ࡻᄬିጯ English Teaching & Learning 32. 2 (Summer 2008): 77-121
Questioning the Author: Effects on Recall,
Inference Generation and Responses
to Questions by EFL Junior High School Students
Yu-chen Liu Hsi-chin Janet Chu
Ming-De Junior High School National Taiwan Normal Universityregina2002@tp.edu.tw hcchu@ntnu.edu.tw Abstract
This study investigated the effects of Questioning the Author (QtA) on the reading comprehension of junior high school students in Taiwan. The participants were sixty-two ninth graders from two classes in a junior high school in Taipei City. The experimental group consisted of 31 students from one class, and the control group was comprised of 31 students in another class. The experimental group received instruction in QtA. For the control group, the instructor utilized a traditional lecture approach.
The intervention was done in two periods per week for five consecutive weeks. Two measures, written recall and open-ended questions, were used in the pretest and posttest to assess three dimensions of comprehension—recall, inference generation in recall and responses to Open-ended Questions. Written recall was used to measure the effect of QtA lessons on the memory of text and on three levels of inference generation: reader-based, text-based and incorrect inferences. The open-ended questions were employed to assess the effects of QtA lessons on students’ answers to factual, interpretive or responsive questions.
ANCOVA analyses on the three dimensions of the measures show that: (1) there was no significant difference between groups on written recall, (2) the experimental group produced significantly more reader-based inferences and (3) the experimental group performed significantly better than the control group on interpretive questions. Key words: Questioning the Author, recall, inference,
78
INTRODUCTION
In traditional reading classrooms, it is commonly observed that teachers dominate the majority of classroom discourse. Most of the questions asked by the teacher are for retrieving facts (Durkin, 1979; Graesser, Pearson, & Hu, 2002; Hyman, 1982), so students remain passive when reading for information. Scarcely are students encouraged to actively engage in interaction with the text and the author, not to mention “to think critically about the underlying messages” (McDaniel, 2004, p. 8). Students respond only when asked or required. Since the scope of their answers is limited to facts, the strictly text-bound responses might prevent students from constructing their own meaning, and shape passive readers waiting for questions to answer.
The gap between teaching and learning has prompted researchers to develop innovative teaching approaches (Almasi, 2003; Baker, Dreher, & Guthrie, 2000; Block & Pressley, 2002; Pressley, 2002; Sweet & Snow, 2003) to narrow the gap. Thus, instructional procedures that guide students to become able readers (Pressley, 2000; Tierney, Readence, & Dishner, 1990), to interact with authors (Beck, McKeown, Hamilton, & Kucan, 1997; Lewin, 2004; Ogle & Blachowicz, 2002; Rosenblatt, 1993) and to take pleasure in English reading (Rosenblatt, 1993) have been proposed.
One of these instructional procedures is Questioning the Author (from here on, QtA) (Beck & McKeown, 2001, 2002; Beck, et al., 1997; Beck, McKeown, Sandora, Kucan, & Worthy, 1996; McKeown & Beck, 2004; McKeown, Beck, Hamilton, & Kucan, 1999;
Liu & Chu: Effects of Questioning the Author
79 McKeown, Beck, & Worthy, 1993; Sandora, Beck, & McKeown, 1999). It is a teaching strategy advocated by teacher-researchers who found frustration in their teaching practice when students did not learn as expected. The teachers studied the phenomena and concluded that students failed because they lacked interaction with the text (Beck et al., 1997). QtA was thus proposed to respond to the problem. QtA is characterized by requiring teachers to segment the text, plan queries, and use discussion moves in class. Teachers help students “grapple with the ideas” (Beck et al., 1997, p. 6) in each segment and design queries to invite students’ interpretation (Beck et al., 1997). Instead of presenting the ideas to students directly, teachers work out the meaning of the text collaboratively with students through discussion moves. Moreover, when participating in the discussion, students take the initiative in presenting discussion issues. Thus, students take more responsibility in the reading process while teachers play the role of facilitators that guide students to go beyond the text.
A series of studies have been conducted by research groups led by Beck (Beck & McKeown 2001; Beck, et al., 1996, 1997; McKeown & Beck, 2004; Sandora, et al., 1999), the results of which showed the positive effects of QtA on the amount of student talk, student-initiated questions, student-to-student interactions, as well as on recall and open-ended questions generated by L1 readers. Despite the favorable results, it is not known whether the effects would be observed with EFL readers, specifically junior high school students in Taiwan. For these students, their language proficiency and their study habits, which have been shaped by an examination-oriented context, might or might not be factors influencing the feasibility and outcome
80
of the QtA approach.
This study was therefore conducted to investigate the effect of QtA on the reading comprehension of junior high school students in Taiwan, in terms of memory of text, inference generation and three types of comprehension (factual, interpretive and responsive). The research questions are:
(1) Do QtA lessons make a difference in students’ memory of texts as indicated by recall?
(2) Do QtA lessons make a difference in students’ inference generation in terms of text-based inferences, reader-based inferences, and incorrect inferences?
(3) Do QtA lessons make a difference in students’ responses to three levels of open-ended questions?
LITERATURE REVIEW
Theoretical Models of ReadingReading comprehension is a complicated mental process. Readers extract meaning through establishing a relationship between print, sound and meaning, and construct meaning through integrating new information with old knowledge. Current reading models generally describe the two phases of reading with different typologies. The construction-integration model (Kintsch, 1998) includes a textbase model and situation model. The transactional model (Rosenblatt, 1993) also depicts two levels of reading—efferent and aesthetic reading. These models point to a reading potential, i.e.,
Liu & Chu: Effects of Questioning the Author
81 reading can go beyond mere reproduction of textual messages. Readers need to bring in their knowledge and experiences to create multiple meanings from the text.
Construction-integration model. In the construction-integration model, comprehension begins with disorganized ideas. These ideas go through a process of integration, become stabilized, and accomplish well-structured coherence. Specifically, in the comprehension process, a mental representation of a text can be described in two phases: the textbase and the text situation. The textbase is composed of the “elements and relations that are directly derived from the text itself” (Kintsch, 1998, p. 103) and includes microstructure and macrostructure. The microstructure represents idea units (propositions) at the local level of the text while the macrostructure refers to the hierarchically structured organization of the text at the global level. The surface structure of the text is the element of the textbase.
The second phase is a text situation. Constructing a text situation may require a variety of knowledge elements, for example, a causal link inferred between implicitly stated propositions, or elaboration implied in the images in the text. A situation model is “a construction that integrates the textbase and relevant aspects of the comprehender’s knowledge” (Kintsch, 1998, p. 107). Due to differences among readers and contexts, the elaborations of the text may depend on the text and on the readers; therefore, a situation varies.
The situation model is the foundation of text interpretation because all knowledge implicitly and explicitly reaches a coherent state in a situation model. To reach full text understanding, mere
82
textbase construction is inadequate because readers need to integrate the text propositions with their own knowledge or personal experience for deep and multiple understandings, that is, constructing a situation model.
In current junior high school classrooms in Taiwan, the approach to reading through word decoding, sentence structure analysis, translation and teacher-directed questioning does not seem to boost the active construction of a textbase, not to mention to reach the level of situational understanding. Hence, an alternative approach that would facilitate active involvement in meaning construction through text-knowledge integration is called for. An approach which involves teacher questioning, with the goal of tapping students’ knowledge and experience to interact with the author, rather than simply addressing textual facts, might help in the understanding of the text situation and in the generation of inferences. Thus, the teaching model, QtA lessons, might be such an alternative.
Rosenblatt’s transactional model. Another model is Rosenblatt’s transactional model. In this model, the reader achieves comprehension in the transaction between the text and the reader. While reading the text, each individual reader brings his/her own “personal linguistic-experiential reservoir, the residue of past transaction in life and language” (Rosenblatt, 1993, p. 381) to the dynamic transaction for meaning construction. Since everyone is unique with his/her own special experience, reading is made from the author’s written form to the reader’s perception.
The transactional theory consists of two readings: efferent and aesthetic (Rosenblatt, 1978, 1993). The efferent reading focuses on
Liu & Chu: Effects of Questioning the Author
83 the intellect. Readers acquire information that the text reveals and they organize and retain it in the memory. The language aspect and the knowledge domain are what readers need to pay attention to, remember, integrate, evaluate and apply after the reading activity. On the other hand, the aesthetic reading centers on the emotion and personal experience that have been ignited during reading. In the reading process, readers make use of their stored knowledge and show reflection toward the text or respond to it based on individual experience. The transacting process is a continuum between these two readings. One text might present two different kinds of reading and elicit various interpretations. Deciding which stance to take upon reading requires readers’ selectivity.
This selectivity of stance forms the core of reader-response theory (Rosenblatt, 1983) which suggests that readers need to be critical of the images that texts reveal and of the premises that they bring to the transaction, and that being critical is being selective. Readers with a critical eye know how to select among possible interpretations and decide the proper one to complete the transaction (Rosenblatt, 1993).
To be a critical reader, thus, involves choice by the reader to go beyond the author’s point of view. Hence, to encourage students to adopt a more aesthetic standpoint and became critical readers, reading classrooms need to help students connect the text with the personal and value the individual voice. Yet, in the current junior high classrooms in Taiwan, this type of reading is seldom an objective. An examination into test questions in general would reveal that students are not expected to go beyond the text level and establish personal
84
significance for reading.
It is therefore proposed that to shape critical readers in the classroom, students need to go beyond the author’s point of view to initiate meaning making. Students need to take not just an efferent stance by organizing and integrating information, but also an aesthetic stance by bringing in individual experience to interact with the text and the author. A teaching model, QtA lessons, is thus proposed to promote reader-text transaction.
Questioning the Author—a Teaching Model
To bolster situational reading and to encourage reader-text transactions, it was thus proposed that teachers adopt Questioning the Author (QtA), a teaching strategy, in order to enhance student engagement with text (Beck et al., 1997). Students need to construct their own meaning while reading, rather than accepting whatever the author presents (Ogle & Blachowicz, 2002). Understanding of the text is achieved through discussion in class.
There are three characteristics that distinguish QtA lessons from traditional reading instruction. First, the use of queries is unique in this approach. Queries are questions, asked by the teacher concerning significant or confusing ideas in the text to help students connect text parts and initiate discussion. Personal interpretation, then, is therefore generated. Second, the role of a teacher is different. Unlike traditional classrooms where the teacher is an authority who dominates any speaking, in QtA classrooms, the teacher is a facilitator who provides students with possibilities that involve decisions, discussion, thinking and interpretation. Third, the notion of constructing is highlighted.
Liu & Chu: Effects of Questioning the Author
85 Rather than receiving ideas from a teacher, learners construct their ideas by active use of knowledge. QtA teachers expect students to produce answers that go beyond what was explicitly stated in the text. To implement a QtA lesson, a teacher follows three steps: (1) segment the text based on the presentation of key or problematic ideas, (2) prior to instruction, plan initial and follow-up queries for each segment, and (3) make good use of discussion moves during instruction (Beck et al., 1997).
Previous Studies on QtA
Several studies have been conducted in the L1 context to assess the effects of QtA. The following is an account of these studies.
Beck, et al. (1996) did a yearlong study with two teachers and 23 fourth graders who implemented QtA in the classroom and noted positive results. Before the study, the two teachers underwent a training session to become familiar with how QtA functions. They then implemented the approach in social studies and reading/language arts classes. When the teacher implemented QtA in the classroom, two distinct features were noted. First, the questions asked by the teachers focused on extending meaning, not retrieving information. Second, the responses by the teachers served to extend the conversation in the classroom rather than to evaluate students’ ideas. Data gathered in the study included videotapes of lessons, notes of the meetings with the teachers, observation summaries, each teacher’s journal entries, debriefings and lesson narratives, students’ responses to individual comprehension tasks as pretests and posttests and interviews with students about their views of QtA.
86
The results showed that the amount of student talk doubled and students began asking more questions. In addition, more student-to-student interactions were observed and evidence showed QtA could “facilitate thinking” (Beck, et al., 1996, p. 411).
Another study done by Sandora et al. (1999) compared the effects of the QtA discussion technique and the Great Books approach. These two approaches both give students chances to interact with the text, construct meaning and share ideas with others. However, QtA has discussions during the initial reading while Great Books has discussion after the whole story is read. Although a questioning format is adopted by both approaches, QtA uses initiating and follow-up queries that go beyond textbase construction whereas Great Books applies a set of questions from fact, interpretation to evaluation.
All students in the sixth- and seventh-grade classes of a school were involved in the study. Students read and discussed four stories from the Junior Great Books series. The sixth graders used the QtA approach, and the seventh graders used the Great Books approach. After reading, students were asked to provide a free recall of the story and to respond to open-ended questions. The results revealed that mean scores on both the recalls and the responses to the open-ended questions were higher for the students who participated in the QtA lessons. Further analyses indicated that students in the QtA discussions, although lower in grade level, gave longer recalls and better interpretations than students using the Great Books approach.
Although the QtA approach has been implemented successfully as a research-based cooperative approach over the years in primary
Liu & Chu: Effects of Questioning the Author
87 schools for L1 readers (Beck & McKeown, 2001; Beck et al., 1997; McKeown & Beck, 2004), whether this approach can be applied in an EFL context is unknown. Moreover, although previous studies have shown the advantages of QtA, the effects investigated were shown primarily on factors other than comprehension. In Beck et al.’s study (1996), the outcome measures were students’ frequency of talk and student-to-student interaction during QtA sessions. In Sandora et al.’s study (1999), although recall and open-ended questions were used to measure the improvement in comprehension ability, these measures might not have addressed the potential impact of QtA on comprehension, i.e., comprehension at higher levels—inference in recall and interpreting text.
In addition to memory for text, this study investigated the effects of QtA on the three levels of inference made in recall, which might better capture the effects of QtA at the level of comprehension beyond the textbase. As recall calls for memory for text, the outcome of text-knowledge/experience integration—inferences, will not be gleaned, if the measure focuses on text replication only. The ideas produced beyond text replication might be the outcome of reader-text integration—inference. Besides, previous studies demonstrated that inferences in recall may be classified into three levels, text-based, reader-based and incorrect (Barry & Lazarte, 1998), as a result of text processing at different levels.
We divided open-ended questions into three types to assess the level of understanding achieved. According to Sandora et al. (1999), merely putting all the answers to different types of questions into one category, open-ended questions, might fail to address the most
88
important impact of QtA lessons—the higher level understanding. Therefore, this study teased out the effect of QtA lessons on comprehension by dividing the open-ended questions into three levels, literal, interpretive and responsive, to better capture the different dimensions of comprehension so as to more precisely capture the effects of QtA lessons.
METHOD
ParticipantsTwo intact classes of ninth graders in a junior high school in Taipei City participated in the experiment. One class with 36 students was the experimental group. The other class with 34 students served as the control group. Since the lessons were carried out as part of the regular curriculum, students in intact classes were involved in the project. They were informed that they would be participating in a study to understand their ability to read stories and were told that the assessment was for reference purposes only and would not affect their semester grade. After data collection, five students from the experimental group and three students from the control group were dropped from data analyses because they failed to complete either the pretest or the posttest, thus leaving 31 for each group. The first author was the English teacher of both groups.
Although the two classes of students were not officially measured to determine their English language ability and thus the initial difference cannot be ensured, three factors helped minimize
Liu & Chu: Effects of Questioning the Author
89 variance between groups. First, these students were from the same geographical region in a narrow age range and hence they shared a similar learning history. Second, they were randomly assigned to classes at the outset of the academic year. Finally, the statistical measure, ANCOVA, was used to partial out the initial differences in the pretest.
Treatment Materials Selection and Development
Four phases were involved in the material selection for treatment: deciding the genre of the text to be used, selecting stories for treatment sessions, developing initiating queries and preparing comprehension check questions.
Determining the genre. Two steps were involved. First of all, it was decided that narrative texts be used because they provide information as well an enjoyable experience, and the narrative text is a familiar text type for junior high students and thus would help generate more interpretation. Second, among narrative genres, the fable was chosen because previous studies used fables as treatment materials (Beck et al., 1997). For the sake of comparison, it was decided that fables would be used as the treatment as well as the test material.
Selecting treatment stories. Five stories from the lowest level of
Aesop’s Fables of Easy Readers Series by Caves Books (Olivier, 1999)
were chosen through the following process. First, six stories among the thirty fable stories at this level were chosen because the content contained controversial points for discussion and the lengths were between 500 and 650 words, which was suitable for teaching in two
90
class periods. Next, one out of the six stories was ruled out because its readability, 2.6 was far below the remaining five, which ranged between 4.1 and 4.6, based on the Flesch-Kincaid Grade Level Index provided by Word software. The range falls within the ranges in the students’ two textbooks for their English course, 3.6-7.2 and 3.6-7.3 respectively. (See a sample reading passage in Appendix A).
Developing initiating queries for treatment. Each treatment story was divided into ten segments at the juncture where there might be a key concept or a confusing idea. An Initiating Query was developed for each segment. Initiating Queries were the first queries that the teacher used to elicit students’ responses. Some focused on theme, some on character, and some on the intent of the story. “What is the author trying to say?” is an example of a typical question. Ten Initiating Queries were prepared for each of the stories for the students to follow. (See a sample of the Initiating Queries in Appendix A.)
Following Initiating Queries, the teacher might ask two types of impromptu queries, Follow-up and Narrative Queries, based on students’ responses or interaction. Follow-up queries focused on the important part of the response, helped students to concentrate on the task and elaborated on students’ responses. The teacher might ask, “Why do you think so?” or “Do you agree with what someone said?” An example of a Narrative Query is “How do things look for this character now?” Narrative queries help students be concerned about characters and “the author’s crafting of the plot” (Beck et al., 1997, p. 42).
Developing comprehension check questions. To motivate students to complete the reading and participate in discussion, nine multiple-choice questions, three factual, three interpretive and three
Liu & Chu: Effects of Questioning the Author
91 reactive, plus one open-ended question for each treatment story were designed. They were used to motivate students in engaging in the reading, but the responses elicited were not used in the study.
Data Collection Material and Instruments
To answer the research questions, the researchers developed the reading materials for the pretest and posttest, and prepared two instruments for data collection: recall sheets and Open-ended Question sheets.
Selecting and modifying pretest and posttest materials. Two Indian fables were selected from the website, Indian Fables, “The Foolish Friend” and “The Prince and the Lion,” to assess reading comprehension in terms of written recall, inference generation and three types of comprehension questions.
These two fables were selected for two reasons. First, based on student judgment in a pilot study with 31 ninth graders, all the students rated the content as unfamiliar. This result meets the criterion for test material—that readers should not be exposed to the content before. Second, both stories shared some similarities in the plot: both are about members of a royal family and an animal, with setting in the first paragraph, a crisis for the characters to deal with, a climax, and a solution in the second paragraph.
Originally, the two texts were of different lengths (The Foolish
Friend, 260 words; The Prince and the Lion, 281 words) and different
readability levels (The Foolish Friend, 4.1; The Prince and the Lion, 6). In order to minimize the differences, a few adaptations were made to the materials. First, the moral lessons given at the end of the stories
92
were deleted because they might influence on students’ interpretations. In addition, the names of the main characters and the kingdom, which were Indian names, were replaced by common English names for ease of reading. Third, words that were not in the 2000-word vocabulary list issued by the Ministry of Education for junior high graduates (2003) were replaced with easier and similar expressions so that they would not affect students’ meaning making. Fourth, complex sentence patterns that might cause difficulty were simplified to ease students’ reading burden. After the adaptations, the length of The Foolish
Friend was 240 words and the length of The Prince and the Lion was
236 words, which was similar to the length of the students’ textbook readings; and the difficulty level was the same, 3.4, which was slightly lower than the difficulty level in their textbook readings.
Three experienced English teachers then read the stories and confirmed the suitability of these materials for ninth graders. Furthermore, in the aforementioned pilot test, more than two-thirds of the students felt the length and difficulty level were appropriate. A sample passage is given in Appendix B.
Pretest and posttest written recall sheets. After reading the texts, students were asked to write down, in Chinese, whatever they remembered. The recall was done in Chinese because recall done in the L1 has yielded better outcomes (Lee, 1986), and according to Brisbois (1992), L1 writing should be employed to interpret L2 reading comprehension.
Open-ended Questions for pretest and posttest. Open-ended Questions included the three types of comprehension questions presented in Gray’s (1960) classification of students’ understanding of
Liu & Chu: Effects of Questioning the Author
93 text: whether they read the lines (literal), whether they read between the lines (interpretive), or whether they read beyond the lines (evaluative or responsive). Three questions were asked for each of the three question types: factual questions about facts, events, or causes of the story; interpretive questions about the theme, character development or judgment of the event and responsive questions about students’ personal responses, evaluation of author’s intent and problem-solving. Altogether, there were nine questions for each of the stories. Appendix C presents a sample set of Open-ended Questions. Treatment Procedure
The students who participated in the study had English class five periods per week. The experiment was done for one period on both Wednesday and Thursday, for five consecutive weeks in April and May, 2005. Two periods were used to teach one story and thus one story was taught per week. The following are the treatment procedures for each group.
Experimental group (hereafter Group E). The treatment procedure followed that of the L1 studies conducted by Beck et al. (1997) and Sandora et al. (1999) except that the query was given in a handout as a written reference, instead of being given orally. In a typical lesson on Wednesday, the teacher, the first researcher of the study, first explained the ten queries. Then she read aloud each segment, stopped to ask the query for the segment and led a whole-class discussion based on the query, for about six minutes for each query. Meanwhile, impromptu questions, Follow-up questions or Narrative questions, were asked to explore students’ reflections
94
further. The process continued in a cycle until the sixth segment was completed.
In the second period on Thursday, the teacher reviewed the content of the previous six segments, explained the four queries for the remaining four segments and then read and discussed the remaining four segments. After the completion of the remaining four segments, about four minutes were devoted to overall discussion. The last seven minutes in the second period were used for testing comprehension.
Control group (hereafter Group C). For the control group, except for the last seven minutes for the comprehension check, the two periods in each of the five weeks were used to cover one story. Each story was roughly divided into two parts. For each part, the teacher first explained new words before she read aloud the story for students to repeat. Then, the teacher explained the sentence patterns, grammar points and translated the sentences.
Data Collection
Data collection was done one week before and one week after treatment.
Design for the pretest and posttest. To collect data, this study employed a split-block design for the pretest and posttest to control for the order effect on recall and comprehension. The experimental group and the control group were each divided into two groups, A & B. For the pretest, Group A read the first story “The Foolish Friend” and Group B read the second story “The Prince and the Lion.” As for the posttest, Group A read “The Prince and the Lion” while Group B
Liu & Chu: Effects of Questioning the Author
95 read “The Foolish Friend.”
Data collection procedure. Pretest and posttest procedures were the same. First, students were given a story to read for ten minutes. After the story was collected, they were asked to write a written recall for fifteen minutes. Second, after the recall sheets were collected, students were given a sheet of Open-ended Questions. The students were given twenty minutes to answer the questions.
Scoring
Scoring retention on the written recall. The scoring of the written recall was based on Bernhardt’s (1991) “pausal unit” system. First, two native speakers of English read both stories to themselves to mark all the places in the text where they took a breath. Wherever there was a discrepancy, the narrower unit was chosen. The resultant template for The Foolish Friend contained 41 units and The Prince
and the Lion 40 units. Each unit was given one point. No partial
points were given. The first researcher and an English teacher in the same school scored all the copies (124 copies) against the pausal unit system. The total points recalled for each story were transformed by percentage of the total number of units in the template. The interrater reliability was .98. When the raters had different opinions, consensus was reached through discussion.
Scoring inference in the written recall. The mismatching part of the recall was classified as inferences in terms of three types: text-based inferences, reader-based inferences and incorrect inferences. Text-based inferences have a strong relationship to the text. They are made to fill the gap between the text ideas. Reader-based
96
inferences are inferences for activation of personal knowledge. They show a causal relationship between the text and the reader’s knowledge. Incorrect inferences refer to the wrong interpretation of a reading or the production of misreading (Barry & Lazarte, 1998). The first researcher re-coded all the idea units that did not match the units in the templates. i.e., the units that were not scored as recalled units. A second rater coded one fifth of the copies to yield inter-rater reliability. The inter-rater reliabilities for text-based, reader-based and incorrect inferences were .92, .85 and .96, respectively. When disputes in coding arose, the raters came to a general agreement through discussion.
Scoring of Open-ended Questions. Responses to each question were scored holistically from 0 to 4. As there were nine questions for each story, the total score was 36 for both pretest and posttest respectively. For each type of question in each test, literal, interpretive and responsive, the total score was 12.
An answer that provided no information or irrelevant information was given a zero. One point was given for scanty information. Two points were given for limited information concerning the text. Three points were given for enough information but no elaboration. When the answer showed fluent expression and clear ideas, the highest score, four points, was given. The first researcher and a second rater, an experienced English teacher, scored all the copies blind. The interrater reliabilities for factual, interpretive and responsive questions were .97, .91 and .91, respectively.
Liu & Chu: Effects of Questioning the Author
97 Data Analysis
To assess the effect of QtA lessons on student recall, the mean percentage of pausal units recalled from both groups were compared using ANCOVA, with Group as an independent variable, Pretest Recall as a covariate variable and Posttest Recall as a dependent variable.
To evaluate the effect of QtA lessons on student inference generation, the frequency of three types of inferences—text-based, reader-based, and incorrect inferences—were again analyzed using ANCOVA, with the same independent and covariate variables as the analysis of recall data.
For the effect of QtA lessons on three types of comprehension questions—factual, interpretive and responsive, again three rounds of ANCOCA were employed respectively with the same independent and covariate variables as the analysis of recall data.
FINDINGS
The results of the written recall, inferences and open-ended questions are presented in the following sections.
Effect of QtA Lessons on Recall
As shown in Table 1, in the pretest, Group E obtained a lower mean score (18.92, SD = 23.99) than Group C (21.80, SD = 23). Groups E also had a lower score on the posttest. For the posttest, the mean score of Group E rose to 27.16 (SD = 25.52); the mean of Group C increased to 29.81 (SD = 25.31).
98
Table 1
Mean Scores for Group and Pre-Posttest on Written Recall
Group Pretest Posttest
Group E (n=31) M = 18.92 (SD = 23.99) M = 27.16 (SD = 25.52) Group C (n=31) M = 21.80 (SD = 23) M = 29.81 (SD = 25.31) ANCOVA analyses on the groups showed that there was no significant difference between groups on the posttest, F (1, 59) = .00,
p > .05. Group E did not outperform Group C on written recall after
the treatment. Thus, QtA lessons did not successfully enhance students’ ability to remember the text.
Effects of QtA Lessons on Inference Generation
The effects of QtA lessons on inference generation are reported in terms of text-based, reader-based, and incorrect inferences.
Effect on text-based inferences. For text-based inferences, Group E generated an average of 2.39 units (SD = 2.85) while Group C generated 2.26 units (SD = 2.25) on the pretest. On the posttest, Group E generated a mean of 3.84 units (SD = 3.05) whereas that of Group C was 3.97 (SD = 2.74) (See Table 2).
Liu & Chu: Effects of Questioning the Author
99 Table 2
Mean Frequencies for Group and Pre-Posttest on Text-based Inferences
Group Pretest Posttest
Group E (n=31) M = 2.39 (SD = 2.85) M = 3.84 (SD = 3.05)
Group C (n=31) M = 2.26 (SD = 2.25) M = 3.97 (SD = 2.74)
An ANCOVA analysis on the posttest, with pretest as a covariate, showed that there was no significant difference, F (1, 59) = .08, p > .05. After the five-week training, QtA lessons failed to aid in the generation of text-based inferences.
Effect on Reader-based Inferences. For reader-based inferences, on the pretest, Group E had a mean frequency of zero (SD = .18) and Group C produced a mean frequency of .13 (SD = .34). After treatment, Group E had a mean frequency of .23 (SD = .5) while the mean frequency of Group C declined to zero (SD = .18) (See Table 3).
Table 3
Mean Frequencies for Group and Pre-Posttest on Reader-based Inferences
Group Pretest Posttest
Group E (n=31) M = 0.00 (SD = .18) M = .23 (SD = .50)
100
Among the sixty-two participants, fifty of them did not generate any reader-based inferences in both the pretest and posttest. Because such a large number of zero frequencies was found, we decided to first examine whether there was any significant difference between the two groups on the pretest, instead of directly conducting an ANCOVA analysis. However, instead of using the number of frequencies of generated reader-based inferences, we used the indicator of whether the participants actually generated any reader-based inferences at all (1 = “yes” and 0 = “no”) on their pretest. A chi-square test (p = .40), with the association measure phi coefficient Φ = -.10, showed that the two groups were not significantly different in whether or not they generated reader-based inferences on the pretest. Therefore, next we can simply compare the posttest of the two groups.
Again because there was a large number of zero and small frequencies on the posttest and a standard independent t-test might not be suitable, the Mann-Whitney test was employed to test the effect of QtA lessons on reader-based inferences. Table 4 displays the result of reader-based inferences and shows that, with a Mann-Whitney U statistic of 402.50 (p =.045 < .05), a significant difference was found between the two groups. In other words, QtA lessons were able to assist participants to effectuate more reader-based inferences.
Liu & Chu: Effects of Questioning the Author
101 Table 4
Summary of Reader-based Inferences
Reader-based inferences
Mann-Whitney U 402.50
Z -2.00
Asymp. Sig. (2-tailed) .045 *
* p < .05.
Effect on incorrect inferences. On the pretest, the mean frequency of Group E was 2.16 (SD = 2.22) while that of Group C was 3.39 (SD = 5.54). On the posttest, Group E made an average of 2.29 inferences (SD = 2.76) whereas Group C had 2.52 (SD = 3.42) (see Table 5).
Table 5
Mean Frequencies for Group and Pre-Posttest on Incorrect Inferences
Group Pretest Posttest
Group E (n = 31) M = 2.16 (SD = 2.22) M = 2.29 (SD = 2.76)
Group C (n = 31) M = 3.39 (SD = 5.54) M = 2.52 (SD = 3.42)
ANCOVA was also conducted to determine the effect of the intervention. No significant difference was observed, F (1, 59) = .31,
p > .05. The treatment made no difference in making incorrect
102
In summary, for inference generation, QtA lessons enhanced the ability to develop more reader-based inferences. The lessons did not show any effect on text-based inferences, nor on incorrect inferences. Effects of QtA Lessons on Open-ended Questions
The effects of QtA lessons on open-ended questions are presented in terms of three types of questions: factual, interpretive, and responsive questions.
Effect on factual questions. As Table 6 shows, on the pretest, the mean score of factual questions was 6.65 (SD = 3.43) for Group E and 6.52 (SD = 3.83) for Group C. The posttest means for Group E and Group C were 7.32 (SD = 3.24) and 7.38 (SD = 3.69), respectively.
Table 6
Mean Scores for Group and Pre-Posttest on Factual Questions
Group Pretest Posttest
Group E (n=31) M = 6.65 (SD = 3.43) M = 7.32 (SD = 3.24)
Group C (n=31) M = 6.52 (SD = 3.83) M = 7.38 (SD = 3.69)
Again, ANCOVA analysis with pretest scores as a covariate showed there was no significant difference between the groups, F (1, 59) = .01, p > .05. Even though Group E made some progress on the posttest, Group C increased their ability to comprehend the textual message as well. Group E did not outperform Group C in answering factual questions.
Liu & Chu: Effects of Questioning the Author
103 Effect on interpretive questions. For the pretest, the mean score of the interpretive questions for Group E was 4.19 (SD = 3.08) whereas for Group C it was 5.58 (SD = 3.64). On the posttest, the mean scores were 5.74 (SD = 2.56) and 5.48 (SD = 3.12) for Group E and Group C respectively. (see Table 7).
Table 7
Mean Scores for Group and Pre-Posttest on Interpretive Questions
Group Pretest Posttest
Group E (n=31) M = 4.19 (SD = 3.08) M = 5.74 (SD = 2.56)
Group C (n=31) M = 5.58 (SD = 3.64) M = 5.48 (SD = 3.12)
ANCOVA analysis on posttest scores, covariating pretest scores, showed that there was a significant difference between groups, F (1, 59) = 4.94, p < .05. Group E outperformed Group C in answering interpretive questions. QtA lessons successfully prompted participants to construct their own meaning and interpret the text.
Effect on responsive questions. On the pretest, the mean score of the responsive questions for Group E was 5.45 (SD = 2.86) and for Group C, 6.16 (SD = 2.46). On the posttest, the mean scores of Group E and Group C were 6.90 (SD = 2.02) and 6.65 (SD = 2.24) respectively (See table 8).
104
Table 8
Mean Scores for Group and Pre-Posttest on Responsive Questions
Group Pretest Posttest
Group E (n=31) M = 5.45 (SD = 2.86) M = 6.9 (SD = 2.02)
Group C (n=31) M = 6.16 (SD = 2.46) M = 6.65 (SD = 2.24)
ANCOVA results on responsive questions indicate that there was no significant difference between groups, F (1, 59) = 1.48, p > .05. After the intervention, Group E did not show greater change.
In summary, QtA lessons contributed to greater improvement on interpretive questions. However, no improvement was found for literal questions and responsive questions.
DISCUSSION
The present study shows the effects of QtA training on two measures, reader-based inferences and interpretive questions. However, measures of recall, text-based inferences, factual questions and responsive questions did not show an effect. The findings will be examined in terms of recall, inference and response to comprehension questions.
Written Recall
The matching of the idea units in students’ written recalls with those in the original text tapped the students’ memory for text. That
Liu & Chu: Effects of Questioning the Author
105 there’s no difference in text memory between the experimental group and control group shows that QtA lessons facilitate text memory no less than traditional lessons do. The control group focused on language aspects of reading. The close examination of language through explanation and translation of words and sentences might help with lower level processing, and thus the reproduction of text. The QtA approach, on the other hand, focuses more on the transaction of text and hence, the construction of a text situation, rather than a textbase. Given the results, we might well conclude that the QtA approach has the same effect on text memory as traditional lessons, albeit with less emphasis at the text level.
The effect of QtA on recall, although not exceeding that of the traditional approach, did go beyond that of other approaches that emphasize reader-text interaction, such as evidenced in Sandora et al.’s study (1999) which showed that the QtA boosted recall more than the Big Book approach. The difference lies in the fact that QtA activates meaning making during reading while the Big Book approach emphasizes meaning making after reading. Active processing during reading might produce a more robust effect on memory than active processing after reading.
Three Types of Inferences
Of the three types of inferences, QtA impacts on the generation of reader-based inferences. Although the number of participants who demonstrated the difference was very small, the significant difference did show the potential of QtA lessons in promoting reader-text interaction. Specifically, it takes more cognitive effort to make
106
reader-based inferences than text-based inferences due to the reasoning and imagining that is required. It is conjectured that the queries in QtA, especially Narrative queries, might push readers to speculate about the text, and this promotes aesthetic reading (Rosenblatt, 1993). On the other hand, Group C who experienced instruction that focused on language might predominantly focus so much on efferent reading that they sought textual information rather than connecting text with endowed knowledge.
Like written recall, QtA lessons did not make a difference on text-based inferences and incorrect inferences between groups because the experimental group focused more on elements beyond the text and the control group focused more on text decoding. It might well be inferred that QtA lessons emphasizing text-knowledge interaction promote as much text-based inference as the traditional approach focusing on text decoding. In a similar vein, students who had QtA lessons, although spending less time on language than students of the control group, produced no more incorrect inferences than the control group. They were not disadvantaged from less exposure to language factors.
Three Types of Comprehension Questions
It might seem an anomaly that QtA lessons increased reader-based inferences yet not responsive questions, for both measures indicate the linking of text and personal knowledge. However, an examination of the processing involved in the two tasks might illuminate the discrepancy. To produce reader-based inferences, students choose the textual segment to reflect on, and students also
Liu & Chu: Effects of Questioning the Author
107 select whatever aspects of their own personal knowledge they wish to elaborate on. Students have freedom in the thinking process, which is more in line with critical thinking (Rosenblatt, 1983). However, to respond to the teacher-cued responsive questions, students are constrained by the segment chosen by the teacher and the direction of thinking presumed by the questions. This does not seem to reflect the purpose of the QtA process which is to facilitate active thinking. Therefore, we could not see the effect of QtA on responsive questions.
When the experimental group’s cognitive potential was curtailed by the constraint in the teacher cues, perhaps more thoughts/reasoning were put into the responses to the interpretive questions, which were text-based. This might be the reason why the experimental group surpassed the control group in answering interpretive questions, which echoes the findings by Sandora et al. (1999).
For the responses to the literal questions, there was no difference between the experimental group and the control group. Although the experimental group was not exposed to as much instruction focusing on the surface language as the control group, they performed just as well. This shows that QtA lessons, despite more emphasis on reader-text interaction, do not hamper processing and retention at the literal level.
QtA Lessons and Aesthetic Reading
The findings in the present study point to the potential of QtA lessons in enhancing aesthetic reading, albeit not in boosting efferent reading—the two contrasting reading stances proposed by Rosenblatt
108
(1983). The positive effects of QtA lessons on reader-based inferences and on responses to interpretive questions speak to us that QtA lessons, with discussion initiated by students, might facilitate aesthetic reading, in which readers bring their unique personal knowledge and experience to transact with the texts, leading to transformation of ideas, and possibly, critical reading. On the other hand, the null effects of QtA lessons on the measures of recall, text-based inferences, factual questions show that QtA lessons might not be able to promote text-bound reading, i.e., reading to replicate textual information, hence efferent reading. The findings also demonstrate that stories, especially fables, come with rich potential for readers to engage their experience and knowledge to interpret and expand the text. If aesthetic reading is the ultimate goal of story reading, then the conventional instructional approach with teacher as a director of meaning making is therefore inadequate, because the readers’ critical thinking, which takes departure from the construction of textbase, (Kintsch, 1998) will be circumvented. QtA lessons with an emphasis on readers’ construction of their own text afford readers’ volition and autonomy to enact on the text elaboration process. Hence, efferent reading is contingent upon self-determined processing and can only be assessed by measurements that trigger student-initiated meaning, such as inferences generated in recall, but not measurement that is constructed by testers, like responsive questions.
Liu & Chu: Effects of Questioning the Author
109
IMPLICATIONS
Based on the findings of the present study, implications are drawn for classroom practice and for future research.
Implications for Classroom Practice
First of all, the results of the study indicate that using a different mode, QtA, to manage a reading class is a good alternative to yield more reader-based inferences and a greater number of responses to interpretation questions. In traditional English classrooms, teachers dominate most of the talk owing to a tight schedule. The QtA lesson, a more interaction-oriented approach, can yield more student talk. It is suggested that English teachers in Taiwan can create a dynamic classroom via questioning and discussing by utilizing the QtA approach.
One disadvantage associated with the QtA approach is that it takes time and effort for the teachers to plan the segments queries and discussion moves in advance. Therefore, it is suggested that teachers collaborate to work on the lesson plans.
Second, the findings of this study suggest that an important dimension of the current model of reading, efferent reading (Rosenblatt, 1993) can be addressed in the EFL classroom. As the results reveal, combining content knowledge with students’ personal experiences may help junior high school readers not only to read the text, but also to construct their own text and begin to make reader-based inferences and respond to interpretation questions better. Through QtA lessons, it is expected that EFL readers can be offered
110
room to experience the product of their transaction with writers (Rosenblatt, 1993), create a situation model for text (Kintsch, 1998), do critical thinking and thus transform themselves from text-bound readers to independent readers.
Third, the results of this study indicate that QtA training might be a feasible approach either as an activity in a regular class or an extracurricular activity in EFL classrooms at the high school level. During regular class time, teachers can conduct a discussion on issues generated from reading and guide adolescents to understand issues and take different perspectives. In extracurricular activities, students with high motivation and proficiency can read more difficult texts to question not only the authors, but also their peers.
Implications for Future Studies
Some limitations cannot be ignored and need to be considered when planning future studies that investigate the QtA approach.
First of all, we do not know whether the effects of reader-based inferences and teacher-cued interpretive questions would be sustained over time. Therefore, it is suggested that an additional posttest be carried out two weeks later or a month later in order to test the effects over time.
Second, how participants from Group E and Group C processed their thoughts during reading is still a puzzle. It is suggested that a think aloud protocol (Kucan & Beck, 1997) can be used to elucidate the difference in mental processes between the QtA training and traditional instruction.
Liu & Chu: Effects of Questioning the Author
111 students were facing the coming National Basic Competency Test, they might not have produced their best performance. Future studies should include students of different grade levels and investigate the interaction effects between QtA lessons and more reader variables, such as gender, grade, or proficiency levels.
Fourth, the present study adopted fables as the main genre. It is recommended that materials from various sources, especially expository texts (Beck et al., 1997) be included. If narrative and expository texts are involved, whether participants take an efferent stance (Rosenblatt, 1993) or whether they read for a study purpose (Van den Broek, Lorch, Linderholm & Gustafson, 2001), can be examined.
Finally, the treatment in this study lasted only five weeks. Five weeks’ exposure to an innovative approach might not be long enough to change the learning habits that have been fostered by the traditional approach over a three-year period (Huang, 2004). Given the assertion by Chang (2002) that “acquisition of new strategies requires persistent practice over time” (p. 9), it is expected that QtA lessons be practiced longer to produce greater improvements in reading comprehension.
ACKNOWLEDGEMENTS
This paper is drawn from the M.A. thesis of the first author, which was modified by the second author, the supervisor of the thesis, into the present version. We would like to thank Professor Rung-ching
112
Tsai in the Department of Mathematics, National Taiwan Normal University, for her statistical consultation.
REFERENCES
Almasi, J. F. (2003). Teaching strategies processes in reading. New York: The Guildford Press.
Baker, L., Dreher, M. J., & Guthrie, J. T. (2000). Why teachers should promote reading engagement. In L. Baker, M. J. Dreher, & J. T. Guthrie (Eds.), Engaging young readers: Promoting
achievement and motivation (pp. 1-16). New York: The
Guildford Press.
Barry, S., & Lazarte, A. A. (1998). Evidence for mental models: How do prior knowledge, syntactic complexity, and reading topic affect inference generation in a recall task for nonnative readers of Spanish? The Modern Language Journal, 82, 176-193.
Beck, I. L., & McKeown, M. G. (2001). Inviting students into the pursuit of meaning. Educational Psychology Review, 13(3), 225-241.
Beck. I. L., & McKeown, M. G. (2002). Questioning the author: Making sense of social studies. Educational Leadership, 60(3), 44-48.
Beck, I. L., McKeown, M. G., Hamilton, R. L., & Kucan, L. (1997).
Questioning the author: An approach for enhancing student engagement with text. Newark, DL: International Reading
Liu & Chu: Effects of Questioning the Author
113 Beck, I. L., McKeown, M. G., Sandora, C., Kucan, L., & Worthy, J.
(1996). Questioning the author: A yearlong classroom implementation to engage students with text. The Elementary
School Journal, 96(4), 385-414.
Bernhardt, E. B. (1991). Reading development in a second language. Norwood, NJ: Ablex.
Block, C. C., & Pressley, M. (Eds.). (2002). Comprehension
instruction: Research-based best practices. New York: The
Guilford Press.
Brisbois, J. E. (1992). Do first language writing and second language
reading equal second language reading comprehension? An assessment dilemma. Paper presented at the Annual Meeting of
the National Reading Conference. (ERIC Document Reproduction Service No. ED354499)
Chang, C. (2002). The reader effect (instruction/awareness of text
structure) and text effect (well-structured vs. bad-structured texts) on first and second/foreign language reading comprehension and recall—What does research teach us?
(ERIC Document Reproduction Service No. ED465180)
Durkin, D. (1979). What classroom observations reveal about reading comprehension instruction. Reading Research Quarterly, 14(4), 481-533.
Graesser, A. C., Person, N. K., & Hu, X. (2002). Improving comprehension through discourse processing. New directions
for teaching and learning, 89, 33-44.
Gray, W. S. (1960). The major aspects of reading. In H. Robinson (Ed.), Sequential development of reading abilities (Vol. 90, pp.
114
8-24). Chicago: Chicago University Press.
Huang, S. C. (2004). The effect of a reading & writing approach to
EFL instruction on students’ motivation. Unpublished master’s
thesis, National Taiwan Normal University, Taipei.
Hyman, R. T. (1982). Questioning for improved reading. Educational
Leadership, 40(1), 307-309.
Kintsch, W. (1998). Comprehension: A paradigm for cognition. Cambridge: Cambridge University Press.
Kucan, L., & Beck, I. L. (1997). Thinking aloud and reading comprehension research: Inquiry, instruction, and social interaction. Review of Educational Research, 67, 271-300.
Lee, J. F. (1986). On the use of the recall task to measure L2 reading comprehension. Studies of Second Language Acquisition, 8, 201-212.
Lewin, L. (2004). Talking back to authors. Educational Leadership,
62(2), 58-63.
McDaniel, C. (2004). Critical literacy: a questioning stance and the possibility for change. The Reading Teacher, 57(5), 472-481. McKeown, M. G., & Beck, I. L. (2004). Transforming knowledge into
professional development resources: Six teachers implement a model of teaching for understanding text. The Elementary
School Journal, 104(5), 391-408.
McKeown, M. G., Beck, I. L., & Worthy, M. J. (1993). Grappling with text ideas: Questioning the author. The Reading Teacher, 46(7), 560-566.
McKeown, M. G., Beck, I. L., Hamilton, R., & Kucan, L. (1999).
Liu & Chu: Effects of Questioning the Author
115
for classroom challenges. Bothell, WA: Wright Group.
Ministry of Education. (2003). Nine-year curriculum guidelines for
elementary and junior high schools: Language domain. Taipei:
MOE.
Ogle, D., & Blachowicz, C. L. Z. (2002). Beyond literature circles: Helping students comprehend informational texts. In C. C. Block & M. Pressley (Eds.), Comprehension instruction:
Research-based best practice (pp.259-274). New York: the
Guildford Press.
Olivier, D. (1999). Aesop’s fables. (Rev. ed.). Taipei: Cavesbooks. Pressley, M. (2000). What should comprehension instruction be the
instruction of? In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research (Vol. 3, pp. 545-561). Mahwah, NJ: Erlbaum.
Pressley, M. (2002). Reading instruction that works: The case for
balanced teaching (2nd ed.). New York, NY: the Guilford
Press.
Rosenblatt, L. M. (1978). The reader, the text, the poem: The
transactional theory of the literary work. Carbondale: Southern
Illinois UP.
Rosenblatt, L. M. (1983). Literature as exploration (4th ed.). New York: the Modern Language Association of America.
Rosenblatt, L. M. (1993). The transactional theory: Against dualisms.
College English, 55, 377-386.
Sandora, C., Beck. I., & McKeown, M. (1999). A comparison of two discussion strategies on students’ comprehension and interpretation of complex literature. Journal of Reading
116
Psychology, 20, 177-212.
Sweet, A. P., & Snow, C. E. (Eds.). (2003). Rethinking reading
comprehension. New York: the Guilford Press.
The foolish friend. (n.d.). Retrieved December 4, 2004, from
http://www.omcetrum.cz/cesky/baiky/anglicke/oo_bajky_ang.ht m
The prince and the lion. (n.d. ). Retrieved December 4, 2004, from
http://www.omcetrum.cz/cesky/baiky/anglicke/oo_bajky_ang.ht m.
Tierney, R. J., Readence, J. E., & Dishner, E. K. (1990). Reading
strategies and practices: A compendium. Needham Heights,
MA: A Division of Simon & Schuster.
Van den Broek, P., Lorch, R. F. Jr., Linderholm, T., & Gustafson, M. (2001). The effect of readers’ goals on inference generation and memory for texts. Memory & Cognition, 29(8), 1081-1087.
ABOUT THE AUTHORS
Yu-chen Liu is an English teacher at Ming-De Junior High School, Taipei. She is interested in the implementation and assessment of innovative approaches to the teaching of reading.
Hsi-chin Janet Chu is an associate professor in the English Department at National Taiwan Normal University. Her teaching and research interests include EFL reading, contrastive rhetoric, and TEFL Methodology.
Liu & Chu: Effects of Questioning the Author
117
APPENDIX A
Sample Reading and Initiating Queries for Treatment
“The Miller’s Donkey”11 One day a miller, who owned a very strong donkey, decided to sell it. He and 2his son set off to the town together, letting the animal walk in front of them.＊ 12 As 3they went they met a group of girls playing along the road. When they saw the old man and little boy walking behind the donkey they began to laugh.
5 “See those two foolish people!” they said to each other. “They’ve got a fine 6donkey, but instead of riding on it, they walk behind it! Who ever heard of such a 7thing?”
8 The old miller, when he heard this, was very surprised.＊2 Quickly he made 9his son get on the donkey and sit on its back. He himself continued to walk behind. 10 Soon they met another party, not of young girls but of old men. They were 11shaking their gray heads and talking of the bad manners of the young.
12 “Look at that!” said one when he saw the miller’s son on the donkey. “Isn’t 13that what we were just saying? The young today don’t care about the old! That 14strong young boy riding easily on that donkey while his poor old father follows on 15foot. Get down, boy, so that your father may get up!”
16 The miller was quite upset by this new talk.＊3 He quickly made his son get 17off the donkey. Then he himself got on the donkey’s back while his little boy 18walked alongside. The child’s legs were shorter than the donkey’s and he had to run 19to keep up. Then some women washing clothes near the road called to the miller.
20 “If that was my child,” shouted one of them, “he wouldn’t be running himself 21to death behind your donkey! You call yourself a father! You’re not worthy of 22having a son at all!”
23 “Oh, dear! Oh, dear!” said the poor miller. “It seems that I’ve been wrong 24again!＊4 Get up here son, and ride behind me! Then perhaps nobody will say 25anything!”
26 So the boy got up and rode with his father. But in a short time they met a 27farmer who looked closely at them.
28 “Tell me, sir,” he said to the miller, “is that poor unhappy little donkey your
1 The initial number represents the line number of the line in the passage. 2 “＊1” means where one segment ends and the teacher starts giving queries.
118 29own?”
30 “Of course it is my own,” replied the miller. “Why do you ask?”
31 “Because only a fool would put so much weight on his animal!” answered 32the farmer.＊5“You and that boy are better able to carry that donkey than the 33donkey is to carry you! Why don’t you both get down and try carrying the 34animal!?”
35 Of course, he did not really mean what he said. But the foolish miller once 36again believed that the last speaker made the most sense.＊6 So he and his son got 37down. Then, with much difficulty, they tied the poor donkey’s legs together; and 38with a long stick which they bought, they were ready to carry the donkey. 39 “Now,” said the already tired miller, “you pick up one end of the stick, my 40boy, and I’ll pick up the other. Then we will carry the donkey across the bridge and into 41the town on the other side of the river.”＊7
42 By this time a crowd had gathered to watch them. They stood and laughed as 43the miller and his little boy tried to pick up the crying donkey, hanging upside 44down.＊8 But the effort was too great for them. They suddenly dropped the stick 45and the donkey very hard. The poor animal then rolled into the river and 46disappeared under the water. It never came up again!＊9
47 The miller began to realize that he had made a fool of himself. Too bad he 48didn’t realize a little earlier.
49 “I have tried to please everyone!” he said loudly, “but I have pleased 50nobody—especially not myself! This is the result of following the opinions of 51others! The next time I want to do a thing I will do it my own way.”＊10
1. What does the author mean here?
2. What does the author mean by using “two foolish people?” 3. What has the author told us about the old miller?
4. How do things look for the old miller?
5. Based on how the author described the old miller, does it connect with what the author told us before?
6. What do you think of that?
7. What is the author trying to say here?
8. Given what the author has already told us about the old miller, what do you think he’s up to?
9. How has the author settled this for us? 10. What’s the author’s message?
Liu & Chu: Effects of Questioning the Author
119
APPENDIX B
Sample Pretest-Posttest Reading Material
“The Foolish Friend”Once upon a time there was a king. He fought, won his battles and beat all his enemies. Once, when the war was over the king went back to his camp; he needed a rest. In the camp there was a gentle ape. It had been taught to carry a sword and it looked so funny that it made the king laugh. The ape did what he was told. The king kept the ape as a good friend and a good body guard. The king told the ape, "Nobody should bother my sleep, take care of it," and the king fell asleep. The ape took the sword and sat down beside the king's bed.
After some time, a bee came in making sounds and stayed on the king's head. It was making sounds in the king's ear. The ape found that the sleep of the king was being bothered. The ape threw the bee away through a window but it came back. Again and again the ape threw the bee away but each time it stayed again on the king's head. At last, the angry ape took the king's sword and hit at the bee. Oh, the hit that was meant for the bee killed the king! Then the ape cried sadly because it loved the king. So the king who had come safely through many wars was killed at last because he made a friend of an ape.
120
APPENDIX C
A Sample Open-ended Questions (English Version)
“The Foolish Friend”Factual:
1. What did the king do before he slept? 2. Where did the bee stay after it made sounds? 3. Why was the ape not happy with the bee?
Interpretive:
1. What did you learn from the sentence “So the king who had come safely through many wars was killed at last because he made a friend of an ape?” and why?
2. The king said, “Nobody should bother my sleep, take care of it?” Do you think he is a lazy king? Why or why not?
3. Finally, the king who won many wars died because of a bee. Is this a good ending? Was the ape a true friend? Why or why not?
Responsive:
1. If you were the king in the story, would you need a body guard? Why or why not?
2. Do you think the king will feel bad if he knows he was killed by a friend? Why or why not?
Liu & Chu: Effects of Questioning the Author 121
