April 2007 Volume 3 Issue 1
Article 5.
PDF File
April 2007 home

Article Title
Are Task Type and Familarity Predicators of Performance on Tests of Language for Specific Purposes?

 Author
Mohammad Ali Salmani-Nodoushan

Bio:
Dr Mohammad Ali Salmani-Nodoushan is Assistant Professor of TEFL at the University of Zanjan, Iran. He has been teaching BA and MA courses at different Iranian Universities for the past fifteen years. Dr Salmani-Nodoushan is member of editorial boards of Asian EFL Journal, The Linguistics Journal, and I-Manager's Journal of Educational Technology. He is also editor-in-chief of Iranian Journal of Language Studies (IJLS). His areas of interest include language testing and English for Specific Purposes (ESP).

Abstract
In a study of the effects of text familiarity, task type, and language proficiency on university students’ Language for Specific Purposes Ability (LSPA) test and task performances, 541 senior and junior university students majoring in electronics took the Task-Based Reading Test (TBRT). Variance analyses indicated that text familiarity, task type, and language proficiency, as well as the interaction between any given pair of these and also among all of them resulted in significant differences in participants’ overall and differential test and task performances. In addition, regression analyses revealed that the greatest influence on subjects’ overall and differential test and task performances was due to language proficiency. Text familiarity had the smallest effect on students' test and task scores. Compared to text familiarity, task type was a stronger predictor of variance in test and task performance.

Keywords: Language for Specific Purposes Ability, Task-Based Reading Test,

1. INTRODUCTION
This study investigates the probable effects of background knowledge (or text familiarity) in English for Specific Purposes (ESP) tests. The assumption is that test-takers' prior familiarity with the propositional content of texts that appear in ESP reading comprehension tests will facilitate their performance on those tests.

2. BACKGROUND OF THE STUDY
English for Specific Purposes (ESP) was a phenomenon that grew out of a number of converging trends after the Second World War. In spite of the fact that these trends have operated in a variety of ways around the world, we can identify three main reasons common to the emergence of what we know as ESP today: (a) worldwide demands, (b)  revolutions in linguistics, and (c) focus on the learner.
The rapid growth of scientific, technical and economic activity on an international scale after World War II created a world which was unified and dominated by technology and commerce. This unification was accelerated by the Oil Crises of the early 1970s, which resulted in a massive flow of funds and Western expertise into the oil-rich countries. All of these generated a demand for an international language, and English was the best choice. As English became the accepted international language, it created a new generation of learners who knew specifically why they were learning it; time and money constraints created a need for cost effective courses with clearly defined goals. At the same time, linguistic studies began to shift attention away from traditional structural treatments of language to discovering the ways in which language was actually used in real communication (Widdowson, 1978). Some linguists began to support the view that the English needed by a particular group of learners could be identified by analyzing the linguistic characteristics of their specialist area of work or study (Hutchinson and Waters, 1987). This view was further supported by  new developments in educational psychology which assigned a central role to learners and their attitudes to learning (Rodgers, 1969). Learners were seen to have different needs and interests, which would have an important influence on their motivation to learn and, therefore, on the effectiveness of their learning. This lent support to the development of courses in which relevance to the learners’ needs and interests was paramount. As a result, ESP was born.

One of the earliest challenges that ESP had to face was the distinction between ESP and English for General Purposes (EGP). ESP differs from EGP in the sense that the content of ESP courses (i.e., words, sentences, and subject matter) relates to a particular field or discipline; English for General Purposes (EGP) is, however, essentially the English language education in junior and senior high schools. In EGP courses, students are introduced to the sounds and symbols of English, as well as to the lexical, grammatical, and rhetorical elements that compose spoken and written discourse. In addition, EGP focuses on applications of English in general situations. Supplementary information about appropriate gestures, cultural conventions, and cultural taboos can also be included in EGP curricula. EGP conducted in English-speaking countries is typically called English as a Second Language (ESL), and EGP conducted in non-English-speaking countries is normally called English as a Foreign Language (EFL). English for Specific Purposes (ESP), however, is research and instruction that builds on EGP; ESP is designed to prepare students or working adults for the English used in specific disciplines, vocations, or professions to accomplish specific purposes. Pedagogically, a solid understanding of basic EGP should precede higher-level instruction in ESP if ESP programs are to yield satisfactory results (Hutchinson and Waters,1987). According to Hutchinson and Waters (1987, p. 19), “ESP is an approach to language teaching in which all decisions as to content and method are based on the learner’s reason for learning.” Dudley-Evans (1998), however, claimes that ESP may not always focus on the language of one specific discipline or occupation. He argues that university instruction that introduces students to common features of academic discourse in the sciences or humanities, frequently called English for Academic Purposes (EAP), is equally ESP.

From its early beginnings in the 1960s, ESP has undergone five main phases of development: (a) Register Analysis, (b) Rhetorical Discourse Analysis, and (c) Target Language Use (TLU) Situation Analysis, (d) Skills-Centered Approach, and (e) Learning-Centered Approach.

Register analysis took place mainly in the 1960s and early 1970s, and was associated in particular with the work of Strevens (Halliday, Mclntosh and Strevens, 1964), Ewer (Ewer and Latorre, 1969) and Swales (1971). It operated on the basic principle that the English needed in one scientific field constituted a specific register different from those of other fields of science, or General English. Rgister analysis sought to identify the grammatical and lexical features of different scientific registers. ESP, in this phase, focused on language at the sentence level.
With the development of Discourse Analysis, however, ESP entered a second phase of development typically known as Rhetorical Discourse Analysis. The basic hypothesis of this stage was succinctly expressed by Allen and Widdowson (1974) who took the view that the difficulties which the students encountered arose from an unfamiliarity with English use rather than from a defective knowledge of the system of English. Allen and Widdowson (1974) argued that ESP students' needs could best be met by an ESP course which developed a knowledge of how sentences were used in the performance of different communicative acts. The tacit assumption of Rhetorical Discourse Analysis is that the rhetorical patterns of text organization differs significantly between specialist areas of use; however, this point was never very clearly examined (Swales, 1985), and indeed paradoxically, the results of the research into the discourse of subject-specific academic texts were also used to make observations about discourse in general (Widdowson, 1978).

The upsurge of interest in communicative language teaching as well as the development communicative syllabi resulted in the emergence of the third phase of ESP (i.e., the stage of Target Language Use (TLU) Situation Analysis). TLU Situation Analysis aimed at establishing procedures for relating language analysis more closely to learners’ reasons for learning. ESP courses designed in this phase proceeded first by an identification of the target situation and then by a rigorous analysis of that situation. The identified features then formed the syllabus of the ESP course. Such a process was usually known as Needs Analysis. However, Chambers (1980) preferred to use the term target situation analysis, since it was a more accurate description of the process concerned. Perhaps the most thorough explanation of TLU Analysis was the system set out in Communicative Syllabus Design by Munby (1978).

The fourth stage of ESP development (i.e., Skills-Centered Approach) was an attempt to look below the surface and to consider not the language itself but the thinking processes that underlie language use (See Chitravelu (1980), Grellet (1981), Nuttall (1982), and Alderson and Urquhart (1984)). The principal idea behind the Skills-Centered Approach was that common reasoning and interpretation processes underlay all language use which enabled the students to extract meaning from discourse regardless of the surface linguistic forms. The tacit assumption in this approach was that students did not need to focus closely on the surface forms of the language; they rather needed to focus on the underlying interpretive strategies, which enabled them to cope with the surface forms. As such, a focus on specific subject registers was unnecessary in this approach, because the underlying processes were not specific to any subject register.

The fifth phase of ESP (i.e., Learning-Centered Approach) emerged out of the shortcomings of its preceding phases. Proponents of the Learning-Centered Approach argue that the four preceding phases of ESP were all flawed in that they were all based on descriptions of language use whereas a truly valid approach to ESP must be based on an understanding of the processes of language learning. In fact, the fifth phase of ESP is concerned with the question of what it really means to know a language.

One of the first scholars to ask this question was Spolsky (1973). Since then, many people have tried to answer this question. Alderson (1991), for instance, has pointed out that the answer to the question of what it means to know a language “depends upon why one is asking the question, how one seeks to answer it, and what level of proficiency one might be concerned with” (p. 12). In the case of Language for Specific Purposes (LSP) testing, Douglas (2000, p. 26) adds the expression, “and in what specific situational context one is interested,” to the quotation from Alderson.

As a result of scholars' attempts at answering Spolsky's question, different models of language ability have been proposed since then. A few of the most influsntial of these models are (a) Hymes's (1972) model of Communicative Competence, (b) Bachman's (1990) model of Communicative Language Ability (CLA), (c) Bachman and Palmer's (1996) reformulation of Communicative Language Ability (CLA), and Douglas's (2000) model of  Language for Specific Purposes (LSP) Ability.

The term communicative competence encompasses the notion that language competence involves more than Chomsky’s (1965) rather narrowly-defined linguistic competence. According to Hymes (1971, 1972), communicative competence involves judgements about what is systemically possible (i.e., what the grammar will allow), psycholinguistically feasible (i.e., what the mind will allow), and socioculturally appropriate (i.e., what society will allow). Moreover, communicative competence affords information about the probability of occurrence of a linguistic event and what is entailed in the actual accomplishment of it. In fact, for Hymes, competence was more than knowledge: “Competence is dependent upon both [tacit] knowledge and [ability for] use” (Hymes 1972, p. 282).

Since the time Hymes proposed the notion of communicative competence, other scholars have reformulated his notion to propose their own models. The current most well-known framework is Bachman's (1990) Communicative Language Ability (CLA), elaborated by Bachman and Palmer (1996). Bachman and Palmer (1996) postulate two components of communicative language ability: (a) language knowledge and (b) strategic competence. Strategic competence serves as a mediator between the internal traits of background knowledge and language knowledge and the external context, controlling the interaction between them.

Douglas (2000) tried to modify Bachman and Palmer's model in such a way as to make it suitable for accounting for ESP competence. He took the position that what was required in Language for Specific Purposes (LSP) testing was an understanding of how Specific-Purpose (SP) background knowledge could interact with language knowledge to produce a communicative performance in SP contexts. As such, Douglas’s framework for LSP ability (a) includes Specific Purpose (SP) background knowledge as a component of communicative language ability, and (b) gives a central role to the cognitive construct of discourse domain. According to Douglas (2000), Language for Specific Purposes (LSP) model requies four considerations: (a) the level of detail necessary in the definition of LSP construct, (b) the treatment of the four skills, (c) whether to include strategic competence or not, and (d) whether to distinguish between language knowledge and SP background knowledge.

In connection to Douglas's last point, it is noteworthy that the distinction between language knowledge (or language proficieny) and background knowledge (or content-area knowledge) has long been a problem for language testers in the interpretation of test results. There are a few studies which suggest that, under some conditions, background knowledge does not influence language test performance. Several other studies have, however, found significant interactions between background knowledge and language test performance. Over the past two decades or so, there have been several studies into the effect of background knowledge on LSP test performance, most of which focused on reading comprehension (See Erickson and Molloy, 1983; Osman, 1984; Alderson and Urquhart, 1985a; 1985b; Koh, 1985; Shoham, Peretz, and Vorhaus, 1987; Alderson, 1988; Alvermann and Hynd, 1989; Hock, 1990; Peretz and Shoham, 1990; Read, 1990; Tan, 1990; Tedick, 1990; Douglas and Selinker, 1993; Chen and Graves, 1995; Jensen and Hansen, 1995; Ridgway, 1997; Papajohn, 1999).

Three articles by Alderson and Urquhart (1983, 1985a, and 1985b) aroused considerable interest in the effect of background knowledge and led to several follow-up studies. In each article, Alderson and Urquhart compared students’ scores on reading texts related to their own field of study with those on texts in other subject areas. The students’ scores on the modules were somewhat contradictory. On the one hand, for example, science and engineering students taking the technology module of ELTS did better than the business and economics students who took the same test, and as well as the liberal arts students, although their language proficiency was lower. On the other hand, the business and economics students did not do better than the science and engineering groups on the social studies module. Alderson and Urquhart concluded that background knowledge had some effect on test scores, but that this was not consistent, and that future studies should take account of linguistic proficiency and other factors as well.

Along the same lines, Shoham, Peretz, and Vorhaus (1987) concluded that, while students in the biological and physical sciences did better at the scientific texts, the humanities and social science students did not do better on the test in their own subject area. In a similar study, Peretz and Shoham (1990) had similar results. Their explanation for this was that the texts were only indirectly related to the students’ specialized fields of study; they suggested that this might support Lipson’s (1984) contention that a totally unfamiliar text is often easier to comprehend than a text with a partially familiar content. Clapham (1996) believes that this contention of Lipson was indeed radical:

If supported by further research, it would be an almost unassailable reason for dropping ESP testing. If Lipson’s idea were taken to its logical conclusion, of course, proficiency tests would have to contain materials outside any candidates experience. The JMB (Joint Matriculation Board) University Test in English for Speakers of Other Languages followed just such an approach, with passages in esoteric subjects such as silver markings and heraldic devices. As a result, item writers had difficulty finding suitable texts and the ensuing materials were often excessively dull. (Clapham, 1996, p. 8)

Ridgway (1997) set out to prove that the background knowledge effect only occurred between two linguistic thresholds. The 69 students from two different disciplines who took part in Ridgway’s study were divided into top and bottom samples according to their levels of L2 proficiency. Ten students with medium L2 proficiency were excluded from the study. Students were asked to read texts from inside and outside their own subject area. According to the two thresholds hypothesis, the effect of background knowledge should have been insignificant for students in the top and bottom groups. For students with low L2 proficiency (‘bottom group’), there were no significant differences between mean scores for the text from their own field of study and the text from an unfamiliar subject area. For groups with high L2 proficiency, however, there were significant differences between mean scores. Ridgway (1997) concludes that the different degrees of text specificity may have contributed to results being as inconclusive as they were.
In another study, Clapham (2000) found that background knowledge had its greatest effect on the performance of subjects who belonged to the medium-proficiency group. In an attempt to explain the findings of her study, Clapham (2000, pp. 515–516) argued:

while lower level students could not take advantage of their background knowledge because they were too concerned with bottom-up skills such as decoding the text, and while high proficiency students were able to make maximum use of their linguistic skills so that, like native speakers, they did not have to rely so heavily on their background knowledge, the scores of medium proficiency students were affected by their background knowledge.

Alderson (2000, p. 104) recognized the potential of Clapham’s findings but also called for clarification:

Needless to say, Clapham’s results need replication and extension. Nevertheless, they suggest that language testers might some day be able to define text difficulty in terms of what level of language abilities a reader must have in order to understand that particular text, and vice versa, what sort of text a learner of a given level of language ability might be expected to be able to read.

3. STATEMENT OF THE PROBLEM

The present study is an attempt at exploring the probable effects of prior familiarity with the content of a text on readers' performance on tests based on that text. The main aims of this study are four-fold: (a) to determine if LSP reading test and task performance is related to language proficiency, (b) to explore if task type is related to LSP reading test performance, (c) to determine if text-familiarity (defined in this study to refer to prior knowledge of the propositional content of texts) affects LSP reading test and task performance, and (d) to determine which factor (text familiarity, task type, language proficiency) is responsible for a greater portion of students’ score variance. As such, the study addrresses the following questions:

1. Is LSP students’ overall and differential test and task performance a function of text familiarity, task type, language proficiency, or the interaction of these variables?
2. Which variable (i.e., task type, text familiarity, or language proficiency) accounts for a greater share of LSP students’ test and task score variance:?
3. Does LSP students’ level of language proficiency significantly affect their overall test and task performance as well as their test and task performance across different levels of the text familiarity cline?
4. Does LSP students’ degree of familiarity with the propositional content of LSP tests significantly affect their overall test and task performance and their test and task performance across different levels of language proficiency?

4. METHOD

4.1. Participants

The total number of the participants in this study is 541 people. The population from which the participants of the present study were drawn included junior and senior university students majoring in electronics at different Iranian universities. These students took the sample version of the IELTS (University of Cambridge Local Examinations Syndicate [UCLES], 2000). They were then classified into four proficiency groups: proficient (93 people), fairly proficient (186 people), semi-proficient (164 people), and non-proficient (98 people). The mean and the standard deviation of the subjects’ IELTS scores were used as the criterion for their classification (i.e., standard deviations from the mean). Students who had scored higher than ‘mean-plus-one’ standard deviation were assigned to the proficient group; those who stood within the ‘mean-plus-one’ standard deviation range were assigned to the fairly proficient group. The semi-proficient group included participants whose scores on the IELTS fell within the mean-minus-one standard deviation range. Finally, the participants who had scored below the mean-minus-one standard deviation range were assigned to the non-proficient group.

4.2. Instruments

Three different instruments were used in the present study: (1) The sample version of the IELTS General Training Reading Module (UCLES, 2000), (2) a Self-report Questionnaire, and (3) the Task-Based Reading Test (TBRT)—which consisted of three modules: Accounting (TBRT-AM), Electronics (TBRT-EM), and General (TBRT-GM).

Determining the participants’ level of language proficiency in general, and their 'reading comprehension ability' in particular, were vital to this study. The sample version of the IELTS General Training Reading Module was, therefore, administered to the participants since it is considered to be the most suitable instrument due to its ‘modularity’ claims; according to the claims made by UCLES (2000), the assumption behind the IELTS is that the ‘link between the reading and writing modules has been lifted’ since 1995, and that each module is a standard gauge for the language skill it measures. Participants' scores on this test were used as an indicator of their level of language proficiency and reading comprehension ability.

Another variable under study was participants' prior familiarity with the propositinal content of texts that appeared in the different modules of the TBRT (i.e., their text familiarity). In this study, text familiarity was taken to refer to the participants' pripor familiarity with content knowledge in general (that is, the knowledge they had acquired from their interests, and hobbies), and with subject or domain specific content knowledge in particular (that is, the knowledge they had acquired from formal schooling). Text familiarity was controlled in two ways. First of all, the participants of the study were drawn from among university students majoring in electronics to make sure that they were familiar with the content of TBRT-EM and unfamiliar with that of TBRT-AM. In addition, a self-report questionnaire was also used since some of the participants might have been familiar with the content of TBRT-AM as a result of personal interest. The self-report questionnaire aimed at showing the subjects’ degree of familiarity with each text that appeared in each of the TBRT modules. This questionnaire was a Likert Scale composed of 20 items (four similar items for each text in each module) through which the subjects indicated their degree of text familiarity with the five passages that appeared in each of the TBRT modules.

To ensure maximum understanding of the questions, the questionnaire was written in the subjects’ native language—Farsi.

The main instrument used in the present study was the Task-Based Reading Test (TBRT) which had three modules with texts from different disciplines: Accounting (TBRT-AM), Electronics (TBRT-EM), and General Digest (TBRT-GM). Each module consisted of five sets of test items: (a) True-False items [N=12], (b) Open-Ended Sentence Completion itmes [N=8], (c) Paragraph Labeling items [N=6], (d) Fill-in-the-Blank Skimming items [N=9], and (e) Multiple-Choice Elicitation items [N=5]. As such, each module included a total of 40 items that measured subjects’ performance of five reading tasks: (a) true-false task, (b) sentence-completion task, (c) outlining task, (d) elicitation of writer’s views task, and (e) skimming task. Each module consisted of five passages of varying lengths, textual complexity, and readability indexes. However, the texts that appeared in the different module where chosen in such a way as to ensure maximum correspondence to the IELTS General Training Reading Module (UCLES, 2000) in terms of such textual features as readability, structural complexity, etc.

Table 1: Comparison of Readability Statistics for IELTS and TBRT

In addition to readability analysis, nine university instructors who were experienced teachers of ESP courses at various Iranian universities were asked to judge whether the texts were of the suitable level of difficulty for the prospective subjects.

The texts that appeared in the TBRT-EM were all taken from the content areas that junior and senior university students majoring in electronics had already studied as part of their academic courses. They included five topics: (a) magnetic flux, (b) vacuum tube diodes, (c) bridge circuits, (d) incandescent lamps, and (e) digital and analog computers. Since the participants of the present study were all majoring in electronics, the assumption was that they were totally familiar with the passages within this module. In the same vein, the TBRT-AM module included five texts. This time, the texts were selected from the materials that were part of the academic courses of university students majoring in accounting. They included the following five topics: (a) chain stores, (b) interest, (c) clearinghouses, (d) assets and liabilities, and (e) corporate finance. It is noteworthy that, since the participants of the present study were all majoring in electronics, the texts within the TBRT-AM module were totally unfamiliar to them. The same procedures were used in the selection of the passages that appeared in the TBRT-GM module. Unlike the two other modules, the texts within this module were expected to contain propositional content with which the participants of the study reported partially familiar. Five passages were selected from the "Microsoft Encyclopedia Encarta" computer package. These texts included such general-digest topics as (a) natural hazards, (b) national parks and sanctuaries, (c) the sensory system of sharks, (d) classification of airplanes, and (e) mission to moon.

Each module within the TBRT included 40 items (i.e., the same number of items as appeared in the IELTS General Training Reading Module). The items measured participants' performance on five different reading tasks. The first group that measured participants’ performance of a true-false task included twelve items. Each item was followed by three answers: true, false, and not given. The participants were expected to read the corresponding passages and to decide whether the propositions expressed in the true-false items were given in the passages or not, and if yes, to make their own choice whether the items were true or false. The second set of items in each module aimed at measuring the participants’ performance of a sentence-completion task. The items in this set were eight open-ended sentences which could be completed in two ways. Following this set of items was a list of possible endings. The participants’ job was to read the corresponding passage and, on the basis of the information presented in the passage, to choose two possible endings from the list to complete each item. A third group of items measured the participants’ performance of an outlining task. This category included six items. The test takers were expected to read a passage. Each paragraph within the passage was labeled with a letter from the English alphabet. The participants were expected to choose from among a list of summaries the one that best represented the main idea of each paragraph. They would then match the summary for each paragraph with the label that signified that paragraph. Participants’ performance of the task of 'eliciting the writer’s views' was also measured. Five multiple-choice items followed a passage in each module; each item had three choices: yes, no, and not given. The test takers were expected to read the passage and to decide whether the propositions expressed in these five items were given in the passage or not, and if yes, whether they represented the views of the writer of the passage or not. The last set of items measured test takers’ performance of a skimming task. The nine items of this category asked the participants to skim the reading passage for two types of information: dates and proper nouns. The former included five items while the latter included four items. The test takers' job was to skim the reading passage and to identify the date or the proper noun that was questioned in the item.

4.3. Procedures

In order to determine whether the items in the different modules of the TBRT were effective, malfunctioning or non-functioning, a pilot administration was of the TBRT carried out. Since the purpose of this process was to screen the items so that the most suitable ones would be included in the final versions of the TBRT, 80 items were included in the pilot version of each module (i.e., twice as many items as were necessary for the final version of the TBRT). The pilot version was then administered to a group of 36 university students majoring in electronics. All of these students took the TBRT-GM pilot module first. Then, these students were randomly assigned into two equal half-groups. One half-group took the TBRT-EM pilot module followed by the TBRT-AM pilot module while the second group took the TBRT-AM pilot module followed by the TBRT-EM pilot module. This procedure was necessary to control for probable practice effect. The results of the administration of the TBRT pilot version were then used for item analysis. After item analysis, from among the 80 items that appeared in each pilot module, the 40 items that had the best item facility and item discrimination indexes were chosen for inclusion in the final version of each corresponding TBRT module.

After the development of the final version of the TBRT, in order to determine whether the TBRT reading modules were suitable for data collection, it was vital that the modules be evaluated through a trial administration. The modules, along with the IELTS General Training Reading module (UCLES, 2000) were, therefore, administered to a group of 20 senior university students majoring in electronics. All these students took the IELTS General Training Reading and the TBRT-GM modules in one administration session, and the TBRT-EM and TBRT-AM modules in another session. To control for any probable practice effect, a counter-balanced design was used in each administration. That is, ten students were randomly assigned to the first-half group and the ten remaining students to the second-half group. In the first session, the first-half group first took the IELTS General Training Reading module and then the TBRT-GM module whereas the second-half group first took the TBRT-GM module and then the IELTS General Training Reading module. In the second administration session, the first-half group first took the TBRT-AM module and then the TBRT-EM module while the second-half group first took the TBRT-EM module and then the TBRT-AM module.

The final versions of TBRT modules and the IELTS were administered to a total of 578 junior and senior university students majoring in electronics. The procedure for the final administration of the tests was similar to that of the trial administration. Here again, for purposes of minimizing any probable practice effect, a counter-balanced design was used for test administration. In addition to these tests, the participants also took the Self-report Questionnaire. On the basis of their responses to the Self-report Questionnaire, and due to the text-familiarity assumptions of the study, 37 subjects who had reported a high-enough prior familairity with the texts that appeared in the TBRT-AM module were discarded from the data.

A Principle Component Factor Analysis was also performed to examine the construct validity of the TBRT modules. The Varimax rotation method (with Kaiser Normalization) was used for this analysis.  The scores of participants on each group of items were included in the analysis. Since five types of items were employed in each module, a five-factor solution factor analysis was performed.  The result of factor analysis showed that the item type in each module loaded under each factor indicating the construct of these items. The results of factor analysis are presented in table 2 below.

As table 2 shows, two different types of items, i.e. skimming and True/False item type loaded under factor one.  This may imply that the participants used a skimming strategy to answer True/False items, or vice versa. Factors 2, 3, and 4 show the construct of sentence-completion items, outlining items, and writer’s view items, respectively.  The outlining items of the accounting module loaded under factor five. It is not known why this happened; this requires further investigation.

Table 2: Varimax Rotatin Factor Analysis for IELTS and TBRT

The reliability analyses revealed that the TBRT modules and the IELTS had satisfactory Spearman-Brown reliability coefficients. The Spearman-Brown reliability coefficient for TBRT-EM was 0.8527, for TBRT-AM 0.8527, and for TBRT-GM 0.8628. The IELTS had a  Spearman-Brown reliability coefficient of 0.8617.

5. ANALYSIS AND RESULTS
The data of the present study were submitted to a number of statistical analyses including (a) one-way analyses of variance, (b) univariate analyses of variance, (c) multi-variate analyses of variance, and (d) multiple regression analyses. The results of these analyses are presented under appropriate headings in the following sub-sections. All of the analyses reported below are based on the 95% confidence interval.

5.1. The Effect of Proficiency
The first aim of the study was to determine if participants’ level of language proficiency affected their LSP test performance at a given level of text familiarity. To this end, test takers' performances across all proficiency levels (i.e., proficient, fairly proficient, semi-proficient, and non-proficient) were compared for significant differences. The results indicated that participants from different proficiency levels performed differently on tests with totally familiar, partially familiar, and totally unfamilar propositional contents. In other words, no matter whether the propositional content of the test was familiar, unfamiliar, or partially familiar, the participants of the study at a given proficiency level performed significantly different from those at any other proficiency levels. Table 3 reports the results of  the post hoc Scheffé test for the participants’ performances on tests with varying degrees of familiar propositional content.

Table 3: Scheffé Test Results for Participants’ Performance on Tests with Varying Degrees of Familiar Propositional Content

The overall performances of the participants on the TBRT were also studied. The sum of their scores on the three TBRT modules (TBRT-EM, TBRT-GM, and TBRT-AM) indicated their total TBRT score. In order to determine if participants’ proficiency levels had an effect on their overall test performances, the main effect analysis of variance was conducted. The results indicated that their overall performances on the TBRT at each of the proficiency levels differed significantly from their performances at any of the the other proficiency levels. Table 4 presents the results of this analysis.

Table 4: Scheffé Test Results for Participants’ Overall Performance on the TBRT across Different Levels of Language Proficiency

A second aim of the study was to determine if participants’ level of language proficiency affected their performances of a given reading task at different degrees of text familiarity. As it was delineated earlier (section 4.2. above), the present study set out to measure participants’ performances of five reading tasks: (a) true-false task, (b) sentence-completion task, (c) outlining task, (d) elicitation of writer’s views task, and (e) skimming task. Participants’ performances on each of these tasks were compared across different levels of proficiency and text familiarity.

The first task studied in this series was the true-false task. An analysis of variance was conducted to determine if the performances of participants on the true-false task at each level of text familiarity varied as a result of their proficiency levels. The results indicated that, in the context of a test with totally familiar propositional content,  participants' performances at each proficiency level differed significantly from those of the participants at other levels of language proficiency. In the context of tests with partially familiar propositional content, the performance differences between participants at each proficiency level with those at any of the other levels were significant except for that of semi-proficient versus non-proficient subjects. A similar finding was observed in the context of tests with totally unfamiliar propositional content. Here again, except for the performance difference between semi- and non-proficient students, the performances of students at any other proficiency level were significantly different in comparison to those of the subjects at other proficiency levels The results of this analysis are presented in table 5.

Table 5: Scheffé Test Results for Participants’ True/False Task Performance on Tests with Varying Degrees of Familiar Propositional Content

The second task studied in this series was a sentence-completion task. Students’ performances on this task across different proficiency levels at each point on the text familiarity cline were analyzed for significant differences. The results indicated that the performances of students at each proficiency level ware significantly different from those of students at any of the the other levels except for that of semi-proficient versus non-proficient subjects at each and every level of the text familiarity (See Table6).

Table 6: Scheffé test for subjects’ sentence-completion task performance on tests with varying degrees of familiar propositional content

Variance analysis was also performed to determine if students’ level of proficiency affected their performances on outlining tasks at a given point on the text familiarity cline. The results of this analysis are presented in table 7.

Table 7: Scheffé Test Results for Participants’ Outlining Task Performance on Tests with Varying Degrees of Familiar Propositional Content

As it can be seen from table 7, students in the non-proficient and semi-proficient groups did not show any significant difference in performance on the outlining task in tets with totally unfamiliar and totally familiar propositional contents. In the context of tests with partially familiar contents, the same students showed a significant difference in performance; however, this difference was so small that it could be neglected (sig=0.046).
Variance analysis was also performed to determine if students' from different proficiency levels performed differently on the 'elicitation of writer’s views' task at any given point on the text familiarity cline (See table 8 for the results).

Table 8. Scheffé Test Results for Participants’ Writer’s View Task Performance on Tests with Varying Degrees of Familiar Propositional Content

The results of this analysis revealed that students at each proficiency level performed significantly different from students at any of the the other proficiency levels. However, there were two exceptions. First, the difference between the performance of non-proficient and semi-proficient students was not significant in the context of tests with varying degrees of familiar propositional content. Second, although proficient and fairly proficient students showed significant performance differences on tests with partially and totally familiar propositional contents at the 0.05 level, these differences were not significant at the 0.015 level in the context of partial text familiarity, and at the 0.025 level in the context of total text familiarity.
Analysis of variance was also conducted to determine if students’ proficiency level significantly affected their skimming task performances at any given level of text familiarity. The results of this analysis are reported in table 9. The results indicated that students at each proficiency level performed differently from students at any of the other proficiency levels at all of the three levels of text familiarity. However, the mean difference between semi-proficient and non-proficient students in the context of tests with totally unfamiliar propositional contents was not significant at the 0.002 level.

Table 9: Scheffé Test Results for Participants’ Skimming Task Performance on Tests with Varying Degrees of Familiar Propositional Content

Another set of analyses performed on the data concerned the effect of students’ proficiency level on their overall task performance. The sum of each students’ scores on a given task across the three TBRT modules (EM, GM, and AM) was taken as an indicator of his total score for that task. A main effect analysis of variance was then performed on the scores calculated in this way to see if students’ proficiency levels influenced their task performance. The results of this analysis are presented in table 10 below.
The results indicated that, in the case of the true-false task, students’ performances at each proficiency level differed significantly from those at any and all the other proficiency levels. In the case of the sentence-completion task, the results indicated that the performances of students at each proficiency level significantly differed from those at any of the other levels; in this case, the mean difference for non-proficient and semi-proficient students was not significant at the 0.023 level.

Table 10: Scheffé Test Results for Participants’ Overall Task Performance

Students’ performances on the outlining task also resembled their performances on the sentence-completion task. Students’ performances at each proficiency level differed significantly from those at any of the other levels. As for semi-proficient versus non-proficient subjects, the mean difference was significant at the 0.05 level but not at the 0.004 level. In the case of the 'writer's view' task, students’ overall performances at each proficiency level were significantly different from those at any of the other proficiency levels; however, the mean differences for semi-proficient and non-proficient students were not significant. Finally, the results of the analysis of variance for participants' overall performances of the skimming task indicated that students’ overall skimming task performances at each proficiency level were significantly different from those at any of the other proficiency levels (See table 10 above).

5.2. The Effect of Text Familiarity
Another assumption of the study was that test takers' degree of text familiarity could affect their test and task performances. Analyses of variance were performed to validate this assumption. The results indicated that text familiarity affected participants' overall test performance, and that the performance differences among students at the three levels of text familiarity were statistically significant. In other words, participants’ test performances at each level on the text-familiarity cline differed significantly from their test performances at any of the other text-familiarity levels. Table 11 presents the results of the post hoc Scheffé test for students’ overall test performances as they relate to the different degrees of text familiarity.

Table 11: Scheffé Test for Students’ Test Performance Across Different Levels of Text Familiarity

Students’ task performances at each text-familiarity level were also evaluated. The results of the analysis of variance for all of the five reading tasks under study revealed a significant difference in students’ performances on these tasks across different text-familiarity levels. Table 12 presents the results of this analysis.

Table 12: Scheffé Test for Students’ Task Performance Across Different Levels of Text Familiarity

5.3. The Effect of Task Type
Analyses of variance were also performed to determine if task type influenced students’ performances in the context of text familiarity. The aim of these analyses was to determine if students’ performances on different reading tasks were influenced differently when the tasks appeared in tests with (a) totally familiar contents (i.e., TBRT-EM), (b) partially familiar contents (i.e., TBRT-GM), and (c) totally unfamiliar contents (i.e., TBRT-AM).
In the first place, students’ performances on different tasks in the context of tests with totally familiar propositional contents were compared for significant differences. The results indicated that only students’ performances on the sentence-completion task differed significantly from their performances on the other tasks (i.e., true-false, outlining, writer’s-view, and skimming). Students’ performances on different tasks in the context of tests with partially familiar propositional contents were also compared. Here again, only the performances of students on the sentence-completion task differed significantly from their performances on each of the other tasks. Along the same lines, students’ performances on different tasks in the context of tests with totally unfamiliar propositional contents were also compared. Once more, the results indicated that students’ performances on the sentence-completion task differed significantly from their performances on each of the other tasks. Moreover, the difference between students’ performances on true-false versus writer’s-view tasks, though significant at the 0.05 level, were not significant at the 0.007 level in the context of tests with totally unfamiliar propositional contents (See table 13 below).

Analysis of variance was also performed to compare students' performances on different tasks over the whole text familiarity. The results of this analysis are presented in table 14 below. As table 14 shows, students’ performances on the sentence-completion task differed significantly from their performances on each of the other tasks. The differences between students’ performances of the true-false task and those of the skimming task, though significant at the 0.05 level, were not significant at the 0.002 level. Moreover, the differences between students’ performances on the true-false task and the outlining task, though significant at the 0.05 level, were not significant at the 0.027 level. The differences among students’ performances on the remaining tasks were not significant (See table 14).

Table 14: Scheffé Test fo Students’ Overall TBRT Task Performance

5.4. Interaction Analyses
Another assumption of this study was that the interaction bewteen any given pair and also all of the independent variables under study would be a source of variance in test results. Therefore, analyses of variance were performed to determine if the interactions between the independent variables of the study (i.e., language proficiency, text-familiarity, and task type) were responsible for variation in students’ performances of individual tasks on the one hand, and their overall TBRT test performances on the other. The results of interaction analyses are presented in tables 15 and 16.

Table 16: Interaction Analysis for Students’ Overall Test Performance

The results of interction analyses led to the following conclsions:

• The interaction between text familiarity and proficiency level led to significant differences in students’ performances on the true-false task.
• The interaction between text familiarity and proficiency level led to significant differences in students’ performances on the outlining task at the 0.05 level but not at the 0.029 level.
• The interaction between text familiarity and proficiency level did not lead to any significant difference in students’ performances on the sentence-completion, writer’s-view, and skimming tasks.
• The interaction between task type and text familiarity led to significant differences in students’ overall test performances.
• The interaction between task type and proficiency level led to significant differences in students’ overall test performances.
• The interaction between proficiency level and text familiarity led to significant differences in students’ overall test performances at the 0.05 level, but not at the 0.009 level.
• The interaction between text familiarity, task type, and proficiency level led to significant differences in students’ overall test performances at the 0.05 level, but not at the 0.027 level.

5.5. Regression Analyses
A more important aim of this study was to determine the relative impact of each of the independent vriables on students overall and differential test and task performances. The assumption of the study was that text familiarity was responsible for the greatest share of variance. The results of data analysis, however, rejected this assumption and revealed that language proficiency had by far the greatest share of variance. The second greatest share of variance belonged to task type. The smallest portion of variance was accounted for by text familiarity.
These conclusions resulted from a set of multiple regression analyses. The first regression analysis compared the relative impact of text familiarity and language proficieny on students’ overall test performances. It was found that language proficiency accounted for 79.5% of the variance whereas text familiarity only accounted for 18.6% of the variance. Moreover, the exclusion of text-familiarity did not affect the relative importance of language proficiency. In addition, the tolerances for proficiency and text familiarity were 01.00 and 01.00 respectively, suggesting that multi-collinearity was unlikely. In other words, the findings were not sample-dependent (See Bryman and Cramer, 1999, p. 263). Table 17 presents the results of this regression analysis.

Table 17: Regression Analysis for Overall Test Performance as the Dependent Variable

The second regression analysis took students’ performances on tests with different degrees of familiar propositional content as its dependent variable. In this case, too, language proficiency had the strongest relationship with the results. In the context of tests with totally familiar propositional contents, language proficiency accounted for 61.2% of the variance in comparison to task type (another independent variable of the study) which accounted for only 18.3% of the variance. Here again, the exclusion of ‘task type’ did not affect the impact of proficiency. Moreover, no evidence of multi-collinearity was observed. In the context of tests with partially familiar propositional contents, language proficiency and task type were found to take care of 61.2% and 15.7% of the variance, respectively. No fluctuation in the impact of language proficiency was observed due to the exclusion of task type from the analysis. Here again, the tolerances for language proficiency and task type were 01.00 and 01.00, respectively, indicating the lack of multi-collinearity. In the context of tests with totally unfamiliar propositional contents, too, the greatest share of variance belonged to language proficiency. While task type accounted for only 16.5% of the variance, language proficiency accounted for 61.2% of the variance. In addition, the impact of language proficiency did not fluctuate after the exclusion of task type from the analysis. No evidence of multi-collinearity was observed either. Table 18 presents the results of this analysis.

Table 18: Regression Analysis for Text Familiarity as the Dependent Variable

The relative impacts of text familiarity, task type, and language proficiency on students’ task performances were also studied. Once more, the greatest share of variance belonged to langauge proficiency. It accounted for 58% of the variance. Task type and text familiarity accounted for 15.9% and 14.1% of the variance respectively. The exclusion of either or both of the other variable(s) (i.e., task type text familiarity) from the analysis did not affect the importance of language proficiency. A more interesting finding was that task type had a greater share of variance than text familiarity. The results also indicated no evidence of multi-collinearity; the tolerances for language proficiency, task type, and text familiarity were 01.00, 01.00, and 01.00, respectively. Table 19 reports the results of regression analysis for task performance as the dependent variable.

Table 19: Regression Analysis for Task Performance as the Dependent Variable

Another regression analysis was conducted to study the relative impacts of text familiarity and language proficiency on students’ performance of each reading task. It was found that language proficiency had the strongest effect on task performance. In relation to the true-false task, language proficiency accounted for 73.4% of the variance while text familiarity accounted for 12.4% of the variance. In relation to the sentence-completion task, language proficiency was responsible for 57.8% of the variance while text familiarity accounted for only 20.9% of the variance. In connection to the outlining task, language proficiency was found to be in charge of 57% of the variance while text familiarity had a share of only 12.5% of the total variance. As for the writer’s-view task, language proficiency accounted for 54.8% of the variance and text familiarity for 14.1% of it. Finally, language proficiency accounted for 68.6% of the variance in relation to the skimming task whereas text familiarity accounted for only 16.6% of the variance. Table 20 presents the results of this analysis.

Table 20: Regression Analysis for Task Type as the Dependent Variable

Language proficiency had the greatest share of variance in connection to the true-false task and the smallest share in relation to the writer’s-view task. Text familiarity, on the other hand, had its maximum influence on the sentence-completion task and its minimum influence on the true-false task. The results of regression analysis for individual reading tasks did not indicate the existence of multi-collinearity. The tolerances for text familiarity and language proficiency in the context of each reading task were 01.00 and 01.00, respectively.

6. CONCLUSION
One of the major aims of this study was to determine if students’ level of proficiency resulted in any significant differences in their task performances at each specific point along the text-familiarity cline. The findings of the study indicated that students’ performances of the true-false and skimming tasks when the tasks appeared in a test with totally familiar propositional contents were a function of their level of proficiency. In the same context, the performances of only the semi-proficient students compared to the non-proficient participants did not show any meaningful difference on sentence-completion, outlining, and writer’s-view tasks. In the context of a reading test with partially familiar propositional contents, only the performance differences observed between semi-proficient and non-proficient students when performing true-false, sentence-completion, writer’s-view, and outlining tasks were not significant. Moreover, in the context of a reading test with totally unfamiliar propositional contents, only the performance differences observed between semi-proficient and non-proficient students when performing true-false, sentence-completion, outlining, and writer’s-view tasks were statistically significant.

A second aim of the study was to determine whether there was any meaningful relationship between students’ level of proficiency and their test performance in the context of a text-familiarity cline. The results of the present study indicated that students’ test performances were a function of their level of proficiency, no matter whether the propositional contents of the tests were totally familiar, partially familiar, or totally unfamiliar. In other words, at all points on the text-familiarity cline, proficiency affected students' test performances.

Moreover, the study aimed at finding out whether students’ level of proficiency affected their test performances regardless of the probable effects of text familiarity;  the results of the study also supported this contention. Students' proficiency levels affected their test performances when the tests consisted of a combination of totally familiar, partially familiar, and totally unfamiliar types of propositional contents.
The study also aimed at determining the probable impacts of students’ degrees of text-familiarity on their test performances. The results of the study supported the existence of such an influence; prior knoweldge of the propositional contents of reading tests affected students' performances on these tests positively. Test takers performed significantly better on tests with totally-familiar propositional contents. This finding lends credence to the existence of a text-familiarity cline. Moreover, the results indicated that the performances of students at each point on the text familiarity cline differed from their performances at each of the other points on the same cline. The results also revealed that complete and partial text-familiarity served as an advantage for students taking a reading test. This finding further supported the claims of Alderson and Urquhart (1985a, 1985b), and Clapham (1996).

Another aim of the study was to find out if text familiarity affected students' reading task performances. The results of the study indicated that students’ performances on a given task at a given point on the text-familiarity cline differed significantly from their performance on the same task at any other point on the text-familiarity cline. This finding also supported Clapham’s (1996) claims.

Another aim of the study was to explain how students’ level of language proficiency affected their performances on a given task across different text-familiarity levels. The results indicated that students’ performances on the true-false, outlining, and skimming tasks varied in accordance to their level of proficiency when these tasks appeared in tests with totally familiar, partially familiar, or totally unfamiliar propositional contents. However, the differences observed between the performances of the semi-proficient and non-proficient subjects on sentence-completion and writer’s-view tasks on tests of varying degrees of familiar propositional contents were not significant. This finding supported the reading threshold hypothesis (Clarke, 1980; Bernhardt and Kamil, 1995) that,  in other words, in order to be able to draw on prior knowledge (that is, to activate schemata), readers need to have already reached a specific level of language proficiency (a threshold level) to be able to disentangle themselves from the web of formal and structural features of the text.

The impact of task type on students’ test performances was also studied in the context of text familiarity. The aim was to determine if students’ performances on a given task was comparable to their performances on other tasks at the same text-familiarity level. The findings indicated that differences between the sentence-completion task and all the other tasks (true-false, outlining, writer’s-view, and skimming) were significant when these tasks appeared in tests with varying degrees of familiar propositiaonl contents. In addition, in tests with totally unfamiliar propositional contents, the differences between the true-false task and the writer’s-view task were also meaningful.

The impact of task type on students’ test performance was also studied in the context of students’ overall test performance (i.e., regardless of the text-familiarity cline). The differences between the sentence-completion and true-false tasks, on the one hand, and all the other tasks, on the other, were significant. The one-to-one comparisons of the remaining tasks also afforded significant results, but, there were three exceptions: (a) outlining versus writer’s-view, (b) outlining versus skimming, and (c) writer’s-view versus skimming. These comparisons afforded no significant results.

Steps were also taken to determine if the interaction between two or more of the independent variables of the study resulted in  any significant difference in students’ test and task performances. Task performance was studied in the context of the interaction between students’ degree of text-familiarity and level of proficiency. The results indicated that this interaction only affected students’ performances of the true-false and outlining tasks. The writer’s-view, sentence-completion, and skimming tasks were not influenced by this interaction. As for students’ overall test performance, the interaction between text familiarity and task type was significant. Students’ overall test performances were also affected by the interaction between text familiarity and language proficiency. Moreover, the interaction between task type and language proficiency caused a meaningful difference in students’ overall test performance. Finally, the interaction among text familiarity, task type, and language proficiency was an important source of variance in students’ overall test performance.

A comparison of the results of regression analyses reported in this study with the findings of Clapham’s (1996) study is intriguing indeed. While Clapham attaches greater importance to text familiarity (accounting for 38% of the variance) in comparison to language proficiency (accounting for 26% of the variance), the present investigation came up with somewhat different results; language proficiency did not appear to account for less than 50% of the variance in any of the comparisons made between any given pair of the independent variables under study in relation to students’ overall as well as differential test and task performances. Moreover, the very high tolerance indexes reported in this study reject any chance for multi-collinearity to occur. This indicates that the findings of the present study are far from being sample-dependent (See Bryman and Cramer, 1999, p. 254). Moreover, the effect of text familiarity on task performance was found to be smaller than the effect of task type. On these grounds, it can safely be argued that perhaps the development and use of LSP tests is out of consideration. As such, the results of this study are somewhat close to Lipson’s (1984) contention that LSP testing is not really justified. The greater impact of task type, in comparison to text familiarity, on students’ performances, however, stands against Lipson’s claims. The findings of the study indicated that, instead of giving students passages with esoteric propositional contents, it might be better to give them a rich variety of reading tasks, and to measure their performances on them.

The findings of the present study were all based on the 95% confidence interval. In other words, all of the significant findings reported above were significant at the 0.05 level (i.e., Alpha=0.05). Some of these findings are not significant if we modulate the confidence interval. The reader’s attention is specifically drawn to the following considerations:

• The interaction between text familiarity and language proficiency did not cause meaninful variation in students’ overall test performances at the 0.009 level.
• The interaction among text familiarity, language performance, and task type did not cause significant variation in students’ overall test performances at the 0.027 level.
• The interaction between text familiarity and students’ proficiency level was not a source of significant variation in students’ overall performances on the outlining task at the 0.029 level.
• The differences observed in the true-false task performances of proficient versus fairly proficient students on tests with totally familiar propositional content were not significant at the 0.001 level.
• The differences observed in the outlining task performances of proficient versus fairly proficient students on tests with partially familiar propositional contents were not significant at the 0.007 level.
• The differences observed in the outlining task performances of proficient versus fairly proficient students on tests with totally unfamiliar propositional contents were not significant at the 0.001 level.
• The differences observed in the outlining task performances of proficient versus fairly proficient students on tests with totally familiar propositional contents were not significant at the 0.002 level.
• The differences observed in the writer’s-view task performances of proficient versus fairly proficient students on tests with partially familiar propositional contents were not significant at the 0.015 level.
• The differences observed in the writer’s-view task performances of proficient versus fairly proficient students on tests with totally familiar propositional contents were not significant at the 0.025 level.
• The differences observed in the skimming task performances of non-proficient versus semi-proficient students on tests with totally unfamiliar propositional contents were not significant at the 0.002 level.
• The differences observed in the overall sentence-completion task performances of non-proficient versus semi-proficient students were not significant at the 0.023 level.
• The differences observed in the overall outlining task performances of non-proficient versus semi-proficient students were not significant at the 0.004 level.
• The differences observed in the students’ overall performances of the writer’s-view and true-false tasks were not significant at the 0.007 level.

Any interpretation of the findings of the present study should consider these points. In addition, the findings of this investigation were based on the performances of the students who took part in the study and should not be overgeneralized to the whole population of Iranian university students majoring in electronics.

REFERENCES
Alderson, J. C. (1988). Testing English for specific purposes: How specific can we get? In A. Hughes (Ed.), Testing English for university study (pp. 16-28). Modern English Publications and British Council.

Alderson, J. C. (1991). Language testing in the 1990s: How far have we come? How much farther have we to go? In S. Anivan (Ed.), Current developments in language testing (pp. 1-26). Singapore: Regional Language Centre.

Alderson, J. C. (2000). Assessing reading. Cambridge: Cambridge University Press.

Alderson, J. C., & Urquhart, A. H. (1983). The effect of student background discipline on comprehension: a pilot study. In A. Hughes, & D. Porter (Eds.). Current developments in language testing (121-127). London: Academic Press.

Alderson, J. C., & Urquhart, A. H. (Eds.). (1984). Reading in a foreign language.London: Longman.

Alderson, J. C., & Urquhart, A. H. (1985a). The effect of students’ academic discipline on their performance on ESP reading tests. Language Testing, 2, 192-204.

Alderson, J. C., & Urquhart, A. H. (1985b). This test is unfair: I’m not an economist. In P. C. Hauptman, R. Le Blanc, & M. B. Wesche (Eds.). Second language performance testing (pp. 25-45). University of Ottawa Press.

Allen, J. P. B., & Widdowson, H. G. (1974). Teaching the communicative use of English. International Review of Applied Linguistics. XII(I).

Alvermann, D. E., & Hynd, C. R. (1989). Effects of prior knowledge activation modes and text structure on non-science majors’ comprehension of physics. Journal of Education Research, 83, 97–102.

Bachman, L. F. (1990). Fundamental considerations in language testing. Oxford: Oxford University Press.

Bachman, L. F., & Palmer, A. S. (1996). Language testing in practice. Oxford: Oxford University Press.

Bernhardt, E. B., & Kamil, M. L. (1995). Interpreting relationships between L1 and L2 reading: Consolidating the linguistic threshold and linguistic interdependence hypotheses. Applied Linguistics, 16, 15–34.

Bryman, A., & Cramer, D. (1999). Quantitative data analysis with SPSS Release 8 for Windows: A guide for social scientists. London: Routledge.

Chambers, F. (1980). A re-evaluation of needs analysis. English for Specific Purposes, 1(1), 25-33.

Chen, H. C., & Graves, M. F. (1995). Effects of previewing and providing background knowledge on Taiwanese college students’ comprehension of American Short Stories. TESOL Quarterly, 29, 663–686.

Chitravelu, N. (1980). Introduction: English for special purposes project. In British Council, The University of Malaya English for Special Purposes Project (V-XVI).

Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Clapham, C. (1996). The development of IELTS: a study of the effect of background knowledge on reading comprehension. Cambridge: Cambridge University Press.

Clapham, C. (2000). Assessment for academic purposes: Where next? System, 28, 511–521.

Clarke, M. A. (1980). The short circuit hypothesis of ESL reading—or when language competence interferes with reading performance. Modern Language Journal, 64, 203–209.

Douglas, D. (2000). Assessing language for specific purposes. Cambridge: Cambridge University Press.

Douglas, D., & Selinker, L. (1993). Performance on a general versus a field-specific test of speaking proficiency by international teaching assistants. In C. Chapelle, and D. Douglas (Eds.)., A new decade of language testing research: Selected papers from the 1990 language testing research colloquium (pp. 235-256).Teachers of English to Speakers of Other Languages (TESOL).

Dudley-Evans, T. (1998). An Overview of ESP in the 1990s. Paper presented at The Japan Conference on English for Specific Purposes, Fukushima.

Erickson, M., & Molloy, J. (1983). ESP test development for engineering students. In J. Oller (Ed.), Issues in language testing research (pp. 280-288). New York: Newbury House.

Ewer, J. R., & Latorre, G. (1969). A course in basic scientific English. London: Longman.

Grellet, F. (1981). Developing reading skills. Cambridge: Cambridge University Press.
Halliday, M. A. K., McIntosch, A., & Strevens, P. (1964). The linguistic science and language teaching. London: Longman.

Hock, T. S. (1990). The role of prior knowledge and language proficiency as predictors of reading comprehension among undergraduates. In J. H. A. L. de Jong and D. K. Stevenson (Eds.), Individualizing the assessment of language abilities (pp. 214-244). New York: Multilingual Matters.

Hymes, D. (1971). Competence and performance in linguistic theory. In R. Huxley, & E. Ingram (Eds.), Language acquisition: models and methods (pp. 3-24). London: Academic Press.

Hymes, D. (1972). On communicative competence. In J. B. Pride, & J. Holmes (Eds.), sociolinguistics (pp. 269-292). Harmondsworth, UK: Penguin Books.

Hutchinson, T., & Waters, A. (1987). English for specific purposes: A learning centered approach. Cambridge: Cambridge University Press.

Jensen, C., & Hansen, C. (1995). The effect of prior knowledge on EAP listening-test performance. Language Testing, 12, 99-119.

Koh, M. Y. (1985). The role of prior knowledge in reading comprehension. Reading in a foreign language, 3, 375-380.

Lipson, M. Y. (1984). Some unexpected issues in prior knowledge and comprehension. The reading teacher, April, 760-764.

Munby, J. (1978). Communicative syllabus design. Cambridge: Cambridge University Press.

Nuttall, C. (1982). Teaching reading skills in a foreign language. London: Heinemann.

Osman, S. (1984). Effects of prior knowledge on ESL reading. In W. K. Byong (Ed.), Reading in Asia: The first yearbook of CCA (pp. 43-61). Hanyang University Press.

Papajohn, D. (1999). The effect of topic variation in performance testing: The case of the chemistry TEACH test for international teaching assistants. Language Testing, 16, 52–81.

Peretz, A. S., & Shoham, M. (1990). Testing reading comprehension in LSP. Reading in a Foreign Language. 7, 447-55.

Read, J. (1990). Providing relevant content in an EAP writing test. English for Specific Purposes, 9, 109–121.

Ridgway, T. (1997). Thresholds of the background knowledge effect in foreign language reading. Reading in a Foreign Language, 11, 151–168.

Rodgers, C. (1969). Freedom to learn. Columbus, OH: Charles Merrill.

Shoham, M., Peretz, A. S., & Vorhaus, R. (1987). Reading comprehension tests: General or subject specific? System, 15, 81-8.

Spolsky, B. (1973). What does it mean to know a language? Or, how do you get someone to perform his competence? In J. W. Oller, & J. Richards (Eds.), Focus on the learner: pragmatic perspectives for the language teacher (pp. 164-176).Rowley, MA: Newbury House.

Swales, J. (1971). Writing scientific English. Walton-on-Thames: Nelson.

Swales, (1985). Episodes in ESP. Oxford: Pergamon Institute of English.

Tan, S. H. (1990). The role of prior knowledge and language proficiency as predictors of reading comprehension among undergraduates. In J. De Jong, & D. Stevenson (Eds.), Individualizing the assessment of language abilities (pp. 214-224). Clevedon, UK: Multilingual Matters.

Tedick, D. J. (1990). ESL writing assessment: Subject-matter knowledge and its impact on performance. English for Specific Purposes, 9, 123–143.

University of Cambridge Local Examinations Syndicate (UCLES). (2000). The IELTS handbook. Cambridge: UCLES.