The knowledge base informing reading instruction/research in Applied Linguistics comes mainly from single-text reading research. To begin to develop a knowledge base informing research into and instruction of multiple-documents literacy, as the immediate current need of L2 students in academic contexts, this study was conducted. The study aimed to explore the contributions of language proficiency and prior knowledge to comprehension across multiple-texts and single-text modes of reading. To this end, a structural model was developed that subsumed hypothesized links between language proficiency, prior knowledge, single-text and multiple-texts reading comprehension. Data were collected from 207 Iranian undergraduate students studying midwifery and paramedical emergencies who took measures of language proficiency, single-text comprehension, multiple-texts comprehension, and a prior knowledge measure. The data were analyzed through structural equation modelling (SEM). The findings revealed that language proficiency and prior knowledge contribute to multiple-texts comprehension both directly and indirectly via impacting single-text comprehension. Moreover, the contributions of these two variables were found to vary across single-text and multiple-texts modes of reading. Implications for theoretical modelling and instruction of reading are presented.

INTRODUCTION

To date, the dominant paradigm guiding reading comprehension research and instruction, both in first language (L1) and particularly in second language (L2) reading contexts, has been the so-called ‘single-text paradigm’ of reading research that focuses exclusively on construction of meaning based on a single information source. This tradition of research has provided a rather sophisticated knowledge base to inform reading research, instruction, and practice. However, the nature of reading in the 21st century is argued to have substantively changed, drifting away from dependence on a single source as best expressed in Bråten, Ferguson, Anmarkrud and Strømsø’s (2013: 322) words:

For students learning academic content, relying on one single source of textual information, such as the science textbook, is increasingly replaced with such intertextual practices as searching for, selecting, processing, and synthesizing information from multiple texts, printed as well as digital.

What is expected of students in the present-day knowledge societies is multiple-documents comprehension defined as ‘the building of a coherent, meaningful representation of a situation or issue discussed across documents’ (Bråten et al. 2013: 880). On a routine basis, in fact, students are required to ‘integrate information from materials expressing diverse, even opposing, perspectives on an issue’ (Ferguson 2015: 731). This is motivated by easy access to an abundance of information documents not least thanks to the ubiquity of technological advancements—such as the Internet—in the present-day knowledge societies.

To probe any scientific issue, a student is routinely required to synthesize, at times conflicting, ideas and integrate pieces of information across a multitude of information sources (Gil et al. 2010) and to manage a meaningful mental representation of the issue from a ‘textual mélange’ (Wineburg 1998, 337). The failure to do so effectively is argued to have drastic consequences on both personal and societal levels (Bråten et al. 2011). On a personal level, lacking the ability to crisscross an issue from a variety of perspectives, the reader may develop a warped and biased position on the issue based on a single information document without assessing the evidentiary value of the position against the evidence presented in other information sources. On a societal level, the ability to form integrated conceptualizations of issues of social concern may prevent readers from ‘genuine participation in democratic discourse concerning their solution’ (Bråten et al. 2011: 49).

In spite of such importance attached to multiple-documents reading literacy in the successful academic functioning of students, much of the knowledge base guiding research and instruction in reading comes primarily from research on reading single texts. This single-text paradigm may not provide an adequate base to inform development of literacy in the context of multiple-documents reading. As posited by Linderholm et al. (2014: 335), ‘to understand how multiple text comprehension occurs and can be facilitated, it would be a mistake to simply apply what is known about reading single texts’. This could be attributed to the level of sophistication required in constructing the mental representation of a topic across multiple documents compared with that required in a single document. The mental representation in the context of multiple-texts comprehension rises above the integration dimension of the construction-integration model of text processing (Kintsch 1988), which argues for the integration of information in a single text with the reader’s prior knowledge to construct a situation model (Bråten and Strømsø 2009). When viewed in the context of multiple-texts comprehension, the situation model is referred to as the ‘documents model’, which is defined as a more ‘highly integrated situation model of the events described in the [multiple] texts’ (Bråten and Strømsø 2009: 3).

Despite the differences in the mental representational demands of the single-text comprehension and multiple-texts comprehension, the general trend of research within the field of applied linguistics has been, as in L1 reading contexts, dominated by the single-text paradigm of reading research. One of the findings from the research on reading in applied linguistics carried out within this single-text paradigm, concerns the relative contributions of language proficiency (LP), on the one hand, and prior knowledge (PK), on the other, to reading comprehension performance. Investigations into LP and PK specifically rode momentum when the view of reading as a purely language-based, bottom-up type of processing shifted to an integrated process involving both bottom-up and top-down processing (McNeil 2012). Reading comprehension, within this perspective, is defined as an interactive process which is primarily linguistic and then cognitive. ‘The processing of linguistic information is central to reading comprehension’ (Grabe 2009: 16) and provides the ground for the utilization of background knowledge that the reader brings to the text. In fact, the reader ‘actively constructs the meaning of the text by comprehending what the writer intends and by interpreting it in terms of the background knowledge activated by the reader’ (Grabe 2009: 15). Both of these processes give rise to inferences which assist the reader in more effective comprehension including ‘(a) within-text inferences that synthesize or summarize ideas found in the text and (b) elaborative inferences that combine concepts from the text with the readers’ prior-knowledge structures to create propositions’ (Barry and Lazarte 1998: 177). Motivated by this shift, a host of studies have examined the link between learners' PK, LP, and reading comprehension (Hammadou 1991; Clapham 1996: Droop and Verhoeven 1998; Chan 2003; Nassaji 2003, 2007; Brantmeier 2005; Usó-Juan 2006; Leeser 2007; McNeil 2011, 2012; among others). While mixed findings have been reported from these studies—which Usó-Juan (2006) attributes to the possible inaccuracies associated with the measurement of both independent and dependent variables—the general conclusion is that the two variables of LP and PK play critical roles in second language reading performance. The critical role of these two variables in comprehension is even more highlighted in reading for specific academic purposes because L2 content-area reading is where the interaction of proficiency and PK is more pronounced. As posited by Usó-Juan (2006), the major characteristic of L2 academic reading which differentiates it from other more general types of reading is the interplay between discipline-specific background knowledge and L2 proficiency.

This interaction is even more highlighted when the new multiple-documents-based nature of L2 students' reading requirements is taken into account. Even though multiple-texts reading literacy has not attracted adequate research attention in applied linguistics, it is, in actuality, the type of reading most L2 students are engaged in within their content-area reading environments. In the current academic environments, a key skill required of college students, including those studying through the medium of L2, is the ability to synthesize across a range of information documents and to critically evaluate the credibility of views across these documents against each other, especially when dealing with a complex scientific problem investigated from a variety of, sometimes contradictory, perspectives (Linderholm et al. 2014; Karimi 2015a). The semantic content integration required of a reader when dealing with multiple texts—especially those of a more scientific nature—should include ‘assertions unique to single texts, assertions upon which multiple texts’ authors agree, and instances where texts’ authors offer contradictory assertions’ (Braasch et al. 2016: 3). This makes the multiple-texts based semantic integration far more challenging than that based on a single text which includes only assertions within a single text. Prior knowledge has been reported as a strong predictor of construction of this multiple-texts based semantic representation (Bråten et al. 2009, 2008; Strømsø et al. 2010). In this multiple-texts-based content representation, the multiple sources are considered to be components incrementally contributing to the developing background knowledge of the readers (Strømsø and Bråten 2002). Thus, every text that the reader reads is integrated into his/her PK structure which might potentially affect reading other accompanying texts (Strømsø and Bråten 2002). This type of reading may, then, require a different PK contribution and, accordingly, a different LP contribution than understanding a single text. This is an issue which has not been investigated in research on reading in applied linguistics. Therefore, the present study aims to use structural equation modelling to examine the contributions of LP, PK, to single-text reading comprehension and multiple-texts reading comprehension.

LITERATURE REVIEW

L2 Proficiency, Prior Knowledge, and L2 Reading

As stated earlier, there has been a long-standing tradition of research within the field of applied linguistics establishing a clear link between LP, PK, and L2 reading comprehension. Different findings have been reported from this body of research. Some of the studies from this line of research have investigated LP and PK in conjunction and some other studies have probed one of these variables in conjunction with a different variable.

For example, Clapham (1996) investigated how discipline-specific background knowledge affected the reading performance of university English language learners. Based on the learners' performance on a grammar test, they were assigned to three groups of LP—low, medium, and high. Results of the study indicated that discipline-specific background knowledge did not exert a significant effect on the readers' comprehension performance at the highest and lowest levels of proficiency. Additionally, subjecting the scores from the three variables to multiple regression analyses, the researcher concluded that for subject-specific texts, LP accounted for 26 per cent of the total reading variance, while discipline-specific background knowledge accounted for only 12 per cent. The results further indicated that LP could be a stronger predictor of reading comprehension in an L2 than discipline-specific background knowledge.

Al-Shumaimeri (2006), along a similar line, investigated how second LP and content familiarity affected the text comprehension performance of Saudi university students while reading content-familiar and content-unfamiliar texts. Results of the study showed significant effects for both content familiarity and LP on the participants' comprehension performance. The results, moreover, demonstrated that knowledge of the content exerted a facilitating effect on reading comprehension, and that LP significantly affected the participants' comprehension performance in both content-familiar and content-unfamiliar texts. Results of the study further indicated a compensatory role for PK among the low-proficiency participants.

Additionally, in a study aimed at determining the contributions of discipline-related knowledge and LP to English-for-academic-purposes reading, Usó-Juan (2006) tested 380 L1-Spanish EFL learners' academic reading comprehension performance. The participants were at various levels of English proficiency and exhibited varying discipline-related knowledge of the topics being tested. Analyses of the data gathered revealed that both LP and discipline-related knowledge acted as significant predictors of reading English for academic purposes. In particular, it was found that LP accounted for a range between 58 to 65 per cent of the English-for-academic-purposes reading variance while discipline-related background knowledge accounted for a range between 21 to 31 per cent of the variance. The results also suggested that at intermediate and advanced levels, LP could compensate for lack of discipline-related background knowledge.

Still in another study, Alptekin (2006) investigated how culture-specific background knowledge influenced Turkish EFL learners' literal and inferential comprehension of an American short story. Results of the study revealed that when there was congruency between content of a text and PK of the readers, they tended to demonstrate a higher level of inferential comprehension. They further concluded that culture-specific background knowledge did not significantly influence the participant readers' literal comprehension.

In a more recent study, McNeil (2011) investigated the contributions of PK and reading strategies to reading comprehension among a total of 20 tertiary-level English language learners. Data collected from the participants were subjected to a pair of regression analyses. The results indicated that while the two variables, in combination, explained a significant portion (56.7 per cent) of variance in the participants' reading comprehension performance, reading strategies accounted for a significantly higher share of the variance (56.3 per cent) than background knowledge, which accounted for only .4 per cent of the observed variance.

As revealed by the review reported above, all studies investigating the relative contributions of PK and LP have been conducted in relation to single-text reading comprehension. No study, to the best of the present researcher's knowledge, has been conducted within the multiple-texts paradigm of reading comprehension.

Multiple-Documents Comprehension and Prior Knowledge

Since there has not been much research conducted on multiple-texts understanding in L2 reading contexts, naturally there is no research history investigating it in relation to LP. A major share of the research in this area has been conducted in L1 reading contexts mostly in relation to a number of variables such as task conditions (Britt and Sommer 2004), readers' epistemology (Strømsø et al. 2008), sourcing (Britt and Aglinskas 2002), among others. One variable which is thought to be firmly associated with performance on multiple-texts comprehension has been PK of the learners, studied both directly and indirectly—as a control variable—in research on multiple-texts understanding. A number of studies investigating variables other than PK have also controlled for the effects of this variable thus implicitly assuming that PK can affect multiple-texts comprehension (Bråten and Strømsø 2006a; Bråten and Strømsø 2010; Bråten and Strømsø 2009; Karimi and Shabani 2013; among others).

One study that has directly studied the effects of PK on multiple-texts understanding is Rouet, Favart, Britt and Perfetti (1997). The researchers in this study examined the role of history discipline expertise defined by them as ‘generalized knowledge of the methods and information sources typical of history problems’ (p. 85) on the participants' reading, use, and evaluation of multiple documents in history. The study participants included both history graduates (discipline experts) and psychology graduates (discipline novices) reading two controversies about the history of the Panama Canal. For each controversy, the participants were required to read a set of documents, develop an opinion essay about them, and evaluate their usefulness and trustworthiness. Results of the study suggested that discipline expertise helped history students construct effective links across multiple information sources and effectively tie their interpretations to the representation of the situation or problem described across the set of documents.

Le Bigot and Rouet (2007) is another study that focused on the link between PK and students' understanding of multiple texts. One of the aims of the study was to find out how PK might affect reading time and performance on multiple-texts reading comprehension. The participants were divided into low PK and high PK groups and were required to read a hypertext comprised of seven component texts about the topic of ‘social influences’. Results of the study revealed an advantage for the high-prior-knowledge students in reading time which was measured in seconds. Additionally, high-prior-knowledge students achieved significantly better scores in the comprehension post-test compared with their low-prior-knowledge counterparts. The better performance of the high-prior-knowledge students was much more noticeable in the macrostructure-based questions than the literal ones.

Strømsø et al. (2008) also investigated the relationship between a number of variables including gender, study experience, word decoding, PK, and topic-specific epistemological beliefs such as certainty of knowledge, simplicity of knowledge, source of knowledge, and justification for knowing and understanding of single and multiple documents. Intratextual and intertextual inference verification tasks were developed for measuring single-text and multiple-texts comprehension, respectively. Results of the hierarchical regression analyses showed that the above-mentioned variables accounted for a significant proportion of the variance in the participants’ performance in both single and multiple texts. PK was found to be a strong predictor of students’ single-text comprehension (β = 0.34) and multiple-texts comprehension (β = 0.28).

The role of PK in explaining the variance in single-text and multiple-texts comprehension was also confirmed in another study conducted by Bråten et al. (2009). The authors examined how source evaluation, trustworthiness ratings of the documents, and PK of the topics predicted Norwegian students’ comprehension of single and multiple documents. As in the previous study reported above, to measure comprehension of single and multiple texts, intratextual and intertextual inference verification tasks were developed. Again, hierarchical regression analyses were conducted with scores on the intratextual and intertextual verification measures as the dependent variables. The results showed that PK of the topic explained a significant share of the variance in both single-text and multiple-texts comprehension.

Also, in another study, Kobayashi (2009) explored the effects of three personal variables including PK of the topic, external strategy use, and college experience on Japanese undergraduate students' comprehension of intertextual relations among multiple controversial texts and recall of intratextual arguments. The results of the study attested to the positive effects of PK of the topic on the comprehension of intertextual relations among multiple texts through enhancing the efficient processing of intratextual arguments by the readers.

Strømsø et al. (2010) investigated how gender, memory for sources, PK, topic interest, and perceived comprehensibility predict performance on single-text and multiple-texts comprehension. To measure comprehension of single and multiple texts, intratextual and intertextual inference verification tasks were designed based on seven separate documents dealing with various aspects of climate change. PK of the topic was also measured through a 15-item multiple-choice test assessing participants’ knowledge of the concepts central to the topic of climate change. A couple of hierarchical regression analyses were run with scores on the intratextual and intertextual inference verification tasks as dependent variables. The results of the study showed that PK of climate change was a strong predictor of comprehension of single (β = 0.35) and multiple texts (β = 0.25) about climate change.

In a study more relevant to the purpose of the present study, Rydland et al. (2012) examined the relative contributions of word decoding skill, L1 and L2 vocabulary, and PK of the topic to L2-Norwegian reading comprehension across two reading tasks: a cloze test of reading comprehension and content-area reading comprehension involving multiple lengthy texts. The study participants were a group of language-minority Turkish and Pakistani fifth-graders. Results of the study revealed that word decoding and L2 vocabulary accounted for a significant share of the variance in the cloze test but a smaller portion of variance in the content-area reading using multiple texts for which PK was the strongest predictor.

THE HYPOTHESIZED MODEL

As shown by this review of the literature, research on multiple-documents comprehension has been conducted mainly in L1 reading contexts. The essential lack of work on this type of reading literacy in L2 reading contexts highlights an obvious need to carry out further research in this area. Thus, the present study sets out to use structural equation modelling to examine the direct and mediated effects (through single-text comprehension) of LP and PK on multiple-documents comprehension. To this end, a structural model of the hypothesized relationships between the above-mentioned variables was formed. The directional links between the variables are presented in Figure 1.

Figure 1:

A Hypothesized Model of LP, PK, Single-Text Comprehension, and Multiple-Texts Comprehension. Note: N = 207; STC = Single-Text Comprehension; MTC = Multiple-Texts Comprehension

Figure 1:

A Hypothesized Model of LP, PK, Single-Text Comprehension, and Multiple-Texts Comprehension. Note: N = 207; STC = Single-Text Comprehension; MTC = Multiple-Texts Comprehension

The paths from LP and PK to single-text comprehension are supported by sufficient empirical evidence in L2 literature on reading (e.g. Clapham 1996; Al-Shumaimeri 2006; Usó-Juan 2006; McNeil 2011). The path from PK to multiple-texts comprehension is supported by findings of studies in L1 reading contexts (e.g. Le Bigot and Rouet 2007; Kobayashi 2009; Bråten et al. 2009). Although there was no study in the literature documenting the link between LP and multiple-documents comprehension, this directional link can be supported by the assumption that any type of reading is affected by proficiency of the readers. Finally, the path from single-text comprehension to multiple-text comprehension is supported by the assumption that content representation based on multiple texts include both assertions within a single text and assertions across multiple texts and, in fact, multiple-documents reading incorporates single-text reading.

METHOD

Participants

The study participants were 207 Iranian Midwifery and Paramedical Emergencies undergraduate students who ranged in age from 18 to 29, with an average of 21.5 years. There were 102 females (49.27 per cent) and 105 males (50.73 per cent). Since Iranian university students often need to read some required or supplementary content-area materials in English, they try to improve their English proficiency as they progress through their university education, either through self-training or attending private language institutes. Therefore, from the first through the second to later years of studies, students with varying levels of English proficiency can be found, a point best suited to the purpose of the present study. Therefore, attempts were made to use a purposive sampling procedure (maximum variation sampling) to select participants with varied years of studies with a proficiency range of elementary to upper-intermediate levels. This issue also fulfilled the researcher's expectation to recruit participants with varying levels of PK of the topic of the texts because, by virtue of being in various years of studies, the recruited participants possessed varying levels of background knowledge regarding technical topics in their content areas. The reason for including Paramedical Emergencies students in the participants was to further make certain of the varying levels of PK of the participants.

Materials

The materials used in the study included a TOEFL-like test of LP, two prior-knowledge tests, a multiple-texts-based intertextual inference verification task, and an intratextual inference verification task.

Intertextual inference verification task

One of the measures used to explore the aim of the study was an intertextual inference verification task developed based on four passages about ‘Multi-Fetal Pregnancy’, which discussed various aspects of the condition. The information about the texts are presented in Table 1.

Table 1:

Information about the texts

Text number Word count Text type Content focus 
957 Technical; from a textbook English for the Students of Midwifery (Alizadeh 2009Fraternal and identical twins, sex ratio with multiple pregnancy, diet, diagnosis, outcome, and management of multiple pregnancies, etc. 
859 Semi-technical Zygosity, chorionicity, and the problems associated with multi-fetal pregnancies 
875 Popular science text The types, causes, symptoms, diagnosis, complications of the condition, and the cares which a multi-fetal pregnant woman needs to take 
1,255 Popular science text Epidemiology of the condition, the risk factors associated with it, presentation of the fetuses, vaginal and operative modes of delivery, complications, and prevention of the phenomenon 
Text number Word count Text type Content focus 
957 Technical; from a textbook English for the Students of Midwifery (Alizadeh 2009Fraternal and identical twins, sex ratio with multiple pregnancy, diet, diagnosis, outcome, and management of multiple pregnancies, etc. 
859 Semi-technical Zygosity, chorionicity, and the problems associated with multi-fetal pregnancies 
875 Popular science text The types, causes, symptoms, diagnosis, complications of the condition, and the cares which a multi-fetal pregnant woman needs to take 
1,255 Popular science text Epidemiology of the condition, the risk factors associated with it, presentation of the fetuses, vaginal and operative modes of delivery, complications, and prevention of the phenomenon 

The texts were from different sources and had a fair amount of information overlap but also contained conflicting information regarding some aspects of the condition. They discussed the condition in slightly different ways employing rather different terminologies and from various aspects which required the readers to integrate and synthesize ideas across them to provide responses to the intertextual questions based on them. The texts were printed on separate sheets of paper and the participants were free to read them in any order they preferred.

The task included 20 statements designed by connecting bits of information across the texts to form an either valid or invalid inference. A statement was assumed to be valid if the idea expressed in it could be inferred from connecting pieces of information across the multiple texts and invalid if the idea in the sentence could not be inferred by synthesizing ideas from the texts. The participants were required to read the texts and mark the inference statements as either right or wrong; their scores were computed based on the number of correct responses to the items. The measure included 14 valid and 6 invalid inferences. Lest the task might not function as a simple knowledge test, the instructions highlighted the necessity of referring to the texts to provide answers to the questions. It was stated in the instructions that the texts may sometimes contain contradictory information; therefore, although some items might seem informationally correct or incorrect, they should be inferred across at least two of the texts. Below is a sample valid inference which should be drawn by synthesizing pieces of information across three of the texts:

About sixty-five percent of all twins are fraternal who develop from a couple of ova and are all dichorionic-diamniotic with the same genetic structure.

In comparison, the following item is an example of an invalid inference which could not be drawn by bridging ideas from the three of the texts.

Knowing if the mother has undergone drug consumption for conceiving is not as effective in predicting multiple pregnancies as knowledge of the family history of twins on the maternal side.

The KR-20 reliability for the scores on the measure with the study sample was calculated to be .83.

Intratextual inference verification task

Based on the 957-word text described above, which was taken from a technical textbook, English for the Students of Midwifery (Alizadeh 2009), an intratextual inference verification task was developed. The task included 10 statements constructed by combining pieces of information across sentences in the text to form an either valid or invalid inference. A sentence was assumed to be valid if the idea expressed in it could be inferred from bridging information bits across sentences in the text and invalid if the idea in the sentence could not be inferred by synthesizing ideas from the sentences. As with the intertextual inference verification task, to keep the test from functioning as a simple knowledge test, the instructions emphasized referring to the texts to confirm or reject the inference statements. The instructions clearly stated that although some items might seem informationally correct or incorrect, they should be inferred across the sentences in the text. Examples of valid and invalid inferences are provided below:

Valid: Unlike twins developed from two ova in a single ovulatory cycle, identical twins develop within a common amnionic sac.

Invalid: Because of increased maternal blood volume, an experienced obstetric attendant should accompany the multi-fetal pregnant mother throughout labor.

The participants marked the statements as either right or wrong, and their scores were computed based on the number of correct responses to the items. The measure included six valid and four invalid inferences. The KR-20 reliability for the scores on the measure with the study sample was calculated to be .81.

PK measure

To assess the participants' PK of the topic of the texts, a PK measure related to ‘Muti-Fetal Pregnancy’ was developed. The test included 30 four-choice items testing the participants' knowledge of the basic concepts related to the condition including symptoms, causes, types of twinning, management and risk factors, etc. The test was in Farsi, the participants' native language. Attempts were made that the items on the test adequately cover the various aspects of the topic discussed in the texts based on which the inference verification tasks were developed. The participants were required to take the test in 15 minutes and their scores were computed based on the number of correct responses to the items of the test. The Cronbach’s Alpha reliability for the test with the sample in the present study was calculated to be .83.

English LP test

To measure the participants' English proficiency, a TOEFL-like proficiency test was used. This type of test, which is a shortened form of TOEFL, is used mostly in Iranian contexts for entrance into post-graduate studies; it undergoes rigorous psychometric scrutiny and enjoys high validity and reliability. Although the test does not cover oral/aural aspects of language, it is assumed to be a reasonable indicator of test-takers' general language knowledge. The test used in the present study included 40 items of grammar, 40 items of vocabulary, and 20 items of reading comprehension. All items were followed by four response options from which the participants were required to select the correct one. The reading items included various item types including inference, negative factual information, author’s purpose, essential information, explicitly stated factual details, and reference. The Cronbach’s Alpha reliability for the test with the present sample was calculated to be .82.

Procedures and Data Analysis

As explained in the previous section, to explore the research question of the study, there were four tests the participants were required to take: a LP test, a multiple-texts-based intertextual inference verification task as a measure of multiple-texts comprehension, a single-text-based intratextual verification task as a measure of single-text comprehension, and a PK test of ‘Multi-Fetal Pregnancy’. To reduce test fatigue, the participants took the tests in three testing sessions. In the first session, the LP test and the PK test of ‘Multi-Fetal Pregnancy’ were administered to the 207 participants. The time allocated to both of the tests was 75 minutes; the participants took the LP test in one hour and the PK test in 15 minutes. During the second testing session, the participants took the intratextual inference verification task. The time allowed for this task was 40 minutes. In the third testing session, the participants took the intertextual inference verification task. The time allowed for this task was 75 minutes. No exam paper was collected till the end of the allotted time. It should also be pointed out that to control for the effect of memory on the participants’ performance in both intratextual and intertextual verification tasks, they had access to the texts while answering the questions. The scores from the four measures provided the data for the purpose of investigating the contributions of PK and LP to multiple-texts reading comprehension compared with single-text reading comprehension. The analyses were done through SEM using LISREL 8.8. According to Usó-Juan (2006), earlier studies investigating the contributions of background knowledge and/or language proficiency to reading comprehension have used statistical techniques such as correlation (Alderson and Urquhart 1983, 1985), one-way ANOVA (Koh 1985), two-way ANOVA (Chen and Donin 1997), among others. She argues that these techniques are carried out by groups not individuals and these studies ‘thus missed a lot of information’ (Usó-Juan 2006: 211). She suggests using other statistical procedures such as regression, which uses individual data. In the present study, SEM was used because it incorporates regression and also because the researcher intended to explore the mediating effects of LP and PK on multiple-documents comprehension through single-text comprehension

RESULTS

The major objective of the study, as discussed above, was to test the hypothesized structural model of the relationships between LP, PK, single-text and multiple-texts reading comprehension. To this end, data were gathered based on the performance of the 207 participants on a set of four measures including the LP test, the PK test of ‘Multi-Fetal Pregnancy’, the intratextual inference verification task, and the intertextual inference verification task. To present a clear picture of the participants’ performance on these four measures, the mean score and standard deviation of the participants’ performance on each of the measures and the correlations among the variables are provided in Table 2:

Table 2:

Correlations among and descriptive statistics (mean and SD) of the variables

Variables M (SDLP PK STC MTC 
LP 37.08 (12.31) 1.00 0.01 0.68 0.63 
PK 16.14 (8.46)  1.00 .26 0.40 
STC 5.06 (1.61)   1.00 0.83 
MTC 9.93 (3.31)    1.00 
Variables M (SDLP PK STC MTC 
LP 37.08 (12.31) 1.00 0.01 0.68 0.63 
PK 16.14 (8.46)  1.00 .26 0.40 
STC 5.06 (1.61)   1.00 0.83 
MTC 9.93 (3.31)    1.00 

Note: STC = Single-Text Comprehension; MTC = Multiple-Texts Comprehension.

As mentioned above, a structural equation modelling analysis was used to test the hypothesized model and to identify how single-texts comprehension mediates the effects of LP and PK on multiple-texts comprehension. As there were more parameters to be identified in the model than the data points, LISREL yielded a saturated model with a χ2 value of 0.00 and a root mean square error of approximation (RMSEA) of 0.00 while a value of ≤.05 is considered to be an acceptable indicator of model fitness. A schematic representation of the yielded model, together with standardized path coefficients, is presented in Figure 2.

Figure 2:

The Structural Model of LP, PK, Single-Text Comprehension, and Multiple-Texts Comprehension. Note: N = 207; STC = Single-Text Comprehension; MTC = Multiple-Texts Comprehension

Figure 2:

The Structural Model of LP, PK, Single-Text Comprehension, and Multiple-Texts Comprehension. Note: N = 207; STC = Single-Text Comprehension; MTC = Multiple-Texts Comprehension

As shown in Figure 2, the paths from LP and PK to single-text comprehension yielded standardized regression weight loadings amounting to .68 and .26, respectively. The standardized regression weight leading for LP shows that it is a stronger causal path than PK. These two paths were statistically significant (p < .05).

The model also reveals that LP and PK have direct effects on multiple-texts comprehension and indirect effects on it through single-text comprehension. The total effect of LP on multiple-texts comprehension was computed using the product of the coefficients in the mediation chains plus the coefficient for the unmediated path from LP to multiple-texts comprehension (i.e. .68 ×. 64 = .43 + .19 = .62). Likewise, the total effect of PK on multiple-texts comprehension was computed using the product of the standardized coefficients in the mediation chains plus the coefficient for the direct path from PK to multiple-texts comprehension (i.e. .26 × .64 = .16 + .23 = .39). Dividing the product of the coefficients in the mediation chains for LP by its total effect (i.e. .43/.62) revealed that 69 per cent of the effect of LP on multiple-documents comprehension is mediated by single-text comprehension. Likewise, dividing the product of the coefficients in the mediation chains for PK by its total effect (i.e. .16/.39) showed that 41 per cent of the effect of PK on multiple-texts comprehension is mediated by single-text comprehension. These computations and a comparison of the standardized coefficients in the unmediated paths from LP and PK to multiple-texts comprehension (.19 and .23, respectively) point to a higher direct effect of PK on multiple-texts comprehension. It should also be pointed out that all the direct and mediated paths from LP and PK to multiple-texts comprehension were statistically significant (p < .05).

DISCUSSION

Results of the study suggest that both LP and PK can play crucial roles in both single-text and multiple-texts types of reading comprehension. Regarding their relative importance, the findings further indicate that LP could be a stronger predictor of both types of reading than PK. However, much of the contribution of LP to multiple-documents comprehension is mediated via impacting single-text comprehension. On the other hand, a smaller share of the contribution of PK to multiple-texts comprehension is mediated through single-text comprehension and a larger share of it is unmediated.

The findings extend L2 reading research by further illuminating the role played by PK in single-text reading comprehension, which has been supported by inconclusive evidence and mixed research findings. The study attempted to rectify some of the problems in earlier research. For example, unlike earlier studies employing statistical tests such as correlation, one-way and two-way analyses of variance, etc. with groups and thus missing a significant amount of information (Usó-Juan 2006), the present study carried out SEM with individuals with wide ranges of LP and PK. Additionally, it employed texts with adequate specificity and ‘suitably specific content in an intended area’ (Usó-Juan 2006: 221). Moreover, in the study, PK was operationalized as direct knowledge of the topic of the texts and not as more general forms of knowledge. With these problems addressed, the results supporting the contributions of PK and L2 knowledge to single-text comprehension are consistent with those of studies that assume a role for these two variables in reading comprehension (Clapham 1996; Pulido 2004; Al-Shumaimeri 2006; Usó-Juan 2006; Leeser 2007; among others). Specifically, the findings are consistent with those of studies (e.g. Tan 1990; Usó-Juan 2006) pinpointing the fact that the variance explained by LP is double the variance explained by PK. In the above-mentioned studies, the beta coefficients for LP were reported to be almost twice as large as those for PK. The findings, however, are inconsistent with those of studies assuming that background knowledge accounts for little variance in reading comprehension (e.g. McNeil 2011). The portion of variance LP explains in single-text L2 reading in the present study, as revealed by the regression weight loadings on the final SEM model, is also higher than that reported in many of the previous studies except for Usó-Juan (2006).

However, a striking finding of the study is that these two variables, PK in particular, contribute differentially to single-text vs. multiple-texts types of reading comprehension as shown by the standardized path coefficients in the SEM model. PK was shown to account for far more variance in multiple-texts comprehension (39 per cent) than in single-text comprehension (26 per cent). A share of it is, however, as stated earlier, mediated through single-text comprehension. This finding is supported by evidence from research on both single-text and multiple-texts reading comprehension. Part of the evidence supporting this finding comes from studies on students reading single texts with varying levels of coherence and the acknowledged interaction of PK and coherence in the comprehension process reported in this line of research (McNamara et al. 1996; Boscolo and Mason 2003). For example, McNamara et al. (1996) found that high-prior-knowledge students were even able to benefit from less coherent texts while low-prior-knowledge students benefited most from maximally coherent texts. Multiple-texts reading could be considered a type of reading involving texts with minimal coherence (Bråten and Strømsø 2006a) because a reader is required to maintain coherence over a series of texts which may, at times, involve conflicting information and may not necessarily follow equally coherent lines of argument. In maintaining coherence over such texts, PK plays a crucial role (Bråten and Strømsø 2006a). The larger unmediated share of the contribution of PK to multiple-texts comprehension also supports this claim, as the employment of PK in relation to single texts, which are often more coherent than a set of texts, seems to be different from the employment of PK in relation to multiple texts. To maintain coherence over a set of texts, the readers need to be equipped with PK that can act as a strong adhesive connecting the texts together. Additionally, part of the evidence supporting this finding comes from L2 reading research which suggests that PK can significantly assist readers to tackle more linguistically challenging texts (McNeil 2012). Dealing with a set of texts on the same topic is certainly a far more demanding task than dealing with a single text (Bråten and Strømsø 2006a; Bråten and Strømsø 2011).

Additionally, the higher total effect of PK on multiple-texts comprehension than on single-text comprehension and the larger share of its unmediated effect on multiple texts comprehension than the mediated share attest to the differences in the application of PK in constructing a mental representation based on a singly presented text compared with that based on a set of texts. The mental representation of content across multiple documents is a higher-order representation than that based on a single text and calls for a more influential role of PK. This mental representation in multiple-texts comprehension goes a step above the integration component of the construction-integration model of text processing (Kintsch 1988), which emphasizes the integration of information in a text with PK to construct a situation model (Bråten and Strømsø 2009), which is defined as the integration of the information from the text with information from the reader's long-term memory (Bråten et al. 2009). The situation model based on multiple texts is called a ‘documents model’ (Britt et al. 1999), which is a more ‘comprehensive representation of information from multiple documents’ (Stadtler and Bromme 2013: 124), which may require a more sophisticated knowledge mechanism. Such a model includes ‘content information presented in a single text as well as either concordant or conflicting information stemming from different texts that is integrated with individual existing prior knowledge’ (Jucks and Paus 2013: 227–8). In constructing such a model, PK equips the multiple-texts readers with a top-down tool to piece together information scattered across different documents and allows them to go beyond and above what is explicitly stated in the documents to generate effective inferences across the documents (Rydland et al. 2012).

CONCLUSION AND IMPLICATIONS OF THE STUDY

Up to the present day, almost all studies on the contributions of different variables, including L1 reading ability, strategic thinking, language competence, and PK, to L2 reading have been conducted in the context of single-text reading comprehension. However, the excessive demands the current academic knowledge societies place upon readers in terms of integrating knowledge components across various sources and combining pieces of information from these sources, more research is required to further explore the various dimensions of this emerging type of literacy to build a knowledge base to inform reading instruction (Karimi and Alibakhshi 2014; Karimi 2015a). The need for further research into this type of reading literacy is recognized even further when we find out that ‘many students ranging from elementary to college level often have difficulties with this task’ (Bråten and Strømsø 2009: 25). Surprisingly however, very little research, if any, has been conducted on this type of literacy in L2 reading contexts. As Karimi (2015b: 54) states—

Although multiple documents literacy is currently the high priority requirement of EFL/ESL students in their academic environments, instructional programs continue to be focused around single text reading literacy. This continued focus on single text reading is rooted either in EFL/ESL professionals’ ignorance of the multiple documents-based intertextual reality of meaning construction in present day knowledge societies or in the naïve assumption that what is done in single text literacy-oriented instructional programs is automatically transferred to the multiple texts reading literacy.

Influenced by this line of thinking, little research, if any, has comparatively explored the contributions of LP and PK to single-text as compared with multiple-texts reading comprehension and the mediation effects of single-text comprehension. This is perhaps rooted in the erroneous assumption that the contributions of LP and PK are similar across types—or levels—of reading comprehension. Although the findings of the present study point to the effects of LP and PK on multiple-documents comprehension through the mediational role of single-text comprehension, the findings also show effects of LP and PK that are unmediated and unique to multiple-texts comprehension. This unmediated effect is particularly noticeable with reference to PK. These findings provide implications for any theoretical model of reading to take into account the differential contributions of LP and PK in its conceptualization of single-text vs. multiple-texts reading comprehension. It should not be assumed that whatever research in single-text tradition of research in L2 has yielded could be equally applicable to the emerging intertextual reality of present-day academic environments. Additionally, the findings of this study provide suggestions for L2 reading instructional programs to consider developing multiple-texts reading, which have been relatively absent, in instructional courses at relatively advanced levels of proficiency to help L2 students tackle the present-day reading requirements they are often faced with in their academic environments. In this process, they should be aware of the vital role of LP and PK for more successful academic functioning in the intertextual reading requirements facing them in present-day knowledge societies. The findings also imply that reaping benefits associated with multiple-documents comprehension might be best reached if this kind of reading comprehension is introduced when L2 learners have been adequately initiated into their academic disciplines and have gained sufficient discipline-related knowledge to facilitate this kind of comprehension.

LIMITATIONS OF THE STUDY

As in any research endeavor, the limitations of the present study should be acknowledged, and the conclusions drawn should be viewed within the constraints imposed on the study. First, the present study focused only on technical texts related to one specific field of study, Midwifery. To provide for adequate generalization of the findings to other texts and topics from other disciplines, further research can be carried out on technical texts from other disciplines or other topics. Second, investigating less scientific texts than the ones used in the present study can show whether the findings can be corroborated with texts with a less academic focus. Third, the nature of the proficiency test which was multiple-choice could be cited as a possible limitation of the study. Use of a more comprehensive proficiency test such as TOEFL or IELSTS could have yielded a better picture of the proficiency level of the participants. Moreover, PK was defined as knowledge of the topic of the texts. PK could have been conceptualized more broadly to include background knowledge of the discipline. Still another limitation of the study could be the lack of information on how long the participants actually spent on the tasks since some of them might have finished working on the tasks before the exam papers were collected.

Mohammad Nabi Karimi holds a PhD in Applied Linguistics and is currently working as an assistant professor in the Department of Foreign Languages, Kharazmi University, Tehran, Iran, where he teaches BA, MA, and PhD courses in applied linguistics and TEFL. His areas of interest include L2 reading, teacher/learner cognition, and teachers’/learners’ beliefs system. He has published on these areas in international journals including Modern Language Journal, System, Australian Journal of Teacher Education, etc. Address for correspondence: Department of Foreign Languages, Kharazmi University, No 43, Mofatteh Street, Tehran, Iran. <karimi_mn@khu.ac.ir>

Acknowledgements

I wish to thank Carita Paradis (Lund University, Sweden) for her valuable advice and help in editing and proofreading the manuscript. I would also like to thank the Editorial Board of Applied Linguistics and three anonymous reviewers for their constructive and thought-provoking comments on earlier versions of the article.

Conflict of interest statement. None declared.

References

Alderson
J. C.
Urquhart
A. H.
.
1983
. ‘The effect of student background discipline on comprehension: A pilot study. In
Hughes
A.
Porter
D.
(Eds.),
Current developments in language testing
  (pp.
121
7
).
London
:
Academic Press
.
Alderson
J. C.
Urquhart
A. H.
.
1985
. ‘
The effect of students’ academic discipline on their performance on ESP reading tests
,’
Language Testing
 
2
:
192
204
.
Alizadeh
P.
2009
.
English for the Students of Midwifery
 .
Boshra Publications
.
Alptekin
C.
2006
. ‘
Cultural familiarity in inferential and literal comprehension in L2 reading
,’
System
 
34
:
494
508
.
Al-Shumaimeri
Y.
2006
. ‘
The effects of content familiarity and language ability on reading comprehension performance of low- and high-ability Saudi tertiary students studying English as a foreign language
,’
Education, Science, and Islamic Studies
 
18
:
1
19
.
Barry
S.
Lazarte
A.
.
1998
. ‘
Evidence for mental models: How do PK, syntactic complexity, and reading topic affect inference generation in a recall task for nonnative readers of Spanish?,
Modern Language Journal
 
82
:
176
93
.
Boscolo
P.
Mason
L.
.
2003
. ‘
Topic knowledge, text coherence, and interest: How they interact in learning from instructional texts
,’
Journal of Experimental Education
 
71
:
126
48
.
Braasch
J. L. G.
McCabe
R. L.
Daniel
F.
.
2016
. ‘
Content integration across multiple documents reduces memory for sources
,’
Reading and Writing
 
29
:
1571
1598
. DOI 10.1007/s11145-015-9609-5
Brantmeier
C.
2005
. ‘
Effects of reader's knowledge, text type, and test type on L1 and L2 reading comprehension
,’
Modern Language Journal
 
89
:
37
53
.
Bråten
I.
Britt
M. A.
Strømsø
H. I.
Rouet
J. F.
.
2011
. ‘
The role of epistemic beliefs in the comprehension of multiple expository texts: Towards an integrated model
,’
Educational Psychologist
 
46
:
48
70
.
Bråten
I.
Ferguson
L. E.
Anmarkrud
Ø.
Strømsø
H. I.
.
2013
. ‘
Prediction of learning and comprehension when adolescents read multiple texts: The roles of word-level processing, strategic approach, and reading motivation
,’
Reading and Writing
 
26
:
321
48
.
Bråten
I.
Ferguson
L. E.
Strømsø
H. I.
Anmarkrud
Ø.
.
2013
. ‘
Justification beliefs and multiple-documents comprehension
,’
European Journal of Psychology of Education
 
28
:
879
902
.
Bråten
I.
Strømsø
H. I.
.
2006a
. ‘
Effects of personal epistemology on the understanding of multiple texts
,’
Reading Psychology
 
27
:
457
84
.
Bråten
I.
Strømsø
H. I.
.
2009
. ‘
Effects of task instruction and personal epistemology on the understanding of multiple texts about climate change
,’
Discourse Processes
 
47
:
1
31
.
Bråten
I.
Strømsø
H. I.
.
2010
. ‘
When law students read multiple documents about global warming: Examining the role of topic-specific beliefs about the nature of knowledge and knowing
,’
Instructional Science
 
38
:
635
57
.
Bråten
I.
Strømsø
H. I.
.
2011
. ‘
Measuring strategic processing when students read multiple texts
,’
Metacognition and Learning
 
6
:
111
30
.
Bråten
I.
Strømsø
H. I.
Britt
M. A.
.
2009
. ‘
Trust matters: Examining the role of source evaluation in students’ construction of meaning within and across multiple texts
,’
Reading Research Quarterly
 
44
:
6
28
.
Bråten
I.
Strømsø
H. I.
Samuelstuen
M. S.
.
2008
. ‘
Are sophisticated students always better? The role of topic-specific personal epistemology in the understanding of multiple expository texts
,’
Contemporary Educational Psychology
 
33
:
814
40
.
Britt
M. A.
Aglinskas
C.
.
2002
. ‘
Improving students’ ability to identify and use source information
,’
Cognition and Instruction
 
20
:
485
522
.
Britt
M. A.
Sommer
J.
.
2004
. ‘
Facilitating textual integration with macro-structure focusing tasks
,’
Reading Psychology
 
25
:
313
39
.
Britt
M. A.
Perfetti
C. A.
Sandak
R.
Rouet
J. F.
.
1999
. ‘Content integration and source separation in learning from multiple texts’ in
Goldman
S. R.
Graesser
A. C.
van den Broek
P.
(eds):
Narrative Comprehension, Causality, and Coherence: Essays in Honor of Tom Trabasso
 .
Erlbaum
, pp.
209
33
.
Chan
Y. H.
2003
. ‘
Cultural content and reading proficiency: A comparison of mainland Chinese and Hong Kong learners of English
,’
Language, Culture and Curriculum
 
16
:
60
9
.
Chen
Q.
Donin
J.
.
1997
. ‘
Discourse processing of first and second language biology texts: Effects of language proficiency and domain-specific knowledge
,’
Modern Language Journal
 
81
:
209
27
.
Clapham
C.
1996
.
The Development of IELTS: A Study of the Effect of Background Knowledge on Reading Comprehension
 .
Cambridge University Press
.
Droop
M.
Verhoeven
L.
.
1998
. ‘
Background knowledge, linguistic complexity, and second-language reading comprehension
,’
Journal of Literacy Research
 
30
:
253
71
.
Ferguson
L.
2015
. ‘
Epistemic beliefs and their relation to multiple-text comprehension: A Norwegian program of research
,’
Scandinavian Journal of Educational Research
 
59
:
731
52
.
Gil
L.
Bråten
I.
Vidal-Abarca
E.
Strømsø
H. I.
.
2010
. ‘
Understanding and integrating multiple science texts: Summary tasks are sometimes better than argument tasks
,’
Reading Psychology
 
31
:
30
68
.
Grabe
W.
2009
.
Reading in a Second Language: Moving from Theory to Practice
 .
Cambridge University Press
.
Hammadou
J. A.
1991
. ‘
Interrelationships among PK, inference and LP in foreign language reading
,’
Modern Language Journal
 
75
:
27
38
.
Jucks
R.
Paus
E.
.
2013
. ‘
Different words for the same concept: Learning collaboratively from multiple documents
,’
Cognition and Instruction
 
31
:
227
54
.
Karimi
M. N.
2015a
. ‘
L2 multiple-documents comprehension: Exploring the contributions of L1 reading ability and strategic processing
,’
System
 
52
:
14
25
.
Karimi
M. N.
2015b
. ‘
EFL learners’ multiple-documents literacy: Effects of a strategy-directed intervention program
,’
Modern Language Journal
 
99
:
40
56
.
Karimi
M. N.
Alibakhshi
G.
.
2014
. ‘
EFL learners’ text processing strategies across comprehension vs. integration reading task conditions
,’
System
 
46
:
96
104
.
Karimi
M. N.
Shabani
M. B.
.
2013
. ‘
Comparing the strategic behavior of more successful vs. less successful readers of multiple technical reading texts
,’
Innovation in Language Learning and Teaching
 
7
:
125
38
.
Kintsch
W.
1988
. ‘
The role of knowledge in discourse comprehension: A construction-integration model
,’
Psychological Review
 
95
:
163
82
.
Kobayashi
K.
2009
. ‘
The influence of topic knowledge, external strategy use, and college experience on students’ comprehension of controversial texts
,’
Learning and Individual Differences
 
19
:
130
4
.
Koh
M. Y.
1985
. ‘
The role of PK in reading comprehension
,’
Reading in a Foreign Language
 
3
:
275
380
.
Le Bigot
L.
Rouet
J. F.
.
2007
. ‘
The impact of presentation format, task assignment, and PK on students’ comprehension of multiple online documents
,’
Journal of Literacy Research
 
39
:
445
70
.
Leeser
M. J.
2007
. ‘
Learner-based factors in L2 reading comprehension and processing grammatical form: Topic familiarity and working memory
,’
Language Learning
 
57
:
229
70
.
Linderholm
T.
Therriault
D. J.
Kwon
H.
.
2014
. ‘
Multiple science text processing: Building comprehension skills for college readers
,’
Reading Psychology
 
35
:
332
56
.
McNamara
D. S.
Kintsch
E.
Songer
N. B.
Kintsch
W.
.
1996
. ‘
Are good texts always better? Interactions of text coherence, background knowledge, and levels of understanding in learning from text
,’
Cognition and Instruction
 
14
:
1
43
.
McNeil
L.
2011
. ‘
Investigating the contributions of background knowledge and reading comprehension strategies to L2 reading comprehension: An exploratory study
,’
Reading and Writing
 
24
:
883
902
.
McNeil
L.
2012
. ‘
Extending the compensatory model of second language reading
,’
System
 
40
:
64
76
.
Nassaji
H.
2003
. ‘
Higher-level and lower-level text processing skills in advanced ESL reading comprehension
,’
Modern Language Journal
 
87
:
261
76
.
Nassaji
H.
2007
. ‘
Schema theory and knowledge-based processes in second language reading comprehension: A need for alternative perspectives
,’
Language Learning
 
57
:
79
113
.
Pulido
D. C.
2004
. ‘
The relationship between text comprehension and second language incidental vocabulary acquisition: A matter of topic familiarity?,
Language Learning
 
54
:
469
523
.
Rouet
J. F.
Favart
M.
Britt
M. A.
Perfetti
C. A.
.
1997
. ‘
Studying and using multiple documents in history: Effects of discipline expertise
,’
Cognition and Instruction
 
15
:
85
106
.
Rydland
V.
Aukrust
V.
Fulland
H.
.
2012
. ‘
How word decoding, vocabulary and prior topic knowledge predict reading comprehension. A study of language-minority students in Norwegian fifth grade classrooms
,’
Reading and Writing
 
25
:
465
82
.
Stadtler
M.
Bromme
R.
.
2013
. ‘
Multiple document comprehension: An approach to public understanding of science
,’
Cognition and Instruction
 
31
:
122
9
.
Strømsø
H. I.
Bråten
I.
.
2002
. ‘
Norwegian law students’ use of multiple sources while reading expository texts
,’
Reading Research Quarterly
 
37
:
208
27
.
Strømsø
H.
Bråten
I.
Britt
M. A.
.
2010
. ‘
Reading multiple texts about climate change: the relationship between memory for sources and text comprehension
,’
Learning & Instruction
 
20
:
192
204
.
Strømsø
H.
Bråten
I.
Samuelstuen
M. S.
.
2008
. ‘
Dimensions of topic-specific epistemological beliefs as predictors of multiple text understanding
,’
Learning and Instruction
 
18
:
513
27
.
Tan
H. S.
1990
. ‘The role of PK and language proficiency as predictors of reading comprehension among undergraduates’ in:
de Jong
J. H. A. L.
Stevenson
D. K.
(eds):
Individualizing the Assessment of Language Abilities
 .
Multilingual Matters
, pp.
214
24
.
Usó-Juan
E.
2006
. ‘
The compensatory nature of discipline-knowledge and English-language proficiency in reading English for academic purposes
,’
Modern Language Journal
 
90
:
210
27
.
Wineburg
S.
1998
. ‘
Reading Abraham Lincoln: An expert/expert study in the interpretation of historical texts
,’
Cognitive Science
 
22
:
319
46
.