978, 1979), who in his analysis of  various types of inference,  maintains that 'the purpose of  many, thought not all, of  the proposed inference types is to explicate the implicit coherence of  the passage' (1979, p.7). Cain and Oakhill (1999) mention two types of  inferencing:  "text-connecting" which make use of  syntactic cues and "gap-filling"  which require the reader to supply background knowledge. But this background knowledge can be as simple as knowing that when one says "he pedaled home" the person in question is understood to be riding a bike. This is still a text base building inference.  The construction of  both the text base and the situation model require inferencing. What distinguishes the processes is the type and degree of  inferencing  they require. McNamara et al (1996) do note that the distinction between what goes on at each level, the construction of  the text base and that of  the situation model, "is not absolute, but a matter of  degree" (p. 19). They differentiate  between "bridging" inferences  which link information  from  two or more sentences within the text and "elaborative" inferences  which "require linking textual and outside knowledge information  which requires some, but not necessarily a very deep, situational understanding" (p. 19), indicating the need for  even further  inferencing  to create a fully  developed situation model. We might see inferencing  then as a process which begins at the level of  making sense of  the text as it is and follows  a gradation that slips into a situation model—building mode. It is in fact,  a jumping off  point, it puts the mind in motion and it is this momentum that is the glisse into gestalt, those "ah-ha" moments which are at the crux of  discovery learning. van Dijk and Kinstch (1983) acknowledge the ambiguity of  the place of  inferencing within the two levels of  comprehension: The biggest problem with discourse inferences  is to determine when they are made: Are they part of  comprehension proper, or do they occur optionally after  comprehension? It is clear that not all possible inferences  can actually be made... The question therefore  is, how can one distinguish those inferences that must occur as an integral part of  discourse comprehension proper? One class of  inferences  that appear to be necessary during comprehension are the bridging inferences  required for  the coherence of  the text (Kintsch, 1974, Clark, 1977; Miller & Kintsch, 1980). (p. 49). Despite the ambiguity, van Dijk and Kintsch choose to assign inferences  to the situation model, using the following  rationale: The text base is a representation of  the text as it is. Bridging inferences  and other types of  inferences  belong to the situation model constructed on the basis of  that text base and knowledge. What is inferred  are therefore  not propositions in the text base but links in the situation model. This shifts  the analysis of  inferences  in discourse processing from  the linguistic level to the conceptual level... (p. 51). Here we see van Dijk and Kintsch assigning inferencing  the role of  the liaison between linguistics and concepts, that is structure and semantics. But we cannot simply focus on structure in our building of  the text base. A reader must construct this base, not simply receive it in its encoded form.  Summarizing a text would be considered evidence of  having constructed the text base. Summarizing is not possible if  the reader has not worked through the propositions and their relations in such a way as to achieved a level of  understanding that allows her to manipulate the propositions to create a format  that is logical to her and in which she may communicate the meaning of  the text. Often  this requires a great deal of manipulation, particularly if  the author has not organized his propositions logically or has left many propositions to be provided by the reader—many gaps to fill  in. The level at which the reader must engage in inferencing  is not a simply theoretical point to be argued for  the sake of  argument. It is vital that we understand the role of inferencing  in the construction of  the text base or we will not be aware of  the huge difficulties  students can encounter at this level of  comprehension and what we can do to help them overcome such difficulties.  Moreover, if  it is true that the conscious manipulation of structure aids the reader in constructing the text base, then it is also true that awareness of structure will aid a reader in inferencing,  and thus aid her in comprehending. The ambiguity may be cleared up in part by separating the first  level into two, creating a three level model of  comprehension. Tapiero and Otero (1999) describe such a model, stating that the reader engages in both micro processing, by which a locally coherent propositional network is set up, and macro processing, by which the text base is "organized in a hierarchical sequence" (p. 341). These two processes result in a text base which is "a representation that is faithful  to the text" (p. 341). Macroprocessing requires determining which propositions are essential to establishing a text base, as a text base cannot include all the propositions contained in the text. It is a summary, and in constructing a summary, a reader must be able to determine which propositions are essential components of  the macrostructure. Here too, attention to connectives can be helpful.  Goldsmith (1982) found that connectives alerted students to the organizational structure of  the text and guided them to delete irrelevant information  and to highlight important information.  They allowed inexperienced readers not only to weed out irrelevant information  within the text while constructing summaries but also to avoid misinterpreting text by misapplying unrelated personal knowledge. Goldsmith recommends modeling using connectives as guides while reading classroom materials. Tapiero and Otero, who were studying the reading of  undergraduates, noted that often both high and low readers develop propositional structures, but only high readers go beyond text based ideas to developing "an appropriate and elaborated situation model evoked by the text" (p.342). Low readers could not assemble a body of  knowledge that could be manipulated. They may have had a clear picture of  what the text said, but only in relation to the text itself,  not in relation to "the real world" or any other theoretical construct. They were able to remember the text, but not to learn from  it. Many younger readers fail  even to achieve Tapiero and Otero's first  and second levels of  comprehension and are unable to summarize in any detail or clarity what they have read. This is particularly true of  complex content area text such as science textbooks. The pretest and post test of  this study is designed to measure the first  two levels of  comprehension, both incorporated in the text base. The third level, the building of  a situation model, will be considered in the discussion of  the intervention. The Discourse Structure, the Domain of  Science and the Young Reader As students progress beyond the third grade, they must meet the demands of expository text. Kent (1984) notes that while a narrative text makes many personal references,  an expository text may make none; while a narrative text is agent oriented, an expository text is subject matter oriented; while a narrative "uses an accomplished time, i. e., either past tense or the historical present", an expository text may vary tenses; while a narrative text is "connected by some chronological linkage", an expository text is governed by sentence topic and parallelism (pp. 234-235). Spiro and Taylor (1980) suggest that the type of  inferencing  required in content area texts tends to be logical, while children's stories tend to demand pragmatic inferences,  and the latter makes fewer  demands on a child. Thus, many of  the elements of  cohesion found  in narrative, such as a consistent agent and verb markings, are often  lacking in an expository text. Palincsar and Magnusson (2001) note that studies paint a bleak picture of  science text, listing as one characteristic a sparseness "with regard to transitions and other devices useful  to attaining cohesion" (p. 154). But Rodgers (1974) found  at least one cohesive prevalent in science text—connectives. Rodgers went so far  as to list the most frequently  occurring connectives in Social Studies and Science textbooks. He notes that authors of  textbooks make liberal use of  connectives, fully  expecting readers to attend to them. Yet it is in the reading of  science text that we see evidence of  the tendency of  students to ignore what they do not understand—whether that be a connective or a concept. Background knowledge is essential for  attending to connectives, making inferences  and constructing a situation model, and many students are either uninformed  or misinformed. Misconception's may be "physically derived (resulting from  interactions with the physical environment), socially derived (based on interactions with family  members, peers, or the media), or instructionally derived (resulting from  formal  instruction)" (Guzzetti et al, 1993, p. 117). It is little wonder that misconceptions abound in that scientific  "fact"  seems to be ever changing due to new discoveries. But even more confusing  is the apparent contradictions between scientific  theory and sense experience—as in the case of  the intuitive knowledge students have about motion. Most students maintain a "pre-Newtonian impetus theory" (Hynd et al, 199!, p. 597) and are loathe to discard their theory (see also Alvermann et al, 1995). When students hold tenaciously to misconceptions, they allow these to govern their reading. If  their underlying assumptions differ  widely from  the macropropositions presented in the text, students may fail  to establish coherence among individual propositions. In fact,  in reading a text refuting  a theory they hold as true they may fixate  only on the sentence containing a description of  that theory, reading this sentence as support for  it. (Alvermann et al, 1995) Otero and Campanario (1990) explain this phenomenon in relation to van Dijk and Kintsch's model: It is possible that subjects do not detect the inconsistencies, that is they do not try to build coherent text bases, because they give up building a situation model for  the text: "A prerequisite for  coherent text representation is the ability to construct a coherent situation model. Without that, memory for  text is stored as incoherent bits and pieces..." (van Dijk & Kintsch, 1983, p.361) (p. 451) Unless students can be made to realize that a text must cohere, they will be lost. Here we see the interplay between the text base and the situation model, but how difficult  it is for  one to establish a situation model if  one is lacking in scientific  knowledge. Still, if  students can be guided to make sound text bases, they can begin to make connections between texts and begin to build up sufficient  knowledge to establish situation models. Finally, for  students to succeed in reading science, they must adopt an attitude of inquiry and discovery. Reading alone will not accomplish this. Palincsar & Magnusson (2001) recommend first  hand investigations exploring a concept before  a student is exposed to text. But providing students with opportunity for  investigation, remains a first  step. Students can't learn all they must learn of  science from  experimentation. At some point, they must read text and at what point will they have sufficient  background knowledge for  reading the text they must read to supply them with the requisite knowledge? Educators cannot hope to provide all the background knowledge necessary to make students naturally sensitive to the cohesive ties of  science argument. It is a catch-22, a vicious circle, rather like using a term to define  the same term. At some point one must go outside the domain of  science to find  a point of  entry to the domain. Educators must present the structure of  language as a portal to the semantics of  science. CHAPTER III METHODS AND PROCEEDURES The Study This is an exploratory descriptive study using both quantitative and qualitative measures. Due to the limitations of  a small sample size, descriptive statistics are used. The study consists of  tests for  selecting and grouping participants, a pretest, an intervention involving both whole class grammar lessons and small group guided reading sessions and a posttest. Scores on tests provided quantitative data and field  notes of  lessons as well as analysis of  individual oral responses on tests provided qualitative data. Participants The experimental group were selected from  a grade 5/6 class of  twenty-eight students (eighteen males, ten females)  in a K-7 elementary school in a working, middle class suburb. The class included 6 severely learning disabled (high incidence) grade six students and one low incidence grade five  student with Down's syndrome. A full  time teaching assistant was assigned to the low incidence student. All students in the experimental class were native speakers of  English. The control group was selected from  a grade 6/7 class of  thirty-one students (twenty males, eleven females)  in the same school. All students in the control class were native speakers of  English. This class included one SLD student in grade seven. Students were selected on the basis of  their scores on the Stanford  Diagnostic Reading Test. In both the experimental and control groups, participants were divided into low, medium and high groups for  analysis. This division was replicated in the guided reading group make-up for  the experimental group. After  one low member from  the control group and one medium of  the experimental group dropped out of  the study, the control group consisted of  eight grade six students (four  girls and four  boys) grouped as follows:  two low (one girl, one boy), three medium (two girls, one boy) and three high (two boys, one girl). The experimental group consisted often  students (three girls and seven boys) grouped as follows:  two low (one grade six girl, one grade six boy), three medium (two grade five  boys . and one grade six boy) and five  high (one grade five  girl, one grade six girl and three grade six boys). For reasons to be given later, the data for  the low groups have been reported separately. Baseline Data The Stanford  Diagnostic Reading Test, 4th edition, published in 1996 by Harcourt Brace and Company yielded baseline data on general reading comprehension which was used to select participants and to establish three subgroups in both experimental and control groups in relation to reading ability: low, medium and high. The Stanford  is a diagnostic tool designed to probe students' strengths and weaknesses in the major components of  the reading process, which yields both norm-referenced  and criterion-reference  information.  The three subsets from  the purple level, form  K administered to participants in this study were comprehension, vocabulary and scanning. Selection and grouping were made on the basis of the overall scores for  the test. The low group students achieved response scores ranging from about grade 2.5 to grade 4, the medium group students achieved response scores ranging from  about grade five  to grade six, the high group students achieved response scores at the grade seven level or above. For each subtest, items are presented in the form  of  multiple choice. In the comprehension test, subjects are asked to read a passage and then answer questions or complete sentences designed to test comprehension of  the passage. The passages vary in genre and include expository, narrative and directions for  a given procedure. In the vocabulary subtest, subjects are asked to choose the definition  of  a given word. The words are given in isolation, so the subtest is a measure of  existing vocabulary knowledge, not the ability to determine meaning through context. In the scanning subtest subjects are instructed to read the questions first  and then scan the passage for  the answers. They are instructed not to read the passage in its entirety. Both the comprehension and scanning subtests test skills which must be applied in reading for  information,  and the vocabulary subtests measure vocabulary knowledge base. A second measure was used after  grouping to provided baseline data on syntactic knowledge. This was the CELF, Clinical Evaluation of  Language Fundamentals, 3rd edition, published by The Psychological Corporation, Harcourt Brace and Company, 1995. This is a diagnostic tool used for  assessing syntactic proficiency.  It consists of  two major parts each consisting of  three subtests. One elicits a Receptive Language Score, the other an Expressive Language Score. It is this latter that was administered to the participants of  the study. The subtests are Recalling Sentences, Sentence Assembly and Formulated Sentences. In Recalling Sentences subjects are asked to listen to and repeat verbatim sentences increasing in difficulty,  assessing "recall and reproduction of  sentence surface  structure as a function  of  syntactic complexity" (p. 56 of  Examiner's Manual). As a straightforward  recall activity, it is the least challenging of  the three subtests. In Sentence Assembly, students are given a group of  words, randomly ordered, and asked to make two sentences, one declarative and one interrogative, assessing "the ability to assemble syntactic structures into grammatically acceptable and semantically meaningful  sentences" (p. 63). It indicates a subject's "syntactic flexibility"  (p. 139) and her ability to manipulate and transform "syntactic structures within the constraints imposed by content words and grammatical markers" (p. 140). Accordingly to Speech Therapist, Esther Baxter, this subset reveals a subject's facility  in grammatical usage and his or her reading experience, or familiarity  with literate language. In Formulated Sentences, students are shown a picture and asked to make up a sentence about it using a particular word or phrase, assessing "formulation  of  simple, compound and complex sentences" (31). Some of  the words provided are content words such as "children", but most are connectives such as "because" and "although". Among the three subtests, this places the most stringent demands upon a subject's knowledge of  grammar and his or her vocabulary. Pretests and Posttests Two measures were designed by the researcher to measure reading comprehension of science text and syntactic knowledge. (See Appendix B.) A pretest was given at the beginning of  the study to provide baseline data and the posttest was given at the end of  the study. The tests were comprised of  a reading passage selected from  a grade six science text, "The Nature of  Matter" from  the Discovery Works series. Comprehension was tested through free  recall and guided recall which consisted of  four  comprehension questions and a meta-cognitive question. Subjects were asked to read a passage silently and answer several comprehension questions in writing. There were no fill  in the blank or multiple choice responses and the type of  questions were varied by degree of  inferencing  required. It was marked by the experimenter and a member of  the school district personnel using the BC Performance Standards for  Reading for  Information. Syntax was tested through chunking isolated sentences taken from  the passages. Subjects were asked to draw in lines to indicate where they would chunk the sentences. They were then asked to give reasons for  where they chose to chunk the sentences. Sentences containing connectives were isolated and subjects were asked to explain the meaning or role of  the connectives within the sentences. This portion of  the test was given orally and responses were transcribed. The Intervention The intervention was conducted over a seven week period. It consisted of  eight whole class lessons taught within the first  five  weeks which focused  on sentence structure. The concepts of  phrases, clauses, subjects, predicates and connectives were introduced and practiced through group activities and individual exercises. During the third week of  this instruction, guided reading groups began to meet, focusing  on reading the science text while applying the concepts introduced in the lessons. They met for  a total of  five  weeks, one or two sessions a week for  a total of  eight sessions each. Classroom observation of  both whole class lessons and guided reading sessions, recorded as field  notes taken by the researcher, also served as a qualitative measure. Results and discussion of  the intervention have been assigned its own chapter, chapter 5. Procedure At the beginning of  October, the two classes from  which the participants were selected were given the Stanford  Diagnostic Reading test. Participants were chosen on the basis of  their performance  on the test, and grouped accordingly. After  selection, the participants were given the CELF test. This is included in the data as a measure of  syntactic competence. The test was given orally and individually by the Resource Room teacher over a period of  three weeks, first  to the experimental group and then to the control group. Testing was not completed before  the reading syntax pretest was given, but it was completed for  the experimental group before  the intervention lessons commenced. For both the pretest and posttests, two passages were selected from  "The Nature of Matter" (Badders et al, 1999). One major consideration in the selection of  the passages was their use of  complex sentences which included connectives. Another was the topic—it had to be something about which students would not likely have a large amount of  background knowledge and something that would contain some counter-intuitive concepts. The passages were rewritten to ensure similarity in both sentence structure and readability. According to the Flesch-Kincaid measure, the pretest was a 7.28 grade level and the posttest was a 7.51 grade level. This is higher than the grade level of  the students and exceeds the comprehension score on the Stanford  Diagnostic of  both low and medium students, but it is either at or below the score of  the students in the high group. Furthermore, it is a recently published (1999) grade level text (regularly used for  grade five  students as well as grade six students in a split 5/6 class). Though the texts used for  the tests were definitely  at frustration  level for  the low group, the intervention passages were at their instructional or independent level. Low group participants were included in the pre and posttesting to see if  the intervention instruction might help them cope with the type of  texts they encounter in their classrooms. The syntax portion of  the test was completed orally and the responses were transcribed. Because of  scheduling difficulties  within the school day, the tests were given in two sittings. After  the pretest was given, the intervention commenced. At the end of  the seven weeks of  the study interval, the posttest was administered. Tests were then marked, data was compiled and the relationship between the two variables, reading comprehension of  science text and syntactic knowledge, was analyzed. The results are presented in chapter 4. CHAPTER IV RESULTS In this chapter, data on both baseline data and data obtained from  the dependent pre and post tests is reported. Baseline Data—Low Group The data on the low group is reported separately from  the results of  the medium and high group. This separation has been made for  two reasons. First, it was found  that after  some omissions from  the group, (absences on the date of  test administration, failure  to complete tests due to unwillingness, etc,) the experimental and control low groups differed substantially from  one another, and second, the absence of  some data limited the comparisons which could be made. One of  the experimental low group (experimental low #2) was designated SLD, yet in performance  he was often  higher than the other group members. He was, however, accustomed to having responses on tests scribed for  him and did not fill  out all response blanks for  the pretest. One of  the control low group (control low #1), scored significantly  lower on the both the Stanford  comprehension subtest and the CELF, bringing down the control average, yet she had no designation. Table  11: Low group baseline data Stanford Comprehension Stanford Total CELF Expressive Language CELF Formulated Sentences experimental low 5.2 3.4 94 8.5 control low 2.8 2.8 86 8 The low data has not been completely omitted, however, because overall scores on the Stanford  Diagnostic (experimental low - 3.4, control low - 2.8) indicate the group is more homogenous than the comprehension subtest indicates (experimental low - 5.2, control low -2.8), and, as mentioned previously, each subtest tests skills necessary in reading for information.  Also, formulating  sentences has been designated as the most stringent indicator of  syntactic knowledge of  the three subtests , and on this measure, the groups are quite similar (experimental low - 8.5, control low - 8). Furthermore, if  one assigns a score of  one to the blanks on the comprehension test according to the BC performance  standards which states that at a level one "responses to comprehension questions or tasks are often  inaccurate or based solely on prior knowledge; often  vague or incomplete" (p. 190), comparisons between before  and after  intervention performance  can be made within and between the two low groups. (The description taken from  the BC Performance  Standards does give an accurate overall "snapshot" of  the student.) Matching of  struggling students is fraught  with difficulty  because the nature of  their difficulties  can be quite diverse, as is the amount of  support they have received or are accustomed to receiving. Such differences  in both strengths and weaknesses exist between students of  more average or high ability as well, but they tend to be more marked and more problematic in struggling students. The data collected in this study remains useful  as an analysis of  the efficacy  of  such syntactic instruction on struggling students. Baseline Data—Characteristics of  Medium and High Groups Table  12: Medium  & High  baseline data Stanford Comprehension Stanford Total CELF Expressive Language CELF Formulated Sentences experimental medium 5.2 5.3 101 8.6 control medium 4.5 5.7 92 7 experimental high 9.8 9.8 104.4 9.8 control high 11.4 9.3 108 11.3 experimental medium & high 8.2 7.6 103.3 9.4 control medium 8 7.5 100 9.2 & high As table 12 indicates, the experimental medium group scores are, in all measures save Stanford  Diagnostic, higher in than the those of  the control medium group. The widest difference  exists in the syntactic measures and will be addressed presently. Conversely, the experimental high group scores are in all measures lower than the control high group save in the Stanford  Diagnostic total where they scored higher by a very slight margin. When the medium and high groups are combined, the experimental and control groups are quite similar with the experimental leading by a slight margin. The disparity in the syntactic scores of  the two medium groups may be in part due to the fact  that the scores of  the experimental medium group included the experimental group's extremes—both the lowest and the highest scores for  syntactic facility. • Table  13: CELF  and  Stanford  scores Part ic ipants Stanford  Diagnost ic CELF Comprehension Vocabulary Scanning Total Formulated Sentences Recalling Sentences Sentence Assembly Expressive Language experimental low 1 5.2 4.5 3.1 3.5 7 10 7 88 experimental low 2 5.2 2.9 3.8 3.3 10 12 8 100 experimental medium 3 7.5 5.7 5.7 6.1 12 13 14 118 experimental medium 4 3.7 4.5 5.7 5.1 9 11 9 98 experimental medium 5 5.2 5 4.8 4.8 5 12 7 88 experimental high 6 13 10.6 8.5 12.1 9 15 10 108 experimental high 7 10.5 10.6 13 13 9 11 12 104 experimental high 8 9.1 5.7 8.5 8.2 '11 13 14 116 experimental high 9 9.1 5.7 9.6 8.6 12 11 9 104 experimental high 10 7.5 5.4 7.6 7.3 8 9 8 90 control low 1 2.3 3.5 3.1 2.4 7 6 6 78 control low 2 3.3 3.6 3.6 3.1 9 9 9 94 control medium 3 3.8 5 7.3 5.4 7 7 9 86 control medium 4 4.1 5.7 7.6 5.6 8 9 10 94 control medium 5 5.5 5.7 7.6 6.1 6 13 9 96 control high 6 13 13 5.3 8.2 10 13 9 104 control high 7 8.3 5.7 9.6 9.6 . 11 10 12 106 control high 8 13 8 9.6 10.2 13 12 12 114 The high score (118) belonged to a grade five  student (experimental medium #3) and the low score (88) belonged to a grade six student (experimental medium #5). Both of  these students characteristically contribute actively to class discussions and did so consistently throughout the intervention. Both proved to have a decided interest and good store of background knowledge in the topic of  study—space exploration, and both were in the same guided reading group for  the intervention, but experimental medium #3 demonstrated a marked ability to reason scientifically  and to express his reasoning. References  will be made to his contributions in later chapters. It is the high score of  the experimental medium #3 that proved uncharacteristic of  the other medium group members both in the experimental and control groups. What was interesting about the scores of  experimental medium #5 is the disparity between the scores he achieved on the three subtests. This student scored lower than any of the other participants in any group and well below the statistical average on formulating sentences, scoring a 5 where the average score is 10 (confidence  interval of  7-12). His score on sentence assembly was also below statistical average and among the lowest in the participants, 7, yet his score for  recalling sentences was in the high average range, 12. Such disparity indicates a definite  weakness. This student is quite a competent reader and has learned to listen, retain and recall. This is particularly true when the student is reading fictional  literature. However, this student generally has more of  a struggle processing content area text. He has difficulty  maintaining a flow  of  thought and will often  grasp onto a concept and use it as a springboard into an imaginative idea fairly  far  removed from  the scientific context of  the discussion. According to the district speech therapist, Esther Baxter, the CELF test analysis suggests that this student would profit  from  practice with sentence frames  in order to build his syntactic flexibility.  This is a focus  later in the intervention during the composition of  sentence summaries. The fact  that these two boys had similar reading abilities, interests and propensities to participate in, if  not dominate, the discussion but differed  substantially in their syntactic ability makes them an interesting pair to observe. It begs the question as to whether the syntactic ability is a correlate of  reasoning ability, and whether it helped experimental medium #5 read, comprehend and respond to both comprehension and syntax questions regarding the text. The gains of  both will be addressed later. Experimental medium #5 has a counterpart within the medium group of  the control. The three CELF subtests scores of  control medium #5 showed a similar disparity and a similar weakness in processing: 6 for  formulating  sentences, 9 for  sentence assembly and 13 for  recalling sentences, control medium #5's overall score was higher, but the spread is similar. Pretest and Posttest Results: Comprehension Comprehension was tested through both free  recall and guided recall. In part A of  the test, free  recall was elicited through the statement: "Tell me everything you can about the information  in this passage". The free  recall responses were graded but the results were not included in the analysis because of  the high tendency of  the participants to simply copy the text. Had the passages been removed from  the students during free  recall the measure would provide more reliable information.  Testing conditions did not permit this flexibility. The guided recall portion of  the test consisted of  four  comprehension questions (see section B in the protocols in Appendix B for  the questions and chapter 6 for  a discussion of the relative difficulty  of  the questions and the degree of  inferencing  required). These four questions were marked using the BC Performance  Standards for  Reading For Information  by both the researcher and a trained member of  district personnel. Interrater reliability was approximately 72%. Scores for  which there was not agreement were discussed and scores were reassigned. A fifth  set of  questions, "What part of  this passage was difficult?  What did you do to help you understand it?" was graded but not included in the analysis as it did not yield any new information  that pertained to the focus  of  the study. Table  14: Comprehension—Medium  & High  Groups score % score % PRETEST experimental medium 23/48 48 control medium 23/48 48 experimental high 48/80 60 control high 34/48 71 experimental total 71/128 55 control total 57/96 59 POSTTEST experimental medium 24/48 50 control medium 18/48 38 experimental high 47/80 59 control high 25/48 52 experimental total 71/128 55 control total 43/96 45 DIFFERENCE BETWEEN PRETEST 0% -14% & POSTTEST TOTALS Table 14 above presents the scores for  the comprehension questions for  both the pretests and the posttests. Note that the total consists of  the total number of  points achieved by all the members of  a particular group out of  the total possible points for  the group, where the total possible for  each individual was 16 ( 4 questions marked out of  4). Thus for  the pretest, the experimental high group achieved 48 points out 80 (16 x 5 group members). All the other subgroups consisted of  three members each rendering the total possible 48. On the comprehension component of  the pretest, the medium groups of  the experimental and the control both scored 48%. This may not sound very high, but one must bear in mind that the rating is on the basis of  a 4 point rubric used in the BC performance standards wherein 4 is exceptional, 3 is a large range encompassing average to high average and 2 tends to be a low average score. It should not be translated into letter grades. On the posttest, the experimental medium group scored 50%, a gain of  2%, while the control group scored 38%, dropping by 10%. On the pretest, the experimental high group achieved an average score of  60%, while the control high achieved an average of  71%. Both high groups dropped in the posttest, but the experimental scored 59%, a decrease of  1% whereas the control scored 52%, a decrease of  19%. Overall, the experimental maintained the same average for  the posttest that they achieved on the pretest, while the control dropped by 14%. It must be noted once again that the posttest proved to be more difficult  than the pretest. The readability level of  the passage was greater (7.51 as opposed to 7.28 for  the pretest passage) and, perhaps more importantly, the nature of  the questions was much more challenging, requiring much more inferencing.  The increased cognitive challenge of  the questions was not by design, and points to a weakness of  the study which will be discussed in chapter 6. Syntactic Knowledge Syntactic knowledge was tested with portion C of  the test. These items were written by the researcher based on her hypotheses regarding the kind of  syntactic knowledge students would employ during reading and regarding the kind of  questions which would allow students to access and express that knowledge. They were developed on the basis of  the results of  the preliminary study (see chapter one) and as means of  testing the concepts the researcher hoped to cover in the intervention. Responses were marked both by the researcher and a trained member of  district personnel. Interrater reliability was approximately 78%. Scores for  which there was not agreement were discussed and scores were reassigned. Several of  the questions failed  to yield new information  and were thus deleted from  analysis. Others addressed concepts which were not presented within the intervention. Though these concepts were part of  the original intervention plan, the unit of  study was modified  to meet the needs of  the students and new avenues emerged during the study. These were followed and observation of  the results are presented in chapter 5. Questions #3 d and #3 e, "What would you say is the main or most important chunk in the sentence?" and "Why" were not included in the analysis as isolating main clauses was not covered in the intervention. Questions #4 a, "Why does this sentence begin with the word 'in fact'?  and #4 b, "What words could you use to replace it?" were not included in the analysis because the term "fact"  caused subjects to focus  on the concept of  fact  and diverted their attention from  the connection made. Furthermore, it signalled an intersentential tie and thus should have been presented within the context of  the text to allow the subjects to recognize the tie. This would be a good focus  for  further  studies—comparing students' ability to recognize intersentential ties with their ability to recognize intrasentential ties. Questions #4 a and #4 b were originally included to examine their ability to recognize these intersentential ties, but failed  to do so because of  the difficulties  noted above. These ties were a focus  in guided reading lessons. Question #5 read as follows:  "Some words and phrases in a passage are there mainly to show how one thought is related to another. They are signals for  relations, just like 'if'  and 'as' above. Can you find  some other words that do that job?" The responses to this question did not enter into the analysis because it yielded no clear information.  The reason for  this was made obvious in the intervention; the exercise of  digging out connectives from  a large block of  text proving to be unhelpful  and confusing.  Finally question #6, "If  you had to choose three key words in this passge, what would they be"? was also deleted from  analysis as the concept of  identifying  key words and how that might interact with the structure was not covered in the intervention. The questions which became the focus  for  analysis were of  three distinct types: chunking sentences for  meaning, giving reasons for  chunking, explaining the meaning and role of  connectives within sentences. Each type of  question will be analyzed separately and then the results of  the syntactic portion as a whole will be discussed. Chunking Performance  on the first  component of  the syntactic tasks, chunking for  meaning (questions #1 a, #2 a and #3 a), is presented in table 4 below. For this task, students were shown a complex sentence taken from  the passage they had read. They were then given the following  instructions: "We can chunk long sentences to help us understand them just as we can chunk long words. Show how you would chunk this sentence by drawing in dividing lines." Their responses were rated from  0-2 based on the logic displayed by their choices. A score of  two was given if  all the divisions were logical—that is, based on syntactic rules, one was given if  some points were logical where others were not, or if  an obvious division (i.e. at a comma) was ignored. Zero was given if  no points were logical. Table  15: Chunking  for  meaning score % score % PRETEST experimental medium 14/18 78 control medium 11/18 61 experimental high 26/30 87 control high 18/18 100 experimental total 40/48 83 control total 29/36 81 POSTTEST experimental medium 14/18 78 control medium 14/18 78 experimental high 30/30 100 control high 14/18 78 experimental total 44/48 92 control total 28/36 78 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 9% -3% Each individual's possible score on this subsection was six; the total possible score for  each group is 6 x the number of  group members. Therefore,  the experimental high's possible total was 30 (for  5 members), while all the other groups had a possible score of  18 (for  3 members). The experimental middle group scored 78% percent both on the pretest and the posttest, demonstrating no change. The control middle group scored lower on the pretest, 61%, but increased to the same score as the experimental middle group, 78%, on the posttest, for  an increase of  17%. The experimental high group scored 87% on the pretest and 100% on the posttest, an increase of  13%. The control high scored 100% on the pretest, but decreased to 78% on the posttest, a drop of  22%. Overall the experimental scores averaged to an increase of  9%, while the control dropped by 3%. Reasons for  Chunking Next, students were asked, "Why would you chunk it (the sentence) at those spots?" (questions #lb, #2b and #3b). Their responses were scored 0-3. This reflected  the syntactical understanding evident in their reasoning. It was more difficult  to assign points to this as it was an open question for  which innumberable responses could be given. A score of  three was reserved for  those who explained that each chunk was made up of  a clause or could be make into a little sentence, two was given for  a mention of  punctuation (e.g. "There's a comma there."), a mention of  connectives (e.g. "There's an "and" there.") or a statement that indicated that the next chunk contained different  information.  A score of  one was given to explanations that mentioned "enough information",  "taking a breath" or "a pause", or a claim that it "sounds right". Zero was given if  the response was nonsensical. Table  16: Reasons for Chunking score % score % PRETEST experimental medium 10/27 37 control medium 09/27 33 experimental high 23/45 51 control high 13/27 48 experimental total 33/72 46 control total 22/54 41 POSTTEST experimental medium 17/27 63 control medium 16/27 59 experimental high 33/45 73 control high 14/27 52 experimental total 50/72 69 control total 30/54 56 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 23% 15% Each individual's possible score on this subsection was nine; the total possible score for  each group is 9 x the number of  group members. Therefore,  the experimental high's possible total was 45 (for  5 members), while all the other groups had a possible score of  27 (for  3 members). The experimental middle group scored 37% on the pretest and 63% on the posttest, an increase of  26%. The control middle group scored 33% on the pretest and 59% on the posttest, also realizing an increase of  26%. The experimental high group scored 51% on the pretest and 73% on the posttest, an increase of  22%. The control high group scored 48% on the pretest and 52% on the posttest, an increase of  4%. Overall, the experimental group increased by 23% and the control increased by 15%. Connectives within Sentences For the final  component of  the syntactic testing, students were asked questions regarding the role or meaning of  the connectives within the sentences they were asked to chunk: "What does the word "if'  do?", "What does "although" mean?" and "Why does it (the sentence) begin with "as"?" (questions #1 c, #2 c and #3 c). These responses were scored 0-2. A two was given if  for  the "if',  subjects explained that it gave a condition under which the other clause or thing would happen. A two was given if  for  the "although", subjects noted that that showed a contrast. A one was given for  each of  the above if  subjects gave an example of  how the "if'  and "although" were used, using them within a new sentence. A two was given for  the "as" if  subjects noted that it indicated simultaneity or cause and effect,  a one if  subjects indicated that it "tells what happens when..." And a zero was given for nonsensical responses. Table  17: Connectives  within Sentences score % score % PRETEST experimental medium 06/18 33 control medium 02/18 11 experimental high 04/30 13 control high 08/18 44 experimental total 10/48 21 control total 10/36 28 POSTTEST experimental medium 10/18 56 control medium 03/18 17 experimental high 13/30 43 control high 06/18 33 experimental total 23/48 48 control total 09/36 25 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 27% -3% Each individual's possible score on this subsection was six; the total possible score for  each group is 6 x the number of  group members. Therefore,  the experimental high's possible total was 30 (for  5 members), while all the other groups had a possible score of  18 (for  3 members). The experimental middle group scored 33% on the pretest and 56% on the posttest, an increase of  22%. The control middle group scored 11% on the pretest and 17% on the posttest, an increase of  6%. The experimental high group scored 13% on the pretest and 43% on the posttest, an increase of  30%. The control high scored 44% on the pretest and 33% on the posttest, a decrease of  11%. Overall, the experimental group realized an increase of 27%, while the control group dropped by 3%. Table 18 below compiles the results of  all components of  the syntactic testing. Table  18: Total  Syntax score % score % PRETEST experimental medium 39/84 46% control medium 31/84 37% experimental high 68/140 49% control high 53/84 69% experimental total 107/224 48% control total 84/168 50% POSTTEST experimental medium 53/84 63% control medium 41/84 49% experimental high 88/140 63% control high 49/84 58% experimental total 141/224 63% control total 90/168 54% DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 15% 4% The experimental middle group scored 46% on the pretest and 63% on the posttest, an increase of  17%. The control middle group scored 37% on the pretest and 49% on the posttest, an increase of  12%. The experimental high group scored 49% on the pretest and 63% on the posttest, an increase of  14%. The control high scored 69% on the pretest and 58% on the posttest, a decrease of  11%. Overall, the experimental group realized an increase of 15%, while the control group dropped by 4%. Table  19: Summary TOTAL SYNTAX COMPREHENSION experimental pretest 48% 55% experimental posttest 63% 55% difference 15% 0% control pretest 50% 59% control posttest 54% 45% difference 4% -14% experimental vs control 11% 14% The above table shows the relationship between the groups' performance  on the two sections of  the tests, comprehension and syntactic knowledge. Overall, in syntax, the experimental scored 48% on the pretest and 63% on the posttest, an increase of  15 %. In comprehension, the experimental group scored 55% on both the pretest and the posttest, maintaining initial performance.  In syntax the control group scored 53% on the pretest and 54% on the posttest, increasing by 1%. In comprehension the control scored 59% on the pretest and 48% on the posttest, decreasing by 11%. In summary then, while the experimental group increased substantially in syntax and maintained initial performance  in comprehension in spite of  the increased difficulty  of  the posttest, the control group increase slightly in syntax and decreased significantly  in comprehension. The Experimental Medium Group A quick comparison of  the two students who scored the extreme low and high scores in the CELF syntax measures can now be made. Experimental medium #5 (who had scored 118 in the CELF), scored 11/16 or 69% in comprehension on both the pretest and the posttest. This was the highest score in the experimental group on the pretest (the range was 4-11) and among the highest for  the posttest (the range was 6-12). In syntax, experimental medium #5 gained 21%, achieving 71% on the posttest. His scores were among the highest for  both the pretest—14/28 (the range was 8-16) and the posttest—20/28 (the range was 14-21). Experimental medium #3 (who scored 88 on the CELF) gained 18% in comprehension, achieving 56% on the posttest. His score was among the lowest in the pretest (6), but was the median for  the posttest (9). In syntax, experimental medium #3 did surprisingly well on the pretest, achieving 15/28. He gained 10% achieving 18/28 or 64% on the posttest. References made to the kind of  responses given by each, particularly in the area of  reasoning, will further demonstrate the gains made by each. The gains of  both these students indicate that the emphasis placed on structure while reading can benenfit  both a reader of  low syntactic ability as well as one with high syntactic ability. (Note that though experimental medium #3 did not improve his comprehension score, he maintained a high score despite the increased difficulty  of  the posttest.) The intervention also allowed the student with high syntactic ability to access his strongly established intrinsic knowledge to more clearly explain his syntactic reasoning. Results of  the Low Group Table  20: Chunking  Sentences—Low  Group score % score % PRETEST experimental low 08/12 67 control low 09/12 75 POSTTEST experimental low 12/12 100 control low 10/12 83 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 33% 8% The total scores for  both the experimental low and the control low have been calculated on the basis of  two groups members as both groups are composed of  two members. In the questions in which subjects were asked to chunk sentences, the experimental low group scored 67% on the pretest and 100% on the posttest, increasing by 33%. The control low group scored 75% on the pretest and 83% on the posttest, increasing by 8%. Table  21: Reasons for  Chunking—Low  Group score % score % PRETEST experimental low 05/18 28 control low 04/18 22 POSTTEST experimental low 12/18 67 control low 10/18 56 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 39% 34% When asked to provide reasons for  their chunking choices, the experimental low group scored 28% on the pretest and 67% on the posttest, increasing by 39%. The control low group scored 22% on the pretest and 56% on the posttest, increasing by 34%. Table  22: Connectives  within Sentences—Low  Group score % score % PRETEST experimental low 02/12 17 control low 03/12 25 POSTTEST experimental low 04/12 33 control low 02/12 17 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 16% -8% When asked to explain the meaning or role of  connectives within sentences, the experimental low group scored 17% on the pretest and 33% on the posttest, increasing by 16%. The control low group scored 25% on the pretest and 17% on the posttest, decreasing by 8%. Table  23: Total  Syntax— -Low Group score % score % PRETEST experimental low 22/56 39 control low 23/56 41 POSTTEST experimental low 33/56 59 control low 28/56 50 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 20% 9% Looking at overall scores on the syntax portion of  the tests, we can see that the experimental low group scored 39% on the pretest and 59% on the posttest, increasing by 20%. The control low group scored 41% on the pretest and 50% on the posttest, increasing by 9%. Table  24: Comprehension-—Low Group score % score % PRETEST experimental low 10/32 31 control low 10/32 31 POSTTEST experimental low 17/32 53 control low 09/32 28 DIFFERENCE BETWEEN PRETEST & POSTTEST TOTALS 22% -3% In the comprehension portion of  the tests, the experimental low group scored 31% on the pretest and 53% on the posttest, increasing by 22%. The control low group scored 31% on the pretest and 28% on the posttest, decreasing by 3%. Table  25: Summary—Low  Group TOTAL SYNTAX COMPREHENSION experimental pretest 39% 31% experimental posttest 59% 53% difference  20% 22% control pretest 41% 31% control posttest 50% 28% difference  9% -3% experimental vs control 11% 25% The above table shows the relationship between the low groups' performance  on the two sections of  the tests, the first  targetting the variable of  comprehension, the second targetting the variable of  application of  syntactic knowledge. Overall, in syntax, the experimental low scored 39% on the pretest and 59% on the posttest, an increase of  20%. In comprehension, the experimental group scored 31 % on the pretest and 53% on the posttest, increasing by 22%. In syntax the control low group scored 41% on the pretest and 50% on the posttest, increasing by 9%. In comprehension the control scored 31% on the pretest and 28% on the posttest, decreasing by 3%. In summary then, while the experimental low group increased substantially both in syntax and in comprehension, the control low group increased less dramatically in syntax and decreased slightly in comprehension. CHAPTER V THE INTERVENTION The intervention consisted of  nine whole class lessons over a five  week period focusing  on sentence structure. The concepts of  phrases, clauses, subjects, predicates and connectives were introduced and practiced through group activities and individual exercises. During the third week of  this instruction, guided reading groups began to meet, focusing  on reading science text while applying the concepts introduced in the lessons. There were three groups, sorted by reading ability, low, medium and high. The ten students who made up the experimental groups, low, medium and high, were assigned to the corresponding guided reading groups. The groups met for  a total of  five  weeks, one or two sessions a week for  a total of  eight sessions each. From beginning to end, the intervention took place during a seven week period. Table  26: Intervention  schedule Weeks lessons guided reading one 1 -two 2 -three 2 2 four 2 1 five 2 1 six - 2 seven - 2 The material for  the lessons were chosen using the same considerations as the selection of  the pre and posttest passages, but the topic had to be distinctly different  from  that of  the test material. Space exploration, a curricular topic, was chosen. The advantages of  this choice was that it was a topic covered by the grade six curriculum, there was a great deal of material on the topic from  which to choose, and there were many counter-intuitive concepts involved in space study. The disadvantage was that space is a popular topic and some students were likely to have a good store of  background knowledge. This disadvantage was compensated for  by focusing  on complex counter-intuitive physic principles. For example, one passage discussed the death of  a star during which the force  of  gravity crushes the core. The passage described the weight of  the core relative to its size, i.e., the density of  the core: At this stage, 15ml of  this matter would weigh the same as four  billion full-grown elephants. But gravity squeezes the core still smaller. In a fraction  of  a second, it may be no larger than a period at the end of  this sentence. (Darling, 1998, p. 62) Furthermore, two of  the passages used in the guided reading sessions for  the high and medium groups ("Saturn: Lord of  the Rings" and "Black Holes and Time Tunnels") were taken from  a grade six anthology, Space, Stars and Quasars in the Collections series. And one passage ("Sojourner and Beyond") used in the low reading group session was taken from  the grade four  Collections anthology. This served to create a more natural setting for  the intervention. Guided reading groups met for  20-30 minute periods. Students were asked to read the selections to themselves in small chunks, either in paragraphs, or at times to the end of  a page. For the first  few  sessions, they were asked to find  connectives and discuss the role of the connective in the passage. In the next few  sessions, they were asked to read smaller portions aloud (1-2 paragraphs) and choose one chunk and ask a question for  which their chosen chunk would provide an answer. In the final  sessions they were asked to read the smaller portions silently and summarize the portion, collaborating on the sentence by providing a clause and thereby building a summative sentence together, supplying connectives and punctuation where necessary. These contributions were made orally with the teacher recording them on a white board. Think Aloud modeling was used throughout the sessions. Emphasis was placed on thinking while one reads. Identifying  chunks was presented as a means of  slowing down and pausing to think about what each chunk means and how it interacts with the rest of  the information  in the passage. In each reading session, the focus  quickly shifted  from  discussing syntax to discussing content. For example, in reading one sentence the teacher stopped to ponder one chunk. teacher: The sun, a mid-sized star.... 'a mid-sized star', it says. What does that imply? experimental high #6: There are a whole lot of  other stars even bigger than the sun. Other students agreed and the discussion went on, revisiting information  learned in previous lessons about the huge dimensions of  the sun relative to those of  earth, and tying it into further  information  from  the passage which was about black holes. The teacher emphasized the importance of  slowing down to catch each chunk because, especially in science text, each phrase and clause can give so much information. While the experimental class was engaged in the intervention, the control class also pursued studies in space exploration. The study primarily consisted of  reading the grade seven prescribed text, Science Probe. The teacher concentrated again on guiding students through the text using strategies listed in Reading  44. Since this teacher was also using the Soar  to Success  program with her grade sevens, she also used the strategies taught in the program with the class during science classes. These whole class lessons occurred approximately twice a week for  45 to 60 minutes. Both the control and experimental class spent two periods in experimentation during the study. Qualitative Data Gathered During the Intervention The first  goal of  the intervention was to develop an understanding that text is made up of  words and that these words play different  roles. The focus  was on the two basic types of words that make up text: form  and function  words. The groups were given a set of  nine word cards: pull, still, shine, come, if,  globe, sun, fire,  hydrogen. They were asked to sort these and name their categories. Four groups put them in alphabetical order. When the request was made that they group, rather than sequence the words, they still grouped by initial letters e.g. sun and shine, fire  and finally,  and then made a miscellaneous group. One group grouped by meaning, for  example, placing "sun" and "shines" together because the sun shines and "hydrogen" and "fire"  because firefighters  sometimes use hydrogen to put out fires.  One group attempted to group by nouns and verbs. When trying to deal with the words that didn't fit,  experimental high #10, said "That's all the type of  words there are, aren't there?". They were unable to come up with further  word types themselves even though they'd had spelling exercises which required them to group adjectives. This group gave up and sorted by initial letters. The frustration  this group suffered  as well as the general inability of  the students to go beyond alphabetic sorting is indicative of  the huge difficulty  students have in classifying words and the lack of  understanding they have in the shifting  roles a word can play. This is a continuing struggle experienced by the class from  which the experimental group was drawn even months after  the study concluded. When asked to brainstorm nouns, verbs, adjectives and adverbs beginning with a designated letter and told they must use one word from  each category to compose a sentence, students used "excited" which was listed under verb as an adjective, e.g. "Eddie was excited about Easter". When it was brought to their attention that they had not used a verb beginning with "e", they argued they had used "excited" and it was on the verb list. They did not understand that by using the word in a certain way, they had given the word a new role—the role of  adjective rather than verb. Students have a great deal of  difficulty  coming to realize that a word does not belong to a certain category, or perhaps categories (e.g. fire  can be both a noun or a verb), but rather that words play roles which are defined  by their relationship to other words within a sentence or text. This renders the teaching of  each part of  speech to young students extremely difficult, and ultimately not worth the effort.  Though Adler recommended the aquisition of  such knowledge, he was writing for  adults, and he concludes that the ability to formally  diagram a sentence is not necessary. What is necessary is an understanding of  how words are grouped in a sentence and the relationships between groups. This is why the emphasis was placed on the difference  between form  and function  in the initial lesson of  the intervention and on subject and predicate throughout the study. It was noted that generally a subject contains a noun and a predicate commences with a verb, but these labels were not focal  points. The chunks that were comprised of  a subject and a predicate, i.e. the clause, and those chunks which modified it, i.e. the phrases, were the focus. The second goal of  the intervention was to establish an understanding of  what a sentence is. In the second session, the students were given a selection of  six phrases and sentences (see below) and asked to identify  which were sentences and which were not and state why. They worked on this in groups. There were six groups. The number of  groups designating each as a sentence is recorded to the left. two 1. Saturday at the pool, one 2. I jumped, one 3. Running to the car. none 4. Roger and me together, five  5. He ate lunch. five  6. Fifty-five  people, all dressed in their best clothes. The groups that voted for  #1 said it "sounded like a sentence". A member from another group pointed out that it "has a thing, but no action". Number two was largely agreed upon as being "too short"; students noted it "doesn't make sense". Number 6 had both a thing and an action and therefore;  they believed, it must be a sentence. Thus, those groups of  words that were long enough and made sense were designated sentences. They also had a sense of  sentences needing a "thing", i.e. noun and an "action", i.e. verb, but they had little sense of  the flexibility  of  these roles. Although they began with an understanding of  the sentence expressing a complete thought and having a noun and a verb, they had a difficult  time assessing the completeness of  a thought and the use of  verbals in the context of  the sentence. The fact  that a word that normally served as a verb could serve as an introduction to an adjectival phrase was misleading to them. The words "noun" and "verb" were supplied by members of  the class. By the time the class got down to discussion question #4, a student, experimental medium #3 noted that it "needs an adjective or verb", displaying a level of  syntactic awareness consistent with his high scores on the CELF subtests. It should be noted that this student had been in the researcher's class in his grade three year when the class did do a short study of  sentences. The third lesson introduced the concepts of  subject and predicate. Students were asked to identify  sentences from  a sheet of  thirty sentences and phrases. They then responded as the sentences were read out with a thumbs up thumbs down. The majority were able to identify  sentences, and then identify  subjects and predicates. The fourth  lesson focussed  on phrases and how they differed  from  clauses. In the lessons that followed,  connectives were presented as words that showed relationships between chunks, i.e. between clauses. A variety of  connectives and the different  types of  relationships they signalled were introduced. After  laying the foundations  for  the concepts to be used, intervention lessons focussed on chunking sentences which required students to isolate clauses and phrases, using commas and connectives as guideposts. This was fairly  simple, as most students had been taught in primary grades that one "takes a breath" or pauses at a comma when one is reading aloud. The comma was now presented at a place to pause for  thinking and the chunks framed  by commas and other punctuation were to be understood as being chunked for  meaning as well as sound. From the first,  students were quite competent, averaging 81% accuracy on chunking individual sentences and identifying  the connectives within these sentences (for  26 papers). This is a fairly  sound average as it included the scores of  the SLD students of  the class. The intervention moved on fairly  quickly to chunking sentences and identifying  the connectives within a paragraph. This proved much more difficult.  The students averaged 48% (26 papers). This might have been partly due to the difficulty  in determining what a connective was. Guided reading groups had commenced by this time, and students were encountering difficulty  picking out connectives from  the passages they were reading. Initial requests for students to pick out one key word and one connective in a given passage met with success. Repetition of  a commonly used connective (e.g. "but" in a passage contrasting earth and Saturn) in a passage also made identification  easy. But when faced  with a passage where there was no distinct pattern, requests for  contributions to a list of  connectives elicited varied responses. Students focused  on small words and often  suggested prepositions such as "it", articles such as "an" and short verbs such as "is". This was particularly true of  the lower reading group. However, when students were instructed more specifically  to identify  only connectives at the beginning of  sentences and after  commas (i.e. those prefacing  clauses which tend to be the ones which clearly show relationships between clauses), student performance  was much better. The average on individual worksheets increased to 86% (23 papers). Again, the use of  punctuation (commas and periods) as guideposts simplified matters. Generally, picking out connectives did not prove a helpful  exercise. Focussing on individual chunks and occasionally referring  to the connectives to see how a chunk related to the entire passage was more helpful. In both individual worksheets and in guided reading sessions, the use of  questioning to identify  the meaning of  chunks and their relationship to the passage of  the whole was introduced. (There is insufficient  data from  worksheets to score and comment upon.) This' was a struggle for  some. In guided reading, the process of  reading a paragraph, targetting a chunk and then posing a question that would be answered specifically  by the chunk was modeled by the teacher. Students were then asked to repeat the process. This met with variable success. Students in the high group were able to do this, but those in the lower groups were prone to asking more general questions, or to attempt to ask questions which would be answered by "yes" or "no", a problem generally encountered with weaker students when one works on questioning. An emphasis was then placed on thinking when one reads, considering what each chunk means. The teacher modeled this through the Think Aloud method, where she read a paragraph chunk by chunk, pausing to note what each one meant in light of  the other. At the end of  the paragraph, the teacher wrote a sentence summary on the board. Students were invited to do this, but thinking aloud, a self-conscious  process, proved embarrassing to students. Students preferred  to read a paragraph, then mention some chunks they noticed and what they meant to them. Students were then asked to help compose group sentence summaries. In working on these, students were asked to contribute ideas in the form  of  clauses. These clauses were fitted  together into a sentence with the help of  connectives provided by the students. This was a collaborative activity utilizing all the concepts studied: subject, predicate, clause, phrase, connectives, in a generative task. Students obviously enjoyed this task and suggested several of  the connectives frequently  encountered in reading, specifically  "but", "although" and "therefore".  Experimental medium #3 was particularly active in these sentence compositions and it was noted by several students that "therefore"  was now his "favourite  word". Thus students within the class went from  the general understanding of  a sentence as a groups of  words that is long, makes sense and should include a thing and an action, to a working understanding that one can combine a number of  clauses to express complex scientific  principles and relationships and a blossoming ability both to read and compose such sentences. Instruction focussed  on syntax—reading chunk by chunk and making sense of  each chunk while using connectives to check for  relationships—quickly developed into lively and engaging discussion of  the passages, engendering questions like "What does it really matter if  we find  out about black holes, anyway. I mean, besides being interesting, can it really affect  us?"(from  a member of  the class not included among the participants of  the test) and comments like "That was a great conversation"(from  experimental medium #3). Structure became a key to developing the semantic flow  of  the text. The building of  the text base led to and was at times simultaneous to the formation  of  a situation model. Following the structure of  the sentences and paragraphs that framed  the passage allowed students to construct meaning from  the text and to manipulate their new knowledge and assimilate their previous knowledge, using both to produce summary sentences. Granted, often  the sessions merely allowed the group to provide sentence summaries for  individual paragraphs. In this way the group was establishing the first  stage—the locally coherent network mentioned by Tapiero and Otero (1999). Often  time did not allow the group to read an entire article in one sitting and they did not progress to the second stage- the hierarchical sequence. But the lack of  the hierarchical sequence did not prevent students from beginning to construct a situation model. They freely  discussed ideas sparked by the text and often  offered  the imaginative "what if".  This supports the theory that the. two levels of comprehension—text base and situation model—are not constructed sequentially. Still an understanding of  the complete text is important, as confusion  arises when important details are missed. This confusion  arose when reading "Black Holes and Time Tunnels". It is a fairly  long (4 pages), fairly  dense article. In one section on the third page, reference  is made to "Cygnus X-l" as a significant  entity to be considered for  the remainder of  the article (Darling, 1998, p. 63). No one (including the teacher) could recall what Cygnus X-1 was as it had been identified  on the first  page of  the article which the group had read two days previous. All the groups members were forced  to do a search through the article to find out who or what was Cygnus X-l. Students need to be able to read science text and construct a sound but flexible  text base, recognizing and incorporating all the main points contained within the text. The goal of reading instruction is to enable students to do this independently. But solitary reading often proves insufficient  to move beyond the text base to constructing a situation model. Such a move will require discussion. Discussion strengthens comprehension for  readers of  all levels of  age and experience. It is the foundation  for  learning in post secondary institutions—one is to read the text and come to class prepared to discuss it. Discussion may well be essential for the younger, less experienced reader. Guided reading has been developed as a format  to guide students in both processes. The text base is collaboratively constructed during reading—often with the teacher modeling the process. The discussion which both accompanies and follows reading should also promote the building of  a situation model. The guided reading sessions in the intervention provided evidence of  both constructive processes and it was the initial focus on language structure which facilitated  the processes. CHAPTER VI DISCUSSION AND CONCLUSIONS In summary, this study was designed to investigate whether teaching students to approach scientific  texts using their knowledge of  language (specifically  syntax) would help them better understand the passages they are reading. It has long been taken for  granted that the study of  grammar is unhelpful  in teaching students to write (Elley, Barham, Lamb, & Wyllie, 1976) . Much of  the research in the past decades has been used to provide evidence that the study of  the structure of  our language restrains the generative, creative process. The reading process, however, must be admitted to be a more analytic process. Constructive, yes, in that the reader does more than merely receive meaning from  the text; rather, he or she constructs meaning from  it. But the text is a whole which must be first broken clown into parts in order for  the reader to reconstruct  the frame.  Simultaneously, the reader is adding to the text frame  from  his or her own store of  knowledge, and the final product of  the act of  reading will be the construction of  something new and unique to that reader. The theoretical framework  upon which the study is based is van Dijk's and Kintsch's concept of  discourse structure which represents reading as a process involving the building of both a text base, the aforesaid  "frame"  and a situation model, the new and unique product of the reader. William Adler's theoretical framework  upon which he based his book How  to Read  a Book  also serves as a point of  reference.  Adler describes reading as an active process that requires taking apart an argument proposition by proposition (a process which he says requires a certain amount of  grammatical knowledge) in order to fully  comprehend it and reconstruct it. Such a complex analytical approach is usually unnecessary when reading fictional narrative, but the reading of  content area text presents different  and greater challenges to the young student, and, of  all content areas, science can present the most complex concepts. Processing science concepts becomes even more difficult  when the concepts are counterintuitive, as in the study of  forces  in physics. The questions addressed in this study were the following: 1. Is direct teaching of  syntax effective  in raising syntactic awareness in young students? 2. Can young students be taught to apply syntactic knowledge during the reading process? 3. Will an increased awareness of  syntax and application of  syntactic knowledge be accompanied by an increase in reading comprehension of  science text? In the course of  the study the experimental group received a pretest and a posttest testing syntactic knowledge and reading comprehension and a seven week intervention which included whole class lessons teaching students to recognize sentence structure, chunk sentences and use connectives as cues for  relationships between propositions. The intervention also included guided reading sessions applying the syntactic concepts to reading science text. A control group received only the pretest and posttest. Results indicate that the experimental group improved in syntactic knowledge and maintained their scores in comprehension, while the control improved slightly in syntactic knowledge and, on the whole, did not improve in comprehension. Furthermore, observation notes from  the intervention indicate that the intervention allowed students to gain a sense not only that they needed to adjust their pace of  reading while reading dense science text, but a notion of  how and where to do so, using commas, connectives and their knowledge of phrases and clauses as points at which to stop and think. What follows  in this chapter is an interpretation of  the findings  and the conclusions that can be drawn in light of  the research questions. The quantitative data composed primarily of  the pretest and posttest results and the qualitative data in the form  of  the observation notes on the intervention will be discussed separately. Question One Is direct teaching of  syntax effective  in raising syntactic awareness in young students? The syntax portion of  the pretest and posttest served to answer this question as well did some of  the quantitative data collected during the intervention. In the intervention, scores on the worksheets covering the concepts of  what a sentence is composed of,  how one chunks a sentence and how one identifies  connectives indicated that direct teaching was effective. Students achieved an 85% average on the scores after  the directions for  identifying connectives was directly tied to the use of  commas as cues. Scores on the syntax portion of the pretest and posttest will be interpreted by subcategory first  and then as a whole. Refer  to chapter 4 for  all the tables mentioned below. Chunking  (Refer  to table 15) Generally the performance  on chunking tended to be fair  to excellent from  the outset, the scores on the pretest ranging from  61% for  the control medium group to 100% for  the control high group. This matches the high performance  of  the grade threes in the preliminary study (90% for  the high and low groups, 70% for  the medium) and supports the conclusions drawn from  that study that students are aware of  commas as signposts. Whereas the experimental medium group realized no gain, they had outperformed  the control medium group substantially on the pretest (78% for  the experimental medium and 61% for  the control medium) and with their gain, the control medium group just managed to match the performance  of  the experimental medium group (achieving 78%). Still, one wonders why the experimental medium group did not improve. The experimental high group, on the other hand, did improve, perfecting  their performance.  The control high group dropped substantially. Such a drop is strange, but is certainly partly accounted for  by the small size of the sample and the reaction of  these particular students to testing. It was noted by both the classroom teacher and the researcher that one of  the control high group demonstrated a fair bit of  resistance to the posttest, feeling  frustrated  with the difficulty  of  the test. Another of  the control high group had a chronic tendency to allow his mind to wander, according to the classroom teacher, and was in that particular space on the day of  the posttest. But the familiarity  with the task resulted in an increased comfort  level within the control medium group (as was expressed by some to the researcher) and should have had a positive effect  with the control high group, particularly since the syntax portion of  the posttest was almost identical to that of  the pretest and therefore  could not be said to be significantly  more difficult. Chunking simply required students to recognize the commas as cues. The questions asking students to give reasons for  their choice of  where they chunked served to determine what meaning subjects understood the commas to be cueing. Reasoning (Refer  to table 16) Students are taught at an early age to pause at commas for  "breaths" when reading aloud. Evidence of  this is provided in the comments of  the grade threes in the preliminary study. Results of  the thesis study indicate that many of  the subjects both in the experimental and control groups came to the study with this understanding. However, the intervention emphasized that these cues also indicate boundaries for  chunks of  words that belong together because they express a whole thought i.e. a clause or proposition, or they contain information modifying  a thought. Results of  these questions indicate that the direct teaching was effective,  with the experimental medium realizing a 26% gain and the experimental high a 22% gain. The control high realized a slight, 4% gain. The surprise this time is that the control medium realized the same gain as the experimental medium a 26% gain. This points to a difficulty  in interpreting and scoring the responses given in the section. There was an effort  made to avoid marking responses simply in relation to one another as was done in the exploratory preliminary study. Furthermore, with the greater cognitive development of  these older students, it would have been simply impossible to assign points for  every combination of  reasons given by the subjects; therefore,  certain words were used to create delineations among levels for  point attribution. Two points were awarded for  the response that a division was made because there was a comma there. It is to be expected that the experimental group is identifying  the comma as a boundary for  a meaning chunk while the control group is still interpreting the commas as a place to pause for  breaths. But there is still a question as to what one is to be doing while inhaling. Does the pause serve as physical preservation, as an opportunity to use silence for emphasis as one would in music, or does it serve as a cognitive cue—time to pause and reflect?  This could be the understanding behind the "pause" or "break" reasoning. In the case of  those who have undergone the intervention, it should be. This is sometimes but not always evident in the oral responses and the benefit  of  the doubt was given to those using the proper term for  that point of  punctuation. The difference  in understanding between the experimental group before  and after  the intervention and between the experimental and control groups at the posttest becomes more obvious when considering individual responses in which subjects are working through the ideas in an effort  to express their reasons. In the control group we see that the pause is indeed largely related to breaths. Control high #6 says on the posttest: "...there should be a pause there. That's where you're supposed to take a breath", and in a second response this student asserts that you chunk "Right where the comma is". Control medium #5 says on the pretest that a comma "Make's you... not like stop, where's... If  there's a period, you stop and take a breath. If  there's a comma, you, like, stop for  a second. Not as long as you would if  there's a period." On the posttest the same subject simply gives as his rationale for  chunking "Because there's a comma there." No further  clarification  as to what a comma does is given. In the case of  control medium #3, there seems to be some development of understanding. On the pretest she gives nonsensical responses when asked to give her reasons for  chunking. ""Cause, "although the specks of  dust", well, you can see specks of  dust, "are small, the particles that make up matter" ... well, 'cause particles are small, so I thought that would go together. And, uhm, I just thought, "are much smaller." On the posttest, she is much clearer, using specific  terms (comma) and thus receives more points. "Because 'although many people think that temperature', I thought there was a good sentence, and then heat 'and heat are exactly the same', since there's a comma there, I think I should put a line there and then." However, though she is identifying  commas, she clearly does not understand how the chunks framed  by commas are grouped to make meaning since she identifies  as a "good sentence" something which splits a relative clause. Among the experimental group, a growth in understanding is evident, even between responses accorded the same point value. Experimental high #8 was fairly  confident explaining chunking from  the outset: "Because I think that's where you should understand it. Because if  you, if  you kept going, it would be, it would be way too long, and you could get confused, like 'the windows are closed' and that... So you could get confused  if  it was too long. So you just chunk it where you think a small sentence would be... That's why I chunked it there." This response was given a three because of  the reference  to small sentences which indicates some understanding of  clauses and sentences as complete thoughts. In the posttest, however, experimental high #8 is able to further  explain the structure of  the clause. Uhm because the first  chunk has—uh, what is it called? It's called the ... Oh! I can't remember what it was called—the... predicate! It has a predicate and a subject. So after  the predicate and the subject you chunk... you chunk it. Later he more fully  explains the combination of  clause and the comma as a clausal boundary marker: Because that—it has a subject and what the subject does. And then after  that it's going on to a new subject, so you chunk it, put a comma or some marker. Experimental high #10 also demonstrates a growth in understanding in that she mentions the use of  a comma for  chunking in the pretest, but is obviously adding to that the understanding that a comma can serve as a clausal boundary marker even though it's still quite shaky. Well, there's a comma right there. And... I don't think I should of  chunked it there because there isn't' really... like the- like there's a subject, like a "rate", but there isn't a predicate. There was, in fact,  a predicate and she had, in fact,  chunked correctly, and though she could not consistently recognize it, she knew what she needed to look for  to make a complete thought—a clause. Experimental medium #3 (the student who has been noted for  his high syntactic ability as reflected  in the CELF measures) was able to identify  both a comma and a connective and to use the language, but because he did not mention clauses, he was awarded only a two for  his response. However, the confidence  with which he applies the principles taught in the intervention is evident. ... Well, I chunked it there because, for  one, there was a comma, and two, there's a connective at the beginning so I know somewhere around there it's gonna be—there's gonna be a chunk, and 'If  the rate at which the particle move changes' makes a lot more sense than 'If  the rate at which the particles' would make. Here he goes beyond the signposts of  the comma and the connective to demonstrate which grouping of  words constitutes a complete thought and which does not. Having conducted this study, the researcher would be better prepared to predict what responses one could expect after  the intervention and would revise the marking categories for the reasoning responses. But to go back and revise the pretest marking on the basis of experimental posttest responses would not be fair  to the control group. Moreover in future studies, it would be helpful  to follow  up the response with a question, "What does the comma tell you?" The comments of  the experimental group members clearly demonstrate an understanding the role of  commas and connectives as markers. Connectives  (Refer  to table 17) These questions dealing with the role of  connectives were the most difficult  for students. The students were more hesitant in their answers and often  asked for  clarification. For some reason, the experimental high scored very low on the pretest and though they realized the greatest increase, 30%, they still did not achieve the same level of  performance as the experimental medium group on the posttest (43% for  the experimental high, 56% for the experimental medium), or that of  the control high group on the pretest (44%). The experimental high group did include two students reticent to make educated guesses. They were unwillingly to go beyond giving succinct responses of  which they were certain. The group also included experimental high #8, a student who is known to be very hard on himself and who found  this category of  questions extremely difficult  on the pretest. .He tried desperately to explain his thinking. In response to "What do the words 'even if'  do?", he responds (Note: square brackets indicate researcher's comments): Makes you think because... that's like even if,  uhm... it could not happen or it could... like if  it happened... I don't even know what I meant by that... even if  it... (sniffing,  close to tears) So that even if  made me think that... ehum, what am I trying to say? It's hard to say... [It is hard to say.] I'm thinking... Even... Holy crap... [You want to pass on that?] Yah. It's hard. However, the overall improvement of  the experimental group, 27%, indicates direct teaching of  the role of  connectives was effective,  even though the end performance  was still a mere 48%. It indicates that the instruction had laid an initial foundation  upon which further instruction could be built. Experimental high #8's remarks demonstrate he had received a degree of  syntactic understanding sufficient  to preserve him from  the discomfort  he experienced on the pretest. He hadn't forgotten  his earlier experience, either. If—uh?  Oh, I had this before!  I think 'if'  says, means that there's two things, like it can be this, but- or... Like 'if  this happens...'—there's a consequence or whatever. Like that. On the pretest, experimental medium #5 (the student who has been noted for  his low syntactic ability as reflected  in the CELF measures) is unable to give a rationale explanation of  the role of  "even if'. Uhm, they like... tell... they, uhm, like tell what the thing is. Or, they like explaining the object that you're talking about or whatever it is that you're talking about, and that's all. On the posttest, he is much clearer in his explanation of  the role of  "if'. It kind of  explains a possibility of  something... because you could, like, 'if  it doesn't work... but 'if'  can also mean... other things, but I don't know what they can also mean. His confidence  wanes, and he struggles to explain the other connectives, but he has an initial understanding. He has benefited  from  the intervention, but he needs more guided practice. Experimental high #10 shows that she understands contrasting role of  "although" by comparing it with another connective also encountered during the intervention, "however". It's like how- like however. Like, even though that people think that, that the temperature and heat are the same, they really aren't, so it's like however... Further evidence of  this blossoming understanding of  connectives continued to arise within the classroom after  the intervention was formally  brought to a close and as the students worked on writing sentences. When students were asked to make sure they were writing interesting sentences, rather than short, simple and uninspired sentences, experimental medium #4 asked, "You mean using connectives?". Total  Syntax  (Refer  to table 18) Overall, the scores among the experimental group rose consistently. The results among the control group differed  between levels. Though the control medium realized a gain (12%) which was not substantially lower to that of  the experimental groups (17% for  the experimental medium and 14% for  the experimental high), the control high dropped (by 11%). All levels of  the experimental group, including the low group which will be considered presently, realized gains whereas the control group growth was erratic. Low Group Data The data from  the low group supports the conclusion that direct instruction can raise syntactic awareness. The gains made in scores on the chunking questions clearly indicate that the experimental low group were able to apply the concrete cues for  chunking, gaining a perfect  performance  along with the experimental high group. The control low also increased their performance,  achieving 83% on the posttest. In the reasoning category, both the experimental low and the control low group made substantial gains (39% and 34% respectively), but again, the difference  of  understanding of  comma made marking difficult.  In both pre and posttests, control medium #1 notes that she chunks because of  commas, using the proper term, and then mentions taking "a breath". For the experimental group, the fact that they are directly applying instruction received in the intervention is obvious. In the pretest, experimental low #1 relies on explanations such as "it just sounds right", but in the posttest, experimental low #1 says she would chunk it at the given spot "Because there's a comma there, and I remember during class when we were chunking sentences, hum, you said that people usually chunk at commas." Experimental low #2 notes on the pretest that "it kinda makes sense". On the posttest he mentions making sense again, but goes on from  there. I think that it may have been able to chunk it in one spot, but I thought that that's where it made the most sense to me. And, why would I chunk it there? I would chunk it there so that you wouldn't get a ton of  information  right, right there. And again the comma, where you, you said you should chunk it. Thus, both subjects in the low group were able to restate the idea that one chunks at a comma, and experimental low #2 was beginning to combine that with the fact  that one must consider the amount of  information  contained within a chunk. The experimental low group also realized a substantial increase in the connectives category (16%), though not as substantial as the other experimental groups. The interesting point here is that, in the experimental group, there is a pattern that emerges which directly correlates with ability levels. Increased understanding of  connectives was ranked 30%, 22%, and 16% for  high, medium and low groups respectively. In the control groups ranking was less consistent with levels, an increase of  6% for  the medium group, a decrease of  8% for  the low group and a decrease of  10% for  the high group. Here the use of  percentages tend to make the difference  sound greater than they are. Since the total score is 18 for  three participants, a 10% drop is only a drop of  two points—the control high scoring 6 on the posttest compared to 8 on the pretest. The medium group gained one point and the low group lost one. One can safely  say that there was no substantial difference—either  growth or loss of understanding—for  the control group regarding connectives, whereas the experimental group did gain in this respect. Overall, the experimental low group benefited  most substantially of  all experimental groups from  the direct instruction, experiencing a 20% increase in the case of  the syntax portion. Regarding comprehension, data was incomplete for  this group. Question Two Can young students be taught to apply syntactic knowledge during the reading process? Though this question is partially absorbed in the third question, "Will an increased awareness of  syntax and application of  syntactic knowledge be accompanied by an increase in reading comprehension of  science text?", it warrants consideration on its own. However, it is primarily answered in observation of  the process demonstrated in the guided reading groups. A good deal of  the interchanges among the medium and high reading groups provide an affirmative  to the answer. (See more detailed discussion of  the intervention in the previous chapter.) Fewer of  these interchanges occurred while the low group met. This is not altogether surprising, as the low group was comprised largely of  students who were far  less sure of  themselves and far  less willing to take risks and try new things. However experimental low #2 began to identify  connectives and use them in sentence summaries by the end of  the intervention, and he proved to realize the second greatest increase in the syntax portion (an increase of  29%, the highest was that of  experimental high #10, 36%), and the highest increase in the comprehension portion (44%—where the average for  the group was 0%) although this increase is partly due to the missing responses in his pretest comprehension. Question Three Will an increased awareness of  syntax and application of  syntactic knowledge be accompanied by an increase in reading comprehension of  science text? The results of  the pretest and posttest are inconclusive, because of  the increased difficulty  of  the posttest. Overall, the experimental group increased in syntactic understanding scores and maintained their comprehension scores in spite of  the increased difficulty  of  this component. The comprehension scores of  the control group dropped considerably. Comparison of  the comprehension questions (see Appendix B) reveals that the first two questions in the pretest request information  that is explicitly stated in the passage, the third question requires the reader to supply a proposition and the fourth  requires a comparison of  two processes described in the passage. The posttest includes no questions which can be answered with propositions explicitly stated in the text. The first  and fourth questions require information  that is implied, but not stated. The second question requires readers to synthesize information  supplied in three sentences. The question is "How does matter change state?" The three sentences in the passage are: The energy of  motion is called kinetic energy. If  the rate at which the particles move changes, they lose or gain energy. When kinetic energy is lost or gained, the water changes (Posttest passage) The third question requires the reader to compare two propositions, one defining  heat and the other defining  temperature. Regarding the levels, the experimental medium group gained one point or 2%, the experimental high lost one point or 1%, averaging to the same percentage as the two groups realized on the pretest. The control groups both dropped substantially, control medium losing 10%, control high losing 19%, averaging to a loss of  14%. The experimental low group gained 20% in syntax and 22% in comprehension, while the control group gained 9% in syntax and dropped 3%in comprehension. The results indicate a relationship between syntactic knowledge and reading comprehension, but further  research in which the levels of  difficulty  of  the pretest and the posttest comprehension items are more closely matched is necessary to further  clarify  the relationship. -Conclusions It must be noted that this type of  comprehension testing was familiar  to both groups as teaching students to read for  information  and give full  responses to questions based on information  text was a schoolwide literacy emphasis. Strategies to encourage students to check for  understanding, etc. were to be a part of  each classroom's study. Such strategies can be very effective.  However, they still need unpacking for  students who are not familiar  with them. To instruct a student to check for  understanding is fine,  but how does a student know what it means to understand and how one can be certain they have understood what the author intended them to understand? It is fine  to ask students to slow down when text becomes dense, but what does it mean to slow down? Does it mean to sound out the words more deliberately, or does it mean to pause and reflect?  And if  they are to pause and think, where should they pause? and what should they be thinking? The results of this study support the use of  punctuation points as places to stop and think, not just general thoughts, but thinking why that group of  words was put together, and what the words mean on their own and how they are related not only to the rest of  the sentence, but the rest of  the passage. The study also shows that in this meaning making endeavor, connectives, the little words which can be found  at the beginning of  chunks, can help to point out these relationships and make them clearer. Commas and connectives served as "buoys", something for  the students to grasp ontoin a sea of  text. These structural elements can also provided a cue for  inferencing  which will make connections both within the text and with information outside the text- a bridge between the text base and the situation model. The atmosphere in the class was one of  inquiry and discussion, where the paragraphs of  texts were taken apart and made sense of.  They were thoroughly discussed for  internal meaning and alongside the text was lain background information  students possessed previous to the study as well as facts  gathered collectively through the reading of  other texts in the classroom. Both the quantitative and qualitative measures support that language structure is a useful  focus  for  reading instruction. Limitations The results of  this study cannot be generalized due to the small number of  18 participants. A wider population would of  course require a greater number of  instructors trained in the intervention methods. The researcher was also limited in time and flexibility  by the demands of  the school curriculum and school schedule. Testing conditions were not optimal (due to limited space and supervision). In future  studies, it must be emphasized to participants that they must fill  in all response blanks. The increased difficulty  of  the posttest comprehension test should be amended in future  studies. It would also be helpful  to have pretests and posttests adapted to the various ability levels of  the participants to ensure the low group was not working at their frustration  level. In both the pretest and posttests, it would be helpful  also to be very deliberate in testing the various levels of  inferencing,  those essential both in building the text base and the situation model, and to ensure that these questions are as closely matched as the syntax items. Recommendations It would be very interesting to pursue these studies at a variety of  different  levels, from  grade four  through middle school, to explore the possibility of  teaching young students to use commas and connectives at a beginning level and then presenting students with more and more sophisticated use of  syntax throughout high school. What effect  might this have on the performance  they would be able to achieve at the post secondary level? Young people are expected to arrive at university equipped to read and understand complex information  text. Perhaps reading instruction which makes use of  the structure of  the language as a tool for analytic reading will allow them to do so. Instruction during the intervention culminated in collaborative sentence summaries. It would be helpful  to go further,  teaching students to write individual summaries of  text, focusing  on creating clear hierarchically organized propositional frameworks  of  the text they read. This is a tool for  further  developing one's understanding of  text, but it is also the beginning of  responding to text and of  communicating one's own original ideas. Literacy encompasses both reading and writing and it is the use of  the structure of  language as a means of  organizing thought that is the theoretical foundation  of  this study. Such organization is essential in producing clear, powerful  exposition. Our language is designed to guide us through the ideas it is intended to express. Students must learn both to read and to write in full  control of  that language. What role does grammatical knowledge play in reading comprehension? The making of  meaning is the goal of  reading. In the constructive process of  making meaning while reading, grammar provides both the building blocks and the tools necessary for  the construction. Construction takes two phases, the frame  of  the house and the finishing  of  the house. Knowledge of  the structure of  language must be applied in both phases of construction in reading: the text base and the situation model. Without such knowledge, the reader may be unable to construct a coherent representation of  complex text. Students begin school required only to build simple houses whose basic elements hold no mystery. But if  they are to construct more elaborate, more sophisticated text models, if  they are to contemplate the construction of  a castle or a cathedral, they will require more sophisticated tools and skills. It must be the goal of  reading instruction to equip students with the necessary tools and training. APPENDIX A Reading  Test  One Every time you move any part of  you, you use your muscles. You even use muscles when you read. There are more than 600 muscles in your body. Some are deep inside. Others lie just under your skin. Some are big and some are very small. Hold your arm straight out in front  of  you with the palm of  your hand facing  up. With your other hand, feel  the upper part of  your arm, the part between your elbow and your shoulder, you can feel  a muscle there. Make a fist  with your right hand. Bring your fist  up to your shoulder. You can feel  the muscle in your upper arm change. It bunches up and gets harder. This muscle is called the biceps. The top end of  your biceps is fastened  to a bone in your shoulder. The other end is fastened  to a bone in your lower arm just below your elbow. When your biceps bunches up and gets hard, we say it is contracting. It is drawing itself  together. When it does this, it pulls on your lower arm, your elbow bends, and your fist  comes up to your shoulder. Your bicep pulls your fist  up to your shoulder. When you want to unbend your arm, you use another muscle. It is called the triceps. Your triceps is on the back side of  your upper arm. It also reaches from  your shoulder bone down to a bone in your lower arm, just below the elbow. When your triceps contracts it pulls your arm straight again. Your biceps and your triceps work together. First one pulls, then the other. Many other muscles in your body work in pairs that way. They pull back and forth. 1. What is the passage about? prompt: Can you tell me more? 2. Show me how your arm muscles work together with your arm and explain as you demonstrate. 3. Where was it hard to understand what you were reading? Show me the first  place that it got hard to understand, where you had to slow down. prompt: How far  back did you go? 4. What did you do to figure  it out? 5. (collapsed with #3) How did you know when you had found  what you were looking for? 6. (omitted) What is the main subject of  the passage? prompt—remember all sentences have subjects and predicates. further  prompts: Is there more than one subject? How many in all? Where does it shift? 7a. Show me the beginning of  the second sentence in paragraph .three. 7b. Show me the end. (pointing) 8a. How would you chunk this sentence? "Every time you move any part of  you, you use your muscles." 8b. Why would you chunk it like that? 9a. Does it make sense if  we read only the first  half?  Why or why not? 9b. Does it make sense if  we read only the second half?  Why or why not? 10. (omitted) Why does it say "you, you"? Is that a mistake? 11a. How would you chunk this sentence? "You even use muscles when you read." 1 lb. Why would you chunk it like that? 12. Is it harder or easier to chunk than the first  sentence? Why? 13. (omitted) What kind of  a sentence is this? "Hold your arm straight out in front  of  you with the palm of  your hand facing  up." (command) Can you find  any others like it? 14a. What is the subject of  this sentence? "Your biceps and your triceps work together." 14b. What is the subject of  this one? "Many other muscles in your body work in pairs that way." Reading  Test  Two Strike a key on the piano. The key is attached to a hammer which will hit a string. The string will vibrate, or move back and forth  very quickly. When something vibrates, it makes the air all around it vibrate. Air can move back and forth  rapidly, just as the strings in a piano can. You hear sounds because air vibrates. Air carries vibrations to your ear. The outer part of  your ear catches the vibrations and carries them into the opening of your ear. Inside your ear there is a tunnel. The vibrations go down the tunnel and strike against a thick wall. This is your eardrum. When vibrations hit your eardrum, th'ey make it vibrate. The eardrum vibrates like the top of  a drum. Behind your eardrum there is a chain of  three small bones. When the eardrum vibrates, these bones vibrate also. They carry the vibrations deep inside your ear. They go to a part of  the ear that is called the cochlea. The cochlea is shaped like a small ball. It has liquid inside it. Nerves that connect the ear to the brain dip into this liquid. When vibrations read the cochlea, the liquid inside vibrates. The vibrations make signals in the nerves, and the nerves carry the signals to the brain. Your brain changes the signals to what we call sounds. All this happens in less than a second. 1. What is the passage about? Prompt: Can you tell me more? 2. Here is a picture of  the ear. Can you explain how the ear hears sounds as you show me on the picture? 3. Where was it hard to understand what you were reading? Show me the first  place that it got hard to understand, where you had to slow down. (Did not use prompt, i.e., how far  back did you go?) 4. What did you do to figure  it out? 5. (collapsed with #3) What part of  the sentence made it hard to understand? 6. (omitted) How did you know when you had found  what you were looking for? 7. (omitted) What is the main subject of  the passage? (prompt—didn't use—What are you tracing through the ear and through the sentences?) 8a. Show me the beginning of  third sentence in this paragraph(#l). 8b. Show me the end of  the sentence. 9. Point to the part of  that sentence with the exact words which tell you what "vibrates". 10a. How would you chunk this sentence: "When vibrations hit your eardrum, they make it vibrate"? I Ob. Why would you chunk it like that? II a. Does it make sense if  we read only the first  half?  Why or why not? 1 lb. Does it make sense if  we read only the second half?  Why or why not? 12. (omitted) What two words in the sentence are referring  to the same thing? 13. (omitted) Why does the author often  start the sentences with "When"? What does it mean here? Does it tell the time? 14a. How would you chunk this sentence: "The outer part of  the ear catches the vibrations and carries them into the opening of your ear"? 14b. Why would you chunk it like that? 15. What is the subject of  that sentence? 16. What is the predicate of  that sentence? 17. (omitted) Why isn't there any comma in this sentence while there is in the other? 18. (omitted) In the last sentence "All this happens in less than a second", what is the "this"? APPENDIX B THESIS STUDY: PRETEST Reading  Passage On a sunny day, beams of  light will stream through your window. Even if  all the doors and windows are closed and there is no hint of  a breeze, you may still notice very tiny specks of  dust darting about. These dust specks can show us a lot about the nature of  matter. Although the specks of  dust are small, the particles that make up matter are even smaller. In fact,  these particles are so small that they can't be seen. Air is made up of  such particles, moving through space. As the particles of  air move about, they collide with each other and with everything in your room. They collide with the specks of  dust. The dust specks move because particles of  air bounce the specks of  dust around. Another way to see the effect  of  moving air particles is to look at inflated  objects, such as balloons and basketballs. They also provide evidence that air is made up of  particles. When you put air in a container, the moving air particles continuously bang against the sides of  the container. It's the pressure of  these collisions that keep objects inflated. Air is a mixture of  several different  gases. Because gases are invisible, it's easy to think of  them as being made up of  tiny moving particles. Liquids and solids are not invisible, but they are also made up of  tiny moving particles. Like air in a room, water in a standing jar seems calm, but if  you gently add coloured liquid to the jar, the water and the coloured liquid mix together on their own. This mixing shows that liquids, like gases, are made up of  moving particles. As the particles of  water and coloured liquid bump into each other, the particles spread out and mixing occurs. A.Tell me everything you can about the information  in this passage, (omitted from  analysis) B. 1. How do dust specks and the particles that make up matter compare in size? 2. According to the passage, why is it easy to think of  gases being made up of  tiny moving particles? 3. Why is it more difficult  to think of  liquids and solids being made up of  tiny moving-particles? 4. How is what happens when you put air in a container similar to what happens if  you gently add coloured liquid to water in a standing jar? 5. (omitted from  analysis) What part of  this passage was difficult?  What did you do to help you understand it? C. Syntactic Awareness-—Chunking " Even if  all the door and windows are closed and there is no hint of  a breeze, you may still notice tiny specks of  dust darting about." We can chunk long sentences to help us understand them just as we can chunk long words. la. Show how you would chunk this sentence by drawing in dividing lines, lb. Why would you chunk it at those spots? lc. What do the words "even if'  do? "Although the specks of  dust are small, the particles that make up matter are much smaller." 2a. Show how you would chunk this sentence by drawing in dividing lines. 2b. Why would you chunk it at those spots? 2c. What does "although" mean? "As the particles of  air move about, they collide with each other and with everything in your room." 3a. Show how you would chunk this sentence by drawing in dividing lines. 3b. Why would you chunk it at those spots? 3c. Why does it begin with "as"? 3d. (omitted from  analysis) What would you say is the main or most important chunk in the sentence? 3e. (omitted from  analysis) Why? "In fact,  these particles are so small that they can't be seen." 4a. (omitted from  analysis) Why does this sentence begin with the words "in fact"? 4b. (omitted from  analysis) What words could you use to replace it? 5. (omitted from  analysis) Some words and phrases in a passage are there mainly to show how one thought is related to another. They are signals for  relations, just like "if'  and "as" above. Can you find  some other words that do that job? 6. (omitted from  analysis) If  you have to choose three key words in this passage, what would they be? THESIS STUDY: POSTTEST Reading  Passage The three common states of  matter are gases, liquids and solids. Most forms  of  matter can be found  in all three states. You have seen matter in different  states. If  you think of  water, you will remember seeing water as a solid (ice), a liquid and a gas (water vapor). In fact,  you have seen water change from  one state to another. As ice melts and puddles "dry up", water changes state, from  solid to liquid and from  liquid to gas. This is how it happens. The particles that make up water move. This movement produces energy. The energy of  motion is called kinetic energy. If  the rate at which the particles move changes, they lose or gain energy. When kinetic energy is lost or gained, the water changes. If  you want to know how fast  the particles of  a material are moving, take the material's temperature! Temperature is a measure of  the average kinetic energy of  the particles in a material. Not all the particles are moving at the same rate, but if  the speed is averaged out, you can find  the temperature. Although many people think that temperature and heat are exactly the same, they are actually quite different.  Think of  a glass of  hot water and a bathtub full  of  water at the same temperature. The particles of  water in each container would have the same average speed. Yet the tub water would have more energy, because there are more particles in the tub. Heat energy includes the total kinetic energy of  the particles in a material. A lot of water will have more heat energy than a little bit of  water, even if  all the water is the same temperature. A. Tell me everything you can about the information  in this passage, (omitted from analysis) B. 1. How can you learn about matter by observing water? 2. How does matter change state? 3. How are heat and temperature different? 4. Why will a lot of  water have more heat energy than a bit of  water? 5. (omitted from  analysis) What part of  this passage was difficult?  What did you do to help you understand it? C. Syntactic Awareness—Chunking " If  the rate at which the particles move changes, they lose or gain energy." We can chunk long sentences to help us understand them just as we can chunk long words. la. Show how you would chunk this sentence by drawing in dividing lines, lb. Why would you chunk it at those spots? lc. What does even if  do? "Although many people think that temperature and heat are exactly the same, they are actually quite different." 2a. Show how you would chunk this sentence by drawing in dividing lines. 2b. Why would you chunk it at those spots? 2c. What does although mean? "As ice melts and puddles 'dry up', water changes state, from  solid to liquid and from liquid to gas." 3a. Show how you would chunk this sentence by drawing in dividing lines. 3b. Why would you chunk it at those spots? 3c. Why does it begin with as? 3d. (omitted from  analysis) What would you say is the main or most important chunk in the sentence? 3e. (omitted from  analysis) Why? "In fact,  you have seen water change from  one state to another." 4a. (omitted from  analysis) Why does this sentence begin with the words "in fact"? 4b. (omitted from  analysis) What words could you use to replace it? 5. (omitted from  analysis) Some words and phrases in a passage are there mainly to show how one thought is related to another. They are signals for  relations, just like "if'  and "as" above. Can you find  some other words that do that job? 6. (omitted from  analysis) If  you have to choose three key words in this passage, what would they be? Worksheets—in order of  presentation to students Worksheet One—Chunking Sentences A. Chunk these sentences by making a slash between chunks. Remember to use commas for  cues. Circle connectives. 1. If  the earth's temperature were to fall  just a few  degrees, many plants and animals would die. 2. If  the the earth's temperature were to rise just a few  degrees, changes in ocean currents could have a devastating effect  on marine life  and the entire food  chain. 3. From our perspective, the sun appears constant and stable, but it is actually a chaotic and violent place. B. Write a question which could be answered by #1 above. Worksheet Two—Chunking Sentences in a Paragraph For the following  paragraph, chunk each sentence with a pencil slash. Underline the chunks that are NOT clauses. Circle the connectives. To discover faraway  planets, astronomers use telescopes in a special way. They do not use them to look directly at planets. Instead, they look for  star wobble. Star wobble occurs when a star moves one way and then the other. Star wobble is movement caused by a planet's gravity. If  the wobble repeats, astronomers know that there is a planet orbiting the star. 1. What does "instead" refer  back to ? In other words, the scientists look for  star wobble instead of  . 2. There is one longer sentence above that has two clauses which are not separated by a comma. Find it and write it on the lines below. 3. Write a question that can be answered by the last sentence in the paragraph. Worksheet Three—Chunking Sentences in a Paragraph Chunk the paragraph below using commas and periods. Underliine all connectives which begin the sentence or follow  a comma. Write the chunks on the lines that give the information  requested below. With a crew on board, piloted-flight  reentry is more controlled than in the case of unpiloted flights.  Rockedts bring the vehicle down into the atmosphere, and in order to protect craft  and crew, special external heat shields absorb the heat generated during reentry. In the case of  the space shuttle, protective insulation is fastened  to the orbiter. Generally, spacecraft  make the final  descent by pararchute, but the shuttle glides down to Earth like a huge, unpowered aircraft. 1. What chunk indicates why piloted-flight  reentry needs to be more controlled than unpiloted flights? 2. What chunk explains why extra protection is needed to absorb the heat generated during reentry? 3. What chunk states how the space shuttle is made safe? 4. What connective marks the contrast between how spacecraft  usually comes down and how the shuttle comes down? REFERENCES Adams, M. (1990). Beginning to read:  thinking  and  learning  about print. Cambridge, Massachusetts: MIT Press. Adler, M. (1972). How  to read  a book.  (Rev. ed.). New York: Simon & Schuster. Alvermann, D., Hynd, C. & Qian, G. (1995). Effects  of  interactive discussion and text type on learning counterintuitive science concepts. Journal  of  Educational  Research, 88, 3, 146-154. Asimov, 1.(1994). Mysteries  of  deep  space: black  holes, pulsars and  quasars. (Rev. ed.). Gareth Stevens Publishing: Milwaukee. Badders, W., et al. (1999). The nature of  matter. In Houghton  Mifflin  Science: Discovery Works.  Houghton Mifflin  Company: Boston. Booth, J. MacWhinney, B. & Harasaki, Y. (2000). Developmental differences  in visual and auditory processing of  complex sentences. Child'  Development 11, 4, 981-1003. Bormuth, J., Carr, J., Manning, J. & Pearson, D. (1970). Children's comprehension of between- and within-sentence syntactic structures. Journal  of  Educational Psychology,  61,5, 349-357. Branley, F. Floating  in Space.  HarperCollins: New York. 1998. Branley, F. Star  Guide.  Thomas Y. Crowell: New York. 1987. Brigance, A. (1983). Comprehension inventory of  basic skills.  Curriculum Associates, Incorporated: North Billerca. Brown, A., Campione, J. & Day, J. (1981). Learning to learn: on training students to learn from  texts. Educational  Researcher, February, 14-21. Bruer, J. (1993). Schools  for  thought:  A science of  learning  in the classroom.  MIT Press: Cambridge. Bullard, J. et al. (1997). Science Probe Seven.  ITP Nelson: Toronto. Cain, K. & Oakhill, J. (1999). Inference  making ability and its relation to comprehension failure  in young children. Reading  and  Writing,  11, 5-6, 489-503. Cote, N. & Goldman, R. (1999). Building representations of  informational  text: evidence from  children's think-aloud protocols. In H. van Oostendorp & S. Goldman, (Eds.), The  construction  of  mental representations  during  reading.  (169-193), Mahwah, New Jersey: Lawrence Erlbaum Associates. Craig, M. & Yore, L. (1996). Middle school students, awareness of  strategies for  resolving comprehension difficulties  in science reading. Journal  of  research and  Development in Education,  29, 4, 226-238. Darling, D. (1998). Black holes and time tunnels. In R. Benson, L. Bryan, K. Newlove, C. Player & L. Stenson (Series Eds.) & S. Stewart (Vol. Ed.), Collections:  Space, stars and  quasars (pp. 61-64). Scarborough: Prentice Hall Ginn, Canada. (Reprinted from Darling, D. (1990). Could  you ever build  a time machine? Dillon Press) Duffy,  G. & Roehler, L. (1982). The illusion of  instruction. Reading  Research Quarterly,  12, 3, 438-445. Ehrlich, M-F. (1999). Meta-cognitive monitoring of  text cohesion in children. In H. van Oostendorp & S. Goldman, (Eds.), The  construction  of  mental representations  during reading.  (281-301), Mahwah, New Jersey: Lawrence Erlbaum Associates. Elley, W., Barham, I., Lamb, H. & Wyllie, M. (1976). The role of  grammar in a secondary school English curriculum. Research in the teaching of  English,  10, 1, 5-21. Ely, R. (1986). Language and literacy in the school years. In Gleason, J.B. (Ed.), The development  of  language  (4th ed). (pp. 398-439). Boston: Allyn and Bacon. Emory, J. (1998). Saturn: Lord of  the rings. In R. Benson, L. Bryan, K. Newlove, C. Player & L. Stenson (Series Eds.) & S. Stewart (Vol. Ed.), Collections:  Space, stars and quasars (pp. 61-64). Scarborough: Prentice Hall Ginn, Canada. (Reprinted from WORLD  magazine, April 1996). Feldt, R. C. & Feldt, R. A. (2001). A proposed method for  learning from  textbooks in the primary grades: use of  text structure to guide self-questioning.  Psychological  Reports, 88, 645-650. Ferstl, E. & Kintsch, W. (1999). Learning from  text: structural knowledge assessment in the study of  discourse comprehension. In H. van Oostendorp & S. Goldman, (Eds.), The construction  of  mental representations  during  reading.  (247-277), Mahwah, New Jersey: Lawrence Erlbaum Associates. Fowler, H. (1984). A dictionary  of  modern  english usage. Hertfordshire:  Omega Books. Frederiksen, C. (1975). Representing logical and semantic structure of  knowledge acquired from  discourse. Cognitive  Psychology,  7, 371-458. Frederiksen, C. (1977). Semantic processing units in understanding text. In R. Freedles, (Ed.), Discourse Production  and  Comprehension.  (57-87), Norwood, New Jersey: Ablex Publishing Company. Fromkin, V., Rodman, R., Hultin, N. & Logan, H. (1997). An introduction  to language. Toronto: Harcourt, Brace & Company, Canada. Geva, E. & Ryan, E. (1985). Use of  connectives in expository texts by skilled and less skilled readers. Journal  of  Reading  Behavior, 17, 4, 331-346. Goldsmith, E. (1982). An investigation  of  the role of  adversative  connectives in helping good and  poor readers  to integrate  information  in text.  (ERIC Document Reproduction Service No. ED 247 525). Grimes, J. (1975). The  thread  of  discourse.  Paris: The Hague. Guzzetti, B., Snyder, T., Glass, G. & Gamas, W. (1993). Promoting conceptual change in science: a comparative meta-analysis of  instructional interventions from  reading education and science education. Reading  Research Quarterly,  28, 2, 117-155. Halliday, M & Hasan, R. (1976). Cohesion in English.  London: Longman. Hynd, C., Qian, G, Ridgeway, V. & Pickle, M. (1991). Promoting conceptual change with science texts and discussion. Journal  of  Reading,  34, 8, 596-601. Irwin, J. (Ed) (1986). Understanding  and  Teaching  Cohesion Comprehension.  Newark, Delaware: International Reading Association. Jones, B. (1990). Space Exploration.  Belitha Press Limited and Gareth Stevens Publishing: Milwaukee. Kameenui, E. & Carnine, D. (1982). An investigation of  fourth-graders,  comprehension of pronoun constructions in ecologically valid texts. Reading  Research Quarterly,  12, 4, 556-580. Karlsen, B. & Gardner, E. F. (1996). Stanford  Diagnostic Reading  Test  (4th ed.). San Antonio: Harcourt Brace & Company. (Forms J/K Purple). Karmiloff-Smith,  A. (1986). From meta-processes to conscious access: evidence from children's metalinguistic and repair data. Cognition,  23, 95-147. Kent, C. (1984). A linguist compares narrative and expository prose. Journal  of  Reading, December, 232-236. Klecan-Aker, J. (1984). A study of  the syntactic skills of  normal middle-school males and females.  Language and  Speech.  27, 3, 205-215. Kurland, D. (1984). Increasing the theoretical base of  developmental education: Case studies in grammar, critical reading and introductory science. Paper presented at the Annual Meeting of  the National Association for  Developmental Education, Philadelphia, PA. Marshall, N. & Glock, M. (1978-1979). Comprehension of  connected discourse: a study into the relationships between the structure of  text and information  recalled. Reading Research Quarterly,  42, 1, 10-56. Mason, J. (1989). Spacecraft  technology.  Sussex: Wayland. Mason, J. & Kendall, J. (1979). Facilitating reading comprehension through text structure manipulation. The  Alberta  Journal  of  Educational  Research. 25, 2, 68-76. McClure, E., Mason, J. & Barnitz, J. (1979). An exploratory study of  story structure and age effects  on children's ability to sequence stories. Discourse Processes, 2, 213-249. McClure, E. & Geva, E. 1983). The development of  the cohesive use of  adversative conjunctions in discourse. Discourse Processes, 6, 411-432. McNamara, D., Kintsch, E. Songer, N. & Kintsch, W. (1996). Are good texts always better? Interactions of  text coherence, background knowledge, and levels of  understanding in learning from  text. Cognition  and  Instruction,  14, 1, 1-43. Mercer, T. (1998). Sojourner and beyond. In R. Benson, L. Bryan, K. Newlove, C. Player & L. Stenson (Series Eds.) & S. Stewart (Vol. Ed.), Collections:  Carving  new frontiers. (pp. 48-53). Scarborough: Prentice Hall Ginn, Canada. Ministry of  Education. (2002). B.C. performance  standards:  Reading.  (Rev. Ed.) Student Assessment and Program Evaluation Branch: Province of  British Columbia. Morice, R. & Slaughuis, W. (1985). Language performance  and reading ability at 8 years of age.' Applied'  Psycholinguistics,  6, 141-160. Muter, V. & Snowling, M. (1998). Concurrent and longitudinal predictors of  reading: the role of  metalinguistic and short-term memory skills. Reading  Research Quarterly.  33, 3, 320-337. Neuwirth, S. (1976). A look at intersentence grammar. The  Reading  Teacher,  October, 28-32. Nye, R. & Ebbitt, W. (1961). Structure  in reading  and  writing.  Chicago: Scott, Foresman and Company. Otero, J & Campanario, J. (1990). Comprehension evaluation and regulation in learning from science texts. Journal  of  Research in Science Teaching.  27, 5, 447-460. Palincsar, A. & Magnusson, S. (2001). The interplay of  first-hand  and second-hand investigations to model and support the development of  scientific  knowledge and reasoning. In S. Carver & D. Klahr (Eds.), Cognition  and  Instruction:  Twenty-Five Years  of  Progress.  (151-187), Mahway, New Jersey: Lawrence Erlbaum Associates, Publishers. Payne, P. (2000). Investigate  space. Vancouver: Whitecap Books. Pinker, S. (1994). The  language  instinct: how the mind  creates language.  New York: HarperCollins. Rodgers, D. (1974). Which connectives? Signals to enhance comprehension. Journal  of Reading,  March, 462-466., Rossi, J. (1990). The function  of  frame  in the comprehension of  science text. Journal  of Educational  Psychology.  82, 4, 727-732. School District No. 44 (1999). Reading  44: A core reading  framework.  North Vancouver. Semel, E., Wig, E. H. & Secord, W. A. (1995). CELF:  Clinical  Evaluation  of  Language Fundamentals  (3rd ed.). San Antonio: Harcourt Brace & Company. (Examiner's Manual) Shara, M. (2002). When  stars collide.  Scientific  American, 287, 5, 44-51. Spiro, R. & Taylor, B. (1980). On investigating  children's  transition  from  narrative  to expository discourse:  the multidimensional  nature of  psychological  text classification. (Tech. Rep. No. 195). Champaign, Illinois: University of  Illinois, Centre for  the Study of  Reading. Stoodt, B. (1970). The  relationship  between understanding  grammatical  conjunctions and /  reading  comprehension. Unpublished doctoral dissertation, Ohio State University, Ohio. Stotsky, S. (1974). Sentence-combining as a curricular activity: its effect  on written language development and reading comprehension. Research in the Teaching  of  English.  30-71. Tapiero, I. & Otero, J. (1999). Distinguishing between textbase and situation model in the processing of  inconsistent information:  elaboration versus tagging. In H. van Oostendorp & S. Goldman, (Eds.), The  construction  of  mental representations  during reading.  (341-365), Mahwah, New Jersey: Lawrence Erlbaum Associates. Tomasello, M. (2000). Do young children have adult syntactic competence? Cognition,  74, 209-253. Van der Henst, J., Yingrui, Y. & Johnson-Laird, P. (2002). Strategies in sentential reasoning. Cognitive  Science,  26, 425-468. van Dijk, T. & Kintsch, W. (1983). Strategies  of  discourse  comprehension. New York: Academic Press. Vogt, G. (2001). Pluto  and  the search for  new planets.  Raintree Steck-Vaughn Publishers: Austin. Warren, B. (1986). Aspects of  skill  in understanding  high-order  semantic relations.  (Tech. Rep. No. 6009). Office  of  Naval Research. (ERIC Document Reproduction Service No. ED 273 927). Wertsch, J. & Tulviste, P. (1992). L. S. Vygotsky and contemporary developmental psychology. Developmental  Psychology,  28, 4, 1-10. Wolfram,  W. (1976). Extended notions of  grammar and reading comprehension. Journal  of Reading  Behavior. 8, 3, 247-258. Zinar, S. (1990). Fifth-graders,  recall of  propositional content and causal relationships from expository prose. Journal  of  Reading  Behavior, 22, 2, 181-199. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078231/manifest

Comment

Related Items