dpgerria AN STATE UNIVERSITY LIBRAR llllllllllllll I! illi illlill‘fiilllllllllllllll L l l 3 1293 00623 6230 LIBRARY Mldilgon State University This is to certify that the dissertation entitled The Significance of Difficult Vocabulary to Reading in a Second Language presented by Monte Gale Salyer has been accepted towards fulfillment of the requirements for PhD degree in _Eng_1_j_sh_ MW v Major professor {#4 /?a MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 r -‘4r PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES Mum on or More date due. “ DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Inditutlon _THE SIGNIFICANCE OF DIFFICULT VOCABULARY TO READING IN A SECOND LANGUAGE BY Monte Gale Salyer A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of English 1990 e; C: ":Q" "Ni wl‘fiq‘: Q) ABSTRACT THE SIGNIFICANCE OF DIFFICULT VOCABULARY TO READING IN A SECOND LANGUAGE BY Monte Gale Salyer There is a gap in second-language reading research regarding whether the activation of topical knowledge enables readers to comprehend text with difficult (unfamiliar) vocabulary. Although researchers (Swaffar, 1988; Barnitz, 1986) have supposed that strengths in top-down processing may compensate for weaknesses in bottom-up processing, few studies have examined this hypothesis. This study investigates the effects of schema- activation and of difficult vocabulary on reading comprehension in a second language. Because schema- activation may facilitate reading comprehension (Taglieber, Johnson, and Yarbrough, 1988), and because difficult vocabulary may disrupt the processing of L2 texts (Carter, 1987), it is assumed in the study that both schema- activation and a lack of difficult vocabulary facilitate L2 reading comprehension. The subjects of the study were international students enrolled in an advanced-level ESL reading class in a small Midwestern university. All subjects read four different Monte Gale Salyer reading passages, in either the easy- or the difficult-vocabulary version. The reading conditions were (1) schema-activated and easy-vocabulary version, (2) schema-activated and difficult-vocabulary version, (3) no schema-activation and easy-vocabulary version, and (4) no schema-activation and difficult-vocabulary version. Schemata were activated by means of scripted introductions to the topic followed by warm-up pre-writing activities. After reading the passages, subjects were administered cloze, written, and oral post-tests. Multivariate analysis of variance tests on the cloze and written tests showed significant effects for vocabulary, but not for schema-activation. Positive correlations between the cloze and written post-tests were found also. These results show that vocabulary development, which has been neglected, is vital to ESL reading. The results of schema-activation, however, may not be as powerful as has been assumed. TABLE OF CONTENTS List of Tables List of Figures Chapter One: Introduction The Nature of the Reading Process Reader- and Text-dominated Theories of Reading The Psycholinguistic Model of Reading The Interactive Model and Preteaching The Over-emphasis on Preteaching Background Knowledge The Importance of Bottom-up Processing Summary Chapter Two: Review of the Literature Schema-theory and Reading in a Second Language The Psycholinguistic Model and Background Information The Compensative Theory The Interactive Model of Reading The Effect of Preteaching Text Structure Barnitz and the Interactive Model of Reading The Effect of Preteaching Background Information Schematizing as Interactive Reading Summary Studies on the Effect of Vocabulary in L1 Reading When Vocabulary Preteaching May Not Help In-depth Vocabulary Knowledge and Reading Comprehension The Effect of Difficult Vocabulary on Reading Comprehension The Effects of Vocabulary and Schemata on L1 Reading Are Schemata and Vocabulary Interactive Variables? Preteaching May Not Overcome Difficult Vocabulary Attempts to Find Interaction Between Vocabulary and Schemata Vocabulary and Prior Knowledge in Second Language Reading An Emphasis on Vocabulary in Earlier L2 Reading Studies /”‘~- . ”a; ,— r 'r ' - i,‘ it"? ‘ iv 4 O‘UINl-‘H 11 13 13 14 15 16 17 19 21 22 24 25 26 27 29 31 32 33 34 37 ‘ 37 ii Do Language Deficiencies ”Short-circuit" L2 Reading? The Importance of L2 Vocabulary to Academic Achievement Why Vocabulary Has Been Underrated Review Summary Purpose of the Study Chapter Three: Method Subjects Reading Materials Adaptation of the Reading Materials Vocabulary Pretests Easy- and Difficult-vocabulary Versions of Passages Posttests The Written-Recall Measure The Oral-Recall Measure Procedures Schema-Activation The Treatment Conditions Overview of the Procedure Chapter Four: Results Results of the Analyses of Variance Analysis of the Cloze Data The Newman-Keuls Multiple Comparisons of Cloze Data Analysis of the Written-Recall Data Comparison of the Cloze and Written Data Analysis of the Written Test Data in the Main Idea Category Analysis of the Written-Recall Data in the Detail Category Summary of the Analysis of Variance Results of Correlational Studies Negative Correlations Between Cloze and Pretest Correlations Between Cloze and Written Tests Correlations Between the Pretest and the Oral Recall Summary of Correlations Between Pretest and Other Measures Correlations Between Cloze and Writteanests Correlations Between Cloze and Oral Tests Correlations Between Written and Oral Recalls Chapter Summary Chapter Five: Discussion The Purpose of the Study Hypotheses Regarding Schema-activation and Vocabulary The Study Results Analysis of the Cloze Data 38 40 41 43 44 46 46 48 49 50 52 54 55 56 58 59 60 61 63 64 65 66 68 69 70 72 72 73 74 75 76 76 77 77 78 79 80 81 82 82 83 85 iii Analyses of the Written Recall Data Correlational Studies Conclusions Why Schema-activation May Prove Ineffective Why Schema—activation May ”Short-circuit" Reading Can Schema-activation Overcome L2 Language Difficulties? Difficult Vocabulary Strongly Affects Reading Comprehension Do Schema-activation and Vocabulary Interact? Why Schema-activation May Be Over-valued Applications for Teaching and Implications for Research Appendix A: The Data Appendix B: The Reading Passages Reading Passage 1: A Theory of Earth's Structure Easy-Vocabulary Version Difficult-Vocabulary Version Reading Passage 2: Learning About Earthquakes Easy-Vocabulary Version Difficult-Vocabulary Version Reading Passage 3: Overpopulation and Social Behavior Easy-Vocabulary Version Difficult-Vocabulary Version Reading Passage 4: Deforestation Easy-Vocabulary Version Difficult-Vocabulary Version Appendix C: The Vocabulary Tests Vocabulary Test 1 Vocabulary Test 2 Appendix D: Cloze Tests Cloze Test 1 Cloze Test 2 Cloze Test 3 Cloze Test 4 Bibliography General References 86 87 88 89 91 93 94 96 98 100 103 104 104 104 106 107 107 109 111 111 114 116 116 118 121 121 124 128 128 129 130 131 134 137 LIST OF TABLES Table 1: Means (and Standard Deviations) for PreTest, Cloze, Written, and Oral Recalls 64 Table 2: Repeated Measures of Analysis of Variance of Cloze Data 65 Table 3: A Newman-Keuls Multiple Comparisons of Cloze Data 66 Table 4: Means (and Standard Deviations) for PreTest Cloze, Written, and Oral Recalls 84 Table 5: Repeated Measures of Analysis of Variance of Cloze Data 85 Table 6: A Newman-Keuls Multiple Comparisons of Cloze Data 86 Table 7: Data from the Study 103 Figure Figure Figure Figure Figure Figure Recall 5: 6: LIST OF FIGURES The Relationship of Reading Theories Schedule of Reading Conditions The Time Schedule of the Study Cloze Test Mean Scores Mean Scores on the Written Test Mean Scores on the Main Ideas of the Written 60 62 67 69 71 CHAPTER ONE INTRODUCTION This chapter summarizes pertinent theories of both L1 and L2 reading so that readers from a variety of backgrounds will be able to appreciate the context of the present research regarding the second-language reading classroom. (Readers knowledgeable of these general backgrounds may wish to proceed directly to the formal review of the literature, which is Chapter Two.) Reader- and text-dominated theories, the psycholinguistic model, and the interactive model will be outlined. Although each model contains useful information, educational practice may be impeded by the extreme application of any theory. Because of an overemphasis on preteaching background knowledge, there has been a corresponding de-emphasis on ESL vocabulary development. Despite the importance of schema-theory, it may be unwise to neglect vocabulary development in ESL reading. THE NATURE OF THE READING PROCESS As a result of the considerable research conducted over the years, we now know a great deal regarding the mechanical (or physical) aspects of reading. For example, reading researchers have conducted numerous studies precisely measuring the movement of the eye so that the number and duration of fixations per page may be known. 1 2 As a result, the externally observable acts of reading are well known. We still lack full insight, however, into the internal meaning-making process of reading. We can confidently suppose that a text, the creation of an author, and external, structured objects somehow touch off meaning in the reader's mind, even as the reader "brings meaning" to text. What occurs between the text and the reader, however, remains only partly understood. To what extent either the text or the reader constructs meaning is unknown. Certain questions arise about this process. What meaning or structures do the reader bring to text? To what extent does text ”dictate” meaning? Also, what choices does the reader possess in responding to visual cues? READER- AND TEXT-DOMINATED THEORIES OF READING One way to view reading theories is to create a continuum on which all theories may be placed, as represented in Figure 1 below. Theories emphasizing the significance of text would range toward one pole of the continuum, whereas views in support of the reader's role would tend toward the opposite pole. FIGURE 1 The Relationship of Reading Theories We Text-dog; nated Theories Theories An earlier view of reading represented it as an adjunct to listening. In this construct, the passive reader, through 3 the text, took in whatever the author might have on his mind. For decades schoolchildren read their lessons out loud. It was thought that the oral production of the words on the page led to a meaningful realization of reading. Many schoolchildren still are drilled extensively in oral reading, presumably for similar reasons. Knowledge that such "reading" practices affect pronunciation more than reading comprehension has not affected educational methodology as yet. In the audio-lingual model, reading was considered secondary to listening. The oral skills were thought primary, whereas reading and writing were considered limited extensions. Writing was referred to as encoding, and reading became known as decoding. Teachers sought to help students learn proper phonics skills so that they might convert encoded script to vital oral form. Reading problems essentially appeared to be seeing problems as this mechanical view dominated. When training in "cracking" the code did not produce capable readers, at times it was supposed that poor readers had visual problems. Although existing at a different place in the educational spectrum, literary critics possess a model compatible with the simpler text-dominated views. In certain critical theories, it appears that there is a near worship of the text, as at least the indoctrinated reader’s mind is said to be dominated by the author’s voice. Close reading is preferred because of the value allotted to text. 4 In fact, the aim is to be converted to the author’s point of view. Slowly the gap between the reader and the text is closed as the diligent disciple reads and rereads text. Although created by the author, once written, the text itself becomes a controlling context. The reader's opinion and context is "surrendered" to the context created by the text itself (Kreiger, 1976). In effect, this particular literary view assigns to the reader much the same role as does the decoding model. Text dominates readers in both cases, and reading deficiencies are believed to arise due to lack of perception. In contrast, alternate theories of reading emphasize the reader. Subjectivists stress the reader’s role in making meaning from text. In fact, one critic has questioned the very existence of text (Fish, 1980). The model subjectivist reader, however, admits the existence of text. Such readers follow text long enough to encounter a condition reminiscent of something in the reader's personal experience. At this point, the reader ”steps in front of text” and constructs a new text which may relate but tangentially to the original text (Bleich, 1978). The end product of this process is a text which diverges radically from the original. Essentially, text serves as a jumping-off point for self-exploration and reflection. Reading only marginally relates to author-written text, but bears strong connections with the activities of the intentionally-creative reader. The subjective reader may 5 radically recreate text according to his or her own will. Subjective theory occupies the end of the reading continuum where the reader’s role is paramount. THE PSYCHOLINGUISTIC MODEL OF READING The psycholinguistic model of reading allows for "behind the eyeball" guesswork as to meaning, but does not permit as much reader-domination as does subjectivism. In this model, the reader's role is to sample a variety of textual cues to meaning. Less textual sampling is required when the reader knows about the linguistic and cultural systems represented by a particular text. Full decoding is neither necessary nor advisable, as overconcern with the graphophonic aspects of text tends to encumber the reader’s comprehension. Possessing limited attention, readers who attend too closely to the details of text may find it difficult to focus on the larger propositions of text as well. Nor is fully linear processing necessary, even at the word level. For example, suppose that a person sees a sign spelling "clos" in the entrance to a cafe. Because we expect to find a sign indicating that the cafe is either open or closed, the person might believe the cafe to be closed on the basis of seeing just the letters ”clos" in the door. Knowledge of the world, then, facilitates reading, and in fact makes text comprehensible. Despite the reader's distance from the text, and the reconstructive activity occurring "behind the eye”, this model avoids 6 intentionally subjective distortion of text. The reader’s task is clearly defined as following the direction of text, although the necessarily selective pose of the reader suggests reliance on text or author as at least an initial guide to thinking. The writer’s implied role is to facilitate the reader’s sampling process by providing amply redundant processing cues. In general, the psycholinguistic model suggests that the reader actively brings considerable knowledge to bear upon text in making meaning. THE INTERACTIVE MODEL AND PRETEACHING More recently, reading researchers (Carrell, Devine, and Eskey, 1988: Swaffar, 1988) have introduced the interactive model (which builds on psycholinguistic theory). This theory may be conceived as carrying the psycholinguistic model one step further, and as filling in blanks only suggested by the former theory. According to the interactive construct, reading is actually a process in which the reader's knowledge and text interact to create meaning. In this way, an L2 reader who knows baseball would then be able to comprehend a passage describing a baseball game. A reader without much prior knowledge about a particular topic, however, would encounter severe processing difficulties. Foreign-language texts may be understood by readers who possess critical background knowledge. To what extent deficiencies in prior knowledge impede reading is not known, but the effect has been 7 validated (Johnson, 1982: Carrell, 1985: Barnitz, 1986). To meet the needs of readers from other cultures, second-language reading teachers have introduced preteaching, with various degrees of success. Generally, readers seem to use pretaught information in processing the main propositions of a passage. Sentence-level processing, however, does not appear to be affected. Some questions do remain, though, as to the real effect of preteaching. Do students learn what they need to know from the preteaching itself? Or do they merely resourcefully gather enough information in the pre-teaching activity to make the reading passage comprehensible, and then go on to fill in the blanks in their mental inventory either by actually extracting meaning from the text, or by informed guessing? The improved comprehension of main ideas by readers who have been pretaught, and the lack of effect on supporting information leave doubts as to the overall reading gains resulting from preteaching background knowledge. On the other hand, having cultural knowledge seems to be very important to reading comprehension. Whether preteaching topical knowledge facilitates L2 reading or not, cultural knowledge is necessary for learners. It may well be that preinstruction in linguistic or cultural information has limited effect on the reading process because cues must be well assimilated before they may be of much use to the reader who is involved in the already complicated bottom-up and top-down reconstruction of text. 8 The attention-consuming factors of reading processing may overwhelm the effective use of additional preinstructional information. The interactive model has been expanded to include not only the interaction of content schema and text, but knowledge of rhetorical patterns and text as well. If readers recognize, for instance, that a paragraph is developed by description, they find more meaning. There is considerable evidence that readers can be taught text-structure patterns, and that reading comprehension is facilitated by such knowledge. Cohesive or connecting terms should be mentioned as well. It has been discovered that a lack of comprehension of connecting words, as well as of vocabulary, negatively affects the reader's comprehension (Barnitz, 1986). THE OVEREMPHASIS ON PRETEACHING BACKGROUND KNOWLEDGE Most reading researchers favor the interactive model. Nevertheless, some researchers (Carter, 1987; Eskey, 1988) express serious concerns over the tendency to focus nearly exclusively on the interaction of prior knowledge and to neglect vocabulary development. Whenever such attention is accorded to schemata, bottom-up processing is slighted. L2 reading researchers, in particular, should be sensitive to the language-related difficulties inherent in reading in a foreign language. A parallel between the complications of borrowing concepts from a dominant culture and from a larger field of research (from L1 to L2 reading research) may be made. Whenever two cultures meet, one becomes 9 dominant. Knowledge, technology, terms, and even ways of life tend to flow from the more powerful culture to the other society. Albeit the exchange often enriches the recipient culture, the interaction can be disruptive as well. Often considerable adaptation is required in order for a borrowed item to become suitable to the new context. To be sure, the importation of schema-theory from L1 reading studies has enriched L2 reading theory. Even so, it may not be in our best interest to accept this theory without adaptation. While the similarities between the process of reading in a first and in a second language are marked, there are differences too, especially concerning aspects of language competence such as syntax and vocabulary. For schema-theory properly to be applied to second-language reading, the model must be adapted to account for not only schemata, but vocabulary difficulty as well. To date, it seems L2 theorists have overlooked this bottom-up process in applying the interactive theory to the conditions of top-down processing alone. No argument is made herein against the application of schema-theory to L2 reading. Rather, it is urged that besides the effect of background knowledge, the L2 model must account also for the role of vocabulary. Studies in L2 reading support the interactive theory. The theory is appealing for several reasons. The term itself suggests the complexity of the reading process, reflecting the involvement of more than one variable. 10 Educational practice in general presently encourages interactive teaching and learning. In addition, the term permits a certain range of interpretation. What interacts with what can be spelled out variously. L2 theorists have adapted the concept recently, so that it has been applied to the relationship between vocabulary and schema. Even this adaptation of the theory, however, has tended not to support vocabulary learning. It has been supposed that a reader's strength in one area may compensate for deficiencies in another. Thus, a reader with appropriate prior knowledge would be able to make up for inadequate vocabulary knowledge in processing a particular passage. Although this view admits that vocabulary strengths may compensate for schema weaknesses this possibility is not considered. Inasmuch as the schemata appear to be more teachable than vocabulary, there has been interest in preteaching prior knowledge to compensate for vocabulary weaknesses, but not in preteaching vocabulary to make up for schemata-deficiencies. The apparent relationship between vocabulary and background knowledge is considered hardly at all. Neither has it been remarked that appropriate vocabulary is necessary to the learning of topical knowledge. THE IMPORTANCE OF BOTTOM-UP PROCESSING Enabling L2 readers to process difficult text by initiating them into the discourse and thought patterns of a culture is useful. Still, such practice need not 11 preclude other basic aspects of language learning, such as grammar and vocabulary. No matter how much an intermediate-level ESL learner may know about American culture, it remains necessary for the language itself to be learned in order for reading comprehension actually to occur. The newcomer to English, who desires to learn vocabulary in connection with reading in a second language, may be dismayed to encounter more exercises in learning schemata than in vocabulary development in the reading textbook. Such an occurrence is not unlikely, given the current emphasis on preteaching background knowledge. Indeed, because of the dominance of schema-theory, more attention has been given to top-down than to bottom-up processing in some ESL reading texts. Because top-down processing in a second language places high demands on the reader’s attention, students may justifiably question why they must learn schema-theory before they have acquired the vocabulary necessary to process a text. A more balanced approach would include the simultaneous development of both vocabulary and schemata. SUMMARY It seems natural that bottom-up processing should receive the reader's primary attention especially at the beginning and to some extent at the intermediate L2 reading levels. It may be that L2 reading practice has surpassed theory. At present, there is no proof that background 12 knowledge compensates for a lack of vocabulary in either L1 or L2 reading. Studies (reviewed in the following chapter) have attempted to find such an interaction, but have not done so (Freebody and Anderson, 1983a: Stahl and Jacobson, 1986). There have been no studies on the interaction of vocabulary and prior knowledge in L2 reading. It seems fitting to investigate the hypothesized interactive effect between prior knowledge and vocabulary. If a relationship can be found, and should readers be enabled to fill in the gaps in their linguistic and cultural knowledge, then so much the better. By means of preteaching background information, readers would then be able to process even texts laden with considerable difficult vocabulary. On the other hand, there remains the possibility that L2 readers may not be able to overcome vocabulary and other bottom-up processing problems by means of explicit instruction in background information. If no evidence can be found for compensation, the theory may require revision. In the following chapter, the relevant literature will be reviewed. Numerous experimental, non-experimental, and review studies of vocabulary, background knowledge, and rhetorical patterns in first and second language yield a wealth of information enlightening our understanding of the reading process. The purpose of this review of the literature will be to ascertain whether there is evidence that background knowledge compensates for deficiencies in vocabulary knowledge. CHAPTER TWO REVIEW OF THE LITERATURE The following review of the literature examines studies from three diverse and, in some ways, contrasting areas of reading research. Each study will be summarized to reveal our present knowledge of the impact of background knowledge and difficult vocabulary in the reading process. First, studies on the implications of schema-theory on second-language reading will be examined, particularly to determine the possible effect of preteaching background knowledge. Second, research which investigates the relationship between native-language reading and vocabulary will be inspected. Finally, studies concerning the impact of vocabulary knowledge on second-language reading will be reviewed to indicate the important role that bottom-up processing plays. SCHEMA-THEORY AND READING IN A SECOND LANGUAGE The work of Frank Smith (1973) regarding the connection between psycholinguistics and reading has affected much of the current work in both first- and second-language reading theory. Smith’s essential argument that very little of what is needed for understanding text comes from the text itself has been quite suggestive for reading researchers. Nor was Smith alone: his work was complemented especially by that of his contemporary, Kenneth Goodman (1973), who 13 14 who noted that effective readers sample text to confirm or reject their predictions as to meaning. These and other theorists proposed a reconstructive model of reading, one in which the reader actively processes text for meaning. Although the ”behind the eye” view had been promoted decades earlier by Huey (1908), literary critics (Krieger, 1976: Scholes, 1985) and linguists (Bloomfield, 1942: Fries, 1962) had given much more emphasis to the primacy of text in reading. The Psycholinguistic Model and Background Information Emergent with the psycholinguistic view of reading, however, was the importance of background information. If reading is to a great extent recognizing old information in new contexts, then reading problems might be attributable to a deficiency in background information. Nowhere would a lack of background information be as likely to occur or as disruptive to reading as in a second language. Few readers in a second language have extensive knowledge of the second culture. Therefore, schemata problems would be expected for L2 readers. Practical classroom experience showed that although readers often felt they ”knew" all the words in a passage, they could not comprehend the total message. This situation prompted an examination of the relevance of schemata to reading, and in the process revolutionized views of the reading process itself. In the course of time, studies delving into the relationship between schemata and reading comprehension appeared in L2 reading 15 literature. Various of these relevant research studies are summarized below. Johnson (1982) investigated precisely this issue in an experimental study of the effects of building background knowledge on second-language reading comprehension. Predicting that most foreign students would lack much information on the American holiday Halloween, Johnson designed a study to measure the reading problems subjects would encounter when called upon to read about this unknown cultural phenomena. Whereas half of the seventy-two subjects were pretaught cultural aspects of Halloween, the thirty-six subjects in the control group were given no background-building exercises. By relying on the pretaught information, the experimental group showed much better reading comprehension of the text than the control group as measured by an ANOVA. There was no measurable effect for preteaching vocabulary (without clues about Halloween) to the control group. The Compensative Theory As second-language researchers began to investigate the effects of background knowledge on reading, it was considered whether preteaching schemata might enable readers to compensate for language deficiencies. Hudson (1982) researched whether the preteaching of schemata could be more significant than language proficiency in reading comprehension. Three measures, or types of intervention were presented to the ESL readers in the study: schemata 16 induction, vocabulary training, and the read-and-test method (which gave the subjects no special help in arriving at meaning). Ninety-three subjects, who were at the beginning, intermediate, or advanced levels, read nine SRA reading passages intended for beginning, intermediate, and advanced readers. Subjects read one passage at each reading level and under each of the three treatments. Comparisons for differences among the means on the ANOVA indicated that beginners scored significantly higher (p<.05) in the preteaching condition than in either the vocabulary or the read-and-test methods. Intermediates achieved almost as well under the read-and-test treatment as they did by the vocabulary method. Advanced subjects scored highest in the read-and-test treatment, although there was no significant difference between any of their treatments. Although the results appear to be mixed, Hudson concludes that preteaching schemata can overcome readers' language or decoding deficiencies. Thus, the experiment supports the compensatory theory. The Interactive Model of Reading The interactive nature of reading was underlined by Carrell and Eisterhold (1983). According to schema- theory, the reader's background information and the information in the text interact in the reading process. It is crucial to reading that the reader be able to relate textual information to the reader's knowledge of the world. The implications for ESL reading include a need for 17 sensitivity to cultural knowledge in a text and previewing texts for students so as to preteach culturally loaded concepts. Essentially, second-language teachers must become aware of the cultural differences implied in text, and be able to teach their students these features so as to avoid cultural interference. Thus, Carrell and Eisterhold’s study provided valuable insight into the cultural aspect of background knowledge in L2 reading. The Effect of Preteaching Text Structure Carrell (1985) investigated the effect of teaching text structure on ESL reading. Noting that the direct teaching of text structure can facilitate reading in a first language, Carrell determined to examine whether similar preteaching might improve ESL reading performance. There were twenty-five subjects in the study, with fourteen composing the experimental group and eleven in the control group. During five hour-long training sessions, the intermediate-level reading students were taught to comprehend four main expository patterns associated with simple, natural reading selections. While the experimental group was receiving explicit training in specific structural patterns, the control group examined the same reading materials. The control group’s preteaching, however, consisted of such linguistic operations as grammar drills, sentence analysis, vocabulary study, and reading comprehension. The control group did not receive preteaching in English rhetorical organization, but 18 otherwise was prepared similarly to the experimental group. The experimental and control groups were given a pretest and a posttest, and the experimental group was given a second posttest to check for continued training effect. The tests consisted of reading, written recall, and organizational identification. The recalls were judged for quality of ideas, whether thesis, main idea, topics, subtopic, or detail. Results show that experimental subjects recognized the structural patterns they had been pretaught, and continued to recognize them in the second posttest, indicating a very definite persistence of effect. The control group scored significantly lower according to Chi-square tests at p<.05. An ANOVA yielded no significant difference between the groups on the pretest. However, the posttest means for the experimental group in recall of theses, main ideas, topics, and subtopics were significantly higher (p<.05) than those of the control group, according to an ANOVA. Hence, the quality of the recall of the experimental group was significantly better than that of the control group. This result strongly indicates that overt preinstruction in rhetorical organization can improve ESL reading comprehension. Accordingly, Carrell recommends the inclusion of training in text structure in ESL reading programs. Her conclusion is that this type of schematic knowledge is important to the development of comprehension in reading as well as an appreciation of content schemata. 19 Barnitz and the Interactive Model of L2 Reading Barnitz's (1986) ”multidisciplinary” review argues for "the roles of cultural schemata and discourse structure in an interactive model of first- and second-language reading" (p. 95). The review of content schemata and comprehension in a second language harks back to an early study (Bartlett, 1932) of how Englishmen recall Indian folk stories differently from Indians. More recently, American college students have been found to recall Grimm's fairy tales more readily than Apache tales. Readers are found to distort stories from unfamiliar cultures, and to remember stories from the native culture more accurately. Hence, cultural schemata can affect reading. Studies (Johnson, 1981: Johnson, 1982) are also reported which indicate that "subjects made more cultural inferences/elaborations in the recall of the text from their own culture, [demonstrating] the interaction of language and culture in text comprehension” (p. 101). Barnitz also reviews the effect of formal schemata on second-language reading. Kaplan’s (1966, 1983) studies in contrastive rhetoric show that the ethnolinguistic thought patterns of various cultures vary. For example, English rhetoric emphasizes linear thought patterns (topic followed by support): Oriental rhetorics avoid direct discussion of the topic, opting to view a topic from several tangential viewpoints. Romance patterns feature many digressions and additions: Semitic rhetoric involves numerous parallel 20 structures: and Vietnamese involves the joining of antithetical compound words. All of these linguistic patterns mirror the values of the culture, with English culture presumably placing a premium on direct logic. Barnitz surveys psycholinguistic processing studies as well. In expository processing, it was found that English readers tend to process Japanese text by English rhetorical patterns and so experience considerable difficulty. When discourse patterns match readers’ expectations, better reading occurs. Barnitz finds ”text and text schemata [to be] independent variables in native and nonnative speakers' models of reading” (p. 106). Even stories, with their apparently universal properties, actually contain some culturally-specific aspects, as for example the lack of goals for heroes in Japanese tales. Cohesion, "the property of text whereby the surface elements of a text are connected" (p. 107) is thought to affect second-language reading also. Cohesion studies show that recall is influenced by cohesion. For instance, the distance between a pronoun and its referrent is a significant variable in second-language reading. The number of possible antecedents may also overload the second-language learner's processor. Barnitz suggests that L2 reading instructors consider both schemata and linguistic factors. Preteaching schemata may facilitate the comprehension of text with difficult vocabulary. Background knowledge—learning strategies 21 must be taught, and techniques in relating schemata to passages are vital. Both content schemata and discourse structure must be viewed as interacting with the elements of language proficiency in the process of reading in a second language. The Effect of Preteaching Background Information Floyd and Carrell (1987) investigated the effect of preteaching cultural content schemata. The question behind the study was: "Could reading be improved by teaching background knowledge?" The twenty intermediate-level subjects were unaware of the two cultural aspects of American culture to be tested. The testing materials consisted of two accounts of a common Fourth of July festivity at Boston. Although both accounts had an identical number of idea units, one was much more complex syntactically. The experimental group had two hour-long training sessions. The text was not taught, although vocabulary was presented. The control group received no special training. To test for the difficulty of the syntactically harder version of the text, half of the control and half of the experimental group read this complex version, while the other half of each group read the easy-grammar text. Even after controlling for variance due to pretesting, there were significant differences between the means of the experimental and control groups on the recall exam, but no significant differences between the syntactically easy and hard scores. The researchers 22 concluded that “in the ESL classroom cultural content is of the utmost importance, and that content may be, and often must be, explicitly taught" (p. 103). Taglieber, Johnson, and Yarbrough's (1988) research yields further insight into the effect of prereading activities in EFL. In the study, three prereading activities, plus a control condition, were utilized. All forty subjects read each of the four passages, but under varying treatments. Before reading, the.subjects not in the control condition received training in either pictorial context, vocabulary, or prequestioning. After reading, the subjects completed both open-ended and multiple-choice examinations. Subsequent statistical analysis revealed that all prereading activities produced markedly improved scores on the multiple-choice measure. It was speculated that the open-ended examination produced no significant results because of the time constraints and difficulty in recalling in a foreign language. The researchers assume that relatively higher mean scores for pictorial context and prequestioning compared with the means for vocabulary preteaching indicate the presence of a compensatory effect. The method of vocabulary teaching was noted as a possible cause for the lower scores. Schematizing as Interactive Reading The interactive nature of the reading process is further reviewed by Swaffar (1988). As with several other reading theorists, by "interactive" Swaffar actually means 23 "schematizing." To be sure, Swaffar mentions the top-down and bottom-up processes. Yet, the emphasis on the role of the reader’s knowledge emerges as central in this study. The boldest assertion in this direction may be: "What is understood depends on the reader rather than on the text" (p. 124). Other strong remarks to this effect include: "One strong interactive component can compensate for a weaker one", and "Limited command of language is not an insurmountable barrier to L2 reading” (p. 125). Citing an instance when a student interpreted the German word ”Wald" (woods) correctly in one context, but mistook it for ”world" in another, Swaffar states that schematizing may overpower word meaning in L2 reading because the surface representation (lexicon and syntax) of a passage is less important than background knowledge to comprehension. The problems posed by a lack of knowledge of appropriate rhetorical patterns are noted as well. Whereas inadequate proficiency in the language poses real reading problems, Swaffar argues that "L2 students rely more heavily on these factors [background and formal schemata] than do native language readers” (p. 129). Nevertheless, Swaffar supposes that a bare minimum vocabulary (2,000 to 5,000 words) seems logically necessary in order for the L2 reader to engage in even "ballpark guessing.” Because so many texts use metaphorical language, L2 readers must also be trained to comprehend semantic extensions (p. 131). There is evidence that syntax and 24 lexicon play virtually equal roles in reading. Subjects with reduced ability in either area showed limited gains in reading compared to subjects with more balanced syntactic and vocabulary knowledge. The two variables must be seen as interactive. In sum, Swaffar finds the L1 and L2 reading process to be highly interactive, and a part of communicative language learning (p. 140). Because reading is above all a reasoning task (albeit connected to a language task), prereading is paramount (p. 141). Summary The application of schema-theory has advanced our understanding of L2 reading. Insight into the role of background knowledge in the reading process has been gained. Even when the L2 reader literally understands every word in a passage, the overall meaning may prove elusive because of insufficient cultural background information. In order to fill this gap, second-language reading teachers have pretaught cultural information. Similarly, L2 readers may lack knowledge of L1 structural patterns. Explicit preteaching of rhetorical patterns has proven successful. On the other hand, preinstruction in vocabulary and grammatical patterns has not proven useful. It has been hypothesized that induced schemata may increase reading comprehension. Further, the compensatory theory specifies that background knowledge gonngnggteg for the reader's vocabulary, syntactic, and other linguistic shortcomings in a second language. There is no evidence, 25 however, to support the compensatory theory. Whereas some studies (Carrell, 1985: Barnitz, 1986) support the idea that preteaching background knowledge and rhetorical patterns may improve reading comprehension, there is a lack of research to show that L2 readers compensate for vocabulary limitations by use of pretaught topical information. Barnitz’ comment that several variables (including language ability) may interact in reading seems valid. Yet there is no positive evidence that vocabulary comprehension and background knowledge, for example, are indeed interactive variables. Swaffar's suggestion that there may be a "bare minimum" language ability requisite before top-down processing may be used appears to be logical. Nevertheless, such comments are merely speculative. Schemata knowlege undoubtedly affects the reading process. Yet this fact does not necessitate the slighting of bottom-up processing in L2 reading models. STUDIES ON THE EFFECT OF VOCABULARY IN L1 READING Marks, Doctorow, and Wittrock (1974) examined more than two hundred sixth graders to determine if varying the frequency of difficult (unknown) words would significantly affect comprehension. The authors hypothesized that "even a few minor variations in meaningfulness, such as inserting a few words that are well understood, can be sufficient to enhance the comprehension of the total story” (p. 259). Word difficulty as measured by word frequency is posited to be vital to readability. Five stories from the sixth grade 26 level of SRA were adapted as materials. Fifteen percent of the content words in the reading materials were altered. Half of the paragraphs became ”difficult”, and half were rendered ”easy” by word substitution. The frequency of the substitution words was determined by referring to Carroll, Davies, and Richman's Word Frequency Book. Even the low frequency words occur in sixth grade readers, however. The hypothesis was supported by a MANOVA at the .05 level. The mean increase in comprehension under the."easy” vocabulary treatment as tested reached twenty-five percent as correct answers averaged forty-two percent on the low-frequency versions and sixty-seven percent on the high-frequency versions of the reading materials. The authors remark that "specific words in sentences [have] important, integral messages within themselves" (p. 262). Such messages are responded to by the reader in a non-linear way. Because significant increases in reading comprehension can be achieved by the use of higher-frequency vocabulary, children's reading materials should be carefully adapted. When Vocabulary Preteaching May Not Help Jenkins et al. (1978) conducted a series of experiments to ascertain the effects of vocabulary instruction on vocabulary and reading comprehension. In three experiments involving learning disabled and non-disabled school-age children, the researcher found that vocabulary could be learned through as little as two hours of direct instruction. Subjects were drilled on synonyms and allowed 27 to make inferences. Most students acquired some vocabulary from the instruction, although the learning-disabled children required more training and showed reduced gains. Although most subjects evidenced increased comprehension at the sentence level, little effect was indicated at the passage comprehension level due to training. Jenkins concludes that the importance of vocabulary to over-all reading comprehension may be overrated. Nevertheless, vocabulary training does appear useful to vocabulary learning. Jenkin’s study may not prove the importance of vocabulary knowledge to reading, but it does underline the depth of vocabulary comprehension necessary in order for reading comprehension to be improved. Mere awareness of synonym matches may not be enough to increase reading comprehension. As the following research reveals, the reader must have several available associations for vocabulary items in order to process text efficiently. In-depth Vocabulary Knowledge and Reading Comprehension Anderson and Freebody (1981) have contributed an excellent summary on vocabulary and reading comprehension. This work particularly investigates readers’ knowledge of word meanings. It is suggested that the number of meanings a reader attaches to a lexical item strongly predicts discourse comprehension. Merely knowing a single synonym match for a word does not facilitate reading. In such an instance, retrieval is slow and uncertain, and the whole 28 reading process readily fails as the reader's attention turns from the topic to the grasping of a difficult or elusive vocabulary item. The authors cite the ”most robust” connection between vocabulary and intelligence, with reported correlations exceeding .90. Studies linking vocabulary to reading exhibit correlations from .41 to .93, regardless of reader- age or language. Readability analyses also underscore vocabulary, and some find it to be eighty percent of variance (r= .90). Why does vocabulary comprehension greatly affect reading? The instrumental hypothesis holds that word knowledge enables readers to comprehend text. Adherents to the aptitude hypothesis claim that superior readers read better and have better vocabularies because these readers are more apt. The knowledge hypothesis reads that persons who know the culture well also know more vocabulary, and it is this understanding of culture which largely enables reading. Reading teachers who believe the aptitude position would be rather fatalistic about their students' prospects. Instrumentalists might favor direct instruction of vocabulary. And knowledge proponents would propose wide reading experiences. Anderson and Freebody (1981) show that learners grasp only part of the meaning of new words at first. For example, young children may know that ”sell" corresponds to ”give", but they may not realize that the seller also obtains something in the bargain. Vocabulary knowledge 29 deepens over time as nuances follow broad distinctions of meaning. Perhaps because of this depth of word knowledge, and also due to the existence of roots and derivates, estimates of individuals' vocabulary size vary wildly. One researcher (Seashore, 1933) estimates college sophomores' vocabulary to be 15,000, while another (Hartman, 1946) cites 200,000. Anderson and Freebody note that another problem in estimating vocabulary size is the unbalanced distribution of words by frequency. Whereas 2,000 words may account for ninety-five percent of words, forty percent of the words in one sampling occurred once. More frequent words are likely to be known. They conclude that although vocabulary knowledge is significantly correlated to reading comprehension, it may be impossible to determine how many words a person knows. Perhaps our static idea of isolated words renders our measurements ineffective. The Effect of Difficult Vocabulary on Reading Comprehension Freebody and Anderson (1983a) have investigated the effect of difficult vocabulary on reading comprehension in a series of studies. In the first experiment, one hundred sixth grade children read in three treatments. Whereas one passage contained high-frequency words, another had low- frequency words substituted for every third content word, and the third version replaced one content word in six with a more difficult synonym. Replacement words were selected by intuition, and the passages were taken from a social studies textbook for sixth graders. Reliable effects for 30 reduced comprehension were not found for the easy word or for the intermediate condition. Only in the condition with one difficult (rare) word per three words was a significant lowering of comprehension scores found. The experimenters guessed that the readers may not have been bothered by encountering one difficult word in six because some of the words occurred in supporting rather than main information positions. To control for this effect, Freebody and Anderson conducted a second experiment on the effect of difficult vocabulary in unimportant and important propositions. When rare vocabulary occured in trivial propositions, students' retrieval of main ideas was facilitated, perhaps because subjects simply did not process sentences with difficult vocabulary. The effect of rare words in important propositions is unclear, although sentence recognition was impaired compared to the easy condition. It was found that a high proportion of low-frequency vocabulary is necessary to significantly impede the reading process. This may be due to insensitivity of measurement and the normal redundancy of text. School text authors, however, may not include rare words unless they are especially needed. Thus, difficult words may not occur often in trivial propositions. The experimenters suppose that readers read around difficult words, if at all possible, in their attempts to reconstruct text. In additon, the readers may not have found the words as difficult as it was supposed. 31 The Effects of Vocabulary and Schema on L1 Reading Freebody and Anderson (1983b) have extended their research to include the effects of difficult vocabulary, cohesion, and schema. Both studies cited herein relate to the interactive theory of reading, yet neither experiment supports it. The hypothesis for the first experiment states that ”difficulty will have minimal effects on comprehension when cohesion is high, but that, with decreasing cohesion, the effects of difficult vocabulary will become less pronounced" (p. 281). Eighty fourth graders read three passages from fifth grade social studies texts. The high-cohesion version of the text was created by using even more repetitions and simplified substitutions than were employed in the natural text: the low-cohesion versions had less explicit ties (produced by using more pronouns and ellipsis and by avoiding the use of connectives). A third "inconsiderate" version was formed by inserting extraneous propositions in the text. Difficult synonyms were then inserted in some versions, thus creating a total of six textual versions (three levels of cohesion and two levels of vocabulary). Half of the subjects read the difficult-vocabulary versions, and half read the easy-vocabulary texts. All students read texts at all three levels of cohesion. After reading, students took a multiple-choice vocabulary test, wrote a brief summary, and completed a sentence-verification task over important ideas, unimportant ideas, and ideas not in the text. 32 The amazing result of the multiple regression analysis is a perfectly null F value (F<1.00) for interaction between the variables of vocabulary and cohesion! Performance improved in the easy vocabulary treatment, and "inconsiderate” text negatively influenced reading comprehension. Lack of cohesion did not yield lowered scores, except in particular sentence measures, showing the consequences of a processing load factor, which makes reading more difficult, but does not seriously impede it. But vocabulary showed a direct effect on comprehension, exhibiting ”no interactions with ability, passage, position, or cohesion levels" (p. 286). Freebody and Anderson speculate that vocabulary and cohesion affect the reader differently. When readers encounter hard words, they may skip them, so that the load factor is not increased. Lack of cohesion, however, is less obvious. So the reader plows through incohesive text, and experiences loading. Are Schemata and Vocabulary Interactive Variables? In the second experiment cited, Freeman and Anderson investigate the relationship of schemata knowledge and vocabulary. Topical and vocabulary knowledge have been found to correlate to reading comprehension. The researchers hypothesized that a significant interaction could be measured between the two variables. Eighty—eight sixth graders read a familiar and an unknown passage, with vocabulary difficulty held as a between-subject variable. 33 The main hypothesis was not supported by analysis, as the vocabulary by familiarity interaction was insignificant in all measures (F<1.00). Easy vocabulary versions were read better, although not generally at a significant level. Familiar versions were also better remembered. The main finding of both experiments is that no support was gained for the compensatory or interactive view of reading. No matter how probable such a theory may seem, this research does not show that readers use strengths in schemata to compensate for vocabulary, or vice-versa. Preteaching May Not Overcome Difficult Vocabulary Stahl and Jacobson (1986) also have examined the connections between vocabulary difficulty and topical knowledge. In this study, sixty-one sixth graders read easy or difficult versions of one text. The research varies from that of Freebody and Anderson in that it included cultural preinstruction to specifically control for amount of prior knowledge. An obscure topic was deliberately selected for the reading passage. A difficult version was created by substituting for about every third content word. Multiple-choice and sentence-verification tests were developed to measure explicit, implicit, and sentence-level comprehension. Subjects were divided into two groups and given either relevant or irrelevant preteaching. A vocabulary-by-preteaching multivariate analysis of variance and one way ANOVA to detect significant sources of variation were utilized. The effect 34 of vocabulary was highly significant on the MANOVA (p<.001), while that of preinstruction yielded significance at the lowest level (p<.05). The authors found that "preteaching improved comprehensioin but did not overcome the effects of vocabulary difficulty" (p. 316). Because no significant interaction between preteaching and vocabulary was found, the compensatory view was not supported. Attempts to Find Interaction Between Vocabulary and Schemata Interest in the interactive effect led Stahl, Jacobson, Davis and Davis (1989) to readdress the issue in three experimental studies. In each of the studies, the researchers attempted to find interaction between vocabulary and prior knowledge, but failed, although vocabulary and background knowledge were found to be independently significant variables. Ninety sixth graders from four classes were assigned to groups for either pertinent or irrelevant preteaching in study one. After reading either an easy- or a difficult-vocabulary version of text, subjects took free-recall, multiple-choice, and sentence-verification tests. The free-recalls were divided into pausal units, and rated as main ideas, details, or distracting ideas. Multiple-choice questions were classified as textually explicit, implicit, or inferred from prior knowledge. In the sentence-verification, subjects merely indicated whether propositions were from the text or not. To test for the effect of preteaching, 35 two of the groups were given correct preinstruction regarding the text, whereas two groups were provided irrelevant preteaching. Analysis revealed that both vocabulary and preinstruction significantly affected the comprehension of main ideas and supporting information, but there was no interaction between vocabulary and preinstruction. Difficulty with vocabulary seems to have disrupted the structuring of a "coherent text base" (p. 36). Incorrect prior knowledge appears to have affected which information readers focus on. Thus, interaction may not occur because the variables operate independently. To determine whether prior knowledge affects sentence-level recall, the researchers created a cloze test. Prior studies had indicated that cloze is sensitive to sentence comprehension, but not to discourse comprehension. It was hypothesized that preinstruction would not affect cloze scores, but that vocabulary would. Again 92 sixth graders were chosen, but this time to complete either difficult- or easy-vocabulary versions from which every fifth word had been deleted. Once more, two groups received pertinent preinstruction, while two others received misinformation. Analysis showed that neither preteaching nor interaction were significant, albeit preinstruction neared significance (p<.10). Vocabulary had a very significant effect (p<.01), as expected. The replacement of function words was also impeded in the difficult -ocabulary condition. This result indicates that 36 vocabulary difficulty can impair the construction of coherent text base because function words signal sentence connections. Perhaps because the difficult vocabulary in this passage was general in nature, preteaching did not affect the replacement of content words. Study three was conducted to examine two further hypotheses. First, it was hypothesized that the recall of the order of events from a passage would be affected by vocabulary difficulty. Second, it was supposed that preteaching would affect the causal relations measure, which depended on inferences. Ninety-nine sixth graders from four classes read the same passages, but were given a sequencing, a multiple-choice assessment of causal- relations test, and an importance-rating task. Analysis found the effect of vocabulary significant (p<.03), but no effect was found for preteaching or interaction for the event-ordering and causal-relations measures. In the importance-rating test, preteaching was significant (p<.02), but vocabulary and interaction were not. The authors conclude that whereas vocabulary and preteaching may be found to affect comprehension independently, there appears to be no interactive or compensatory effect. The researchers conclude that "vocabulary difficulty affects a very literal comprehension of the text" (p. 40). On the other hand, preteaching may have more affect on summarization, or the more selective process of tagging propositions as important, as it affects knowledge. 37 The Relationship of Difficult Vocabulary to L1 Reading Studies in first language reading support that the presence of difficult (low-frequency) vocabulary may affect reading comprehension significantly. Unknown vocabulary correlate strongly to low frequency. It may be that missing the meaning of difficult words, some of which bear "compressed meaning", causes comprehension problems. Such words are utilized because one word serves to convey meaning which otherwise would have to be expressed in several words. Schemata knowledge does not seem to help readers process difficult-vocabulary text. VOCABULARY AND PRIOR KNOWLEDGE IN SECOND-LANGUAGE READING As in first-language studies, there has been considerable interest in the effect of vocabulary and background knowledge in reading in a second language. The parallel is not exact, however. For whereas first-language researchers continue to examine the role of difficult vocabulary as well as schemata in reading, in L2 reading such research has become scarce. There are some early papers on vocabulary in ESL which shall be cited first. An Emphasis on Vocabulary in Earlier L2 Reading Studies Twaddell (1973) recommended that vocabulary be expanded through extensive experience in the target language, and not by artificial vocabulary exercises. His basic argument was that massive vocabulary learning is practical only by this natural method, and that "the indispensable [aspect] 38 of FL [foreign-language] acquisition” is vocabulary expansion (p. 64). Given the polysemy of words, glosses provide at best merely a beginning point for second- language reading. Because of the large number of low-frequency words, "sensible guessing" is the best way to acquire vocabulary. Teachers should accept and even expect a "temporary vagueness" in learner's answers. Eskey (1973) remarked that the reading teacher must give indirect instruction in reading components such as syntax, lexicon, cultural information, and rhetoric, "since no one knows what reading is or how readers do it” (p. 169). His comments are quite progressive, as they anticipate current interest in preteaching both prior knowledge and rhetorical text patterns in L2 reading. Eskey underlines the importance of vocabulary in reading, and suggests that it be directly taught and then reinforced by extensive reading. Cloze is also recommended to promote the acquisition and testing of various reading skills. Do Language Deficiencies "Short-Circuit” L2 Reading? Clarke (1979) introduced the concept of the "short-circuit" effect in his study of good L1 readers who became poor readers in English due to language deficiencies. Good reading skills may not transfer into the L2 when the reader’s basic competence is below a critical level. The research Clarke investigated emphasized the importance of teaching reading as ”a psycholinguistic guessing game”. However, ”attempting to 39 teach someone to use the phonemic, morphemic, syntactic, semantic, and discourse cues of the language before he has learned what they are . . . seems unrealistic" (p. 139). Clarke finds that L2 reading teachers must teach both language skills and reading skills. He remarks on the weaknesses of assuming that oral miscues produced in oral reading parallelled those made in silent reading. Flaws in retelling procedures constitute another methodological problem, as subjects simply may not tell all they know. Cloze tests appear to be very useful tools in determining reading comprehension. Perkins (1983) recognized the potential hazards of linguistic interference to the second-language reading process. Perkins researched the degree of inferential misrecognitions of forty simple locative/spatial sentences. The forty-three beginning, intermediate, and advanced level adult subjects were asked to mark sentence matches as true statements, false statements, true inferences, or false inferences. An ANOVA indicated that the degree of misrecognition was significant (p<.01) between groups. Beginning and intermediate readers made significantly more misrecognitions than advanced readers. The implication from the research is that (advanced) readers who completely process a text understand its implications more than those (beginning- and intermediate-level readers) who process less fully. Devine (1987) indicates that overall language 4O proficiency correlates positively with L2 reading ability. Low general competence in the target language imposes a "linguistic limit” on second-language reading. To verify this assumption, an on-going study of the general L2 proficiency as well as of the L2 reading competence of twenty beginning-level ESL students was conducted during the 1982-83 school year. The subjects were given three tests at three-month intervals. The findings support that "increasing language competence correlates positively with increases in syntactic and semantic acceptability of oral reading miscues” (p. 83). Students who improved in holistic test scores also tended to show increased reading ability. Interestingly, the good readers in this study did not change from dependence on syntactic cues to reliance on semantic ones. They continued to utilize both types of information. Furthermore, subjects did not attend less to processing visual cues as their ability to consider textual meaning improved. Both findings support the view that, despite emerging L2 reading proficiency, bottom-up processing is not replaced by top-down processing. The Importance of L2 Vocabulary to Academic Achievement In a study of which variables count most in second- language learning, Saville-Troike (1984) rates vocabulary as very important for academic achievement. Although the children in this study began the school year with similar English competencies and were from comparable socioeconomic groups, by the end of the year their academic achievement 41 varied considerably. Several variables were examined to determine the causes of this result, including oral proficiency, home-language influence, personality, and attitude. The subjects were nineteen children attending ESL and content area classes at the elementary school level. Tests show little correlation between subjects’ spoken grammaticality and academic achievement. The one oral feature which did match achievement is ”the number of different vocabulary items each child used" in the interviews (p. 207). Tapings of ESL sessions throughout the year reveal that the childrens’ expressions were often ungrammatical, yet "grammar was never corrected when real communication was at stake" (p. 211). Saville-Troike concludes that if teachers are most concerned about students’ school achievement, then they must know that vocabulary is paramount for content learning. Spoken practice, grammar exercises, and social interaction may not be as valuable as learning the appropriate content terms for L2 reading and listening. Furthermore, the subjects who learned the most had the chance to talk about what they were learning in their native language, which suggests a strong need for a bilingual approach whenever feasible. Why Vocabulary Has Been Underrated Carter’s review (1987) generalizes that ”for many years vocabulary has been the poor relation of language teaching. Its neglect is in part due to a specialization 42 in linguistic research on syntax and phonology" (p. 3). Especially in Great Britain, however, there is something of a revival of interest in vocabulary teaching. Direct learning is appropriate for beginners, but after learning the core 2-3,000 words, the learner must use indirect, context-based, inferential strategies to learn the vast range of infrequently encountered items. Lexical cohesion, the relations between surface features in the text, and coherence, the relationship of ideas represented in the text, have been distinguished. Also, cloze is being extended to the level of discourse by the deletion of communicative units. Although some work is being done in vocabulary studies, much more remains before it will reach the level of sophistication of other areas of linguistic research. Thus, it is likely that L2 studies will continue to highlight variables other than vocabulary in the reading process. On the basis of his own research, Diller (1978) contends that college students recognize more than 240,000 words. He finds that a base vocabulary of 10,000 words forms an adequate matrix for learning new vocabulary in context. An advocate of direct methods of language learning (and a critic of structural methods), Diller believes that vocabulary can be taught and learned. He proposes that too much time is devoted to developing sentence patterns and too little attention is given to learning vocabulary in meaningful and grammatical context. 43 Eskey (1988) also has taken note of the disproportionate interest in top-down rather than bottom-up processing in L2 reading studies. This researcher supports the psycholinguistic model, but believes its advocates have gone too far in deemphasizing ”the perceptual and decoding dimensions of that process” (p. 93). Although encouraged by the appearance of the interactive model, Eskey has been equally disappointed in the practical implications of this position. Once again, the proponents of this model emphasize the role of top-down processing to the exclusion of any investigation of bottom-up processing. Eskey notes that "language is a major problem in second language reading, and . . . even educated guessing at meaning is no substitute for accurate decoding" (p. 97). REVIEW SUMMARY Although second-language reading teachers probably have maintained their interest in bottom-up processing, most recent research has highlighted top-down processing. The terms have changed from the Psycholinguistic Model to Schema-theory to Interactive Reading, but the basic emphasis on reading as educated guessing has not. Of course, L2 readers process text "behind the eyeball". Cultural background information as well as knowledge of text structure are undeniably useful to the reader. The good second-language reader, similar to the good first- language reader, uses several cues to process text. The Compensative Theory, however, lacks support. 44 Because background information may be pretaught successfully, it has been assumed that prior knowledge enables readers to overcome linguistic deficiencies. Repeated attempts to find evidence of L1 readers’ making up for linguistic limitations by recourse to top-down processing strengths have failed. There is no evidence of interaction between vocabulary and either cohesion or schemata. Furthermore, preteaching content information does not enable Ll readers to overcome problems in comprehending text with difficult vocabulary. No such studies have been conducted in L2 reading. There is evidence for a linguistic ”short—circuit" in L2 reading. Readers who lack sufficient bottom-up processing skills may be unable to read even texts about which they possess topical knowledge. Vocabulary comprehension in the target language has been correlated to both reading comprehension and overall academic success. The popularity of educated guessing over accurate decoding, however, has resulted in the devaluation of vocabulary as a cue in L2 reading. Purpose of the Study There is a gap in L2 reading research as to whether the activation of topical knowledge enables readers to comprehend text with difficult vocabulary. Although researchers (Swaffar, 1988; Hudson, 1982: Barnitz, 1986) have supposed that strengths in top-down processing may compensate for weaknesses in bottom-up processing, there 45 have been no investigations of this hypothesis. It is the purpose of the present study to investigate the effects of schema-activation and difficult vocabulary on reading comprehension in a second language. Schema-activation is believed to facilitate L2 reading comprehension (Taglieber, Johnson, and Yarbrough, 1988). Difficult vocabulary may disrupt the processing of L2 texts (Barnitz, 1986: Carter, 1987). In this study, it is assumed that both schema- activation and a lack of difficult vocabulary facilitate L2 reading. The following hypotheses regarding the relationship between schema-activation and difficult vocabulary in L2 reading were formulated: 1. L2 readers who receive schema-activation will achieve significantly higher reading comprehension scores than readers who do not receive schema- activation. 2. L2 readers who process passages with difficult vocabulary will attain significantly lower reading comprehension scores than readers who process passages without difficult vocabulary. 3. L2 readers who receive schema-activation and read passages without difficult vocabulary will achieve significantly higher reading comprehension scores than readers who process texts under all other conditions (with schema-activation and with difficult vocabulary: without schema-activation and without difficult vocabulary: without schema- activation and with difficult vocabulary). CHAPTER THREE METHOD It is the purpose of the study to investigate the effects of schema-activation and difficult vocabulary on reading comprehension in a second language. Previous thinking suggests that schema-activation via pre-reading activities may facilitate L2 reading comprehension. It also has been thought that texts with difficult vocabulary may "short-circuit" the reading process. To probe these assumptions, a study with a Repeated Measures, two-way design (subjects by experimental treatments) was used to compensate for the relatively low number (twelve) of subjects in the study. In this method, all experimental treatments are administered to all subjects. The four variables in this study are with or without schema- activation (warm or cold) and with or without difficult vocabulary (difficult or easy). The subjects read passages under four treatment conditions: (1) Easy/Warm, (2) Easy/Cold, (3) Difficult/Warm, and (4) Difficult/Cold. SUBJECTS The subjects in the study comprised the entire population of an advanced-level ESL reading class in a small Midwestern university. Their TOEFL scores ranged between 450 and 500. Advanced-level L2 subjects were 46 47 selected purposely to examine the use (or disuse) of bottom-up and top-down processing by good readers. It has been supposed that beginning and intermediate-level L2 learners rely heavily on bottom-up processing. Advanced learners, however, are thought to depend more on top-down processing (Swaffar, 1988). All subjects had been admitted into the university on the basis of their English proficiency scores and other qualifications. Each of the students was taking both ESL and content area courses in the university, either at the undergraduate or graduate level. As in many other intensive ESL classes in a college environment, members represented a variety of linguistic and cultural groups. A subject who spoke French came from Haiti. There were two speakers of Spanish: one from Puerto Rico, the other from Mexico. Six of the subjects hailed from Japan. The three Portuguese-speaking subjects were Brazilians. The subjects had been in the United States less than two years. The shortest period of stay was just four months. Most of the students had been living in America for at least seven months at the time of the study. All subjects had taken at least one intensive, ten-week session of ESL at the university previous to the time of this study. Although there were 18 class members in all, the scores of those subjects who did not attend all four sessions connected with the study were dropped. Thus, the number of subjects in the final data analysis was twelve, as indicated above in the nationality information. 48 Subjects ranged in age from eighteen to twenty-seven. There were four males and eight females in the final count. In the opinion of the classroom instructor, none of the students had a ”reading problem”. All were judged to be competent ESL readers. The subjects were aware that they were volunteering for a research project. None, however, knew the exact nature of the study, or of the research hypotheses until after all data had been gathered. READING MATERIALS The nature of the study required that both difficult- and easy-vocabulary versions of all reading materials be available. It was determined to adapt materials from existing ESL reading passages. The general procedure followed was modeled on the procedure used by Stahl, Jacobsen, Davis, and Davis (1989) with native speakers. To control for vocabulary differences alone, it was necessary to adapt content-identical versions of one text. It would be advisable also that the reading passages be taken from an ESL reading textbook, because such materials would be at approximately a suitable linguistic and conceptual level for the typical ESL reading student. To this end, four passages from Beggeen the Ligee (Zukowski/Faust et al, 1983), a commonly used, intermediate-advanced level ESL reading text, were selected. This text had previously been used with success with intensive ESL classes at the advanced reading level: it was assumed that the material would present no special problems to the readers. 49 Adaptation of the Reading Passages Without special vocabulary training, the passages in the particular units selected might be difficult for the subjects because these materials were used as a basis for learning advanced vocabulary. Because the vocabulary exercises found in the text were not included, it was expected that the original passages would provide a suitable starting point for the creation of a difficult vocabulary reading version. To control for topical knowledge, passages about which prior classes of students had displayed little background knowledge were chosen. As expected, students’ comments during the course of the study revealed that they had little prior knowledge of the topics of each of the four passages. The passages were technical, rather than general or narrative. Topics included the structure of the earth, earthquake prediction, deforestation, and the social consequences of overpopulation. It was believed that such topics would engage the young adult readers’ attentions, and permit a valuable check of their reading comprehension potential. Similar previous first-language studies had used selections from social studies also. This area of study had been chosen to control for prior knowledge. In the L1 studies, however, the materials had not been consciously created to teach vocabulary. Both the selection of reading materials and the manner of their adaptation were patterned after the earlier study, despite the slight differences noted. 50 The original passages were judged to be somewhat lengthier than appropriate to the time constraints involved in the study. Since several steps had to be executed within the time-space of a fifty-minute class period on each of the four days needed in the study, the passages needed to be short enough to be read in twenty minutes or less. Therefore, the original texts were condensed to fill one single-spaced page, and contained approximately 400 to 500 words each. The language, basic structure and style of the originals were maintained in general. Nevertheless, some syntactic and rhetorical simplification inevitably resulted from the summarizing process. The number of paragraphs per passage was reduced from as many as eleven to a maximum of six: the goal being to produce a comprehenisible five-paragraph essay. Paragraphs in the adapted version were limited to no more than seven sentences. At times, complex sentences were made simple by the deletion of dependent clauses. All the same, the adapted versions of the original texts retained much the same style and even wording as the original. To ensure that the adapted texts maintained overall coherence, the passages were checked for coherence and overall readability by a native ESL professional. Vocabulary Pretests After the four passages had been condensed, they were scanned for apparently difficult words. The vocabulary in the original text were assumed to be difficult, as were any 51 other items which had been found troublesome in previous instructional use of the text. Next, an objective instrument was developed to verify assumptions regarding difficult (unknown) and easy (known) vocabulary. The two pretests were both multiple-choice vocabulary measures. The first multiple-choice examination contained vocabulary from the ”Earth Structure” and "Earthquake” reading passages. The second vocabulary test measured the subjects’ knowledge of difficult lexical items in the "Deforestation" and "Overpopulation" selections. On the tests, subjects were required to read sentences in which the word in question had been underlined. Next, students were to select the equivalent word from among the four alternatives listed below it. The underlined item was believed to be a difficult vocabulary term, but the correct alternative was a common vocabulary item which the subjects would be likely to be familiar with. The twelve subjects completed both vocabulary measures. Analysis of results from the vocabulary pretest showed that over 50% of the responses were incorrect matches for vocabulary items which had been assumed to be difficult. This low rate of correct matches was interpreted as support that these vocabulary items were difficult for the subjects. Easy vocabulary items were verified by correct matches of at least 60% of responses. This higher rate of correct responses was taken as an indication that these vocabulary items were less difficult for the subjects. 52 Only three of the items assumed to be difficult (unknown) were found not to be so on the objective measure. Where differences occurred between the researcher’s judgements and testing, the objective measure was assumed to be correct. Experience suggests that L2 subjects at times possess knowledge of uncommon words, yet exhibit complete unfamiliarity with words common to native speakers. Thus, the vocabulary items identified as ”difficult" are simply the words which the subjects in this study did not know. No attempt was made to find why one word was known while another word was not known. Easy- and Difficult-Vocabulary Versions of Passages Having thus identified difficult vocabulary items, it remained to design both easy- and difficult-vocabulary versions of all four reading passages. Because the original passage was written to teach difficult vocabulary and was at a high vocabulary level, the first condensed version served as the difficult-vocabulary form, with very minor adjustments to ensure the pre-selected spacing of difficult vocabulary items. The difficult-vocabulary version was found to contain an unknown vocabulary item at about every sixth content word. This condition closely paralleled that of the difficult-vocabulary version in previous L1 studies, although some researchers had not found significant effects for vocabulary difficulty at less than a one-in-three ratio (Freebody and Anderson, 1983a). Unknown words for L1 readers may be technical as well. It 53 was determined to use unknown, but not necessarily technical vocabulary in this study. Such items were selected so as to create a check which was not dependent on a narrow-range of vocabulary. Whereas some difficult words were technical ("plate tectonics" and ”deforestation"), in most cases the difficult word could be classified as general (”transpire" and ”bountiful”). Moreover, the passages themselves had been written with a general rather than a technical audience in mind. The design of the non-difficult vocabulary versions was accomplished by substituting words rated as commonly-used in place of most of the difficult words. Words which were intuitively rated as "difficult” by the agreement of two ESL professionals were counted as difficult. Subsequently, a weak mixture of difficult vocabulary words (one in fifteen) was inserted into the ”easy”-vocabulary versions, and a somewhat higher proportion of difficult words (one in six) was utilized in the difficult-vocabulary versions. The resulting easy-vocabulary version differed from the difficult version in that it had a difficult vocabulary item at intervals of fifteen content words, as opposed to the difficult version’s ratio of one-to-six. Easy- vocabulary versions thus had a lower density of low-frequency (unknown) lexical items than did the difficult-vocabulary forms. Thus, two content-identical, but lexically variant forms of each reading passage were developed. Each easy- vocabulary passage was matched with a difficult-vocabulary version. 54 Each set of passages contained an identical number of thought or pausal units, and indeed had much the same wording overall. The differences between the paired passages were due to vocabulary alterations only: no other changes were made. In their final form, the reading materials consisted of four pairs of (or eight) reading passages on four different topics. POSTTESTS Three tests were developed to measure reading various aspects of reading comprehension in a second language. The first was a cloze measure. The cloze was created by selecting two or three of the body paragraphs from all four passages. The subjects who had read the easy version might score higher due to better reading comprehension. But it was not believed that such advantage would be substantial, since both versions contained identical content. To guard against giving readers of either version an advantage, cloze deletions were selected from items which appeared on both versions. In all cases, deleted cloze items were content words. The purpose of the study was to measure the effect of difficult vocabulary on L2 reading comprehension. Therefore, it was not intended that the cloze test itself should be an intrinsically ”difficult" exercise. A "difficult" cloze bearing a deletion ratio of one-word-in-five was not used. On the other hand, using a blank ratio of one-word-in-ten might lengthen the measure, and the time consumed, unnecessarily. Cloze deletions were 55 made at every seventh content word as this frequency mediates between the normal boundary rates of one-in-five and one-in-ten. The cloze instrument had been found useful especially in measuring sentence-level comprehension in L1 studies (Stahl and Jacobsen, 1986: Freebody and Anderson, 1983b), and was expected to serve the same purpose in this study. To control for judgement disagreements, only the words listed in the text (exact matches) were accepted as correct on the cloze tests. Nevertheless, mis-spelled words were allowed if they were decipherable. The Written-Recall Measure The second posttest was a written-recall measure. Whereas it was believed that the cloze would provide a very particular measure of the subjects’ recall at the sentence level, the written recall was intended as a more global check of comprehension of main propositions and of details. To this end, students were specifically enjoined to write what they deemed to be central to the passages, as well as pertinent supporting information as time would allow. No particular length of written recall was prescribed, but the proctor, who was also the classroom instructor, asked the subjects to write as much as possible during the allotted time. Ten minutes were allowed for the written recall. Apart from these conditions, the written recalls were open-ended measures of reading comprehension. To prepare to judge the written recalls, it was necessary to evaluate first the four passages to 56 distinguish main ideas from supporting details. Using a procedure suggested by Omanson (1982), two readers independently divided the passages into pausal units, and then rated the units as main ideas or supporting details. Differences were reconciled by the decision of a third rater. When the two raters disagreed as to the nature of the pausal unit, the case was sent to arbitration. The third rater’s opinion was then matched with that of another judge, and the case was decided. Next, the subjects’ protocols were examined. The protocols also were divided into pausal units, which in turn were rated similarly to the reading passages. Besides identifying main ideas and details in the protocols, however, the raters also found irrelevant information due to incorrect recall or schematic reconstruction. The Oral-Recall Measure The third posttest consisted of an oral-recall measure. In this case, subjects were instructed to recall once more any main ideas or supporting information from the reading passages. The subjects readily responded to this prompt. Yet, after offering from two to four pieces of information, most of the subjects abruptly stopped. The non-verbal message of the subjects at this point seemed to indicate that they were uncertain whether they had "fulfilled" the requirements of the study. It seemed that the subjects were following ordinary oral-discourse patterns in which a single question does not elicit 57 numerous or extended responses. As the subjects’ store of information did not appear to be exhausted, students next were prompted to supply more data. Every attempt was made to avoid turning the student’s answer in any particular direction. Prompts were made purposely vague, such as "Can you remember anything else?” and ”What do you think was important in the passage?" At times the prompter repeated the subject’s last statement, and asked what else might be recalled. It was noticed repeatedly during these oral- recall sessions that the subjects seemed to be attempting to please or satisfy the prompter by giving just enough, without seriously stretching the subjects’memories. Whereas the writer may pause to think after writing a sentence, oral respondents took little time in formulating answers to prompts. An oral-recall measure had been included in the research design because subjects are known to respond variously to written and oral testing. It was supposed that subjects who supplied little information on one measure might provide more and better data on the other. Obviously, some L2 subjects would possess better spoken proficiency than written, whereas others would have an opposite combination of skills. The use of these forms of recall represented a deliberate attempt to permit subjects to exhibit comprehension by any means possible. In any event, some comparison of written and oral responses thus would be possible. Oral recalls were reduced to transcriptions, 58 which in turn were analyzed using the same procedure as had the written recalls. Thus, both oral and written recalls were divided into pausal units, compared to the reading passages, and then rated as to number of main ideas, details, and irrelevancies. PROCEDURES As this study represents an attempt to analyze the effects of vocabulary and prior knowledge on second- language reading comprehension, it was not only necessary to control for vocabulary difficulty, but also for amount of background information. This was partly accounted for by the use of non-general reading materials, with which the reader likely would be unfamiliar. To further control for prior information, it was determined to supply the experimental group with preteaching, while denying the control group any special warm-up to the topic. This design tests as well the possible effect of the widely practiced technique,in either a first or a second language, of introducing a topic prior to reading . Accordingly, a script intended to introduce the reading topic was written for each passage. Scripts contained fifty to sixty words relevant to the main ideas in each reading. The preteaching information was scripted so that the researcher could replicate the study, and to control the amount and nature of the warm-up. Approximately three minutes were devoted to the reading of the script at the beginning of each day. Brief explanations of the scripted 59 material were provided at the subjects’ request. Schema Activation Upon the completion of this first-phase of the warm-up, subjects were directed to explore the topic by engaging in a writing exercise for the remaining seven minutes of the preteaching period. Without the help of the proctor, subjects wrote anything which seemed pertinent to the topic in these prereading exercises. It was expected that the brief remarks prior to the writing activity would be adequate to activate the subjects’ schema-processors. As the students contemplated the topic, however, it was anticipated that some irrelevant information would appear in the writing samples. Although irrelevant ideas were not introduced in the warm-up script, it was believed that some writers would include distracting information inadvertently in their warm—up essays. These writing samples were collected promptly at the expiration of the ten-minutes allotted, and were analyzed later in the same manner as the written and oral recalls had been. Pausal units were judged as representing main ideas or details predicting what would be in the reading passages, as well as irrelevant to (or not contained in) the reading materials. While the randomly-assigned experimental group of six subjects were participating in this pre-reading warm-up activity, the control group of six subjects were told to meet outside and were permitted to engage in conversation in the hall outside the classroom. The control group was 60 informed that the group in the classroom was receiving information about the reading topic. Despite the obvious curiousity of the control subjects, they were not given any information regarding the nature of the reading passage. The Treatment Conditions Each day, each subject read a difficult version without a warm-up, a difficult version with a warm-up, an easy version without a warm-up, or a difficult version with a warm-up. This schedule is represented in Figure 2 below. FIGURE 2 SCHEDULE OF READING CONDITIONS 0 0 Day One Easy Vocabulary Difficult Vocabulary with Warm-Up without Warm-Up Day Two Difficult Vocabulary Easy Vocabulary without Warm-Up with Warm-Up Day Three Easy Vocabulary Difficult Vocabulary without Warm-Up with Warm-Up Day Four Difficult Vocabulary Easy Vocabulary with Warm-Up without Warm-Up On various days, the populations of the experimental and control groups were altered. As noted in Figure 2, the six subjects in Group One served as the experimental group by receiving pre-instruction on Days One and Four. This same population served as the control group by not receiving the warm-up activity on Days Two and Three. Similarly, Group Two functioned as the control group on Days One and Four, but assumed the role of experimental 61 group on Days Two and Three. To control further for possible differences between the abilities of the groups, the composition of the subjects in the groups was altered after Day Two. That is, half of the subjects in Group One entered Group Two, and three of the subjects of Group Two joined Group One, beginning on the third day of the experiment. OVERVIEW OF THE PROCEDURE Apart from the vocabulary pretesting, the experimental design required that the subjects attend four fifty-minute class periods meeting twice weekly for two weeks. Each day, the following procedure was used. First, the class was divided into a control and an experimental group, which division was varied so that all subjects would have participated under all treatments by the end of the study. For ten minutes, the control group idled in the hallway adjacent to the room while the experimental subjects for that day experienced the warm-up activity. Next, the control group re-entered the classroom. The classroom instructor collected all warm-up papers, and passed out the reading passage for the day. One group of subjects received either the easy- or the difficult-vocabulary version as the other student group received the alternate version of the reading passage. Twenty minutes were allowed for reading. No dictionaries or other reference materials were permitted. Proctors were not permitted to give the subjects any help other than testing directions. 62 The time schedule followed on the four days of the study is outlined below in Figure 3. DAXLAEIIXIIIX 1 Z l A 1. Group One: Warm-Up Easy Reading Cloze Written Group Two: None Diff Reading Cloze Written 2. Group One: None Diff Reading Written Cloze Group Two: Warm-Up Easy Reading Written Cloze 3. Group One: None Easy Reading Written Cloze Group Two: Warm-Up Diff Reading Written Cloze 4. Group One: Warm-Up Diff Reading Cloze Written Group Two: None Easy Reading Cloze Written After the passages were collected, the posttests were administered. one and four, but the written recall was administered first The cloze measure was given first on days on days two and three to test whether the order would affect recall. allowed ten minutes each. The cloze and written-recall tests were The oral-recall measure was conducted simultaneously to the written-recall procedure. Subjects were taken one at a time from the classroom to complete the oral-recall procedure. Between three and five minutes were required to perform each oral recall. To compensate for the students’ consequent loss of time to do the oral recall, subjects were permitted five extra minutes after class to complete their written recalls. CHAPTER FOUR RESULTS The data gathered in the study were subjected to three forms of analyses. This initial discussion overviews the more detailed explanation provided hereafter. First, the data from the cloze and written posttests were exposed to a Repeated Measures Analysis of Variance. The ANOVA of both the cloze and written test data reflected that there were indeed significant differences between treatment populations. Further analysis of the data was required, however, in order to ascertain the direction of the treatment variances. Therefore, a second measurement of the cloze and written test data was employed. The Newman-Keuls Multiple Comparisons Test was used as it enables the identification of several possible differences between combinations of mean scores. This measure yielded the existence of separate treatment populations in both the cloze and written test data. Specifically, the Newman-Keuls test supported that the "easy" vocabulary variable, whether with or without schema-activation, resulted in higher mean scores for reading comprehension on both the cloze and the written posttests. The data of the pretest and the oral recall 63 64 were insufficient for either the ANOVA or the Newman-Keuls analyses. RESULTS OF THE ANALYSES OF VARIANCE The mean scores and standard deviations of the pretest, cloze, written-recall, and oral-recall measures are presented below in Table 1. The apparent trend for subjects to score higher in the easy-vocabulary treatment, regardless of the presence or absence of schema-activation, will be discussed in connection with the various following analyses. TABLE 1 Means (and Standard Deviations) for PreTest, Cloze, Written, and Oral Recalls Code: Easy=Easy Vocabulary: Difficult=Difficult Vocabulary: Warm=With Schema Activated: Cold=No Schema-activation Meeeere Mein Idea Degeil Irrelevant PreTest 1.08 (1.00) 2.75 (1.06) 2.67(l.92) Cloze Test Easy/Warm 15.42 (2.71) (no categories available) Difficult/Cold 12.33 (2.06) Easy/Cold 16.50 (2.39) Difficult/Warm 12.50 (2.24) Written Recall Easy/Warm 1.50 (1.17) 3.08 (1.68) 0.25(0.62) Difficult/Cold 0.67 (0.65) 3.58 (1.73) 0.08(0.29) Easy/Cold 3.50 (1.78) 2.92 (1.68) 0.25(0.62) Difficult/Warm 2.08 (1.62) 2.58 (1.73) 0.58(1.00) Oral Recall Easy/Warm 2.00 (1.00) 3.00 (0.00) 0.33(0.66) Difficult/Cold 2.67 (0.58) 2.33 (1.58) 0.67(0.33) Easy/Cold 1.67 (1.53) 2.67 (1.53) 2.67(1.53) Difficult/Warm 1.67 (0.58) 1.33(0.58) 1.00(0.00) 65 Analysis of the Cloze Data In order to test the hypotheses of the study, the data in Table 1 was subjected to several analyses. First, a Repeated Measures Analysis of Variance was applied to the cloze data. This form of analysis was selected as most appropriate for an experiment in which all of the subjects are administered all of the experimental treatments (Willemsen, 1974). The data from this measure are reported in Table 2 below. If the analysis performed had been a» common analysis of variance, the relatively high sum of squares score within groups and the lower sum of squares score for between groups would be interpreted as supporting the null hypothesis. However, in a repeated measures experiment, such individual differences have little meaning (Willemsen, 1974). The critical scores in this case are the very high F score, which is well above the .05 significance level of 2.53, and the very robust probability indication of 0.00. B Repeated Measures of Analysis of Variance of Cloze Data Source Sum of Squares D.F. M.S. F p Between Groups 5 48.56 11 Within Groups 354.75 36 Repeated Factor 157.73 3 52.58 8.81 0.00 Error 197.02 33 5.97 This analysis of the cloze test data strongly rejects the null hypothesis. All the same, it does not of itself 66 reject the various alternate hypotheses in that it merely establishes that statistically significant differences occur between the treatment populations. The Repeated Measures of Analysis of Variance alone does not distinguish the particular direction of the treatments, or which variables are more important. Thus, it was necessary to follow up on the initial analysis by performing a test of multiple comparisons. The Newman-Keuls Multiple Comparisons Test of Cloze Data The Newman-Keuls Multiple Comparisons measure was selected as a conservative procedure. Newman-Keuls is recognized as facilitating the identification of several possible differences between combinations of mean scores (Willemsen, 1974). The procedure is particularly useful in accounting for the rank order of means. The Newman—Keuls test was used in order to identify the separate treatment populations in the study. In fact, the measure reflected that there were identifiable differences between the treatment populations. The comparisons of the cloze means of the four treatments is summarized below in Table 3. TABLE 3 A Newman-Keuls Multiple Comparisons of Cloze data Mean p Q q value (.05) Mean (3) - Mean (2) = 4.16 4 5.91 3.83 Mean (3) - Mean (4) = 4.00 3 5.67 3.47 Mean (3) - Mean (1) = 1.08 2 1.54 2.88 Mean (1) - Mean (2) = 3.08 3 4.37 3.47 Mean (1) - Mean (4) = 2.92 2 4.14 2.88 Mean (4) - Mean (2) = 0.17 2 0.24 2.88 67 The result of the Newman—Keuls procedure is that groups 2 and 4 are found to be significantly different (at .05) from groups 1 and 3. The Easy/Warm and Easy/Cold treatment groups ranked as homogenous and separate populations from the Difficult/Cold and Difficult/Warm treatment groups. It is noteworthy that regardless of preinstruction, it was the treatment involving the easy-vocabulary version which yielded the highest mean scores on the cloze test. The relative rankings of the mean scores on the cloze test are particularly telling. Figure 4 reflects the strong differences among the mean scores of the treatment groups. The Easy/Cold treatment yielded the highest means. There is little difference between the means for the Easy/Cold and Easy/Warm treatments. Yet it was the group without a warm-up which scored higher. The means for both difficult-vocabulary treatments are nearly equal, again regardless of schema- activation. This result supports that, at least in this FIGURE 4 Cloze Test Mean Scores 68 study, the self-activated schema-activation had no positive effect on reading comprehension as measured by the cloze test. The data are not in full agreement with the hypotheses. Especially, it does not support the prediction that the warm-up activity would facilitate reading comprehension. The differences between variables were significant only for vocabulary. This measure indicates that whereas vocabulary has a significant effect on reading comprehension, schema-activation has none. Analyses were performed on other measures to examine this result. Analysis of the Written-Recall Data The second measure performed on the data was another Repeated Measures Analysis of Variance. In this case, the data analyzed was the combined mean score of the main ideas and details from the written posttest. This combined mean score was realized by adding each main idea and detail score, and then treating the combined score as a score of the number of ”correct" responses made by each subject. Because there was less variance in the means between the treatment groups on the written than on the cloze test, the reliability of the written test is uncertain. This contrast is reflected in Figure 4. The results of this procedure differed from those of the measure of the cloze means. The analysis found the Easy/Cold reading group to belong to a different population from the other three groups. The Easy/Warm treatment again ranked second, but the difference between the means on the two treatments was 69 much greater. The mean scores from the written posttest are presented below in Figure 5. FIGURE 5 Mean Scores on the Written Test Comparison of the Cloze and Written Data The result of the analysis of the mean scores on the written test was significant. The mean scores of the Easy/Cold treatment were varied significantly (at .05) from the mean scores for the other treatments. In both tests, the highest mean scores were obtained by the readers in the Easy/Cold condition. The Easy/Warm scores ranked second once again, although the contrast between this group and the third ranked group was less marked in the written test data than in the cloze data. The mean scores from the Difficult/Warm treatment also were slightly above those from the Difficult/Cold condition on both the cloze and written measures. A similar rank-order of scores by treatment can be established readily for both posttests. 70 This similarity is summarized in the formulas below: Cloze: mean2 > meanl > mean3 > mean4 Written: meanz > meanl > mean3 > mean4 The rank order of the four treatments isidentical on both the cloze and the combined written measures. While this result may be ascribed to happenstance, the probability of this particular result occurring by mere chance is less than one in one hundred (.01). In both tests, the treatment which is optimal for reading appears to be Easy/Cold, followed by Easy/Warm. This result supports hypothesis 2 (that difficult vocabulary impairs reading comprehension). In contrast, it supports neither hypothesis 1 (that schema-activation improves comprehension) nor hypothesis 3 (that schema-activation egg easy-vocabulary text result in better comprehension). Analysis of the Written Test Data in the Main Idea Category As a significant difference was found between the mean combined scores of the treatment groups on the written recall, the data were subjected to an additional analysis. The data from the main ideas and the details were analyzed separately to provide a closer look at the results. A Repeated Measures Analysis of Variance on the mean scores of only the main ideas of the written recall yielded a significant difference between the Easy/Cold and all other groups, as found in the prior combined analysis. A much higher F score (8.16) than that of the combined measure, 71 and one very similar to that of the Cloze was produced. The p<.01 was also identical to that of the Cloze test. The distribution of the mean scores in this analysis are represented graphically below in Figure 6. These results broadly parallel, but do not replicate, those of the previous measures. As before, the mean scores of the Easy/Cold treatment are superior. The treatment groups occupying rankings 2 and 3 have been exchanged. Still, the differences between the Difficult/Warm and Easy/Warm conditions are negligible. No special advantage is shown to have been gained by subjects receiving preinstruction, and the only significant treatment again includes the easy-vocabulary variable. This result as well supports hypothesis 2, but not hypotheses 1 and 3. Although difficult vocabulary may impede reading comprehension, there is little positive effect for schema-activation. FIGURE 6 Mean Scores of the Main Ideas on the Written Recall 4.0 3.5 U 3.0 U 2.5 U 2.0 U U 1.5 U U U 1.0 U U U 0.5 U U U U 72 Analysis of the Written-Recall Data in the Detail Category Having completed the analyis of the main idea category of the written recall, a subsequent analysis of the data gathered in the category of ”details" on the written recall was performed. This analysis, however, produced no significant results as the mean scores for all treatments ranged between 3.08 for the Easy/Warm condition and 2.58 for the Difficult/Warm condition. The F score of 0.67 and the approximate probability of 0.58 indicate that, on this analyis, all four groups can be regarded as parts of a homogenous population. The results of a Repeated Measures of Analysis of Variance of the written irrelevant data also proved insignificant. Very similar mean scores resulted in an F score of just 1.13 and a low probability factor (p=.35) in this case. No analyis was performed on the data from the oral recalls due to an inadequate number of subjects. The lack of control group data for the pretest ruled out the use of a Repeated Measures of Analysis of Variance for this measure as well. Summary of the Analyses of Variance The salient result of the Repeated Measures Analysis of Variance and the Newman-Keuls Multiple Comparisons on the data is that on the cloze and written measures, the Easy vocabulary variable stands out as separate. The data analysis does not align with hypotheses 1 and 3, but supports hypothesis 2. In general, the results show that when even advanced-level ESL readers process text with as 73 few as one low—frequency vocabulary item per six content words, reading comprehension suffers. There is no indication, however, that schema-activation improves reading comprehension scores. Furthermore, vocabulary and schema-activation were not found to interact as variables. That is, neither strengths in vocabulary nor schemata serve to compensate for corresponding weaknesses. These results challenge the predictions made at the outset of the study regarding the effect of schema-activation. Nevertheless, the anticipated effect of difficult vocabulary was supported. The significance and implications of these results will be discussed in the forthcoming chapter. THE RESULTS OF CORRELATIONAL STUDIES In addition to the preceding analyses intended to investigate the effect of the several treatments on second- language reading comprehension, another series of measurements were performed on the data. The analyses performed thus far on the data facilitated an objective assessment of the value of the three primary hypotheses in the study. It also was speculated, however, that certain comparisons between the various types of evaluation instruments might be made possible. It was speculated prior to the study whether the subjects’ reponses would vary not only according to the reading treatment, but also depending on the evaluative measure administered. Several comprehension checks, including pretests, cloze, written, and oral posttests, were employed. Would subjects score 74 similarly on all measures? Or would scores for cloze markedly differ from scores for pretests, for example? Or would oral and written posttest scores be consistent due to their open-ended nature? To obtain at least tentative answers to such questions, a series of correlational studies was performed on the data. Correlational analyses permit a determination of the extent to which scores on one test are associated with another. Such measures point out whether a subject who scores low on one test also scores low on another. The findings from these comparative analyses will be expressed in the commonly-used Pearson’s Correlation Coefficient (R). The magnitude for R ranges from -1 to +1. Negative Correlations Between Cloze and PreTest The overall results from the comparison of the data imply a slightly negative relationship between the pretest and the cloze measure. Subjects who performed better on the warm-up may have become over-confident of their comprhension of the topic. As a result, such readers may have processed the text less carefully than other subjects. The overall negative correlation between cloze and pretest scores is quite low (r= -.11). Nevertheless, the negative correlation between the two measures also is consistent in nine out of the twelve subcategories. In the Difficult/Cold treatment, for example, the correlation of main idea mean scores between the cloze and the pretest is markedly negative (r= -.64). This result is not surprising 75 because subjects in the "cold" condition do not receive schema-activation. Hence, their mean scores would be expected to be lower than those who participated in the PreTest, and who received schema-activation. The negative correlation (r= -.50) between pretest and cloze detail mean scores in the Easy/Warm treatment, however, is not explained easily. In general, schema-activation did not promote higher scores on the cloze. This result is supported not only by the negative correlations between the mean scores of the cloze and preTtst, but also by the Repeated Measures of Analyses of Variance reported above. Correlations Between Cloze and Written Tests Correlational studies of the pretest to written test data also resulted in several negative scores, although not so many as in the pretest to cloze analyses. These results suggest that subjects who scored high on one measure tended not to score high on the other test. As in the pretest to cloze comparison study, all correlations between the pretest and the main ideas section of the written test were negative. As before, the highest single negative correlation (r= -.65) was shared by the written main idea and the pretest main idea response sections in the Difficult/Cold treatment. Positive correlations were once more made in the detail section, but in this case only when the written test treatment was difficult. Irregularities in correlational scores may be due to the small number of subjects in the study. 76 Correlations Between the Pretest and the Oral Recall The pretest data was compared next to the results of the oral recall. The outcome of these correlational studies was mixed. There was a zero correlation (r= 0) between the pretest and the main ideas of the oral test. A comparison of the pretest to the details of the oral test, however, produced a notable positive correlation (r= .59). The positive correlations between the pretest and other tests in the category of details seems to represent a trend. It may be that schema-activation enables the comprehension of details. Conversely, schema-activation does not correlate to improved comprehension of main ideas in this study. In the final comparison in this category, there was a negative score (r= .28) between the pretest and the oral test irrelevancies as well. Summary of Correlations Between Pretest and Other Measures This series of correlations concluded the comparisons between the pretest and the other three tests. In general, the relationship between these measures appears to be weakly negative. Certain of the r scores were relatively high. There was a tendency for the stronger negative scores to occur in connection with the main ideas category. Most positive scores were found in the details and irrelevant idea sections. Because correlational studies merely indicate relationships, no cause-and-effect condition can be claimed. When combined with the results from several analyses, however, the trends in the data 77 appear more weighty. The results of the analyses of variance as well as of the correlational studies between the pretest and other measures support that schema- activation was ineffective in improving reading comprehension scores. Correlations Between Cloze and Written Tests Further correlational analyses were performed to determine whether relationships exist between the subjects’ scores on other measures. A series of correlational studies was conducted to compare cloze to written test scores. The scores were found to be weakly positive overall. Five r scores were positive, and three were negative. As in the previous results, the Difficult/Cold correlation proved notably positive (r= .57). The trend for negative scores in the detail category was also followed. Scores in this classification were all positive, except for a very weak (r= -.15) score in the Difficult/Cold treatment. In fact, the single highest correlation (r= .81) in all the data occurred in this section under the Easy/Warm treatment. These results support that the subjects with high mean scores on the cloze tended to have high mean scores on the written recall as well. Correlations Between Cloze and Oral Tests The data from the cloze and oral tests were compared next. These findings are generally inconsistent with those 78 previously examined. A positive relationship was found to exist overall between these two measures. In this case, four robustly positive scores (of at least .50), and just two negative scores were discovered. The minimally positive section of this correlation study was the main ideas, which had been found to be negative in the main in other categories. Certain consistencies did, however, emerge in this study. Most notably, the Difficult/Cold treatment again produced the highest negative score (r= -.50) in a comparison of main idea scores. The detail category yielded three positive correlations, too. Overall, it seems that the subjects who scored well on the cloze also scored well on the oral recall. This result is somewhat surprising. It had been supposed that subjects who were stronger in oral than in written skills would weaken correlations between these two measures. Correlations Between Written and Oral Recalls The remaining correlational study was performed to compare the written to the oral recalls. The results of these analyses were varied. Both the detail-to-detail and the irrelevant-to-irrelevant categories yielded zero (r= O) correlation scores. A positive correlation (r= .42) was found for the main ideas classification. Thus, there was a tendency for subjects to score either high or low on both measures in supplying main ideas. No such pattern, however, was evident among detail and irrelevant responses. The mixed results in this correlation make 79 generalizing difficult. It is unclear whether subjects with strong oral skills scored higher on the oral recall than on the written recall. Similarly, it is not certain whether subjects who possess better writing abilities scored higher on the written recall than on the oral recall. CHAPTER SUMMARY The results of the analyses utilized in this study support hypothesis 2. It seems that difficult vocabulary, even when introduced at a rate of just one-item-per-six- content-words, negatively affects second-language reading comprehension. Hypotheses 1 and 3, however, were not supported in this study. Schema-activation was not found to improve reading comprehension scores. Neither did a combination of schema-activation and "easy"-vocabulary text produce better reading comprehension. The results of the analyses of variance strongly supported hypothesis 2, and served to reject hypotheses 1 and 3. The results of the correlational studies, however, were less clear. In general, correlations were weak, or non-existent. A tenuously negative relationship between the pretest and the other tests may reflect also that schema-activation does not facilitate reading comprehension. Nor did the results support that students score better on measures which utilize their better writing or oral skills. CHAPTER FIVE DISCUSSION There is a general assumption in second-language studies that readers may use background knowledge to overcome vocabulary problems. To be sure, ESL teachers have long recognized that limited language proficiency and vocabulary deficiencies can "short-circuit” the reading process (Clarke, 1979: Eskey, 1987). It has, however, become unpopular in L2 research to attend to bottom-up processing in general, and to vocabulary in particular (Carter, 1987). As bottom-up processing and vocabulary have been devalued, top-down processing has become emphasized increasingly in models of reading in a second language. The recent introduction of the term "interactive reading" may be seen as bearing promise to restore a balance between top-down and bottom-up processing in reading theory. Unfortunately, although the expression itself may be accurate, it is not yet being rigorously applied. In practice, however, the interactive model seems to favor heavily top-down schematizing and to belittle bottom-up processing (Eskey, 1988). Schema theorists have improved our knowledge of reading by showing that reading comprehension depends on knowledge of rhetorical patterns, the topic, and the target language. 80 81 The benefits of top-down processing are so impressive that it has been supposed that L2 readers who possess superior schematic knowledge may compensate for linguistic deficiencies (Barnitz, 1986: Swaffar, 1988). There is no research, however, which bears out this compensative theory. Despite repeated attempts, first- language researchers have been unable to find interaction between schemata and vocabulary (Freebody and Anderson, 1983b: Stahl and Jacobson, 1986: Stahl et al, 1989). Neither has schemata been found to compensate for vocabulary deficiencies in L2 reading, although such an effect has been supposed (Hudson, 1982). The obvious question to be answered by L2 reading researchers is: ”Do L2 readers in fact use schemata to overcome problems in processing text with difficult vocabulary?” THE PURPOSE OF THE STUDY It is the purpose of the present study to fill this gap in second-language reading knowledge by investigating the effect of activating schemata and of introducing difficult vocabulary in L2 reading. In this study, it is assumed that whereas schema-activation facilitates L2 reading, the introduction of difficult (unknown) vocabulary to passages negatively affects reading comprehension. It is supposed also that the "optimal" L2 reading condition occurs when the reader’s schemata has been activated and when the text has little or no difficult vocabulary. The following DYDQLDESQE regarding the relationship between 82 schema-activation and difficult vocabulary in second- language reading were formulated: Hypotheses Regarding Schema-activation and Vocabulary 1. L2 readers who receive schema-activation will achieve significantly higher reading comprehension scores than readers who do not receive schema- activation. 2. L2 readers who process passages with difficult vocabulary will attain significantly lower reading comprehension scores than readers who process passages without difficult vocabulary. 3. L2 readers who receive schema-activation and read passages without difficult vocabulary will achieve significantly higher reading comprehension scores than readers who process texts under all other conditions (with schema-activation and with difficult vocabulary: without schema-activation and without difficult vocabulary: without schema- activation and with difficult vocabulary). THE STUDY In order to test these hypotheses, four reading treatments were designed for a two-way, Repeated Measures study. The reading treatments were with or without schema- activation, and with or without difficult vocabulary. Twelve advanced—level ESL students read and were tested for their comprehension of four reading passages on four separate days in January, 1989. The four "difficult"- vocabulary versions were adapted by substituting unknown vocabulary items for every sixth content word. Parallel "easy"-vocabulary versions utilized known words at intervals of every 15th content word. Half of the subjects received daily schema-activation and read either difficult or easy vocabulary—passages, whereas the alternate group 83 did not receive schema-activation and read text in the opposite (whether difficult or easy) treatment. Thus, all subjects had performed in all treatments by the end of the four days. In order to measure the subjects’ performance, four measures were taken. After the experimental group for the day had listened to a brief introduction to the reading, they were asked to engage in prewriting on the topic. When the subjects had finished reading the passage for the day, they were administered a cloze test, a written recall, and then an oral recall. Multiple measurements of the subjects’ anticipation and recall of the passages were utilized in the analysis. The use of multiple measures (pretest, cloze, written recall, and oral recall) also made possible a comparison of the different measures themselves. Using a procedure suggested by Omanson (1982), the reading passages and all measures (pretest, cloze, written and oral recall) were analyzed to identify main ideas, supporting details, and distracting or irrelevant information. This analysis was used to make a quantitative measurement of the subjects’ responses possible. RESULTS The data in the study were subjected to three forms of analyses. First, the data from the cloze and written posttests were exposed to Repeated Measures Analysis of Variance. The MANOVA of both the cloze and written test 84 data reflected that there were indeed significant differences between treatment populations. Further analysis of the data was required, however, in order to ascertain the direction of the treatment variances. The Newman-Keuls Multiple Comparisons Test yielded the existence of separate treatment populations in both the cloze and written test data. Specifically, the Newman-Keuls test supported that the "easy"-vocabulary variable, whether with or without schema-activation, resulted in higher mean scores for reading comprehension on both the cloze and the written posttests. The data of the pretest and the oral recall were insufficient for either the ANOVA or the Newman-Keuls analyses. Means (and Standard Deviations) for PreTest, Cloze, Written, and Oral Recalls Code: Easy=Easy Vocabulary: Difficult=Difficult Vocabulary: Warm=With Schema Activated: Cold=No Schema-activation Eeasurs Main Idea Detail I evant PreTest 1.08 (1.00) 2.75 (1.06) 2.67(1.92) Cloze Test Easy/Warm 15.42 (2.71) (no categories available) Difficult/Cold 12.33 (2.06) Easy/Cold 16.50 (2.39) Difficult/Warm 12.50 (2.24) Written Recall Easy/Warm 1.50 (1.17) 3.08 (1.68) 0.25(0.62) Difficult/Cold 0.67 (0.65) 3.58 (1.73) 0.08(0.29) Easy/Cold 3.50 (1.78) 2.92 (1.68) 0.25(0.62) Difficult/Warm 2.08 (1.62) 2.58 (1.73) 0.58(1.00) Oral Recall Easy/Warm 2.00 (1.00) 3.00 (0.00) 0.33(0.66) Difficult/Cold 2.67 (0.58) 2.33 (1.58) 0.67(0.33) Easy/Cold 1.67 (1.53) 2.67 (1.53) 2.67(1.53) Difficult/Warm 1.67 (0.58) 1.33(0.58) 1.00(0.00) 85 Analysis of the Cloze Data In order to test the hypotheses of the study, the data in Table 1 was subjected to several analyses. First, a Repeated Measures Analysis of Variance was applied to the cloze data. The critical scores in this case are the very high F score, which is well above the .05 significance level of 2.53, and the very robust probability indication of 0.00. This analysis of the cloze test data strongly rejects the null hypothesis, but does not distinguish the particular direction of the treatments, or which variables are more important. Source Sum of Squares D.F. M.S. F p Between Groups 48.56 11 Within Groups 354.75 36 Repeated Factor 157.73 3 52.58 8.81 0.00 Error 197.02 33 5.97 The Newman-Keuls measure was applied to the results of the MANOVA because it is recognized as facilitating the identification of several possible differences between combinations of mean scores (Willemsen, 1974). The result of the Newman-Keuls procedure is that groups 2 and 4 are found to be significantly different (at .05) from groups 1 and 3. The Easy/Warm and Easy/Cold treatment groups ranked as homogenous populations, and the Difficult/Cold and Difficult/Warm treatment groups ranked as a separate 86 population. Regardless of preinstruction, it was the treatment involving the less difficult vocabulary version which yielded the highest mean scores on the cloze test. The relative rankings of the mean scores on the cloze test are particularly telling. An analysis of the cloze means is summarized below in Table 6. Mean p Q q value (.05) Mean (3) - Mean (2) = 4.16 4 5.91 3.83 Mean (3) - Mean (4) = 4.00 3 5.67 3.47 Mean (3) - Mean (1) = 1.08 2 1.54 2.88 Mean (1) - Mean (2) = 3.08 3 4.37 3.47 Mean (1) - Mean (4) = 2.92 2 4.14 2.88 Mean (4) - Mean (2) = 0.17 2 0.24 2.88 Analysis of the Written Recall Data The result of this analysis was significant. The F score in this case was significant at 3.39, and p=.029. This data analysis may be seen as comparative to that of the cloze. In both cases, the highest mean scores were obtained by the readers in the Easy/Cold condition. The Easy/Warm scores ranked second once again, although the scores were not equidistant from other rankings. The mean scores from the Difficult/Warm treatment also were slightly above those from the Difficult/Cold condition on both the cloze and written measures. Thus, a similar relationship of scores by rank-order for treatment can be established for the cloze and written posttests. 87 The salient result of the Repeated Measures Analysis of Variance and the Newman-Keuls Multiple Comparisons on the data is that on the cloze and written measures, the easy- vocabulary variable stands out as separate. The data analysis does not align with hypotheses 1 and 3, but supports hypothesis 2. In general, the results show that when even advanced-level ESL readers process text with as few as one low-frequency vocabulary item per six content words, reading comprehension suffers. There is no indication, however, that schema-activation improves reading comprehension scores. Furthermore, vocabulary and schema-activation were not found to interact as variables. That is, neither strengths in vocabulary nor in schemata serve to compensate for corresponding weaknesses. These results overthrow the predictions made at the outset of the study regarding the effect of schema-activation. Nevertheless, the anticipated effect of difficult vocabulary was supported. Correlational Studies The data were subjected to a series of correlational studies to determine what relationships exist between the various measures utilized in the study. The overall results of these studies were unclear and tenuous. As might be expected, the correlation between the cloze and written posttests yielded a weakly positive score (r= .32), indicating that students who scored well on one measure tended to score well on the other also. 88 The result of correlational study between the written and oral posttests seems to be rather unusual. Students who excell at writing, for example, but who possess weaker oral skills might be supposed to score differently on these two measures. The results, however, indicate rather low correlations between these two measures. In fact, there is a positive correlation (r= .42) between the mean scores of the main idea category of the two tests. Students apparently are able to communicate at least their main ideas, whether orally or in print, regardless of their particular strengths and weaknesses. The results of the several correlational studies between the mean scores of the pretest and the various posttests are noteworthy as well. It had been predicted that students who received schema-activation would produce better reading comprehension scores. No such result, however, is indicated in the data. The correlations between the pretest and the posttests are insignificant, and negative generally. The most noteworthy correlations (r= -.31 and -.33) were found to exist between the pretest and main idea category of both the cloze and written posttests. CONCLUSIONS The findings of this study do not confirm that schema activation is most important to L2 reading comprehension. Neither does schema-activation appear to overwhelm problems with comprehending passages which have difficult, 89 unknown vocabulary items. In fact, difficult vocabulary appears to trouble second-language readers much more than it has been supposed. At least in this study, readers comprehended easy-vocabulary passages significantly better than difficult-vocabulary versions, whether or not schemata was activated. The three hypotheses of the study will be discussed in light of the results. Why Schema Activation May Prove Ineffective Hypothesis 1, predicting that schema-activation would facilitate better reading comprehension, is not supported by the data. Indeed, none of the analyses performed on the data uphold this hypothesis. Truly, this is an astounding, yet clearly revealed conclusion in this study. Although common sense and prior research (Taglieber et al, 1988) suggest that schema-activation facilitates improved reading comprehension, in this research there is no such evidence. Schema-activation may be ineffective for several reasons. First, it is always possible for a warm-up activity to be flawed, and for schemata not to be activated. Second, inappropriate information may be triggered inadvertently during schema-activation. Such irrelevant information may serve to ”derail” comprehension. Finally, even successful schema-activation may not be sufficient to enable the L2 reader to process text laden with difficult vocabulary. These potential stumbling blocks to schema-activation will be examined in view of the present study. 90 Schemata may be more difficult to activate properly than it appears. The students’ background information must be estimated. If students lack adequate topical knowledge, it must be learned. (It was the purpose of this study to test only the effectiveness of schema-activation, and not the results of teaching schemata.) Schema-activation might have been blocked in this study by the particular nature of the passages themselves. At least in some cases, the students knew little or nothing about the reading topic. Reading 1, ”A Theory of Earth’s Structure," may serve as an example. When required to prewrite about the topic, the six students in the experimental group produced more irrelevant than pertinent information. The students wrote twenty-four pausal units in all. Fourteen of these units were classified as ”irrelevant" to the topic, and just ten units were deemed to be on the topic. One of the student’s prewriting samples exemplifies the groups’ apparent lack of knowledge regarding plate techtonics: ”The earth is not round completely. It has artic and antarctic.” Another ESL student produced five pausal units, four of which were inapplicable to this topic. In addition to the earth’s being round, this student offered that it "goes around the sun.” Obviously, the schema-activation did not work for this subject. Nor were these responses exceptional: only four of the main ideas in the passage were anticipated in this set of prewriting samples. The prewriting results from the second day were not 91 much better. Although the topic was less specialized, nearly half of the responses were not relevant to the reading passage. One of the students offered that earthquakes come due to a "lack of rain.“ Another (non-Japanese) wrote, ”I don’t know nothing about the phenomenon [because] it’s not in the news.” Merely three main ideas related to the ”earthquake” passage were found in the samples of these six subjects. The students found it difficult to provide much appropriate information regarding the two remaining reading passages as well. A Japanese subject observed that deforestation is a problem because "human life concerns animal life.” This information was judged irrelevant because the passage does not discuss animal habitats. Thus, the student was contemplating an issue which was not contained in the reading passage itself, and thus was judged irrelevant. The prewriting samples for the fourth topic, overpopulation, contain many details, but more irrelevant pausal units than main ideas. Why Schema-activation May "Short-circuit” Reading Throughout the prewritings, it is evident that the students sometimes were able to anticipate certain of the forthcoming details in the passages. Very often, however, the schema-activation exercise triggered irrelevant information as well. Whether this irrelevant information "short-circuited" subsequent reading comprehension is unknown, but this would account for the lower mean scores 92 realized by subjects who received schema-activation and yet performed worse than those without a warm-up. It is difficult to determine whether the students supplied irrelevant information in the prewriting samples because of a lack of appropriate information or because of their erroneous beliefs. Every writer is aware that tangentially-related ideas slip altogether too easily into one’s paragraphs. The student who offered that the earth is round may have only partly grasped the significance of the topic concerning plate techtonics. Additionally, the subject may have supplied the information due to an oversight. Regardless of how irrelevant information entered the prewriting samples, such information may have had a detrimental effect on the students’ comprehension of the reading passages. The reader does not process cues in L2 text as well as in the first language. When the reader who has activated irrelevant information approaches a foreign-language passage, certain distortions may creep into the reader’s re-creation of the text as reading comprehension is impeded. The reader who offered that the earth goes around the sun, for example, produced just two pausal units in the post-writing sample. Interestingly, both clauses concerned the EQYQEEDL of the earth as well: ”The earth is alive and moving [and] the ground is moving." This student’s cloze results were among the lowest also. This subject wrote several (four) irrelevant units on the 93 pretest, and just one correct detail. The weakly negative correlations (r= -.32 and -.22) between the mean scores of the pretest irrelevant category and both the cloze and the written posttest mean scores suggest that irrelevant information in the prewriting samples may have negatively affected reading comprehension. It probably operated as merely one among many factors in the reading process, however. Can Schema-activation Overcome L2 Language Difficulties? According to the results of this study, schema- activation does not overwhelm, and may be of little help in solving the problems of processing difficult-vocabulary texts. It is especially remarakable that students who participated in schema-activation did not achieve higher mean scores for reading comprehension on any of the posttests. Especially on the cloze test, students in the treatment condition scored notably lower than those who did not participate in the activation of schemata. This result is frankly puzzling. Should not readers perform better when they have been cued in to the topic? Why would not L2 readers benefit from schema-activation, and be able to "read around" difficult vocabulary? Perhaps the problem can be explained by noting that information is organized differently by various cultures. The schemata which actually is activated in a warm-up activity vary from the intended schemata. In addition, it may be difficult for L2 readers to apply schemata activated in an L2 94 context. Regardless of the cause, however, schema- activation has not been shown to overcome problems in comprehending L2 text with difficult vocabulary. Difficult Vocabulary Strongly Affects Reading Comprehension Anyone who has attempted to read in a foreign language appreciates the impasse created by difficult vocabulary items in a second-language passage. Despite one’s best intentions to "read around” these obstacles, such "road blocks" are difficult to ignore. It is as though one were wearing dark sunglasses while driving through a forest on a moonless night. The cues are there, but they are not at all easy to pick up. All systems operate at reduced-power, depending on the L2 proficiency of the reader. The units of the language do not ”flow”. Often the reader’s attention is caught up in the word-by-word chase after meaning. It is troublesome to consider the global implications of the passage when the words themselves are intractable. Whereas the reader of native-language material usually can predict the next word, phrase, or even development in the plot, such anticipation is treacherous in foreign-language passages. Under such conditions, one word appears to count as much as the next, and word-skipping may lead the reader into one of countless dead-ends. The reader simply does not know what requires the most attention. Hence, the L2 reader’s preoccupation with difficult vocabulary is very much a practical consideration. 95 Apart from the reader’s interest in difficult vocabulary, there is the inherent importance of the words themselves to consider. Low-frequency vocabulary are not retrieved and used unless they are required. These words tend to bear considerable significance, often carrying the meaning of whole phrases in one term. Of course, this is the great tease of vocabulary. The words which are most common are quite devoid of content: they are function words such as articles and prepositions. Thus, the very words which the reader most needs to know are the ones which are least likely to be known. And the vocabulary which are known may not carry the critical meanings which the irksome ones do. The result from the L2 reader’s perspective can be a very grammatical text of uncertain meaning rather akin to the following nonsense: ”The very large fremkin slowly eeggeg a pair of LLQEQQee by means of his handy zeeheg." The pieces are there, but even the best guesses do not add up to anything substantial. Difficult vocabulary poses a thorny problem to reading comprehension, especially in a second language. Not only do difficult vocabulary items distract the reader, they also leave the reader at a definite loss for important information. Surely, at times the reader is able to discover the meaning of the word without intervention. But difficult words are not frequently encountered, so it takes time to learn them. Also, when the reader does manage to guess the intended 96 meaning of a passage despite difficult vocabulary, the reader may well realize that such gained information is merely tentative as the language itself was not processed thoroughly. Difficult vocabulary, then, presents to the L2 reader a formidable task. Do Schema-activation and Vocabulary Interact? The results of this (or any other) study do not support that schema-activation and vocabulary are interactive variables. The highest mean scores were attained by readers who read easy-vocabulary versions EIEDQQL schema-activation. Comprehension was better when the reading passages had less difficult vocabulary: processing was not improved by schema-activation. No interaction between vocabulary and schema-activation was found. The current view of reading as an interactive process implies the interaction of more than one variable. A variety of cues are accessed by the reader in making meaning from text. These cues are morphophonemic, lexical, syntactic, and rhetorical. Because there is a surfeit of information, the experienced reader samples only those cues necessary to confirm or modify the reader’s expectations. The reader may be reassured by fitting information--or shocked by detours, gaps, or apparent incoherencies. How the readers in this study utilized the information made available by schema-activation is unknown. The students may have responded to the warm-up activity in any of several ways. It may be that the readers did not 97 bring the information accessed in the schema-activation to bear on the reading process. Although the prewriting samples contain considerable information judged to be irrelevant to the reading topic, much pertinent information is evident in the prewriting as well. Perhaps the readers later recognized that some of the information accessed in the warm-up activity was irrelevant, and regarded their thoughts prior to reading the passages as useless or even suspect. Or the reading task itself may have proven so laborious as to preclude the synthesis of prior knowledge and textual features. The reader’s attention may have been so focused on bottom-up processing that little attention could be devoted to ”behind-the-eyeball” operations. Thus, even "advanced” L2 readers may find it impossible to apply top-down processing techniques to difficult L2 reading passages. Why Schema-activation May Be Over-valued Studies (Taglieber et al, 1988: Johnson, 1982: Barnitz, 1986) have been performed to show that background information and schema-activation affect L2 reading comprehension. The importance of background information may have been over-emphasized, however, in some cases. The first problem is that the reading materials utilized in these studies were culturally-loaded. Fairy tales, legends, fables, and accounts of holidays and other customs were selected to favor the significance of prior knowledge. Obviously, background knowledge is vital to 98 understanding such passages. Nevertheless, L2 readers must process many other less culturally-loaded texts as well. To imply that cultural insight or prior knowledge is a considerable part of the L2 reading problem is to beg the question. How is it that information is to be gained? Do not many foreign-language learners acquire cultural and other knowledge by reading? If in fact reading involves learning, then the meaning that the reader extracts from a text primarilly depends on the adaptations which the text itself imposes on the reader’s expectations (Britton, 1970). It also may be questioned whether prereading activities that preteach content actually improve reading comprehension or merely improve comprehension itself because the student knows the content even before reading. Undeniably, prior knowledge facilitates L2 reading. A reader who knows the gospel by heart may read it in a second language with ease. But how much of this process is really reading and how much is simply knowing? To be sure, the reader must have some topical knowledge in order to have suitable expectations. Nevertheless, an over-emphasis upon pre-teaching text tends to devalue the reader’s inherent ability to make meaning by means of direct interaction with text. If reading is indeed "guided meaning-making” (Scholes, 1985), then the classroom instructor must avoid substituting oral instruction for private meaning-making. 99 APPLICATIONS FOR TEACHING AND IMPLICATIONS FOR RESEARCH Three research hypotheses regarding the nature of reading in a second language were tested in this study. Although it was predicted that schema-activation would improve L2 reading comprehension, the results did not support this prediction. On the contrary, readers who participated in schema-activation received lower mean scores on the posttests than did readers without schema-activation. This result occurred in connection with all four reading passages, and whether the readers processed difficult- or easy-vocabulary versions. It was predicted as well that readers who participated in schema-activation and read the easy-vocabulary text would achieve comprehension scores superior to readers in all other treatment conditions. This prediction was not upheld by the results. Readers who read easy-vocabulary versions and who did gee undergo schema—activation produced the highest mean scores for reading comprehension. The easy-vocabulary and schema-activated treatment group ranked second to the previous treatment on the post-tests. Notwithstanding the rejection of the two hypotheses regarding schema-activation, one prediction did prove reliable. It was hypothesized that readers who processed passages with difficult vocabulary would attain lower reading comprehension scores than readers who read easy-vocabulary versions. This prediction was supported by the result as the mean scores for readers who processed 100 easy-vocabulary versions were markedly superior to the scores for those students who read passages with difficult (unknown) vocabulary. This result occurred with or without schema-activation. The salient result of this study is, then, that vocabulary difficulties may be far more troublesome to the second-language reader than has been thought. Even when the density of low-frequency vocabulary items is no more than one content word in six, L2 reading comprehension may be impeded seriously. Further studies regarding the effect of difficult vocabulary on second-language reading comprehension are warranted. Relatively little is known concerning the potential interaction between difficult vocabulary and the several other cues available to the L2 reader. The precise nature of the reader’s trouble in dealing with difficult vocabulary is little understood as well. Besides the implications of difficult vocabulary, this study also brings into question the usefulness of schema- activation in second-language reading. Although L2 reading instructors have assumed that strengths in top-down processing may compensate for weaknesses in bottom-up processing, no research supports that background knowledge and vocabulary are interactive variables. Unquestionably, schemata play an important role in the reading process. First- and second-language readers are able to read because they have both pragmatic and linguistic knowledge. This 101 does not indicate, however, that readers may compensate for linguistic deficiencies by depending on schematic knowledge. It may be safe to allow that second-language reading is interactive. But it may be imprudent to believe that top-down processing enables readers to process linguistically difficult text. Language teachers naturally wish to hasten their students’ learning process. Learning to read in a second language is a laborious task. Even for ”advanced" ESL readers, most unadapted reading materials fairly bristle with difficult vocabulary. The syntax, cohesive devices, rhetorical pattern, and other potential cues as well may appear formidable. Despite the instructor’s most earnest desire to expedite learning, students must receive guidance in both bottom-up as well as top-down processing. It is a disservice to second-language students to assume that they will pick up basic linguistic knowledge (such as vocabulary and syntax) on their own, and to suppose that mere pre-teaching will obviate all reading problems. To be effective, L2 readers need to process both schemata and difficult vocabulary. APPENDICES A Datafmcnthestudy guzgz Quenfi Ennglnt MM EDGE" # MDI 110 220 211 00 1 3‘ 1 34 2 101 1‘2 323 0 01 3 33 1 23 010000010201 034425130443 50221‘01231‘ 120000000000 ‘2‘443110‘33 1315426523“ 000000000001 171524543443 100001021111 201000000000 615452243212 230102211011 mumnnumnnuun nnmuuuuumnmm nmumnuuuuuuu mnnnmuunmnuu ‘21117‘411‘2 1‘2‘2‘322‘32 03212010211 Lzmhmmmmmmmm No Ian I»: mifficult Vocamlary i miffiwlt Vocabulary a No I—Irrelevant Info 102 APPENDIX B THE READING PASSAGES READING PASSAGE 1: A THEORY OF EARTH’S STRUCTURE Easy-Vocabulary Version Few people who have experienced an earthquake or witnessed a volcanic eruption doubt that our planet is active, alive. When the earth a house stands on begins to move, when a mountain suddenly swells, then there is proof that our Earth is a changing place. Geologists, the scientists who investigate the earth, have put together a possible explanation of the earth’s structure. Their theory, called plate tectonics, is that the surface of the earth is hard rock like a crust. This cool superficies floats on magma that is below the surface. The surface of the earth moves like ice on water. Weak places become cracks in the ice: similarly, there are faults in the Earth’s rock. When activity inside the planet causes mounting pressure, plates of rock move to cause an earthquake, or the tension is released by a volcano. There is bountiful evidence to support the theory of 103 104 plate tectonics. Aerial photographs and satellite pictures show profound cuts in the Earth. These trenches are clues for the arrangement of the surface plates. Experts have found places where ground motion is likely to transpire. Along these cracks, they measure the shock waves that are caused by the small tremors from deep underground. The measurements buttress plate tectonics today. One implication of plate tectonics is that all of the earth was once a single mass. By plate realignment, this mass separated into plates moving in different directions. Before the most recent movement, South America touched Africa just under its western bulge, and North America wrapped around North Africa. Europe and most of Asia were far to the north. Africa’s movement into Europe forged the Mediterranean Sea and some mountains: the Alps and the Atlas range. The Himalaya range came from the movement of the subcontinent of India away from southeast Africa and into Asia. The westward drift of America accounts for the mountain regions of the Sierras and Andes. This theory also foreshadows a changing arrangement of continents. California will separate from the mainland. North and South America will at some point disengage. Africa, Australia, Japan, and India will continue to drift toward varying points. While the Atlantic Ocean widens, the Pacific will narrow. 105 Difficult-Vocabulary Version Few people who have experienced an earthquake or witnessed a volcanic eruption doubt that our planet is active, alive. When the terra firma a house stands on begins to move, when a mountain suddenly swells, then there is proof that our Earth is a changing place. Geologists, the scientists who investigate the earth, have put together a possible explanation of the earth’s framework. Their theory, called plate techtonics, is that the surface of the earth is hard rock like a crust. This cool superficies floats on magma that is below the surface. The surface of the earth glides like ice on water. Weak places become cracks in the ice: similarly, there are faults in the Earth’s rock. When activity inside the planet causes mounting pressure, plates of rock move to cause an earthquake, or the tension is unleashed by a volcano. There is bountiful evidence to support the theory of plate tectonics. Aerial photographs and satellite pictures show profound trenches in the Earth. These trenches are clues for the arrangement of the surface plates. Experts have found places where ground motion is likely to transpire. Along these cracks, they measure the shock pulses that are caused by the small tremors from deep underground. The measurements buttress plate tectonics theory. One implication of plate tectonics is that all of the 106 earth was once a single mass. By plate realignment, this mass separated into plates moving in different directions. Before the most recent movement, South America nudged Africa just under its western bulge, and North America wrapped around North Africa. Europe and most of Asia were far to the north. Africa’s movement into Europe forged the Mediterranean Sea and some mountains: the Alps and the Atlas range. The Himalaya range was the issue of the movement of the subcontinent of India away from southeast Africa and into Asia. The westward drift of America accounts for the craggy regions of the Sierras and Andes. This theory also foreshadows a changing arrangement of continents. California will dislodge from the mainland. North and South America will at some point disengage. Africa, Australia, Japan, and India will continue to drift toward varying points. While the Atlantic Ocean widens, the Pacific will shrink. READING PASSAGE 2: LEARNING ABOUT EARTHQUAKES Easy-Vocabulary Version Earthquakes occur along the lines of the flaws in the Earth’s crust. Scientists have known this for a long time. Of late, however, the study of earthquakes-- especially predicting them--has changed. The Chinese study earthquakes because they have suffered some of the world’s most horrific earthquakes. 107 Two times in Chinese history, in 1556 and 1976, more than half a million people died in quakes. Because of their success in forecasting quakes, many lives have also been saved. In the terrible winter of 1974 and 1975, there were signs that could be read on two separate occaisions. In December, 1974, the experts posted a warning in the Liaoning Pennisula. All the people left the cities, but the earthquake did not happen. Two months later, ominous events were predicted again. Some were unwilling to leave their homes. They thought it was another blunder. However, the police cleared the dangerous areas. This time a severe earthquake did occur. Because the cities had been emptied, thousands of lives were saved. Before the catastrophic 1976 earthquake, there were readable, but unclear signs, so a warning was not given. The way that the Chinese learned about the warning signs was engaging. They tried to record everything. They noted as much information as they could. Then, after a quake, the information was studied. They looked for changes that occured just before the quake. They found some novel and unusual facts. The Chinese noted that animals behave strangely before an earthquake. Chickens, which usually roost at night, keep both feet on the ground. Snakes, which usually hid in holes, come up out of the ground. Dogs, even those that are usually quiet, bark incessantly. American scientists are noticing earthquakes signs, 108 too. A Colorado expert found that diminutive earthquakes happened when water was forced into a deep well. Engineers were trying to remove oil by using water. Each time they pushed water into the well a little quake transpired. This action led to experiments with water. By pushing water into cracks, they hope to control the increase in pressure. Weak earthquakes can reduce the stress between the plates that agitate the surface of the earth. Less pressure means fewer quakes. Another theory of earthquakes is intertwined with the moon. Reid A. Bryson of Wisconsin says the Earth’s tides have been stronger recently: that is, the levels of the ocean’s high and low watercycles have been higher and lower. One result of the moon’s stronger tug on the earth could be an increase in volcanic activity. The Smithstonian observatory in Arizona set out that the Earth’s crust reacts to the moon’s pull. The liquid, molten rock may be contracted like the water in the oceans. If the pull becomes imposing, the weakest places in the surface are likely to break. The results would be a surge in volcanos and more earthquakes. Difficult-Vocabulary Version Earthquakes occur along the lines of the flaws in the Earth’s crust. Scientists have known this for a long time. Of late, however, the study of earthquakes-- especially predicting them--has been refined. 109 The Chinese study earthquakes because they have suffered some of the world’s most horrific earthquakes. Two times in Chinese history, in 1556 and 1976, more than half a million people perished in quakes. Because of their success in forecasting quakes, many lives have also been preserved. In December, 1974, the experts posted a warning in the Liaoning Pennisula. All the people forsook the cities, but the earthquake did not happen. Two months later, ominous events were predicted again. Some were unwilling to leave their homes. They thought it was another blunder. However, the police evacuated the dangerous areas. This time a severe earthquake did befall them. Because the cities had been emptied, thousands of lives were saved. Before the catastrophic 1976 earthquake, there were readable, but unclear signs, so a warning was not dispensed. The way that the Chinese learned about the warning signs was engaging. They tried to record everything. They noted as much information as they could. Then, after a quake, the erudition was studied. They looked for changes that occured just before the quake. They found some novel and unusual facts. The Chinese noted that animals behave strangely before an earthquake. Poultry, which usually roost at night, keep both feet on the ground. Snakes, which usually lurk in holes, come up out of the ground. Dogs, even those that are usually quiet, bark incessantly. American scientists are noticing earthquakes signs 110 too. A Colorado expert found that diminutive earthquakes happened when water was forced into a deep well. Engineers were trying to remove oil by using water. Each time they compelled water into the well a little quake transpired. This action led to experiments with water. By pushing water into crevices, they hope to control the increase in pressure. Weak earthquakes can diminish the stress between the plates that agitate the surface of the earth. Less pressure means fewer quakes. Another theory of earthquakes is intertwined with the moon. Reid A. Bryson of Wisconsin says the Earth’s tides have been stronger recently: that is, the levels of the ocean’s high and low watercycles have been higher and lower. One result of the moon’s stronger tug on the earth could be an increase in volcanic activity. The Smithstonian observatory in Arizona set out that the Earth’s crust reacts to the moon’s pull. The liquid, molten rock may be contracted like the water in the oceans. If the pull becomes imposing, the weakest places in the surface are likely to break. The results would be a surge in volcanic activity and more quakes. READING PASSAGE 3: OVERPOPULATION AND SOCIAL BEHAVIOR Easy-Vocabulary Version A classic series of experiments to determine the effects of overpopulation on communities of rats was 111 reported in February of 1962 in Seleneifie Amerieen. The experiments were conducted by a psychologist, John Calhoun, and his associates. In each of these experiments, an equal number of male and female adult rats were placed in an enclosure and given an adequate supply of food, water, and other necessities. The rat populations were allowed to increase. Calhoun knew from experience approximately how many rats could live in the spaces without experiencing stress due to overcrowding. He allowed the population to increase to approximately twice this number. Then he stabilized the population by removing babies that were not dependent on their mothers. He and his associates then carefully observed and recorded behavior in these overpopulated communities. At the end of their experiments, Calhoun and his associates were able to see that overcrowding causes a breakdown in the normal social ways among rats. The females in the rat population were the most seriously affected by the population. They showed deviant behavior. In fact, many of the babies died as a result of poor maternal care. For example, mothers sometimes abandoned their babies to die. However, the experiments verified that in overpopulation communities, mother rats’ behavior may be considered pathological. The dominant males in the rat population were the least affected by the overpopulation. Each claimed an area of his own. Therefore, these individuals did not experience 112 the same degree of overcrowding. Since the dominant males had adequate space, they were not as seriously affected. However, dominant males did behave oddly at times. Their vexatious behavior consisted of attacks on weaker male, female, and immature rats. Nondominant males in the experimental rat communities also exhibited deviant social behavior. Some moved very little and ate and drank at times when the other rats were sleeping in order to shun contact with them. Other nondominant males were hyperactive, chasing other rats and fighting each other. This segment of the rat population, like all other parts, was affected by overpopulation. The behavior of the nondominant males and of the other parts of the rat population has similarities to human behavior. People in densely populated areas exhibit odd behavior similar to that of the rats in Calhoun’s experiments. In large urban areas such as New York City, London, and Mexico City, there are abandoned children as well as mean, powerful individuals. There are also people who withdraw and people who become hyperactive. The quantity of other social problems such as murder, rape, and robbery also frequently occur in densely populated communities. Is the principal cause overpopulation? Calhoun’s experiments suggest that it might be. In any case, social scientists and city planners have been influenced by the results of this series of experiments. 113 Difficult-Vocabulary Version A classic series of experiments to determine the effects of overpopulation on communities of rats was reported in February of 1962 in SQifiDLiIiQ Amerieen. The experiments were conducted by a psychologist, John Calhoun, and his associates. In each of these experiments, an equal number of male and female adult rodents were placed in an enclosure and given an adequate supply of food, water, and other necessities. The rat populations were allowed to increase. Calhoun knew from experience approximately how many rats could live in the spaces without experiencing stress due to overcrowding. He allowed the population to increase to approximately twice this number. Then he stabilized the population by removing offspring that were not dependent on their mothers. He and his associates then carefully observed and recorded behavior in these overpopulated communities. At the end of their experiments, Calhoun and his associates were able to conclude that overcrowding causes a breakdown in the normal social routines among rats. The females in the rat population were the most seriously affected by the population density. They showed deviant maternal behavior. In fact, many of the pups died as a result of poor maternal care. For example, mothers sometimes abandoned their babies to die. However, the experiments verified that in overpopulation communities, mother rats’ behavior may be considered pathological. 114 The dominant males in the rat population were the least affected by the overpopulation. Each claimed an area of the enclosure as his own. Therefore, these individuals did not experience the same degree of overcrowding. Since the dominant males had adequate space, they were not as seriously affected. However, dominant males did behave oddly at times. Their vexatious behavior consisted of attacks on weaker male, female, and immature rats. Nondominant males in the experimental rat communities also exhibited deviant social behavior. Some withdrew completely: they moved very little and ate and drank at times when the other rats were sleeping in order to shun contact with them. Other nondominant males were hyperactive, chasing other rats and fighting each other. This segment of the rat population, like all other parts, was affected by overpopulation. The behavior of the nondominant males and of the other components of the rat population has similarities to human behavior. People in densely populated areas exhibit deviant behavior parallel to that of the rats in Calhoun’s experiments. In large urban areas such as New York City, London, and Mexico City, there are abandoned children as well as cruel, powerful individuals. There are also people who withdraw and people who become hyperactive. The quantity of other social pathology such as murder, rape, and robbery also frequently occur in densely populated communities. Is the principal cause of these dilemmas 115 overpopulation? Calhoun’s experiments suggest that it might be. In any case, social scientists and city planners have been influenced by the results of this series of experiments. READING PASSAGE 4: DEFORESTATION Easy-Vocabulary Version Deforestation has only recently been recognized as a global problem. But the tropical regions are where it is occuring most rapidly. This is because most forests in other regions have already been reduced centuries ago. Jungles are being cut down today, however, because of the increased use of machinery and the surging demand for wood products. Tropical forests produce a nearly endless variety of plants. People often do not realize that the jungle has a very fragile balance of nature. The soil is very rich in nutrients, but only because of the decaying plants and animals. When the plants and animals disappear, the earth quickly loses its richness. For example, in the Amazon region of Brazil called Bragantina, the forest was cleared around 1900. By 1940, the earth there had turned into a hard, useless surface. Bragantina was left as a wasteland reminder of the danger of meddling with the balance of nature. Besides losing productive earth, another reason to keep 116 tropical forests is the abundance of plants and animals they support. Many of these life forms have not been scientifically analyzed yet. Scientists wish to gain as much information as possible before they are destroyed. This information can be valuable. For instance, many drugs are manufactured from the raw materials found in these forests. Quinine for malaria and Dioscorea for birth control are tropical forest products. Scientist are convinced that these forests hold many other secrets, such as the key to natural pesticides. The third argument for keeping tropical forests concerns their effect on worldwide climate patterns. Plants give off water vapor into the air, which becomes rain. Plants also take in carbon dioxide and give off oxygen. If the forests are cut down, then there will be less good air to breathe. And because the carbon dioxide blanket over the earth tends to warm the environment, and increase in this gas would create a warming effect. Without large rainforests, our world faces foul, hot, dry earth and air. The rainforests are disappearing at lightening speed. Every second hundreds of trees fall before saws. Of course, there is some good news. In countries such as Korea and Australia, wise forest management provides for planting trees and for the setting aside of protected forests. Some trees should not be cut because they take hundreds of years to reach maturity. Finally, something 117 must be done about poverty, for it is the very poor who cut down so many of the trees for land, homes, and fuel. It is necessary that we act now to make sure that the world will be hospitable in the future. Difficult-Vocabulary Version Deforestation has only recently been recognized as a global problem. But the tropical regions are where it is occuring most rapidly. This is because most forests in other regions have already been decimated centuries ago. Jungles are being cut down today, however, because of the increased use of machinery and the surging demand for wood products. Tropical forests produce a nearly interminable variety of plants. People often do not realize that the jungle has a very fragile ecosystem. The soil is very rich in nutrients, but only because of the decaying plants and animals. When the plants and animals disappear, the earth quickly loses its fertility. For example, in the Amazon region of Brazil called Bragantina, the forest was cleared around 1900. By 1940, the earth there had turned into a hard, useless surface. Bragantina was abandoned as a wasteland reminder of the danger of meddling with the balance of nature. Besides losing productive earth, another reason to keep tropical forests is the cornucopia of plants and animals they support. Many of these life forms have not been 118 scientifically categorized yet. Scientists wish to garner as much information as possible before they are destroyed. This information can be invaluable. For instance, many drugs are manufactured from the raw materials found in these forests. Quinine for malaria and Dioscorea for birth control are tropical forest products. Scientist are convinced that these forests hold many other secrets, such as the key to natural pesticides. The third argument for conserving tropical forests concerns their effect on worldwide climate patterns. Plants give off water vapor into the atmosphere becomes rain. Plants also take in carbon dioxide and give off oxygen. If the forests are cut down, then there will be less good air to breathe. And because the carbon dioxide blanket over the earth tends to warm the environment, a surge in this gas would create a warming effect. Without large rainforests, our world faces foul, hot, parched earth and air. Nevertheless, the rainforests are disappearing at lightening speed. Every second scores of trees fall before saws. Of course, there is some good news. In countries such as Korea and Australia, wise forest management provisions provide for planting trees and for the designation of protected forests. Some trees should not be cut because they take hundreds of years to reach maturity. Finally, something must be done about pauperism, for it is the very poor who cut down so many of the trees for land, 119 homes, and fuel. It is necessary that we act now to ensure that the world will be hospitable in the future. APPENDIX C THE VOCABULARY TESTS Vocabulary Test 1 Score VOCABULARY CHECK UP Name Directions: Put the letter representing the appropriate answer in the blank at left. Newspapers are filled with a. newshype, b. articles, c. people, d. incidenary information. work with each other. a. sexists, b. racists, c. associates, d. rivals. Prisons are composed of many small a. playpens, b. enclosures, c. intellects, d. jokes. The pilot the falling airplane. a. pulled out, b. ejected, c. exploded, d. stabilized. All parents have human a. neuroses, b. psychoses, c. offspring, d. torture. From the data, we can several things. a. involve, b. conclude, c. impress, d. confine. How thick is it? What is its a. overlap, b. proton, c. resistance, d. density? Those who react unusually are considered a. pliant, b. rigid, c. deviant, d. resilient. We believe in the care of children. a. maternal, b. patterned, c. indigent, d. prorated. All systems are up and their proper functions have been by the support team. a. verified, b. spiked, c. detonated, d. undone. The dog licked its a. pups, b. up, c. collar, d. servant. 120 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 121 This man is not just mentally sick, he is a. conformist, b. individualistic, c. scholarly, d. pathological. Leaders tend to be in personality. a. dominant, b. bizarre, c. impolite, d. weak. Did the patient any signs of recovery? a. discern, b. exhibit, c. create, d. face. The weak and fearful man to the corner. a. extracted, b. withdrew, c. collapsed, d. drank. This kid isn’t just moving around, he’s a. superordinate, b. conductive, c. loudly, d. hyperactive. How many is this machine made of? a. components, b. glasses, c. tips, d. puffs. This man is mean and a. kind, b. obscure, c. cruel, d. beneficial. Who did the criminal sexually? a. gripe, b. obfuscate, c. regress, d. rape. This is a problem, but the cause is that. a. disputing, b. royal, c. principal, d. outrage. This is a very bad, problem. a. horrible, b. OK, c. concentrated, d. resolved. He is wrong: he made a a. observation, b. tremor, c. blunder, d. go. This is very bad. It’s a. catatonic, b. capable, c. catastrophic, d. it. When a bird comes down, it a. expires, b. roosts, c. flies, d. soars. What happened? What over there? a. transpired, b. related, c. obscured, d. did. He causes trouble by them into a riot. a. agitating, b. retelling, c. serving, d. knowing. These things are together. a. touched, b. separated, c. indolent, d. intertwined. When the comes in, the water rises. a. tide, b. whale, c. ride, d. sun. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 122 Pull it: Give it a a. put, b. rip, c. sip, d. tug. He’s very big and a. immaculate, b. weak, c. imposing, d. inferior. The space travelers are now back on a. terra firma, b. the tree, c. the map, d. air. What are you using to make the house? a. shower, b. convex, c. framework, d. show. The sailboat lightly over the water. a. glides, b. plows, c. squats, d. stows. The mean dog has been on the robber. a. relished, b. fed, c. unleashed, d. true. The diggers have the lost treasure. a. lost, b. resurfaced, c. shot, d. unearthed. The workers are the weak walls with wood. a. sounding, b. buttressing, c. going, d. pithing. The boy the girl to get her attention. a. nudged, b. killed, c. shouted, d. scaled. The mountains are sharp and a. devastated, b. ripped, c. cropped, d. craggy. We are the bear from the house right now. a. killing, b. snapping, c. dislodging, d. doing. If you put pure cotton in water, it a. dirties, b. outs, c. regrets, d. shrinks. Rats, mice, and beaver are called a. mealtickets, b. rodents, c. vulgar, d. down. There’s not much, but it will be a. sufficient, b. gone, c. large, d. leaden. He does the same thing everyday according to his a. rutting, b. plane, c. level, d. routine. The two men exhibit , or similar behavior. a. parallel, b. deviant, c. obvious, d. dumb. The of the world are too many to solve. a. dopes, b. cracks, c. vitals, d. dilemmas. The method was bad but now has been a. defined, b. clarified, c. refined, d. done. The poor children in the disaster all a. flourished, b. perished, c. ate, d. got out. 123 Vocabulary Test 2 score Vocabulary Check Up Name Directions: Put the letter representing the appropriate answer in the blank at left. 1. 10. 11. 12. 13. 14. 15. 16. A person’s face has many distinctive 221:.- a. eyes, b. features, c. clues, d. bulges. A railroad track can eeperere a town into two parts. a. cut, b. move, c. arrange, d. shift. There is a deep ereek in the earth in the hills. a. fault, b. bulge, c. pattern, d. plate. The lines and circles on the cloth make a gesig . a. picture, b. crack, c. pattern, d. plate. There is one leng_line_2f_mguntains. a. peak, b. range, c. valley, d. subcontinent. The higheer_peinr in the mountains is over there. a. peak, b. range, c. valley, d. floor. The man found infernetion ro eelve the mystery. a. clues, b. views, c. news, d. dues. French bread has a hQiQ_§QII§§§_QE_ED§_QE§§1Q§o a. crack, b. cut, c. crust, d. coast. There are many cities along the ocean side. a. rail, b. island, c. crust, d. coast. Dig a leng_uarr9!_hele to Put the wall in- a. tension, b. tremor, c. trench, d. tightness. Ships are able to stsx_gn_t92_2f_uater- a. witness, b. float, c. fault, d. fly. Faucets are Eh1DQ§_EDQ§_I§l§§§§_!éL§£_E£§§§B£§- a. valves, b. stems, c. plates, d. drums. The greer_lepg_me§e actually moves. a. part, b. evidence, c. continent, d. shift. India is nart_2f_a_larger_land_ma§s. a. surface, b. subcontinent, c. satellite, d. den. Taiwan 18 a land_mass_§urreunded_bx_gatero a. fault, b. torpedo, c. tropic, d. hot island. The genrer of the earth is hot, liquid rock. a. core, b. clue, c. cut, d. outside. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 124 This hard job creates preeenre. a. tremor, b. work, c. bulge, d. tension. The preefi is in the book. a. clue, b. truth, c. lie, d. evidence. Earth_ssientists study earthquakes- a. geologists, b. deists, c. cultists, d. dudes. The eneek_yeyee were strong. a. cores, b. tremors, c. trenches, d. tensions. You may eee_eng_£ee1 what I have learned. a. pressure, b. change, c. witness, d. sensitize. The quake nexed_te_ene_side the road- a. shoved, b. shifted, c. cracked, d. resurfaced. Volcanoes have eruptions. a. magma, b. heat, c. blow ups, d. dangers. They believe eerrn ie made 32 of plates en water. a. tectonics, b. platonism, c. migranes, d. Budda. If rocks in get very hot, they become liggig roc . a. lava, b. magma, c. pumice, d. crack. That man has a good igee. a. tendency, b. prototype, c. proton, d. theory. They took pictures in_rhe_eir. a. serially, b. floating, c. aerially, d. now. The sun is likely to come out tomorrow. a. fortunate, b. bright, c. probable, d. unknown. Some indieerions are that he did it. a. riots, b. drifts, c. trends, d. implications. Sailboats may grit; when there is no wind. a. be still, b. arrive, c. spin, d. float about. After a big meal, your middle may beige. a. burn, b. hurt, c. be large, d. float. The story may eeeeenr_fier his story. a. explain, b. record, c. count, d. emphasize. I doubt it, but there may be a elighr change. a. radical, b. moderate, c. little, d. surprising. Despite the tragedy on board, all lives were a. preserved, b. taken, c. lost, d. put out. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 125 The frightened people all the burning house. a. ran into, b. went under, c. did, d. forsook. The police all persons from the waterfront. a. excited, b. evacuated, c. killed, d. looked. What good fortune has you that you are rich? a. befallen, b. totalled, c. hurt, d. tickled. The drugs were to the sick people. a. shot, b. drug, c. tugged, d. dispensed. She has gotten great in the college. a. fate, b. germs, c. loss, d. erudition. Chickens, ducks, and geese are all a. poultry, b. heavy, c. featherless, d. brutal. Evil persons may in the dark places. a. float, b. lurk, c. haven, d. replenish. The army the enemy to go back. a. helped, b. suggested, c. compelled, d. asked. The mountain is not smooth, it has many a. crevices, b. ice spots, c. dollars, d. inches. His kind speech our fears. a. increased, b. got to, c. upped, d. diminished. The woodcutters the trees in the forest. a. developed, b. decimated, c. polished, d. ate. The ocean water into the inlet. a. surged, b. ripped, c. floated, d. compelled. The wait seemed a. new, b. old, c. out, d. interminable. Every animal lives in its own a. hut, b. hole, c. ecosystem, d. dispensary. The woman with children has no trouble with a. money, b. room, c. fertility, d. scores. There is something for all in the a. cornucopia, b. ledger, c. anvil, d. cave in. The scientist all the animals by size. a. read, b. got, c. categorized, d. detached. The farmer all the fruit on the trees. a. ate, b. set out, c. reviled, d. garnered. 53. 54. 55. 56. 57. 58. 59. 60. 126 The gold and silver are nearly a. valuable, b. rich, c. invaluable, d. moldy. The ground is so dry that my mouth is a. parched, b. wet, c. moist, d. pulsating. It’s here. , I wonder who brought it. a. Then, b. Which, c. Nevertheless, d. Been. The animals came by handfuls and a. scores, b. millions, c. digits, d. drums. What have you made for the future? a. excision, b. delusion, c. guts, d. provision. The ill have been to the hospital. a. split, b. designated, c. replaced, d. done. The poor suffer from a. pauperism, b. nutrition, c. insomnia, d. time. He tied it to that it wouldn’t fall off. a. take, b. tough, c. ensure, d. despoil. APPENDIX D CLOZE TESTS Cloze Test 1 Score Comprehension Check Name Directions: Use your understanding and memory of the reading passage to identify the correct word to fill in the blank below. Write the letter of the word next to the appropriate number. A Theory of Earth’s Structure There is bountiful evidence to support the 1 of plate tectonics. Aerial photographs and satellite 2 show profound trenches in the Earth. These 3 are clues for the arrangement of the 9 plates. Experts have found places where 5 motion is likely to transpire. Along these § , they measure the shock pulses that are 2 by the small tremors from deep 8 . The measurements support plate tectonics theory. One implication of plate tectonics is that 2 of the earth was once a 19 mass. By plate realignment, this mass 11 into plates moving in different directions. 12 the most recent movement, South America nudged 12 just under its western bulge, and 14 America wrapped around North Africa. Europe and the 15 of Asia were far to the 1§ . Africa’s movement into Europe forged the Mediterranean 11 and some mountains: the Alps and the 1g range. The Himalaya range was the 12 of the movement of the subcontinent of 29 away from southeast Africa and into Asia. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. a. single, b. Sea, c. North, d. South, e. trenches, f. all, 9. India, h. theory, i. ground, j. interior, k. cracks, 1. surface, m. underground, n. bulk, 0. issue, p. Atlas, q. caused, r. Africa, s. north, t. separated, u. before, v. pictures. 127 128 Cloze Test 2 Score Comprehension Check Name Directions: Select the appropriate word from the list below to fill in each blank in the passage. Then put the letter for the word in the blanks on the left below. Learning About Earthquakes The Chinese study earthquakes because they have 1 some of the worst 2 . Two times in Chinese history, in 5 and in 1976, more than half of a 5 people perished in quakes. Because of their 5 in forecasting quakes, many lives have also been 5 . In the winter of 1974-75, there were 2 signs twice. In December, the experts 2 a warning in the Liaoning Peninsula. All the people 2 the cities, but the earthquake did not 10 . Two months later, the 11 again warned the people. Some were 12 to leave their homes. They 12 it was another false alarm. However, the 15 evacuated the dangerous areas. This time a 15 earthquake did befall them. Because the 15 had been emptied, thousands of 12 were saved. Before the terrible 1976 15 , there were readable, but unclear signs, so a 12 was not given until too 29 . ___1. a. police ___2. b. cities ___3. c. severe ___4. d. thoughtless __5 . e . earthquake ___6. f. million ___7. 9. success ___8. h. saved ___9. i. unwilling ___10. j. occur ___11. k. lives ___12. l. sent ___13. m. thought 14. n. left 129 ___15. O. 1556 ___16. p. suffered .___17. q. readable __18. r. 1776 ___19. s. good ___20. t. posted u. people v. late w. experts x. warning y. quakes Cloze Test 3 Score Comprehension Check Name Directions: Select the appropriate word from the list below to fill in each blank in the passage. Then put the letter for the word in the blanks on the left below. Overpopulation and Social Behavior The females in the rat population were the most 1 affected by the population. They showed deviant 2 . In fact, many of the pups 2 as a result of poor maternal A . For example, mothers sometimes 5 their babies to die. However, the 5 verified that in overpopulation communities, 7 rats’ behavior may be considered pathological. The dominant males in the 5 population were the least affected by 2 . Each claimed and area of the enclosure as his won. Therefore, these 19 did not experience the 11 degree of overcrowding. Since the dominant males did behave oddly at 14 . Their vexatious behavior consisted of attacks on 15 male, female, and immature rats. Nondominant males in the experimental rat 16 also exhibited deviant social behavior. Some 17 very little an ate and 15 at times when the other rats were 19 in order to shun contact with them. Other _25__ males were hyperactive. 130 ___1. a. abandoned ___2. b. behavior (___3. c. care ___4. d. died ___5. e. experiments ___6. f. affected ___7. g. nondominant ___8. h. have '___9. i. individuals ___10. j. same .___11. k. communities (___12. 1. moved ___13. m. mother .___14. n. males ___15. o. overpopulation ___16. p. order ___17. q. drank ___18. r. rat ___19. s. seriously 20. t. times Cloze Test 4 Score Comprehension Check Name Directions: Select the appropriate word from the list below to fill in each blank in the passage. Then put the letter for the word in the blanks on the left below. Deforestation Besides losing productive earth, another reason to preserve 1 forests is the abundance of plants and 2 they support. Many of these life 1 have not been scientifically categorized yet. 5 wish to gain as much 131 information from the tropical 5 as they can before theyh are 5 . This information can be invaluable. For 1 , many drugs are manufactured from the 5 materials found in these forests. Quinine for 2 and Dioscorea for birth control are tropical forest 1Q . Scientists are convinced that these forests 11 many other secrets, such as the 12 to natural pesticides. The third argument for conserving tropical forests 13 their effect on worldwide climate patterns. 15 give off water vapor into the atmosphere, which becomes 15 . Plants also take in 15 dioxide and give off 17 . If the forests are 15 down, then there will be less good 12 to breathe. And because the carbon dioxide 20 over the earth tends to warm the environment, a surge in this gas would create a warming effect. ___1. a. animals ___2. b. plants ___3. c. concerns ___4. d. destroyed ___5. e. cut .___6. f. forms (___7. 9. ground ___8. h. hold ___9. i. instance ___10. j. carbon ___11. k. key ___12. 1. air ___13. m. malaria ___14. n. rain .___15. o. oxygen .___16. p. products ___17. q. blanket ___18. r. raw 19. s. scientists 20. 132 t. tropical u. forests BIBLIOGRAPHY Allen, Virginia French. (1983). Teennignee in Teeening yeennn1ery. New York: Oxford. Anderson, R. C., & Freebody, P. (1981). Vocabulary knowledge. In J. Guthrie (Ed.), Qennreneneien end 19228109: Researsh.aexiexs (pp. 77-117). Newark. DE: IRA. Bartlett. F. c. (1932). Remembering: A Study in Experimental ang §Qe151 Esyeno1ogy. Cambridge, MA: Cambridge University Press. BarnitZ. J- G- (1985)- Reading 2222192090; 2f Hennatixe Speakers 51 Eng1ien. Orlando, FL: Harcourt, Brace, Jovanovich. Barnitz, J. G. (1986). Toward understanding the effects of cross-cultural schemata and discourse structure on second language reading comprehension. lgurnal.2f Beading fiehexiex. 18. 95-116- Bleich, David. (1978). EQQIQQLiYQ gririeien. Baltimore, MD: The Johns Hopkins University Press. Bloomfield, L. (1942). Linguistics and reading. The Elementary English 822192. 19. 125-130. 183-186. Britton, J. (1970). Lengnege ang Learning. London: Penguin. Carrell, P. L. (1985). Facilitating ESL structure by teaching text structure. IE§QL Qnerrerly, 19, 727-752. Carrell, P. L. & Eisterhold, J. C. (1983). Schema theory and ESL Reading pedagogy. IgfigL Qnerrer1y, 17, 553-573. Carter, R. (1987). Vocabulary and second/foreign language teaching- Language Teaching. 20. 3-16- Clarke, M. A. (1979). Reading in Spanish and English: evidence from adult ESL students. Lengnege Leerning, 29, 121-143. 133 134 Devine, Joanne. (1987). General language competence and adult second language reading. In Devine, J.: Carrell, p.; and Eskey. D- (Edeo). Beeeareh in Reading in English as a Seeend.Lengeege. (pp- 73-86)- Washington. o.c.: resor. Diller, Karl Conrad. (1978). The Lengnege Teeening genrreyerey. Rowley, MA: Newbury House. Eskey, D. E. (1973). A model program for teaching advanced reading to students of English as a foreign language. Language Learning. 23. 169-184. Eskey, D. E. (1988). Holding in the bottom: an interactive approach to the language problems of second language readers. In Carrell, P.: Devine, J.: Eskey, D. (Eds.), Intereetixe Appreeehee t9 Seeendphengeege Beading. (pp- 93-100). New York: Cambridge University Press. Fish. Stenley- (1980). Ie.There a Text in This glass? Cambridge, MA: Harvard University Press. Floyd, P. & Carrell, P. L. (1987). Effects on ESL reading of teaching cultural content schemata. Lengnege Learning. 37. 89-108- Freebody, P. 8 Anderson, R. C. (1983a). Effects of differing proportions and locations of difficult vocabulary on text comprehension. genrne1 91 Begging Behexier. 15. 19-40- Freebody, P. & Anderson, R. C. (1983b). Effects of vocabulary difficulty, text cohesion, and schema availability on reading comprehension. Begging Research Quarter1y, 18, 277-305. Fries, C. C. (1962). Lingnieriee en; Begging. New York: Holt. Goodman, K. (1973). Psycholinguistic universals in the reading process. In Smith, F. (Ed.), £§y9n511ngnistics and Reading (pp. 21-27). New York: Holt. Hartman, G. (1946). Further evidence of the unexpected large size of recognition vocabularies among college students. leurnel ef Edeeetienel Rexehelegx. 37. 436-439. Huey. E- (1908). The Bexehelegx and Redegegx 9f Beading. New York: MacMillan. Hudson, T. (1982). The effects of induced schemata on the "short circuit" in L2 reading: non-decoding factors in L2 reading performance. Lengnege Leerning, 32, 1-25. 135 Jenkins, J. R. et al. (1978). Vocabulary and reading comprehension: instructional effects. (Technical Report No. 100). Bolt, Beranek an Newman, Inc., Cambridge, MA: Illinois University, Urbana: Center for the Study of Reading. (ERIC). Johnson, P. (1981). Effects on reading comprehension of language complexity and cultural background of a text. IE§QL QBQISEILY: 15. 169-181- Johnson, P. (1982). Effects on reading comprehension of building background knowledge. TE§QL Quarterly, 16, 503-516. Kaplan, R. B. (1966). Cultural thought patterns in intercultural education. Language Learning, 16, 1-21. Krieger. H- (1976)- Theerx 9f gritieiem= A Traditiea and lte 5yetem. Baltimore, MD: Johns Hopkins University Press. Marks, C. B.: Doctorow, M. J.: & Wittrock, M. C. (1974). Word frequency and reading comprehension. The Jenrnel of Eeaeatienal Beeeareh. 67. 259-263- Omanson, R.C. (1982). An analysis of narratives: Identifying central, supportive, and distracting content. .Dieeeuree Ereeeeeee. 5. 195-224- Perkins, K. (1983). Semantic constructivity in ESL reading comprehension. T555L Qnerterly, 17, 19-27. Saville-Troike, M. (1984). What reelly matters in second language learning for academic achievement? TESQL Quarterly. 18. 199-219. Scholes, R. (1985). Tertnel finger. New Haven, MA: Yale University Press. Seashore, R. (1933). Measurement and analysis of extent of vocabulary Eexehelaaieal Bulletin 30 709-710. Smith. F- (1973)- Eexehelinaaieties and Beadina- New York: Holt. Stahl, S. A. & Jacobsen, M. G. (1986). Vocabulary difficulty, prior knowledge, and text comprehension. Journal of Reading Behaxier. 18. 309-323. Stahl, S. A. et a1. (1989). Prior knowledge and difficult vocabulary in the comprehension of unfamiliar text. Beading Beeeareh Quarterly 24 27-43. Swaffar, J. K. (1988). Readers, texts, and second languages: the interactive processes. The Modern Language learaal. 72. 123-149- 136 Taglieber, L. K.: Johnson, L. L.: 8 Yarbrough, D. B. (1988). Effects of prereading activities on EFL reading by Brazilian college students. TE5QL Qnerterly, 22, 455-472. Twaddel, F. (1973). Vocabulary expansion in the TESOL classroom. TEfiQL Qnerterly, 7, 61-78. Willeneen. mm (1974). Wine Statistieal Beneening. San Francisco: Freeman. Zukowski/Faust, Jean: Johnston, Susan: 8 Atkinson, Clark. (1983). Betyeen the Linea. New York: Holt. 137 GENERAL REFERENCES Carrell, P. L. (1981). Culture-specific schemata in L2 comprehension. In R. Orem 8 J. F. Haskell (Eds.), Seleeted Banere Eran the.ninth Illineie TESQLLEE (pp- 123-132). Chicago, IL: Illinois TESOL/BE. Carroll, J. 8.: Davies, P.: Richman, B. (1971). here Eregneney 5553. Boston, MA: Houghton Mifflin Company. Eco, U. (1979). The Bele er the Benger. Bloomington, IN: Indiana University Press. Gibson. E. & Levin. H- (1975)- The Eexehelegx of Reading- Cambridge, MA: MIT Press. Hirsch, E. D. (1987). gnltnrnl Litereey. Boston: Houghton. Ingarden. R- (1973)- The techitiah of the Literary Kerk of Art. Trans. by Grabowicz, G. C. Evanston, IL: Northwestern University Press. Iser, W. (1978). The Aet er Begging. Baltimore, MD: John Hopkins University Press. Mezynski, K. (1983). Issues concerning the acquisition of knowledge: effects of vocabulary training on reading comprehension. Beyley er Egneetlenel Researgh, 53, 253-279. Neisser, U. (1967). Qegnltlye Eeyehelegy. New York: Appleton. Olson, D. (1977). From utterance to text: the bias of language in speech and writing. fieryerg Edueetlenal Beyleg, 47, 257-281. Rosenblatt. L- (1978)- The Reader. the Text. the Eee_. Carbondale, IL: Southern Illinois University Press. Steiner, G. (1975). fitter fiehel. New York: Oxford University Press. N STATE V lllllllllli R HICHIG WI 31293