Sign up

How explicit instruction makes a difference: metadiscourse markers and EFL learners' reading comprehension skill.
Article Type:
Report
Subject:
Reading comprehension (Evaluation)
Indirect discourse (Study and teaching)
English education (Methods)
English education (Curricula)
Authors:
Jalilifar, Alireza
Alipour, Mohammad
Pub Date:
09/22/2007
Publication:
Name: Journal of College Reading and Learning Publisher: College Reading and Learning Association Audience: Academic Format: Magazine/Journal Subject: Education Copyright: COPYRIGHT 2007 College Reading and Learning Association ISSN: 1079-0195
Issue:
Date: Fall, 2007 Source Volume: 38 Source Issue: 1
Geographic:
Geographic Scope: United States Geographic Code: 1USA United States

Accession Number:
171686943
Full Text:
This study attempted to determine the effect of explicit instruction of metadiscourse markers on pre-intermediate EFL learners' reading comprehension skill. To this end, 90 students were selected following a pre-test and divided randomly into three groups. Then three versions of the same test, original, modified and unmodified metadiscourse-free texts, were provided. Results revealed that the group receiving the original version outperformed the group with the unmodified version, but their performance was about equal with the group receiving the modified version. In addition, two of the groups answered a questionnaire on how they judged the texts. Next, one of the groups received instruction on metadiscourse. Finally, a posttest was administered. Results revealed the positive influence o f form-focused instruction of metadiscourse. It also revealed that metadiscourse markers are primarily responsible for cohesion rather than coherence.

**********

For many researchers, the concept of metadiscourse is defined as discourse about discourse, and it is based on a view of writing as a social and communicative action between the writer and the reader. It deals with the study of textual resources at above-sentence levels and can be defined as the linguistic resources used to organize a discourse or the writer's stance toward either its content or reader and includes a heterogeneous series of cohesive and interpersonal features, which help the readers connect, organize, and interpret materials in a way preferred by the writer and with regard to the understandings and values of a particular discourse community (Hyland, 1998, 2000; Vande Kopple, 1985). As noted by Hyland and Tse, through metadiscourse, a writer is able not only to transform a dry, difficult text into coherent, reader-friendly prose, but also relate it to a given context and convey his personality, credibility, audience-sensitivity, and relationship to the message (2004). Metadiscourse markers, according to Vande Kopple, help readers organize, classify, interpret, evaluate, and react to propositional content (1985).

Metadiscourse markers are discourse elements that help both the writer and the reader in two ways. On one level, they help the writer organize the propositional content and his ideas mainly through using connectives such as and, therefore, in addition, etc. Additionally, connectives, such as the sequencers first, then, finally, etc., guide the reader through the maze of the writer's units of thoughts by indicating the organization of the text. On the other level, metadiscourse markers build an interaction between the reader and the writer and account for the atmosphere and reader-friendliness of the text. As an example, the writer reveals his own ideas when he uses illocution markers like I recommend that, or I believe that, or, by using hedges like perhaps, might, and apparently and emphatics such as clearly, undoubtedly, and surely, which allows the writer to send signals to the reader about whether he is doubtful or sure about what he is pointing out. Such markers, however, may not signal coherence on their own. Once metadiscourse ties are removed from the text, slight modifications can still make the text coherent. In the following example, taken from one of the texts in the present study, the second version appears incoherent; however, slight modifications on it made it coherent and meaningful even though the metadiscourse elements were removed.

Many researchers have made attempts to scrutinize the contributions that metadiscourse can make to language pedagogy. Traditionally, the focus of an immense part of the literature on metadiscourse has been on writing skill. However, only a few of these studies have been experimental in nature and have tried to tease out the effect of metadiscourse on the students' writing; rather, they are chiefly comparative studies that have applied cultural and discipline specific viewpoints as their point of departure. Conceding that a great number of theoreticians have brought metadiscourse knowledge within the context of writing to the center of attention, the role that this crucial part of discourse plays in reading comprehension skill seems to have sunken into oblivion. On the grounds of this conspicuous disregard, there is an opaque picture of the correlation between the recognition of metadiscourse markers and the promotion of students' reading comprehension skill.

A number of researchers have conducted studies on metadiscourse markers in relation to different modes of language and have substantiated the facilitative role of metadiscourse markers (Dahl, 2004; Hyland & Tse, 2004; Ifantidou, 2005; Lee, 2002; Perez & Macia, 2002; Steffensen & Cheng, 1996; Suau, 2004). Steffensen and Cheng (1996) studied the propositional and pragmatic functions of metadiscourse in writing. Lee (2002) focused on coherence in writing in terms of five features including metadiscourse. Perez and Macia (2002) investigated the role of metadiscourse in lecture comprehension. Dahl (2004) studied the role of national culture and academic discipline in textual metadiscourse in research articles. Suau (2004) attempted to bridge the gap between research articles and popular science articles from the viewpoint of metadiscourse. Camiciottoli (2003) studied the effect of metadiscourse on reading comprehension ability.

While reading a text, second language learners all too often confront difficulties comprehending the author's line of reasoning due to not being acquainted with the organizational content and interactive features of metadiscourse markers that shape the text. Therefore, this is a common problem from which many reading comprehension classes are suffering. However, research has not addressed the role of explicit instruction of metadiscourse markers for the pre-intermediate language learners. Steffensen and Cheng assert that "in spite of its importance, metadiscourse as a pragmatic system is rarely taught and is typically approached even by skillful writers on an intuitive basis" (1996, p. 3). Concerning the dearth of experimental studies on the influence of metadiscourse markers on reading comprehension in EFL contexts, it is evident that the experimental research into the impact of the knowledge of metadiscourse markers on reading comprehension skill is still in its infancy. Moreover, previous studies have not offered convincing evidence on the functional role of metadiscourse markers, whether they help achieve coherence or cohesion. Therefore, the present research aims to uncover the existence of any relationship between the knowledge of metadiscourse markers and the EFL learners' reading comprehension ability. To this aim, the present study is motivated to answer the following questions:

1. Are metadiscourse markers helpful to readers in recognizing the ways by which writers style their texts?

2. Does reference to metadiscourse markers during reading affect reading comprehension ability?

3. Does the direct teaching of metadiscourse markers enhance the EFL learners' reading comprehension ability?

4. How do metadiscourse-free texts appear to the readers?

5. How does the removal of metadiscourse markers from a text affect its comprehensibility?

6. Does modification or reorganization of sentences on a metadiscourse-free text contribute to its comprehensibility?

Methodology

Participants

Initially, 130 male and female TEFL (Teaching English as a Foreign Language) students, who were enrolled in a two-credit Reading Comprehension course as a requirement for their BA degree at Masjed Soleiman Islamic Azad University, took a sample proficiency test. Ninety of these students were selected as the homogeneous pre-intermediate sample based on their scores on the English language proficiency test. Then, they were randomly divided into three groups, 30 subjects representing each group. Only one of the groups received instruction on metadiscourse markers.

Instruments

For the purpose of this experiment, six reading comprehension passages including 36 multiple-choice test items from a sample proficiency test were selected as the pretest. The test was used to enable the researchers to select a homogeneous sample. As calculated through the Kudar and Richardson's formula (KR-21 Method), the reliability of this test was 0.823. The same test was again administered at the end of the semester to trace possible changes in the students' reading comprehension ability.

Another test that included four reading comprehension passages from a sample TOEFL test (Broukal & Nolan-Woods, 2001) was also designed. The reliability of the texts based on KR21 was 0.961. Each passage included 132 words on average, and the semantic complexity of these texts averaged out at about 21 propositions per text.

In order to determine the effect of metadiscourse markers on the participants' reading comprehension ability, three versions of this latter test were designed. The original version included metadiscourse markers. In the second version, metadiscourse markers were removed, and the passages were structurally modified to account for possible distortions as a result of the omission of metadiscourse markers. Due to the removal of metadiscourse markers from the texts, the unmodified texts appeared loose and incoherent because connectives that constitute a huge part of metadiscourse markers and logically bond the propositions together were unavailable. To make up for this unwanted incoherence, the researchers modified the texts as follows:

1. Long sentences were broken into smaller ones in order to be much less confusing. For example, note the following two passages.

But when language characters develop and an alphabet and number system have reached a certain stage, there comes a demand for some formal teaching and so schools are established for a select few, prospective rulers and priests, to supplement the education given by family and tribe. [The original text] (Broukal, & Nolan-Woods, 2001, p. 253) Language characters develop. An alphabet number system has reached a stage. There comes a demand for some formal teachings. Schools are established for a select few, prospective rulers, priests, to supplement the family-tribe education. [The modified text]

2. To sound clearer, some ambiguous sentences were paraphrased.

The family is the first educator and a life-long influence. [The original text] (Broukal, & Nolan-Woods, 2001) The family is the first educator with a life-long influence. [The modified text]

3. Because the canonical word order was disrupted, words were combined to make compound nouns or adjectives and some parts of speech were also changed. For example, the unmodified sentence "family tribe provide all the education"(Broukal, & Nolan-Woods, 2001, p. 253) in the first text was modified into the phrase "family-tribe provision of all the education".

4. Punctuation was modified as well.

In the third version of the test, metadiscourse markers were removed leaving the four passages metadiscourse-free, but no modification was made on the passages. Finally, a questionnaire with five questions on metadiscourse markers was designed for the participants in two of the groups (see Appendix I for a sample questionnaire). The purpose behind administering this questionnaire was to find out to what extent the omission of metadiscourse markers from a text affects comprehension from the language learners' point of view. We also wished to see how they judged the text difficulty and whether they found the writers' styles helpful.

In order to locate metadiscourse markers in the texts, we used the comprehensive classification by Vande Kopple (1997). The rationale for the selection of this classification was that it is based on Hallidayan's (1985) macro-functions of language including textual and interpersonal categories. Moreover, the classification has been subject to many revisions and modifications by Vande Kopple himself and other theoreticians in the field (Vande Kopple, 1985, 1997; Crismore, Markkanan & Steffensen, 1993; Longo, 1994, Mao, 1993; Rahman, 2004). Vande Kopple's (1997) classification, which is a revision of his two previous proposals, is as follows:

Textual Metadiscourse

1. Connectives reveal organization and intertextuality (e.g., first, next).

2. Code Glosses are parenthetical definitions within sentences.

Interpersonal Metadiscourse

3. Illocution Markers identify discourse acts (e.g., we recommend, we claim, we conclude, to sum up).

4. Validity Markers assess the probability of truth of the propositional content:

a. Hedges (e.g., perhaps, may, might, often, usually, apparently);

b. Emphatics (e.g., clearly, undoubtedly, it is obvious that, of course, very crucial);

c. Attributors (e.g., according to Einstein).

5. Narrators let the readers know who said what (e.g., Mrs. Jones said).

6. Attitude Markers reveal the attitudes of the writer toward propositional content (e.g., surprisingly, unfortunately).

7. Commentaries direct comments to the reader (e.g., the reasons for these choices are simple, most of you will oppose the idea that).

Procedure

Having selected the groups, in order to test the effect of metadiscourse markers on the subjects' reading comprehension ability, the researchers administered the three versions of the same test. Group One received the original version of the passages. Group Two received the modified metadiscourse-free version, and Group Three took the unmodified metadiscourse-free version. Next, the participants in Groups Two and Three were asked to provide responses to the questionnaire and indicate how smoothly they read the passages. Following this, the researchers made comparisons between the results of the different versions of the test and the responses to the questionnaire.

The next phase of the experiment lasted for one academic semester and began with training the learners. Throughout the semester, the instructor, who was one of the researchers, provided Group Three with explicit instruction on how metadiscourse markers are used by the writers and interpreted by the readers, following the presentation, practice, production (PPP) model, whereas the other groups did not receive any instruction on the significance of metadiscourse markers. The teaching procedure followed these steps:

The researcher first described the concept of metadiscourse and then gradually presented the above classification. He selected one category of metadiscourse devices and explained and exemplified the category with examples so that a vivid picture of metadiscourse, its importance and the functions that it performs in a text, was created in the students' minds. This was the presentation phase. To make sure that the students had a good grasp of metadiscourse markers, in each session, the researcher provided more explanation on metadiscourse markers, and as the learners were reading the course-required passages, they were required to find the metadiscourse devices and recognize to which category they belonged. This was the practice phase. They were also given metadiscourse-free texts to complete by suggesting the right elements. This stage was considered the production phase.

Groups One and Three, finally, took a reading comprehension test (the pre-test in the study) that included metadiscourse markers. Since at this stage, we wished to see only the role of explicit instruction on metadiscourse, one group was excluded from this phase of the study. Later, we applied the statistical t test to compare the mean scores obtained by the students in the two groups on the posttest.

Data Analysis

Comparison of the original and the unmodified texts

Group One took the reading comprehension test that included metadiscourse markers, whereas Group Three took the version of the same reading comprehension test in which metadiscourse markers were left out with no modification made on the texts.

The statistical t test was administered. The t value of 7.857 observed for a degree of freedom 58 was significantly above the t critical value of 1.671 at the .05 level of significance. Therefore, there was a meaningful difference between the two mean scores, and the t value was significant enough to confirm that Group One outperformed Group Three. This led researchers to conclude that structurally the omission of metadiscourse markers from the extracts and the lack of any modification on the texts naturally made the participants fail to comprehend the texts and this resulted in their lower performance.

Table 1 shows the data related to the results of the original and the unmodified versions of the subjects' reading comprehension tests on the effect of metadiscourse markers on their reading comprehension ability.

Comparison of the original and the modified texts

The results did not indicate a significant difference in the mean scores of Group One, taking the original reading comprehension texts, and Group Two, taking the modified versions. The statistical t test was administered. The t value observed for a degree of freedom of 58 was 1.063, which was less than the t critical value of 1.671 at the .05 level of significance. Therefore, there was not a meaningful difference between the two mean scores, and the t value confirmed that the groups performed rather equally on the reading comprehension test. It is likely that the omission of metadiscourse markers from a text does not hinder the comprehensibility of the propositional content presented in the text once enough structural modifications are made on the text.

Table 2 reveals the statistics related to the results of the original and the modified versions of the participants' reading comprehension tests on the effect of metadiscourse markers on their reading comprehension ability.

Comparison of the modified and the unmodified texts

Comparisons of the two mean scores of the second and third groups revealed a drastically significant difference. In order to make sure that the difference in the mean scores was statistically significant, the statistical t test was administered. The t value observed was 7.353 for a degree of freedom of 58, which was high above the t critical value of 1.671 at the .05 level of significance. The results, therefore, confirmed that Group Two performed much better than Group Three on the reading comprehension test. This showed that the omission of metadiscourse markers from a text made it difficult to comprehend unless enough modifications were made on it. The results, on the one hand, support the idea that the removal of metadiscourse markers from a text hinders its comprehensibility. On the other hand, the results reinforce the idea that structural modifications made on a metadiscourse-free text increases its comprehensibility significantly.

Table 3 illustrates the results of the unmodified and the modified versions of the subjects' reading comprehension tests and the effect of metadiscourse markers on their reading comprehension ability.

Results of the groups on the questionnaire

Table 4 details the number and percentage of the third group's responses to the questionnaire on the unmodified version of the reading comprehension test of metadiscourse markers. This questionnaire encompassed five questions.

The results made clear that a great proportion of the students, 83.5%, admitted that the text was difficult. In addition, more than two-thirds of the students reported that they could not follow the main points of the texts in the test, which could possibly imply that metadiscourse markers create a link between the idea units in a text and that once they are absent, the whole passage becomes less clear.

In regards to the author's style of writing and his attempt to create a dialogue with the reader, half of the students denied the author's helpfulness, and this response was due to the absence of metadiscourse markers, which in turn made the passages dry and incohesive. That is, metadiscourse markers account for the cohesion and reader-friendliness of the text. In addition, more than half of the students agreed that the extracts in the test did not include any introduction to help the reader comprehend the texts with more ease. This observation may indicate that a metadiscourse-free passage is difficult to understand because it is disorganized. The conclusion to be drawn from this questionnaire is that metadiscourse markers are responsible for creating organization in the propositional content of the text.

Table 5 shows the number and percentage of the second group's responses to the same questionnaire on the modified version of the reading comprehension test of metadiscourse markers.

The results made clear that still a great proportion of the students, 43.5%, found the text difficult, while slightly above the same proportion found the text to be of average difficulty, although the texts were modified. The dearth of these metadiscourse devices failed to link the propositions in the text and the sentences, as such, they appeared less related to one another. Here, 56.5% of the students reported that they had an average understanding of the main points of the extracts in the test, but nobody claimed a complete grasp of the main points. This result might imply that once metadiscourse markers are absent, the sentences in the passage remain less cohesive. However, the propositional content of the passage remains stable, but the function of organizing these propositions is not fulfilled properly. Therefore, many of the students could at least partially comprehend the texts due to the reshaping of the propositional content.

Regarding the author's style of writing and his attempt to create a dialogue with the reader, more than three-fourth of the students agreed on the author's helpfulness. Because of the structural modifications made on the sentences, the texts did not appear too hard or incoherent, although they were less cohesive. Therefore, although metadiscourse markers account for the cohesion and reader-friendliness of the text, modifications partially compensated for their absence. Furthermore, a large number of the students claimed that the extracts in the test hardly included enough introduction to help the reader comprehend the texts with more ease. The conclusion to be drawn from this questionnaire is that metadiscourse markers are primarily responsible for creating organization in the propositional content, but they do not add to the propositional content of a text.

Results of the explicit instruction of metadiscourse

The results revealed a significant mean difference. The statistical t test was also administered. The t value observed was 4.405, which was significantly above the t critical value of 1.671 at the .05 level of significance. This substantiated that Group Three outperformed Group One on the sample reading comprehension posttest. These results, as such, not only rejected the third null hypothesis (H03) of the study that explicit instruction of metadiscourse markers has no effect on EFL students' reading comprehension skill but also confirmed the idea that explicit instruction of metadiscourse markers promoted EFL students' reading comprehension skill.

The findings of this study are in line with other works on metadiscourse markers such as those by Perez and Macia (2002) and Camiciottoll (2003), who have substantiated the effective role of metadiscourse markers in spoken and written texts.

Table 6 shows the descriptive and inferential statistics related to the results of the subjects' posttest on the influence of instruction on metadiscourse markers to the first group.

Discussion

This section reflects on the results by providing answers to the questions raised in the beginning of the paper.

(1) Are metadiscourse markers helpful to readers in recognizing the ways by which writers style their texts?

Through the use of textual metadiscourse markers, a writer fashions his text and organizes the propositional content that he wishes to share with the readers. In the same vein, interpersonal metadiscourse markers allow him to create a dialogue with the readers and to style his writing. However, the students answering the questionnaire in relation to metadiscourse-free passages denied the writer's helpfulness to the readers and, as a result, they failed to understand the main points of the passages. Therefore, owing to the experimental results of this study and others (Hyland & Tse, 2004; Lee, 2002, and Steffensen & Cheng, 1996), metadiscourse markers make the texts syntactically complete, and thus, they facilitate comprehension by helping readers recognize the writer's style of writing.

(2) Does reference to metadiscourse markers during reading affect reading comprehension ability?

A comparison of the results of the pre and post tests showed that the reading comprehension ability of the first group did not improve. This finding could be described as the subjects' weakness applying metadiscourse markers in an efficacious manner on their own. It is also in line with Schmidt's Noticing Hypothesis that language items that learners are supposed to pick up by themselves had better be brought to their attention through activities that focus on form, repeated noticing, and continued awareness of the language features like metadiscourse markers (1990).

(3) Does the direct teaching of metadiscourse markers enhance the students' reading comprehension ability?

Following studies by Spada and Lightbown (1993), Perez and Macia (2002), and Camiciottoli (2003) and based on the results of the posttest of this study, it was revealed that explicit instruction of metadiscourse markers proved effective in enhancing the subjects' reading comprehension ability. That is, the subjects in the instructed group performed far more successfully on the posttest. The reason for this phenomenon can be that the participants in the instructed group were able to recognize different kinds of metadiscourse markers, interpret the functions performed by them, and predict the following events from the preceding propositions in the text due to being instructed on metadiscourse markers and consciously drawing on them while reading.

(4) How do metadiscourse-free texts appear to the readers?

As indicated by the data collected from the questionnaire, more than half of the students who read the unmodified metadiscourse-free passages stated that the texts were difficult to understand and that they could not follow the main points of the texts. The distortion in comprehension might well be ascribed to the dearth of textual metadiscourse that is responsible for organizing the propositions and "managing quite extended stretches of discourse" (McCarthy, 2005, p. 49) in the text. These connectives "presuppose a textual sequence" and establish "a relationship between segments of the discourse" (McCarthy, 2005, p. 46). The removal of these markers from the texts broke the bonds existing among the propositions and, consequently, the texts' sentences appeared confusing.

Moreover, the students denied the friendliness of the texts. Undoubtedly, since interpersonal metadiscourse accounts for the atmosphere of the text, the lack of these markers resulted in the texts' dryness. What is more, the participants declared that the texts appeared incohesive and telegraphic.

Previous studies did not include modified texts as a proof for the cohesive function of metadiscourse markers. The findings also challenged the arguments made by Lee (2002), who claimed that metadiscourse markers account for coherence of the texts. The results of the modified versions of metadiscourse-free texts proved that the coherence of the texts was not reduced by the lack of metadiscourse. This is because, in accord with Widdowson (1978), coherence, which is meaning-bound, has little to do with cohesion, which is grammar-bound in nature; metadiscourse markers are associated, primarily, with cohesion not coherence, which indicates "how the information is relevant, or what the point of saying it is" (Salkie, 1997, p. 94). What makes a text coherent is the semantic and not structural relatedness. Note the following example in which the text is incoherent even though cohesive ties are present:

Therefore, although the texts in this study were incohesive in terms of the binding relationship between the sentences, they were not incoherent and the students could make sense of the propositional content. Also, the incohesiveness of the passages referred to the lack of textual metadiscourse markers, with connectors being a major part of them.

(5) How does the removal of metadiscourse markers from a text affect its comprehensibility?

That the third group failed to answer the questions in the unmodified metadiscourse-free version of the test revealed that the removal of metadiscourse markers was a stumbling-block, and it strongly influenced the readers' comprehension of the text. Therefore, the distortion in comprehension resulted from the removal of metadiscourse markers that bind the propositions together. The immense magnitude of this incohesion made the passages hard to comprehend. However, once modifications were made on the text, despite the absence of these ties, coherence was enhanced and participants could follow the text. Such cohesive ties are, therefore, part of the resources of a language which are neither necessary nor sufficient for coherence on their own.

(6) Does modification made on a metadiscourse-free text contribute to its comprehensibility?

The response to this question would be positive. Compared with the first group of subjects, results of the performance of the subjects in the second group, who took the modified version of the metadiscourse-free test, revealed only a trivial difference. The conclusion to be drawn is that modifications made on the metadiscourse-free passages made up for their incohesiveness. Therefore, the students were capable of comprehending the extracts which were not deemed incoherent but less connected.

Conclusion

The aim of this study was to fill the existing gap in the experimental work on metadiscourse markers. To this end, we attempted to find out the effect of metadiscourse markers on reading comprehension skill. Analysis of the data enabled us to arrive at the following conclusions.

1. Pre-intermediate language learners fail to employ metadiscourse markers in their reading comprehension skill on their own. That is, they hardly rely on metadiscourse information unconsciously. They must be made aware of the important roles that metadiscourse markers play in a text to improve their language learning in general and their reading comprehension skill in particular. Therefore, it is necessary not to let the students be oblivious of these crucial devices through bringing metadiscourse markers to their consciousness. Metadiscourse markers should be taught explicitly in reading comprehension classes as a means to enhance students' reading comprehension ability. As noted by VanPatten and Cadierno (1993) and Fotos (1993), instruction that focuses on output raises learners' consciousness and is sometimes necessary as it encourages learners to be accurate while they attend to meaning.

To provide instruction on metadiscourse markers, following the results of this study, language teachers are suggested to apply the well-established PPP (Presentation, Practice, and Production) approach in their treatment. To this end, the teacher may introduce one category of metadiscourse markers each session through sufficient explanations and examples. Then, learners may be required to practice locating those elements in texts. Next, they may be given texts in which metadiscourse elements have been removed, and they suggest or select appropriate metadiscourse markers to complete the gaps. Finally, learners may be encouraged to produce metadiscourse markers in a free writing task to have a good grasp of that specific category. This process can be conducted for all the categories of metadiscourse markers.

2. The paucity of metadiscourse markers in a text, to a great extent, affects the text syntactically rather than semantically. In other words, metadiscourse-free passages are hard to understand due to incohesion, and not incoherence. Therefore, it is claimed that the removal of metadiscourse markers from a text does not influence its coherence as much as it does influence the text's cohesion and propositional organization.

The research into metadiscourse markers is useful in itself. We set our sights on how readers establish the relationship between the propositions and interpret the text as a unified whole. The findings of this study may be considered as tendencies and a starting point for further investigation.

References

Broukal, M., & Nolan-Woods, E. (2001). NTC's preparation for the TOEFL. Lincolnwood: NTC Publishing.

Camiciottoli, B. C. (2003). Metadiscourse and EAP reading comprehension: An exploratory study. Reading in a Foreign Language, 15(1), 28-44.

Crismore, A., Markkanen, R., & Steffensen, M. S. (1993). Metadiscourse in persuasive writing: A study of texts written by American and Finnish university students. Written Communication, 10(1), 39-71.

Dahl, T. (2004). Textual metadiscourse in research articles: A marker of national culture or of academic discipline. Journal of Pragmatics, 36(10), 1807-1825.

Fotos, S. (1993). Consciousness raising and noticing through focus on form: Grammar task performance versus formal instruction. Applied Linguistics, 14, 385-407.

Halliday, M. A. K. (1985). An introduction to functional grammar London: Edward Arnold.

Hyland, K. (1998). Persuasion and context: The pragmatics of academic metadiscourse. Journal of Pragmatics, 30, 437-455.

Hyland, K. (2000). It might be suggested that...: Academic hedging and student writing. Australian Review of Applied Linguistics Series, 16, 83-97.

Hyland, K., & Tse, B. (2004). Metadiscourse in academic writing: A reappraisal. Applied Linguistics, 25(2), 156-177.

Infantidou, E. (2005). The semantics and pragmatics of metadiscourse. Journal of Pragmatics, 37(9), 1325-1353.

Lee, I. (2002). Helping students developing coherence in writing. English Teaching Forum, 32-39.

Longo, B. (1994). Current research in technical communication: The role of metadiscourse in persuasion. Technical Communication, 4(2), 348-352.

Mao, L. R. (1993). I conclude not: Toward a pragmatic account of metadiscourse. Rhetoric Review, 11, 265-289.

McCarthy, M. (2005). Discourse analysis for teachers. Cambridge: Cambridge University Press.

Perez, M. A., & Macia, E. A. (2002). Metadiscourse in lecture comprehension: Does it really help foreign language learners. Atlantis, 14(2) 3-21.

Rahman, M. (2004). Aiding the reader: The use of meta-linguistic devices in scientific discourse. Nottingham Linguistic Circular, 18, 30-48.

Salkie, R. (1997). Text and discourse analysis. London: Routledge.

Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2) 17-46.

Spada, N., & Lightbown, R M. (1993). Instruction and the development of questions in L2 classrooms. Studies in Second Language Acquisition 15, 205-224.

Steffensen, M. S., & Cheng, X. (1996). Metadiscourse and text pragmatics: How students write after learning about metadiscourse. (ERIC Document Reproduction Service No. ED400709)

Suau, F. (2004). Metadiscourse in research and popular science articles or how to please the audience: A cross-generic and intra-generic analysis in search for a common meta-discursive core. Paper presented at the 9th International Pragmatics Conference in Riva del Garda, Italy.

Vande Kopple, W. J. (1985). Some exploratory discourse on metadiscourse. College Composition and Communication, 36, 82-93.

Vande Kopple, W. J. (1988). Metadiscourse and the recall of modality markers. Visible Language, 22, 233-72.

Vande Kopple, W. J. (1997). Refining and applying views about metadiscourse. Paper presented at the 48'h Annual Meeting of the Conference of College Composition and Communication, Phoenix, AZ. (ERIC Document Reproduction Service No. ED411529)

VanPatten, B., & Cadierno, T. (1993). Explicit instruction and input processing. SSLA 15, 225-243.

Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University Press.

Alireza Jalilifar holds a Ph.D. in applied linguistics. He has been teaching for more than 20 years in secondary schools and universities. He is now a member of Shahdi Chamran State University of Ahvaz, Iran. His main interests are genre analysis, metadiscourse and thematic structure. He has published in the International Journal of Applied Linguistics and local Iranian Journals. He has also presented lectures in LSP conference in Malaysia (2005), IPrA (Gothenborg 2007) and Odense (2007). He may be contacted atjalilifar20@yahoo.com Mohammad Alipour graduated from the Center for Science and Research in Ahvaz majoring in English Language Teaching. His main interest is metadiscourse.
... curious as it may sound, this relationship in many respects
   comes close to that between history and literature. Economics,
   after all, is the science (in the broad meaning of the term) of
   something which men actually do: even if the science did not exist,
   men would still make economic decisions, economic predictions and
   participate in the various forms of economic organization which, in
   part, it is the economist's function to describe. [The original
   text] (Broukal, & NolanWoods, 2001, p. 255)

   ... this relationship in many respects comes close to between
   history literature. Economics is the science of something men do:
   the science did not exist, men make economic decisions, economic
   predictions participate in the various forms of economic
   organization it is the economist's function to describe. [The
   unmodified text]

   This relationship is similar in many respects.
   Economics is the science of something men do. Men make economic
   decisions, economic predictions, participate in the various forms
   of economic organization without science. It is the economist's
   function to describe. [The modified text]


While David Beckham was busy playing for Real Madrid, even Jennifer
   looks stunning in these pictures. Just in case you weren't aware of
   it, it's Christmas time. You must have nearly had a heart attack,
   so what's the deal with your girlfriend now you're a pop star?


Table 1 Results of the reading tests of metadiscourse

Groups   N      M       SD     test

G1       30   29.58   10.306   7.857
G3       30   10.73     .528

Table 2 Results of the reading tests of metadiscourse

Groups   N      M        SD     t test

G1       30   29.58    10.306   1.063
G3       30   26.825    8.528

Table 3 Results of the reading tests of metadiscourse

Groups   N      M        SD     test

G1       30   10.73     8.528   7.353
G3       30   26.825   10.306

Table 4 Number and percentage of the group's responses to the
questionnaire

Responses in percentage
terms                        Not at all   Average      Almost

Questions
1. Extracts were easy        25 = 83.5%   5 = 16.5%    0 = 0%
2. Understand the main       20 = 67%     8 = 26.5%    2 = 6.5%
   points
3. Author's style is         16 = 53%     10 = 33.5%   4 = 13.5%
   helpful
4. Author creates a          15 = 50%     11 = 36.5%   4 = 13.5%
   dialogue with reader
5. Extracts include intro-   17 = 56.5%   7 = 23.5%    6 = 20%
   ductory passages

Responses in percentage
terms                        So much

Questions
1. Extracts were easy        0 = 0%
2. Understand the main       0 = 0%
   points
3. Author's style is         0 = 0%
   helpful
4. Author creates a          0 = 0%
   dialogue with reader
5. Extracts include intro-   0 = 0%
   ductory passages

Table 5 Number and percentage of the responses to the questionnaire

Responses in percentage
terms                        Not at all   Average      Almost

Questions

1. Extracts were easy       13 = 43.5%    14 = 46.5%   3 = 10%
2. Understand the main       5 = 16.5%    17 = 56.5%   8 = 27%
   points
3. Author's style is         4 = 13.5%    8 = 26.5%    12 = 40%
   helpful
4. Author creates a          0 = 0%       7 = 23.5%    12 = 40%
   dialogue with reader
5. Extracts include intro-   4 = 13.5%    8 = 26.5%    12 = 40%
   ductory passages

Responses in percentage
terms                        So much

Questions

1. Extracts were easy        0 = 0%
2. Understand the main       0 = 0%
   points
3. Author's style is         6 = 20%
   helpful
4. Author creates a         11 = 36.5%
   dialogue with reader
5. Extracts include intro-   6 = 20%
   ductory passages

Table 6 Results of the sample reading comprehension posttest

Groups   N      M        SD     t test

G1       30   28.56    10.691   4.405
G2       30   38.826    8.723
Gale Copyright:
Copyright 2007 Gale, Cengage Learning. All rights reserved.