不同圖片類型對於看圖寫作表現之影響

全文

(1)Chapter I Introduction 1.1. Background Picture writing is currently one of the formats of writing assessments adopted in Taiwan. In this type of writing assessment, examinees are required to compose paragraphs by looking at and understanding a picture sequence, interpreting the picture sequence, and then transferring the visual information into the verbal mode of expression. With different types of writing assessments, when it comes to the question which is the best way to measure a student’s ability to express ideas in written words, the question always leads to an unsettled dispute; that is, no one may make bold statements, without controversy, claiming one or another writing measurement to be the best. Written expression assessment, according to Cole et al. (1999), has long been plagued by a lack of theory, even though it is generally suggested that when testing writing, examiners should set writing tasks which may elicit valid samples of writing and that the samples of writing should be scored both validly and reliably. Using pictures for writing gains a place among various kinds of ways in writing assessments such as in the GEPT (General English Proficiency Test), the JCEE (Joint College Entrance Examination), and the SAT (Scholastic Achievement Test). Pictures provide writers with ideas for common tasks like describing people, objects, 1.

(2) places, or even processes. Picture writing is also likely to require writers to describe events and incidents happening in daily life. In this way, on the one hand, writers can focus their full attention on using the written language rather than strain their brains for ideas to write about. On the other hand, if writers fail to express in written words the content of a picture, examiners may well conclude that the writers lack the language ability, which is necessary for the writing purpose. 1.1.1. English Composition in the Joint College Entrance Examination (1981-2001) English composition was added to the English test in the JCEE in the academic year 1981. Since then, the item types mainly involve free writing, guided writing, and topic sentence writing. In “free writing,” only one topic is offered; in “guided writing,” clues together with a given topic are offered for developing details in each paragraph; and in “topic sentence writing,” examinees are provided with topic sentences to develop each paragraph. In 1993, the College Entrance Examination Center conducted research on different item types of English composition. Among them, five have been found appropriate for the purpose of the JCEE (Chen et al., 1993). These five types are: (a) free composition with instructions, in which examinees are required to develop paragraphs around one single given topic with some provided clues or requirements to help expand ideas, and (b) composition based on pictures, in which a single 2.

(3) picture is given to facilitate examinees’ writing ability, and (c) composition based on a given situation, in which a situation is offered and examinees have to discuss or express opinions about the target situation, and (d) composition based on one topic sentence, in which examinees have to develop paragraphs with given topic sentences for each paragraph, and (e) composition in the form of letters, in which examinees are required to respond to a given letter. Although picture writing is among the suggested item types, it was adopted only once in the College Entrance Examination for Evening Programs in 1984. In this picture writing, one single picture was offered along with seven sentence patterns, four of which were required in writers’ composition. 1.1.2. English Composition in the Scholastic Achievement Test and the Department Required Test (2002-) By the academic year 2002, the JCEE has developed into two separate tests. One is the Scholastic Achievement Test (SAT) and the other is the Department Required Test (DRT). Picture writing, since the academic year 2004, has been used to evaluate the writing ability of those taking the Scholastic Achievement, while topic sentence writing and guided writing are mainly used in the Department Required Test. Table 1.1 shows the item types and content of each year in these two tests.. 3.

(4) Table 1.1 English Composition Topics in the SAT & the DRT of Each Year Item Type. Subject. Academic year. 2002 SAT. Guided Writing. DRT Topic Sentence 2003 SAT. Guided Writing. DRT Topic Sentence. The Most Precious Thing in My Room How to Use the Jackpot (Topic sentence: If I won two million dollars in the lottery, I would help…) Music Is an Important Part of Our Life Exams (Topic sentence: 1st paragraph: Exams of all kinds have become a necessary part of my high school life. 2nd paragraph: The most unforgettable exam I have ever taken is…. 2004 SAT. Picture Writing. DRT Guided Writing 2005 SAT. Picture Writing. DRT Guided Writing 2006 SAT. Picture Writing. DRT Guided Writing 2007 SAT. Picture Writing. DRT Guided Writing 2008 SAT. Guided Writing. A Drunken Man (Besides, there is a given start: One evening…) Traveling Is the Best Teacher Dozing Off in an English Class (Besides, there is a given start: In the English class last week,…) Class Reunion A Surprise Being Misunderstood/Feeling Wronged A Girl and a Cat A World Without Electricity An Apology Letter. 4.

(5) From 2004 to 2007, picture writing was used as the format for writing assessments of the SAT, even though in 2008 the format which was employed was guided writing. In the years from 2004 to 2007, it was the picture sequence rather than one single picture as suggested by Chen et al. (1993) that was used for the assessment of writing capability. Unlike the picture writing format used in the College Entrance Examination for Evening Programs in 1984, the picture prompts used in the SAT from 2004 to 2006 were all with three sub-pictures rather than only one single part. In taking the English composition test, writers were required to create a coherent and logical story by describing and producing ideas from the content of each picture frame. However, the SAT of 2007 was a different case. Two new elements were added to the picture writing in 2007, bringing some minor changes to this format. Firstly, the number of picture frames was not three, as before, but four. Secondly and more significantly, the picture sequence was no longer fixed but flexible. In other words, writers were allowed to re-order the picture sequence if they liked to exert their own imagination to create a complete, coherent and logical story. 1.2. Motivation and Objectives In a research conducted by the College Entrance Examination Center (Lin, 2005), it concludes that one reason for using picture sequences as a writing prompt is that. 5.

(6) picture writing can solve the existing problem—memorizing expressions or reproducing model essays in answering narrative composition questions. Picture writing does not allow examinees to start or end their paragraphs with memorized scripts, for the content of the picture sequence is certain and clear. The memorized expressions or the reproduction of model essays that writers try to utilize are no longer appropriate for different picture contents. This explains the test result that 9.59% of those taking the SAT in 2002 scored zero in the English composition of guided writing; however, about 11% to 12% of those taking the SAT from 2004 to 2006 scored zero in the English composition of picture writing. The increase in the number of those getting zero reveals a fact that writers no longer get points by utilizing memorized expressions or by the reproduction of model essays in picture writing as they do in guided writing. Using picture writing as the format of English composition may prohibit examinees from memorizing expressions or reproducing model essays as a shortcut to writing. Besides, the test results in this study reflect the striking fact that one-fifth of those taking the SAT have trouble writing a narrative with the aid of picture sequences. Is it the examinees themselves that lack the language ability which is necessary for the writing purpose? Or is it the picture sequence prompts that fail to effectively stimulate those examinees’ writing proficiency? These questions arouse an interest in the. 6.

(7) current study to figure out what kind of picture sequence prompt can successfully facilitate students’ writing ability. Moreover, the washback effect of the practice of using picture prompts leads to the fact that more emphasis is now put on picture writing in senior high schools than on the ones having formerly been adopted, like topic sentence writing and guided writing. Senior high school teachers, while preparing students for the SAT, often collect picture prompts from picture composition books and from magazines, newspapers, and comics in order to acquaint students with picture writing. Senior high school students, while preparing themselves for the SAT, usually expose themselves to a writing environment with lots of picture prompts. With the dominance of picture writing in teaching and learning activities, there is a need of large numbers of pictures that are appropriate to serve this purpose. Nevertheless, as more and more materials on picture prompts are being published, which features eliciting writing, the real concern is what effect the different types of picture sequences have on the writing performance. Do picture sequence drawers/designers give careful consideration to the effect of picture prompts? Do they deeply ponder which type of picture sequence is the most efficient in facilitating the writing ability? If they do not, the visual prompt is nothing more than a hindrance to the development of the writing ability of a student or an examinee. In this way,. 7.

(8) writing assessments, failing to smooth the writing process and to facilitate writing proficiency, are nothing more than an unfair puzzle, where a student or an examinee loses out, not because of his/her inability, but because of invalid or unreliable clues. Therefore, it is essential and vital that we find the appropriate types of picture sequences that foster, rather than hinder, the writing process. 1.3. Purpose of the Study and Research Question The purpose of this study is to classify the picture sequences and to examine their effect on stimulating the writing ability. This study collects writing prompts of picture sequences and then builds up the classification of picture sequences. The classification may exclusively apply to picture prompts with three frames rather than those with only one single part or those with four or more parts. With the classification, the study turns to look at the effect of each type of three-frame picture sequence on writing performance. The study analyzes the produced writing samples in terms of the length of writing products, lexical richness, and syntactic complexity in order to find out which type most effectively fosters the writing process. The research questions are: 1.. Which kind of three-frame picture sequence elicits the richest content?. 2.. Which kind of three-frame picture sequence elicits the most varied lexicon?. 3.. Which kind of three-frame picture sequence elicits the greatest syntactic. 8.

(9) complexity? 1.4. Significance of the Study It is hoped that this study will arouse the awareness, among picture prompt material designers as well as EFL/ESL teachers, of the importance of carefully designing and choosing an effective and appropriate picture sequence prompt for examinees and students. Besides, picture prompt designers and EFL/ESL teachers may be provided with a reference, helping them to design and choose an appropriate and effective three-frame picture sequence as a writing prompt.. 9.

(10) Chapter II Literature Review The literature review has two subsections: (a) picture prompts, including the effect, the possible defect, their use in English writing assessment in the JCEE, and their characteristics of effective picture prompts; (b) classification of picture sequences with three frames. 2.1. Picture Prompts It is often a difficult task to get writers to write, and the difficulty may arise from the writing process affected by several factors, including the attitude toward writing, writing capability, interest in the topic, background knowledge, and experience. In order to help writers to write, many ways have been suggested like outlining, modeling, etc. Among these ways, using prompts to elicit writing is quite common. Picture prompts refer to the pictures serving as a prompt for a story. The setting and characters portrayed in the picture suggest an event or relationship that writers are invited to develop and describe through a story. 2.1.1. The Effect of Picture Prompts on Writing Pictures can be used to provide concrete spurs to form imagery, which is a key strategy in producing writing (Sinatra, 1981). In addition, picture sequences can be used to train writers for the underlying structures of different types of discourse. 10.

(11) (Sinatra, 1986). According to Chiang (2003), the strength of picture aids generally falls into four categories: to assist young writers in producing a better-organized story frame; to provide linguistic cues that help to start the flow of thought, which contributes to a richer content; to elicit personal response and evoke imagination and creativity; to alleviate young learners’ writing anxiety and to be used as a motivational vehicle. In Chiang’s investigation on the effectiveness of picture prompts on EFL writing (2003), with the help of the given pictures, the students can construct a complete story format with a beginning, a body, and an ending. The students can write slightly longer texts with a richer vocabulary. Besides, there is a positive change of the students’ attitude and motivation. Pictures do provide something to write about. Juliebo & Edwards (1989, p. 26) stated, “Our best writing comes from situations where we feel a need to communicate, for without a need there is no point in writing.” Picture prompts stimulate ideas, convey information that can be shared, and remind one of his/her own background knowledge style (Stevick, 1986; Ur, 1988). In this way, pictures can be used to provide the content or the topic in order to make writing easier and better. Similarly, Alejandro (1994) mentions that pictures and other forms of visuals are popular because pictures evoke ideas and thoughts for children to write about, thus creating. 11.

(12) descriptive writing. With the above review of the literature, it is obvious that picture prompts are held highly as successful stimuli to encourage writing. Nevertheless, there are pros and cons of the picture prompts. In some studies, pictures are thought to limit imagination and creativity and make writers refrain from elaborating the content. 2.1.2. Possible Defects of Picture Prompts on Writing Picture prompts are used to get writers to write. However, some researchers have found that there are some possible defects in using picture prompts. Hough, Nuss, & Wood (1987) attempt to answer the question: what types of stimuli encourage story-telling. They give story elicitation tasks using a single picture, a wordless picture book, or a verbal prompt to write a story. The first two stimuli elicit an extensive vocabulary but with ill-defined story structures or disconnected discourse. They reason that pictures may hinder the story-telling process by limiting imagination and by adding verbal messages to the picture rather than allowing for elaboration. This statement is confirmed in a study of 165 elementary students, which yielded similar results. Joshua (2007) holds that although pictures can provide context and background information for student writers of all ages, they may also constrain creativity and that the deciding factor appears to be students’ background experience and knowledge of the English language. She expands on her thought that once. 12.

(13) students develop language proficiency, pictures may hamper creativity in writing development. In a word, it is no wonder that the picture prompt is mainly used as a stimulus not for advanced writers but for beginners, who struggle for the content to be written down. 2.1.3. The Use of Picture Prompts in the English Writing Assessment in JCEE Picture writing is one of the item types of English composition suggested in the report “The Pamphlet of Item Writing Skills and Scoring Guides for the Writing Ability Test on the JCEE,” which is published by the College Entrance Examination Center in 1993. The report also points out that an ideal picture prompt for the elicitation of writing capability is one single picture with an unambiguous central idea. Besides, the single picture should be provided with some cues or hints to help students grasp the content without misunderstanding, thus rendering the scoring unbiased and impartial. Picture writing was first adopted in the College Entrance Examination for Evening Programs in 1984. One single picture was offered with seven sentence patterns, four of which were required in writers’ composition (see Appendix A). However, according to the report (CEEC, 1993), these cues or hints failed to elicit writing as expected and the picture sequence, compared with the single picture, was then considered a more difficult writing format due to a more limited content and more constricted organization.. 13.

(14) However, according to a survey conducted in the report (CEEC, 1993), both college professors and high school teachers have held picture writing in high esteem as a better item type for JCEE English composition. In Table 2.1, it is revealed that professors and teachers who favor picture writing account for 44.6% and 37.2%, ranked respectively the first and the third among the item types, including free writing, guided writing, picture writing, question-answering writing, situation writing, writing with offered words and phrases, and others or no comments. Table 2.1 Opinions on Better Writing Item Types Collected from College Professors and High School Teachers College Professors (N=139 ). High School Teachers (N=371). Person. Percentage (%). Person. Percentage (%). Free Writing. 61. 43.9. 162. 43.7. Guided Writing (Topic Sentence). 58. 41.7. 178. 48. Picture Writing. 62. 44.6. 138. 37.2. Question-Answering. 25. 16.5. 81. 21.8. Situation-Based Writing. 50. 36.0. 195. 52.6. Writing with Offered Words and Phrases. 34. 24.5. 93. 25.1. No Comment. 5. 3.6. 17. 4.6. Others. 11. 7.9. 2. 0.5. Items. Groups Number. Note. More than one item type could be chosen; From The Pamphlet of Item Writing Skills and Scoring Guides for the Writing Ability Test on the JCEE (p. 5), by Chen, K. T., Huang, T. S., Lin, S. E., Lin, C. I., & Li, C. L., 1993, Taipei: College Entrance Examination Center.. In the paper (CEEC, 1993), many professors and high school teachers offer their 14.

(15) own opinion on picture writing. One professor suggest that picture prompts should be provided with a title, which not only prevents students from memorizing formulaic expressions or reproducing model essays but also helps them grasp the content. Another professor mentions that the writing ability of senior high graduates in Taiwan is supposed to be lower than that of 4th-to-6th grade native speakers. In view of this, controlled writing, including picture writing, should be much more appropriate for senior high students in Taiwan. A professor, however, holds a different view, mentioning that picture writing is usually for description or narration, not for expository and argumentation. Therefore, he suggests that picture prompts should include simple charts to make up for the lack of practices for expository and argumentation. All these above are mostly studies conducted in 1993, and relatively little research about picture writing has been put forward in recent years even though picture writing has been a mainstream for the SAT since 2004. Picture prompts with three picture frames were used as an established format for writing assessment of the SAT from 2004 to 2006. Three picture frames were given in sequence and the picture prompts in 2004 and 2005 were given along with a requirement that writers should start their composition with a given unfinished sentence. One reason for this adoption of picture writing is that the College Entrance. 15.

(16) Examination Center (CEEC) is trying to prevent writers from memorizing formulaic expressions or reproducing model essays. According to an analysis conducted by the CEEC (Lin, 2005), the SAT examinees who scored zero in the English composition respectively account for 12.83% and 11.12% in 2004 and 2005 respectively. Probing into the cause of the poor score, Lin infers that in picture writing, examinees can not get even 2 to 3 points by beginning and ending their composition with ambiguous expressions as before in such item types as guided writing. To be more specific, the content of the picture prompts is exact and explicit, making it difficult for examinees to describe the pictures with similar and ambiguous memorized expressions. The CEEC has indeed an intention to remind the English educators that English teaching is by no means training students to memorize formulaic expressions or to reproduce model essays but to use English to express themselves. 2.1.4. Characteristics of Effective Picture Prompts As the proverb says, “A picture is worth a thousand words,” a picture does convey certain messages. However, an author’s intended message may be interpreted in different ways by different readers. Sometimes readers do not even understand the message of the picture at all. In either case problems arise as a result. In other words, pictures can be effective writing aids, but the mere presence of intriguing pictures does not ensure the success of writing. In fact, not all the picture prompts are of equal. 16.

(17) value. Seaman (1993) mentions that the task of writing in a second language from a picture prompt involves a number of variables, including the characteristics of the picture, the native language and culture of the student, and the nature of writing task itself. It is essential to understand what quality or characteristic makes a picture prompt comprehensible and successful in eliciting writing. Picture content is a prominent characteristic in selecting a picture prompt (Schweizer, 1999). The content involves the object, the subject matter, and complexity. The object refers to a thing that is depicted in picture prompts and is important to the main characters. It could be a toy, a stuffed animal, or a book. By describing a proper object, a writer may come up with stories and ideas, producing writing samples with richer content. Therefore, picture prompts should clearly depict the objects they are meant to represent, or they may elicit content with visually similar objects. Moreover, the object of picture prompts should be relevant and age-appropriate to the background knowledge and experience of the target writer (Cycowicz, Friedman, Snodgrass & Rothstein, 1994). The subject matter or the topic should consider familiarity and writers’ prior knowledge, triggering scripts that writers can use to generate writing (Bates, 1991). However, pictures with recognizable figures produce more descriptive writing, which is possibly because writing about a known figure places a burden of being correct on. 17.

(18) children and in turn constrains their creativity (Bates, 1991). Complexity should also be taken into account in picture selection: high levels of complexity in terms of interactions between objects within the picture produce more and better writing than low levels of complexity (Samson & Wescott, 1983). However, visual complexity can hinder object identification particularly in young children (Cycowicz et al., 1994). Petterson (1986) stresses finding a balance in level of complexity; too little complexity results in too little information transfer. As for the type of prompt, it should relate to the type of writing one wishes to generate; single pictures will generate more descriptive passages, sequences of picture will generate more narratives (White, 1978). Regarding measurement of written language expression, Hooper et al. (1994, p. 388) provide an extensive review and criticism of the literature in writing assessment. They make suggestions and raise questions regarding the best ways to assess writing. One of their suggestions is to carefully select the picture prompts used to elicit a narrative response. They recommend using a real photograph as a visual stimulus rather than a line drawing, and suggest that the photo should contain:. 1. At least two characters, with perhaps one potential protagonist and one potential antagonist although this is not absolutely necessary, depending on the situation depicted. 18.

(19) 2. A depiction of some kind of “interesting” or “novel” scene or event (e.g., children lost in a cave). 3. Some kind of potential conflict between the antagonist and protagonist that necessitates that the protagonist engaging in a goal-based sequence of events to resolve the conflict or a scene in which a main character must resolve a problem using some goal-directed sequence of events. (Hooper et al., 1994, p. 388). According to Hooper et al. (1994, p. 388), though in the written expression of narrative structure one would expect developmental differences, the major components of initial events, attempts, and direct consequences should be elicited from most competent writers, while setting, internal response, and reaction may not be obligatory. Hooper et al. assert also that the aforementioned criteria should elicit a more cohesive, goal-directed response, which has been supported by Cole, Muenz, Ouhchi, Kaufman, & Kaufman (1997) with several of their studies. Cole et al. (1997) attempted to test Hooper et al.’s (1994) aforementioned stimulus hypotheses. They initially developed the “cliff” prompt, which met all of Hooper et al.’s criteria. They required 50 participants (older adolescents and adults) to write two stories: one story used the cliff prompt as a stimulus, whereas the other used a line-drawing. They found that the “Hooper” photographic stimulus would yield higher scores than a conventional line drawing stimulus on items that assessed. 19.

(20) structure (unity of ideas, fluidity of story, and so forth); however, on writing mechanics items (capitalization, punctuation, and so forth) it exhibited no significant difference. This study was replicated and supported by Cole et al. (1999) with a sample comprised of middle school students, rather than older adolescents and adults as the former one. Nevertheless, what they focused on was the photographic picture versus the line-drawing picture. They did not expand their research in each specific aspect of Hooper et al.’s (1994) theory versus the total effect to examine and analyze the effectiveness of each feature suggested, such as the novel event and the conflict. Building on Cole et al.’s work, Joshua (2007) developed a criterion for selecting a picture prompt that can be used for classroom writing. Picture prompts should be thought provoking and engaging, imaginative, and grade-level appropriate. However, the research focused on the classroom writing activity rather than on writing assessment, and on the native language rather than on the second or foreign language. 2.2. The Text Organization Patterns, the Story Grammar, and Expectation of the Consequence In the past, not many research studies have investigated which picture prompt type best facilitates students’ writing ability. With the absence of related studies on the classification of picture prompts with multiple frames, a review of literature on the text organization patterns, story grammar, and expectation of the consequence is. 20.

(21) expected to be helpful for a study on multi-frame picture prompts. 2.2.1. Sinatra’s Text Organization Patterns Text organization patterns play an important role in writing (Sinatra, 2000; Wong, 1997). Sinatra (2000) has put forward six major structure maps for expository and persuasive text patterns and types of thinking involved. Since text structure analysis exists for narrative story texts as well (Gordon & Braun 1983; Lodge, 1992; Van den Broek, 1989), the idea occurs to the researcher that it could be workable to apply these six text structures as the foundation of her classification of picture prompts. Picture prompts are in themselves a visual way to represent verbal texts and therefore definitely can be analyzed on the basis of text organization patterns. Moreover, Sinatra’s construct involves also thinking styles, which enhances the belief in its application to the classification. With Sinatra’s six major structures, which are shown in great detail in Table 2.2, and with the picture prompts having been collected from composition books, the researcher develops an initial classification of three-frame picture prompts. This initial classification includes (a) the sequence type (the time-based type and the steps-in-a-process type), (b) the topic-development type, and (c) the cause-and-effect type. First, the researcher derives the time-based type from the description of the text structure pattern of sequence maps. That is, episodes and events are arranged in. 21.

(22) temporal order and chronological order. Besides, the reason why classification maps and comparison maps are not included in the classification is that these two maps are mostly applied in one single picture with two parts. The reason why persuasion maps are not included is that persuasion maps are suitable for argumentation rather than narration and story-telling in picture sequences.. 22.

(23) Table 2.2 Maps for Expository and Persuasive Text Patterns and Types of Thinking Involved Map Name. Text Structure Pattern. Type of Thinking. Sequence Map;. Episodes and events arranged in. Selecting major episodes of expository sequence. Steps-in-a-Process. temporal order; sequence of events,. and then sequencing of subordinate events;. Sequence Map. chronological order; array and. detailing process sequences as with instructions,. organize steps of a process.. recipes, directions.. Topic Development Map. Telling about, describing, or. Visualizing components or views of a central topic. (also called Spider Map). explaining a central topic. Spatially. and relating/connecting subordinate and explicit. arranged around topic idea are. detailed information.. information groups showing coordinate ideas and then subordinate/detailed information. Classification Map. Information classified in a. Classifying or categorizing one central topic into. hierarchical layering of ideas related. class, subclass, attribute, and/or example. to one central concept; ideas arranged. relationships; clear hierarchy of information.. from top to lower level; often called a taxonomy. Comparison/Contrast Map;. Information about two or more topics. Indicating or inferring features about two or more. Same/Different Map. is compared and contrasted. Major. aspects of a central topic and listing or telling. features or characteristics related to. about likenesses and differences relative to each. each topic are indicated in center;. feature.. outer bands show the specific information being compared or contrasted relative to a feature. Cause and Effect Maps:. Causal patterns are connected; events. Addressing aspects of causality whereby. Chain of Events Pattern;. or happenings cause additional events. something happens as a result of something else,. Many Causes to One Effect. or outcomes to occur; a chain or. usually leading to major effects; or. Pattern; Major Event to. cascading array.. chain-of-happenings occurs in which each outcome. Outcomes Pattern. of a previous event precipitates a new outcome; or a major event occurs, leading to outcomes and additional causal/sequential events.. Persuasion Maps:. A persuasive argument is developed. Presenting a belief or thesis in a convincing way by. Persuasion by Arranging. by presenting reasons for one’s belief;. explaining reasons, with accompanying evidence. Reasons; Persuasion by. reasons are supported by data, by. and information to support one’s particular position. Rational Argument;. information, by examples, and by. on a topic.. Persuasion by. what others have said about the topic.. Point/Counterpoint Note. From “Teaching learners to think, read, and write more effectively in content subjects,” by Sinatra, R. C., 2000, Clearing House, 73 (5), 266-73. 23.

(24) 2.2.2. Story Grammar The structure of a story has been described as story grammar. “Story” refers to a narrative written discourse, and “grammar” here does not mean the study of morphology or syntax but the elements of a story. According to proposed theories (Mandler & Johnson, 1977; Rumelhart, 1975; Stein & Glenn, 1979), the story elements include, but are not limited to, the title, author, setting, main characters, conflict and resolution, events, and conclusion. Some additional elements that may be specified under “events” are the initiating event, internal response, attempt, consequence, and reaction. Stories may vary in context, but these story grammar elements are conveyed whether directly or inferably. Moreover, Stein & Glenn (1979) describe story grammar as a well-developed story inclusive of two main parts, setting and episode. In the setting, background knowledge about the leading characters is offered; in the episode are initiating events, initial responses, an attempt to obtain a goal, the outcome of the attempt, and characters’ reaction to the outcome. Liles (1989) makes it more specific by expanding that the behavioral sequence contains an initiating event or internal response motivating a character to formulate a goal-directed plan, an action or attempt at resolving the situation, and a direct consequence marking attainment or nonattainment of the goal.. 24.

(25) Judging from the above, the salience of those story elements can not be ignored when it comes to classifying picture prompts, which in themselves are mainly narratives. The story elements, especially the goal, the action, and the consequence, should be considered an important criterion in the classification of picture prompts. In view of this, the significance is reconfirmed of the cause-and-effect type. However, does the outcome or characters’ reaction of a story, i.e., the effect of the cause-and-effect type, make any difference in the picture sequence type? 2.2.3. Expectation of the Consequence Writers often suspend a reader’s interest with well-timed surprises. There is no doubt that surprise is a vital and central element of many stories. Surprise, for instance, has been shown to be essential in determining the perceived interestingness of a piece of text (e.g., Iran-Nejad, 1987; Kim, 1999). Similarly, the fun of a surprising ending can be applied in a picture sequence, which is used as a prompt for eliciting writing texts. A surprise occurs with the occurring of something unexpected in an event sequence. Yet, the relationship between expectation and surprise is not as clear-cut as it is supposed to be. Meyer, Reisenzein, & Schützwohl (1997) proposed a Expectancy-Disconfirmation Hypothesis, suggesting that surprise is experienced when expectations based on pre-existing schemas fail to be confirmed. According to. 25.

(26) Grimes-Maguire & Keana (2005), the following two distinct types of knowledge may drive comprehension, determining whether or not a forward inference can be made. The first is knowledge about the causal structure of the world derived from direct and personal experience, and the other is the knowledge of story structure. Likewise, readers of picture prompts may operate these two types of knowledge to form expectations of the consequence depicted in the picture prompts. Readers are sensitive to depicted details that may set up subsequent surprises. The picture prompt, just like an author, has the power to control the amount of information given to the reader about an upcoming surprise, which in turn has a great influence on the reader’s representation of the following events and the outcome. Moreover, an ending of a surprise may provide readers with a new perspective to ponder over the picture sequence, which in turn may elicit more ideas, making the prompt a successful writing incentive. Keenan, Baillet, & Brown (1984) have mentioned that readers are apt to make inferences involving causal relations, where one event explicitly stated in a text is connected by a causal inference to another explicitly stated event. In consideration of this, the expectancy of the consequence should play an indispensable part in classifying picture sequence prompts, the ones with causally sequenced plots. Therefore, regarding expectation of the consequence as a criterion, the researcher. 26.

(27) takes a further step to classify the cause-and-effect type into two sub-groups, i.e., the cause-and-effect type with an expected consequence and the cause-and-effect type with an unexpected consequence. 2.3. Summary Through the above review of the literature on the effect and use of picture prompts in writing assessment in the JCEE, it is found that no wonder picture prompts have been a generally-recognized mainstream in writing assessment. Though picture prompts used in writing may come with some possible defects, these demerits can not outweigh the merits of picture prompts. For the lack of classification of picture prompts, the researcher establishes her own classification on the basis of the text organization pattern, story grammar and the expectation of the consequence. The researcher derives five different types while classifying three-frame picture prompts, which are shown in Table 2.3. The classification includes (a) the time-based type, (b) the steps-in-a-process type, (c) the topic-development type, (d) the cause-and-effect type with an expected consequence, and (e) the cause-and-effect type with an unexpected consequence, which will be described in detail in Chapter III.. 27.

(28) Table 2.3 The Classification of Three-Frame Picture Prompts Type. Pattern. Time-Based Type. Each part of the picture sequence within a theme is developed with the passing of time.. Steps-in-a-Process Type. This type focuses on the ordering of steps to carry out a goal.. Topic-Development Type. A theme dominates each frame of the picture prompts which tells about or explains a central topic.. Cause-and-Effect Type with an Expected Consequence. Events or happenings of the previous parts cause some events or outcomes to occur in the following parts. Or a problem-solution sequence is embedded in the event sequence. The following events, solution, or reaction are expected.. Cause-and-Effect Type with an Unexpected Consequence. This type has a causally sequenced plot or a problem-solution sequence embedded; however, the effect is an event, solution or reaction out of the expectation of the characters or readers, namely, a surprising ending.. 28.

(29) Chapter III Research Design and Methodology In this study, differences in written responses to different kinds of three-frame picture prompts will be examined. The design of the study is to be described in the following sections: (a) materials, (b) subjects, (c) procedure of data collection, and (d) data analysis. 3.1. Materials 3.1.1. Development of Types of Three-Frame Picture Prompts This research focuses on three-frame picture prompts. Therefore, the following classification will not apply to picture prompts with only one single part or those with four or more parts. On one hand, picture prompts with only one single part does not require the arrangement of a sequence or a plot. On the other hand, after browsing through the picture sequences collected, the researcher finds that a picture sequence with four or more parts usually has more than one storyline and may involve more than one type of development. For the sake of simplicity in the present study, the following classification mainly applies to the picture writing structure with three frames. 3.1.1.1.. Criteria in the Classification of Three-Frame Picture Prompts:. Thinking Map, Story Grammar, and Readers’ Expectancy. 29.

(30) With Sinatra’s six major structures proposed in the thinking map (2000), the researcher develops an initial classification of three-frame picture prompts. This initial classification. includes. the. sequence. type. (the. time-based. type. and. the. steps-in-a-process type), the topic development type, and the cause-and-effect type. Moreover, with the story elements including the goal, the action and the consequence, put forward in story grammar (Mandler & Johnson, 1977; Rumelhart, 1975; Stein & Glenn, 1979) and studies on readers’ expectancy of the consequence, the researcher further proposes the following two types: the cause-and-effect type with an expected consequence and the cause-and-effect type with an unexpected consequence. As a result, the researcher derives five different types of the development of three-frame picture prompts: the time-based type, the steps-in-a-process type, the topic-development type, the cause-and-effect type with an expected consequence, and the cause-and-effect type with an unexpected consequence. In the time-based type, each part of the picture sequence within a theme is developed with the passing of time. In the steps-in-a-process type, the focus is on the ordering of steps to carry out a goal. In the topic-development type, a theme dominates each frame of the picture prompts which tells about or explains a central topic. In the cause-and-effect type with an expected consequence, events or. 30.

(31) happenings of the previous parts cause some events or outcomes to occur in the following parts, or a problem-solution sequence is embedded in the event sequence with expected following events, solution, or reaction. And in the cause-and-effect type with an unexpected consequence, there is a cause-effect plot or a problem-solution sequence embedded with events, solution or reaction out of the expectation of the characters or readers, namely, a surprising ending. 3.1.1.2. Collection and Adaptation of Multiple-Frame Picture Prompts Picture prompts were examined and collected from writing textbooks like Creative & Guided Composition (2004), Beginning Composition through Pictures (2005), Composition through Pictures (2006), Creative Writing Book 1 (2005), and Creative Writing Book 2 (2005). In the process of collecting, it was amazing to find that the format and content of picture prompts would vary from one to another. The number of picture frames should range from one to six, and expressed in each picture prompt were different ideas or experiences ranging from daily life to an imagined future world. In view of the variety of those collected picture prompts, the researcher first sorted out picture prompts, based on the following criterion－a picture prompt in which figures and moves of the characters are vividly and explicitly depicted, facilitating the comprehension of the picture content and arousing the motive to write,. 31.

(32) without any of the explanation of the teacher or the examiner, and in which there is a clear-cut central idea, allowing writers to think coherently and leaving space for imagination. Then the researcher adapted these selected picture prompts to her target format, that is, three-frame picture prompts, as those which were used in the SAT from 2004 to 2006. In order to adapt the different picture prompts with four or more parts to the target format, the researcher had to do without the frames describing subordinate detailed information and simply kept the three left parts containing the central thesis. For instance, in Figure 3.1, the picture frame 3, 4, and 5 describe the chase, which is depicted or can be inferred from the picture frame 2.. Figure 3.1 Sample of the Original Picture Prompt “Chase”. Note. Adopted from Beginning Composition through Pictures (p. 157), by Heaton, J. B., 2006, Essex, England: Longman.. 32.

(33) Therefore, the researcher cut out the picture frame 3, 4, and 5, to turn the six-frame picture prompt into a three-frame one and keep the same central thesis at the same time (see Figure 3.2). Though the adapted three-frame picture prompt loses the element of the suspense to some degree, the readers can still clearly acquire the content and the main idea of the original six-frame picture prompt. (For all the adaptation, see Appendix B and Appendix C.). Figure 3.2 Sample of the Adapted Picture Prompt “Chase”. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman.. With the adaptation, the researcher could classify these three-frame picture prompts into five different groups on the basis of the thinking maps presented by Richard C. Sinatra (2000), story grammar (Mandler & Johnson, 1977; Rumelhart, 1975; Stein & Glenn, 1979) and the studies on surprise endings. 3.1.2. Validation. 33.

(34) To make the classification more valid and reliable, the researcher decided to have these categories examined and validated by three experienced and qualified English teachers in Taichung First Senior High School. The researcher first explained the rationales and the criteria of building up the classification as well as clarified all the details about the classification of picture prompts. After making sure of all teachers’ comprehension of the classification, the researcher offered them picture prompts (For details, see Appendix C) in random, asking them to categorize these picture prompts based on the classification. These three teachers, having scanned and reviewed these picture prompts, clearly and precisely fixed most of these prompts into proper categories. However, they had met with some trouble when classifying certain picture prompts, especially in the cause-and-effect types with or without an expected consequence. First, all of them mentioned that picture prompts, especially those with words in the picture content, should be of a better overall quality, making the understanding of the picture prompts direct and easy. Take Figure 3.3 for example. In the first picture frame, there are two signs on the wall, one saying, “Danger,” and the other saying, “Mountain Inn.” Because of the poor printing quality, the words aren’t clear, which hampers the understanding of the subsequent picture frames. All of these three teachers felt confused, wondering whether the character, regardless of the danger,. 34.

(35) insists on going mountain climbing, or he is required to kill the tiger as a skilled hunter. Different initial events may make a big difference in the interpretation of the outcome of the picture sequence. If the character is a common mountain climber, it would be a surprising ending that he should kill the tiger bravely. However, if the character is an expert, promising to deal with this wild dangerous predator, then it would be within expectation that he kills the tiger.. Figure 3.3 Picture Prompt “The Hiker and the Tiger”. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman.. Second, though there is a flow of actions depicted in the picture sequence, it still takes readers’ imagination to make the flow of actions coherent and logical. A picture sequence, which leaves large space for readers’ imagination, may lead to different interpretations, probably owing to the different personal experience or background knowledge. For example, in Figure 3.4, a lady finally gets her bag back with the help. 35.

(36) of the cab driver. However, it is left unknown how the cab driver is aware of the lady’s eagerness. If she calls for help, the cab driver’s action is expectable; still, if the cab driver voluntarily sends the bag back, the outcome should be out of the lady’s expectation. In fact, the space for imagination between the second and the third picture frame made the three teachers hold different ideas about its category. Two fixed it into the cause-and-effect type with an expected consequence while one the cause-and-effect type with an unexpected consequence.. Figure 3.4 Picture Prompt “Gratitude to an Honest Driver”. Note. Adapted from Creative & Guided Composition, by Siew, T., 2004, Singapore: Asian Publications.. Interestingly, a teacher regarded a novel way to solve a problem as unexpected. The novel way, in his opinion, was out of the readers’ expectation. In Figure 3.5, a fairy goddess comes to the poor girl’s rescue, using magic power to get her out of trouble. The witchcraft is beyond the teacher’s expectation; therefore, he fixed the 36.

(37) picture into the cause-and-effect type with an unexpected consequence. Yet, with the researcher’s explanation that it is the consequence of the event, not the solution to the problem, that makes a difference in the picture prompt type, the teacher comprehended the rationale more and re-classified the picture into the cause-and-effect type with an expected consequence type.. Figure 3.5 Picture Prompt “The Fairy Godmother”. Note. Adapted from Creative Writing Book 2, by Newaskar, M., 2005, Singapore: Asian Publications.. Besides, another teacher viewed the first picture frame as the cause and the last as the outcome, neglecting the effect of the second picture frame. For example, Figure 3.6 depicts how a horrible dinosaur makes himself accepted. There is a transit of emotion from fear of the dinosaur (Picture Frame 1) to friendliness to the dinosaur (Picture Frame 2). The teacher conceived that the result was beyond the characters’ initial thought. However, as the researcher explained that the second picture offered a 37.

(38) cause of the transit, making the outcome logical and expected, the teacher changed her mind and re-fixed the picture prompt into the cause-and-effect type with an expected consequence.. Figure 3.6 Picture Prompt “A friendly Dinosaur”. Note. Adapted from Creative Writing Book 2, by Newaskar, M., 2005, Singapore: Asian Publications.. In addition to the discussion about the cause-and-effect types, there are two other pictures arousing the researcher’s attention. One is a picture describing raising a pet bird (see Figure 3.7). This picture was fixed into the topic-development type by two teachers while one teacher, with rich imagination, linked these three frames with the causal relation, making this picture a cause-and-effect type with an expected consequence. She believed the fruit made the seemingly unhappy bird start to chirp.. 38.

(39) Figure 3.7 Picture Prompt “The New Pet”. Note. Adapted from Creative Writing Book 1, by Newaskar, M., 2005, Singapore: Asian Publications. The other picture, describing the splash on a rainy day, aroused a heated discussion (see Figure 3.8). One teacher linked these frames with the passage of time; another teacher noticed the central topic in the theme; and the other teacher mentioned the surprise ending. The researcher gave her explanation that first, there is a time marker－rain; second, these three frames can not be re-arranged in a different order, violating the feature of the topic-development type; and third, even if there is a surprise ending, no cause-effect plot exists. All of these teachers accepted the account and re-fixed the picture into the time-based type. Figure 3.8 Picture Prompt “Splash”. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman. 39.

(40) With this validation, the classification of picture prompts was reconfirmed. 3.1.3. The Classification of Three-Frame Picture Prompts 1.. The Sequence Type: 1.1. The Time-Based Type: Each part of the picture sequence within a theme is developed with the. passing of time. The lapse of time can be defined by a chronological marker－an image in the picture or a situational inference such as the sun and the moon or waking up and going to bed, etc. Exposed to the time-based type, writers are inclined to chronologically organize a narrative into a sequence of actions. For example, the time-based type task (see Figure 3.9) requires students to describe the character preparing for school, including getting up, brushing his teeth, washing his face, putting on his uniform, combing hair, and then walking to school.. Figure 3.9 Sample of the Time-Based Type. Note. Adapted from Creative Writing Book 1, by Newaskar, M., 2005, Singapore: Asian Publications. 40.

(41) 1.2.. The Steps-in-a-Process Type:. This type focuses on the ordering of steps to carry out a goal. Exposed to this type, writers are inclined to detail process sequences for telling how something is done or how to do something as procedures in instructions, recipes, or manuals. For example, the steps-in-a-process type task (see Figure 3.10) requires students to describe the process of making tea: to place a kettle filled with water on the stove to heat the water, to scoop the tea leaves into a teapot, and then to pour the hot water from the kettle into the teapot.. Figure 3.10 Sample of the Steps-in-a-Process Type. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman.. 2.. The Topic-Development Type: A theme dominates each frame of the picture sequence which tells about or. explains the central topic. There is neither plot nor chronological sequence in this type. In other words, all parts are related to the shared central topic idea and evolve around it, sharing no other relationship with one another. Furthermore, it. 41.

(42) makes no difference in the picture sequence if the order of the three picture frames is re-arranged. For example, the topic-development type task (see Figure 3.11) requires students to describe a useful animal－the cow. First, the cow, which appears to be gentle and obedient, feeds on grass. Second, the cow is used to pull the plough or carts. Third, the cow produces milk, which comes to human beings in many forms: fresh milk, condensed milk, milk powder, cheese, or butter.. Figure 3.11 Sample of the Topic-Development Type. Note. Adapted from Creative & Guided Composition, by Siew, T., 2004, Singapore: Asian Publications.. 3.. The Cause-and-Effect Type: This type of picture sequence has a cause-effect plot. Events or happenings. of the previous parts cause some events or outcomes to occur in the following parts. There could be many causes leading to one effect or there is possibly one major event resulting in outcomes and additional causal/sequential events. In other cases, there is a problem-solution sequence embedded in the event 42.

(43) sequences. The problem is explicitly depicted in the previous parts of the picture sequence. What follows is the protagonist’s way to solve the problem and the consequence or the reaction to solution attainment and endings. Exposed to the cause-and-effect type, writers are inclined to causally organize a narrative into a sequence. 3.1 The Cause-and-Effect Type with an Expected Consequence: There is a cause-effect sequence embedded in the event sequences. What follows the cause is an expected consequence, reaction or ending. For example, the cause-and-effect type with an expected consequence task (see Figure 3.12) requires students to describe how a hat salesman successfully takes back his hats from naughty monkeys. Coming to realize the monkeys copy every move he makes, the old hat salesman throws his own hat on the ground. All the monkeys imitate him, and he takes all of his hats back as a result.. Figure 3.12 Sample of the Cause-and-Effect Type with an Expected Consequence. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman.. 43.

(44) 3.2. Cause-and-Effect Type with an Unexpected Consequence:. There is a cause-effect sequence embedded in the event sequences. However, the effect is an event or reaction out of the expectation of the readers, namely, a surprise ending. Or there is a problem-solution sequence embedded in the event sequences. The problem is explicitly depicted in the previous parts of the picture sequence, initiating the solution to the problem. Nevertheless, what the solution results in is beyond either the characters’ or the readers’ expectation. In this type, the readers may review the story from a different perspective. For example, the cause-and-effect type with an unexpected consequence task (see Figure 3.13) requires students to describe one fishing experience－a mischievous swimmer attaches a battered boot to an old fisherman’s hook, making the old fisherman shocked and disappointed.. Figure 3.13 Sample of the Cause-and-Effect Type with an Unexpected Consequence. Note. Adapted from Composition through Pictures, by Heaton, J. B., 2006, London: Longman.. 44.

(45) These five picture sequences the researcher singles out are closely-related within frames and have been precisely and correctly fixed by all the three validation teachers. Moreover, different types though these five chosen three-frame picture prompts are, they share other qualities, including potentially used vocabulary in the same word level, everyday life topics which are familiar to senior high school students, and overall quality of each picture frames. 3.2. Subjects Paragraphs were written by a senior high senior class of 45 students in Taichung First Senior High using five different types of picture sequences: the time-based type, the steps-in-a-process type, the topic-development type, the cause-and-effect type with an expected consequence, and the cause-and-effect type with an unexpected consequence. Subjects were assigned to write paragraphs with the aid of every type of the above-mentioned picture sequences. This study was conducted during the first semester of the academic year 2007 in Taichung First Senior High School. The researcher selected a senior class of 45 students as the subjects. The reason why the senior class was chosen is that seniors were in the face of the approaching SAT in the coming year and therefore would take a serious attitude toward picture writing, a required part in the SAT. As one of natural-science oriented classes (自然組) in TCFSH, this class has seven hours of. 45.

(46) English per week, five for textbook teaching and two for English composition training. 3.3. Procedure of Data Collection The writing samples were collected during the regular school day through English classes. That is, the task was designed as an integral part of the regular class activity, and all students in regular English classes participated. The exercise was used to give students writing practice and to provide feedback to the teacher and students. As an incentive, students were told that they would receive credit for completing each task to the best of their ability. To make subjects familiar with the writing process, the researcher would prepare seven picture prompts of five different types for the subjects. The seven picture prompts also allowed the researcher to single out the ones eliciting writing samples of the same word level with a view to excluding its potential interference in the result of this study. The same instruction was followed each time a writing sample was going to be collected. Subjects were given 30 minutes to write from each picture prompt. In addition to the picture prompts, a single sheet of lined paper with the instructions, “Write a paragraph about this picture sequence. You will have 30 minutes to complete this assignment. Do not use any reference materials.” was used to collect the writing. 46.

(47) samples. Moreover, the researcher as well as the English teacher read the instruction from a prepared script, making sure that students would read the directions carefully, would be aware of the 30-minute time limit, and would not use any reference materials or other resources to complete the task. Subjects were then told when to begin writing and were given a 5-minute warning at the end of 25 minutes. At the end of 30 minutes, the subjects were instructed to stop writing and hand in their writing samples. The writing samples were collected in seven consecutive weeks with each subject writing every week a composition based on the five different picture prompts. And the whole schedule of the research project is listed in Table 3.1.. Table 3.1 Schedule of the Research Project Week. Procedure. Week 1. The Time-Based Type (Going to School). Week 2. The Steps-in-a-Process Type (Making Tea). Week 3. The Cause-and-Effect Type with an Unexpected Consequence (Fishing). Week 4. The Cause-and-Effect Type with an Unexpected Consequence (A Smart Dog). Week 5. The Topic-Development Type (Cow). Week 6. The Cause-and-Effect Type with an Unexpected Consequence (Playing Table Tennis). Week 7. The Cause-and-Effect Type with an Expected Consequence (Hat Salesman). Note. The three-frame picture prompts used are as mentioned in 3.1.3.. 47.

(48) 3.4. Data Analysis 3.4.1. Instruments The instruments used for dada analysis of the different types of picture sequence are Web Vocabulary Profiler (Web VP) and Microsoft Word. The Web VP is applied to obtain the results of the word level, the text length, and the type token ratio of each writing sample; Microsoft Word is used for the counting of the T-unit. 3.4.1.1.. Web Vocabulary Profiler. The Web VP, from the Compleat Lexical Tutor at http://www.lextutor.ca/, is an on-line computer program that performs lexical text analysis. It is an adaptation from Heatley and Nation’s (1994) Lexical Frequency Profiler by Tom Cobb, UQAM Canada. Either a text with less than 2000 words can be typed or pasted, or a larger text file can be uploaded into the Web VP (see Figure 3.14). With a click on the “Submit_window” check box, the program can immediately analyze the target text and show the frequency profile as an output (see Figure 3.15). The frequency profile includes data of word families, types (different words), tokens (words in a text), and percentages for the four word frequency categories. In minute detail, the percentage of function words and content words in the K1 word category is offered. Besides, more data is displayed about words in text (tokens), different words (types), type-token ratio, tokens per type, and lexical density (content. 48.

(49) words/total) for the whole text. This relatively simple tool has been useful in understanding the lexical acquisition and performance of second language learners. It has been used by researchers like Laufer & Nation (1995) in Vocabulary size & use: Lexical richness in L2 written productions; Meara, P. & Fitzpatrick, T. (2000) in Lex30: An improved method of assessing productive vocabulary in an L2.. 49.

(50) Figure 3.14 Sample of the Web VP Input (Adopted from Compleat Lexical Tutor, by T. Cobb, n.d., http://www.lextutor.ca/vp/eng/).. Figure 3.15 Sample of the Web VP Output (Adopted from Compleat Lexical Tutor, by T. Cobb, n.d., http://www.lextutor.ca/vp/eng/output.pl?).. 50.

(51) 3.4.1.2.. Microsoft Word. When Microsoft Word finishes checking spelling and grammar, it can display information about the reading level of the document. The readability statistics includes the counting of words, letters, sentences and paragraphs for the whole text; the average number of sentences per paragraph, words per sentence and letters per word; Flesch Reading Ease score and Flesch-Kincaid Grade Level score (see Figure 3.16). To derive the data about the T-unit, the researcher reduces each sentence in writing samples into the smallest unit and ends each unit with a dot, making the unit an independent sentence. With the modified texts and the employment of readability statistics in Microsoft Word, the figure shown as the counting of sentences is the number of T-units and the average number of words per sentence is the number of words per T-unit, i.e., the mean length of T-units.. 51.

(52) Figure 3.16 Sample of the Readability Statistics of Microsoft Word. With these two tools, specific word distributions of the writing samples were worked out and the results from Web VP and Microsoft Word were provided for further discussion. 3.4.2. Data Analysis To obtain objective evidence, the researcher decided to turn to quantitative measures. Each writing sample was scored using three different measures: the length of writing products, lexical density (the type-token ratio), and the mean length of T-units. Larsen-Freeman & Strom (1977) and Reid (1990) endorse this measure of writing by using the length of writing products. As to lexical density, Retherford 52.

(53) (2000) cites Miller (1981; 1991) and acknowledges that the values compared when determining the type-token ratio, the total number of words in a specified language sample and the total number of different words in the same language sample, are most valuable for evaluating the appropriateness of the child’s vocabulary development. As for the T-unit, the mean length of T-units has been considered to be among the most objective ways in examining writing data (Hunt, 1965; Gaies, 1980). Originally, the T-unit was designed to assess development in first language acquisition; however, it has been widely adopted for analysis in second language acquisition. Wolfe-Quintero et al. (1998), reviewing across thirty-nine L2 writing studies, concludes that the mean length of T-units is one of the most satisfactory measures. To sum up, these three measures, the length of writing products, lexical density (the type-token ratio), and the mean length of T-units, can be utilized to assess fluency, vocabulary diversity, and linguistic complexity of sentences respectively (see Table 3.2).. Table 3.2 Research Items and Measures Items for Examination Research Measure Word Range. The Length of Writing Products. Lexical Richness. Lexical Density (The Type-token Ratio). Syntactic Complexity. The Mean Length of T-units. 3.4.2.1. Word Frequency. 53.

(54) Before examining the data with the three above-mentioned research measures, the researcher had to exclude possible interference of the factor of word frequency. To be more specific, the researcher had to make sure that each three-frame picture sequence used in the present study might elicit vocabulary of the same word level. In this way, it would be not the vocabulary elicited but the different picture types that make a difference in the result of the research. With regard to the frequency counts of vocabulary, according to Nation & Waring (1997), 2000 high-frequency word families account for about 80% coverage of all written texts. Besides, it is about 2000 words that are used by most native speakers in their daily conversation and used in the dictionary’s definitions (Thornbury, 2002). Therefore, to count the word frequency of the writing samples, the researcher. decided. to. focus. on. the. 2000. high-frequency. words.. As. undergraduates-to-be, the subjects were supposed to have mastered the 2000 high-frequency words. In Web VP, which the researcher applied for the result of word frequency, words are divided into four categories by frequency: (a) the most frequent 1000 words of English, (b) the second most frequent thousand words of English, i.e., 1001 to 2000, (c) the academic words of English (the AWL, 550 words that are frequent in academic texts across subjects), and (d) the remainder which are not found on the other lists.. 54.

(55) After pasting texts onto the Web VP, the researcher collected the results of the two categories (a) and (b). The researcher then conducted one-way ANOVA to see whether there would be any significant difference in the word frequency of different picture types of writing samples. From the seven different picture prompts: the Time-Based Type (Going to School), the Steps-in-a-Process Type (Making Tea), the Cause-and-Effect Type with an Expected Consequence (Fishing), the Cause-and-Effect Type with an Unexpected Consequence. (A Smart. Dog),. The. Topic-Development. Type. (Cow),. the. Cause-and-Effect Type with an Unexpected Consequence (Playing Table Tennis), and the Cause-and-Effect Type with an Unexpected Consequence (Hat Salesman), the researcher collected totally 315 writing samples. Comparisons then were made across samples elicited from these seven different picture prompts for the word frequency (2000 high-frequency words). Table 3.3 displays the significance of the difference in word frequency. Except the Cause-and-Effect Type with an Unexpected Consequence 2 (CEUE 2) and the Cause-and-Effect Type with an Unexpected Consequence 3 (CEUE 3), there is no significant difference in the writing samples of the other five, in which 2000 high frequency words account for from 93% to 94%. In other words, the rest five picture prompts elicit writing samples of the same word frequency.. 55.

(56) After excluding the possible interference of the factor of word frequency in the study, the researcher derived the following five as her target research samples: the time-based type (TB), the steps-in-a-process type (SP), the topic-development type (TD), the cause-and-effect type with an expected consequence (CEE), and the cause-and-effect type with an unexpected consequence 1 (CEUE 1), each of which respectively represents the different picture type in the researcher’s classification.. Table 3.3 The Significant Difference in Word Frequency (Student-Newman-Keulsa) Picture Type. N. Subset for alpha = .05 1. 2. CEUE 3. 45. 90.5362. CEUE 2. 45. 91.0189. SP. 45. 93.0547. CEUE 1. 45. 93.1118. TD. 45. 93.2511. TB. 45. 93.7280. CEE. 45. 94.6656. Sig.. .432. .068. Note. Means for groups in homogeneous subsets are displayed; a Using Harmonic Mean Sample Size = 45.000; TB = The Time-Based Type; SP = The Steps-in-a-Process Type; CEUE = The Cause-and-Effect Type with an Unexpected Consequence; TD = The Topic-Development Type; CEE = The Cause-and-Effect Type with an Expected Consequence.. 3.4.2.2.. Analysis of Word Range: The Length of Writing Products. The number of English words in each sample is used to measure fluency. Both. 56.

(57) Koda (1993) and Henry (1996) use the length of writing products in their foreign language writing studies as an indicator of fluency. The length of writing products is also one of the evaluation methods in Way et al.’s study (2000) to assess the fluency of the writing samples from novice learners of French. 3.4.2.3.. Analysis of Lexical Richness: Lexical Density. Lexical density refers to the proportion of new and repeated words in a text. A text which has a low lexical density will have a relatively small number of different words which are often repeated. A text with a high lexical density will use a great number of different words. Lexical density is also referred to as the semantic range of a text or writer. A common tool to measure lexical density is the ratio of different words (types) to the total number of words (tokens). This is known as the type-token ratio. The definition is as follows:. Theoretically, a high type-token ratio means that the text involves a wider range of vocabulary, while a low type-token ratio means the text contains a more restricted set of vocabulary items (Baker, 2000). It is noteworthy that type-token ratios are not comparable when calculated for different sample sizes. Based on the results of Hess, Sefton, & Landry’s study (1986), 57.

(58) with samples that differ in number of words, type-token ratio measures on the language of young children should not be compared. Further, type-token ratio measures for sample size of 50 and 100 words show reliabilities that are judged to be inadequate for research or clinical purposes. Greater reliabilities require larger word-sample sizes. Many researchers, however, have mistakenly assumed the ratio to be constant over a given text (Chipere, Malvern, Richards, & Duran, 2001). Richards (1987) shows that the ratio is closely related to the text length. To be more specific, larger samples will give lower values for type-token ratios and smaller samples will give higher values for type-token ratios. Several researchers have long been aware of the problem and have tried to correct it, and the solutions have taken the form of controls on the text length or transformations of the type-token ratio (Chipere, Malvern, Richards, & Duran, 2001). In this study, the researcher would use ANCOVA (analysis of covariance) to test whether the factor (the type-token ratio) has an effect after removing the variance (the length of writing products) on the writing samples of different picture types. Since ANCOVA is a regression-based method, the relationship of the dependent variable to the independent variable must be linear in the parameters, as in regression analysis. Therefore, in order to derive an error-free result, the researcher had to test whether. 58.

(59) there was a linear relationship to make sure of the need to remove the variance (the length of writing products). 3.4.2.4.. Analysis of Syntactic Complexity: The Mean Length of T-units. In Dictionary of Language Teaching & Applied Linguistics (Richards et al., 1992), the T-unit is defined as follows:. “…the shortest unit (the Terminable Unit, Minimal Terminal Unit, or T-unit) which a sentence can be reduced to, and consisting of one independent clause together with whatever dependent clauses are attached to it. For example, the sentence After she had eaten, Kim went to bed would be described as containing one T-unit…Compound sentences contain two or more T-units.” (Richards et al., 1992, p. 488). In other words, the T-unit is the shortest unit which can stand alone as a sentence. The length of a sentence can then be measured by counting the number of words per T-unit. Hunt (1965) found that the mean length of T-units, which is obtained by dividing the total number of words by the number of T-units in the writing sample, is a reliable measure of syntactic complexity and that gradual lengthening of T-units occurs from grade to grade. One who can produce more words per T-unit has a higher language proficiency than those with fewer words per T-unit. The definition of the T-unit is as follows:. 59.

(60) total number of words Mean Length of T-units =. number of T - units In order to compare the different effects of these five kinds of pictures, scores for each writing sample, including the length of writing products, the type-token ratio and the mean length of T-units, were entered into the SPSS12.0 (SPSS Inc., Chicago, IL, USA). With three mean scores (the length of writing products, the type-token ratio and the mean length of T-units) for each picture type (the time-based type, the steps-in-a-process type, the topic-development type, the cause-and-effect type with an expected consequence, and the cause-and-effect type with an unexpected consequence), data analysis were conducted by the utilization of one-way ANOVA and ANCOVA to see whether the results were significant or not.. 60.