735

TheBritish

British Journal of Educational Psychology (2009), 79, 735-747 l ^ g H r ^^ 0 ogica© 2009 The Brítísh Psychological Society * * * ^ bOCiety

www.bpsiournals.co.uk

The comprehension skills of children learningEnglish as an additional language

K. Burgoyne'* J. M. Kelly (née Hutchinson)^ H. E. Whiteley^and A. Spooner^'Down Syndrome Education International, Portsmouth, UK^University of Central Lancashire, Preston. UK^Edge Hill University. Ormskirk. UK

Background. Data from national test results suggests that children who arelearning English as an additional language (EAL) experience relatively lower levelsof educational attainment in comparison to their monolingual. English-speaking peers.

Aims. The relative underachievement of children who are learning EAL demandsthat the literacy needs of this group are identified. To this end. this study aimedto explore the reading- and comprehension-related skills of a group of EAL learners.

Sample. Data are reported from 92 Year 3 pupils, of whom 46 children arelearning EAL.

Method. Children completed standardized measures of reading accuracy andcomprehension, listening comprehension, and receptive and expressive vocabulary.

Results. Results indicate that many EAL learners experience difficulties inunderstanding written and spoken text. These comprehension difficulties are notrelated to decoding problems but are related to significantly lower levels of vocabularyknowledge experienced by this group.

Conclusions. Many EAL learners experience significantly lower levels of Englishvocabulary knowledge which has a significant impact on their ability to understandwritten and spoken text. Greater emphasis on language development is thereforeneeded in the school curriculum to attempt to address the limited language skills ofchildren learning EAL.

A significant number of pupils in UK schools learn English as an additional language(EAL): in 2008, 14.4% of pupils in primary schools and 10.8% of secondary schoolpupils were recorded as having a first language that is known or believed to be otherthan English (Department for Children. Schools and Families [DCSF], 2008). A recentstudy of reading attainment (TWist. Schagen. & Hodgson. 2007) reported that reading

* Correspondence should be addressed to Dr Kelly Burgoyne, Down Syndrome Education International, The Sarah DuffenCentre, ße/mont Street, Southsea, Portsmouth POS INA, UK (e.mail: Kelly.Burgoyne@downsed.org).

DOI: 10.1348/000709909X422530

736 Kelly Burgoyne et ai

achievement is lower for this group of learners than for pupils whose first language isEnglish. Data from national test results support this, indicating that children who haveEnglish as a first language experience higher rates of educational attainment thanchildren learning EAL at each stage of education (DCSF, 2006/07). This pattern ofrelative underachievement suggests that current educational practice is failing to meetthe literacy needs of children learning EAL (NALDIC, 2008). This paper aims to explorethe literacy skills of this group of children in order to understand and identifyappropriate support for their educational needs.

Successful reading requires that children acquire good decoding skills, and for manychildren, reading difficulties stem from problems at this level (Perfetti, 1985). A largebody of research identifies phonological awareness as critical for the development ofword reading skills (e.g. Bryant, Maclean, Bradley, & Crosslands, 1990; Hatcher,Hulme, & Ellis, 1994; Muter, Hulme, Snowling, & Taylor, 1997) and the efficacy ofphonics teaching in boosting reading skill is well documented (e.g. Hatcher et al., 1994;Stuart, 1999, 2004). In accordance with this evidence, recent recommendationssupport an integral role for the teaching of phonics within the national literacy strategy(Rose, 2006).

Increasing understanding of the key role of phonic skills in literacy acquisitionpotentially leads to the assumption that high-quality phonics teaching will automaticallylead to improvements in reading attainment (Stuart, 2003). Effective reading, however,involves more than decoding print; indeed, the primary purpose of reading is tounderstand the text (Nation & Angelí, 2006; Stuart, 2003). Successful reading thereforedepends on the development of both word reading and comprehension skills, skillswhich are related but are independent of one another (Oakhill, Cain, & Bryant, 2003).Many reading tests, including those by which national reading standards (StandardAttainment Tests, SATs) are assessed, confound word recognition and textcomprehension skills; consequently, it cannot be assumed that poor performance onthese measures are necessarily a consequence of problems with word reading. Indeed,research evidence suggests that for some children, low reading attainment refiectsproblems in understanding, rather than decoding, text (i.e. poor comprehenders;Yuill & Oakhill, 1991). Previous work with EAL learners suggests a similar patternof comprehension difficulties in the absence of word reading problems (Erederickson& Frith, 1998; Hutchinson, Whiteley, Smith, & Connors, 2003).

Literacy experiences occur outside of the classroom as well as in school. In thisstudy, the children learning EAL are of Asian heritage, and speak a variety of MotherTongue languages. In addition to attending English-speaking schools, many of thesechildren attend Mosque where they are learning to read the Quran. In this context, theaim of reading is not to understand the meaning of the text, as very few Arabic readersprogress to understanding the Arabic language (Rosowsky, 2001), but rather to decodethe text, and recite the text from memory (Robertson, 2002). One potentialconsequence of this is a concept of reading as essentially a process of decoding, tothe relative neglect of meaning. This approach to reading the Qur'an may then transferto the reading of English text in the classroom (Rosowsky, 2001). Whilst this approachmay support the development of strong decoding skills (Rosowsky, 2001),comprehension of the text is likely to be relatively poor

Failure to identify comprehension difficulties and provide appropriate support forthe development of comprehension will lead to reduced access to the curriculum andan inability to reach the individual's full potential. This was recently demonstrated in astudy by Cain and Oakhill (2006b): children with poor comprehension at the age of 8

The comprehension skills of children learning EAL 737

cotitinued to experience cotnprehensioti problems at age 11, leaditig to sigtiLñcantlylower SAT scores than good comprehenders in Maths, English, and Science subjects.Difficulties with comprehension require specific and targeted support; however,children identified as struggling readers on the basis of SAT scores at age 7 are likelyto be targeted with phonics-based training with the aim of improving word-recognition skills, which will do little to support the development of comprehension(Stuart, 1999, 2004).

Comprehension skills are important not only for understanding written text, but alsofor understanding spoken language, and there is considerable evidence of a singlecomprehension system which underlies the comprehension of both written and oraltext (e.g. Gernsbacher, Varner, & Faust, 1990). Before children begin to read, thelanguage comprehension system develops through oral language experiences. Learningto read involves learning how to access this system via the written code (Stuart, 2003).Thus, in the early stages of beginning reading, the level of understanding that a child canattain from reading text is highly dependent on developing decoding and wordrecognition skills (Sticht & James, 1984). In this way, the level of comprehension that achild can attain on reading tests can be restricted as a result of word reading problems(Spooner, Baddeley, & Gathercole, 2004). Assessing comprehension of both written andspoken text may therefore identify whether comprehension difficulties reflect a generallanguage comprehension problem, or difficulties related specifically to reading (Cain &Oakhill, 2006a; Carlisle, 1989).

Comprehension is a multidimensional skill that involves a number of processes atseveral different levels. As a consequence, difficulties in any one of several componentskills may contribute to comprehension failure (August, Francis, Hsu, & Snow, 2006;Cain & Oakhill, 2006b). Components that have been consistently identified as importantto the comprehension process include background knowledge (including knowledge ofstory structure) and vocabulary knowledge, inference and integration skills,comprehension monitoring, and working memory (e.g. Nation, 2005; Nation & Angelí,2006; Perfetti, Landi, & Oakhill, 2005; Perfetti, Marron, & Foltz, 1996; Stuart, 2003). Thecurrent study focuses on one factor known to be critical for comprehension; that is,vocabulary knowledge.

Many children who are learning EAL experience significantly lower levels of Englishvocabulary knowledge relative to their monolingual, English-speaking peers (Cameron,2002). Weaker language skills are likely to have a significant impact on the ability tounderstand text (Stuart, 2004). Previous work has focused on the role of receptivevocabulary in explaitiing comprehension. However, tests of reading and listeningcomprehension often require children to formulate and express answers to questions,whether verbally or in writing. Consequently, the ability to retrieve appropriatelanguage to express understanding is likely to contribute to performance oncomprehension tasks (Hutchinson et al, 2003). It is therefore important to assessboth receptive and expressive language skills.

This study examines the reading- and comprehension-related skills of childrenlearning EAL relative to their monolingual peers. Tests of reading comprehension andword reading are completed alongside measures of listening comprehension andvocabulary by pupils in school Year 3. It is predicted that children learning EAL willexperience difficulties comprehending both written and spoken text, and that thesedifficulties are related to the weaker vocabulary skills of this group of children incomparison to their monolingual, English-speaking peers, rather than to problemsdecoding text.

738 Keiiy Burgoyne et al.

One further factor considered in this research is gender. Although recent evidencesuggests that gender differences may not exert the level of impact that is commonlybelieved (Connolly, 2006), boys are often found to do less well in terms of literacy skillsthan girls (e.g. DCSF, 2006/07; Sammons, Elliot, Sylva, Melhuish, & Siraj-Blatchford,2004; Whitehead, 2006). It is therefore important to explore the potential effects ofgender on measures of reading ability. For this purpose, gender is included as a variablewhen analysing scores on the measures of vocabulary and reading skill.

Method

ParticipantsNinety-two Year 3 children from four schools in the North-west region participatedin the study. In this year of school, there is a significant shift in emphasis of classroomtasks from word-based skills to language comprehension (DCSF, 2006). It is thereforeat this age that comprehension difficulties will begin to have a significant impact on theability to achieve in school. In line with this, this age group has been the focus ofmuch of the research exploring comprehension skills in children (e.g. Hutchinson,Whiteley, Smith, & Connors, 2004; Hutchinson etal, 2003; Yuill & Oakhill, 1991).

Of the monolingual children, 20 were male and 26 female; there were 21 boys and25 girls learning EAL. Children were defined as learning EAL if a Mother Tonguelanguage was spoken to some extent in the home environment. This was initiallydetermined by teacher indications of Mother Tongue use, and confirmed prior to testingby the language preference questionnaire (Beech & Keys, 1997), which examineschildren's language experiences across a range of contexts. This identified a range ofMother Tongue languages spoken by the EAL pupils, including Punjabi (18 EAL pupils),Urdu (13), Gujerati (6), Pushto (4), and Bengali (3). Mother Tongue language wasunconfirmed for two of the EAL pupils. The majority of children learning EAL (87%)reported that they attend Mosque at least 5 days a week.

Materials and procedureChildren completed several standardized reading-related measures. Form 1 of theNARA-R (Neale, 1997) provided measures of text reading accuracy, comprehensionand fluency. This test is one of the most commotily used measures of reading abilityin the UK (Cain & Oakhill, 2006a). In the NARA-R, children progress through a seriesof six graded reading passages. The child reads the passages aloud and readingaccuracy errors are corrected. Following each passage, the child is asked a number ofopen-ended questions which tap both literal and inferential understanding of thestory. Children are permitted to refer back to the text when answering the questions.Test administration ends when a prescribed number of reading accuracy errors havebeen made.

Single-word reading was measured using the Wide Range Achievement Test 3(WRAT3; Wilkinson, 1993). ChUdren read a list of 42 words of increasing difficulty,presented out of context. Testing ends when the child makes 10 consecutive readingaccuracy errors, and the number of words a child reads correctly is recorded.

The Expressive and Receptive One-Word Picture Vocabulary Tests (Brownell, 2000)measured lexical knowledge. These tests provide comparable assessments of expressiveand receptive vocabulary and have been co-normed allowing for meaningfulcomparisons of expressive and receptive language. Research supports the concurrent

The comprebension skills ofctiildren learning EAL 739

and predictive validity of these tests (Prezbindowski & Lederberg, 2003), and they areused elsewhere in literacy research (e.g. Groen, Yaski, Laws, Barry, & Bishop, 2008;Hargrave & Senechal, 2000; Jiminez, 1997; Swanson, Rosston, Gerber, & Solari, 2008).To measure receptive vocabulary, children select the picture that most accuratelyrepresetits a given word from a choice of four alternatives. In the expressive test,children are presented with a series of pictures depicting an object, action, or concept,which they are then required to name. The tests were developed in the USA and originalversions contain some stimuli that were considered to be less familiar to British children;these were anglicized for use in the study (e.g. an outline of the USA was replaced with acomparable representation of the UK).

Two measures of listening comprehension were included. Form 2 of the NARA-R(Neale, 1997) is a parallel assessment to Form 1; passages 1-4 of Form 2 wereaudiotaped and presented as a test of spoken language comprehension. In addition,children completed levels B and C of the Listening Comprehension Test Series (LCTS;Hagues, Siddiqui, & Merwood, 1999). In the LCTS a set of four recorded clips are playedto the class. The clips aim to reflect the range of listening contexts in which pupils' skillsmay be assessed, and include, for example, a chüd's description of the route home, aclassroom discussion, and a narrative. Multiple choice and true/false questions are usedto assess comprehension of the clips; these are heard twice, with responses to questionsmade following the second presentation.

ResultsData obtained on each of the measures were analysed by separate 2 x 2 between-groupANOVAs with factors language and gender.

Reading accuracyTable 1 presents the raw scores for WRAT3 single-word reading and NARA-R text readingaccuracy.

Table I . Mean scores (and standard deviations) fo r reading accuracy as a function of language

and gender

WRAT3

NARA-R

Male

31.00(4.95)

43.20(18.70)

Monolingual

Female

29.92 (6.02)

43.38(21.34)

Mean

30.39 (5.55)

43.30 (20.02)

Male

34.52 (4.90)

57.38(16.88)

EAL

Female

31.92(6.10)

49.16(23.04)

Mean

33.11 (5.68)

52.91 (20.66)

A significant main effect of language in both analyses indicates that children learningEAL achieved significantly higher scores than monolingual children for both single-wordreading (f(l,88) = 5.549, p=.O21, TJ = .059) and text reading accuracy(P(l, 88) = 5.458,p = .022, TJ = .058). There were no other significant effects.

In the NARA-R, testing ends on the basis of reading accuracy errors; reading accuracytherefore influences comprehension scores, as more accurate readers can progressfurther through the passages, and may consequently attempt more comprehension

740 Kelly Burgoyne et al.

questions, than weaker readers. Indeed, a comparison of the number of NARA-R storiescompleted by the two groups of children revealed significant differences in favour of thechildren learning EAL (monolingual mean = 3 74 stories; EAL mean = 4.76 stories;i(90) = 3.125. p = .002). In order to control for these differences, an ANCOVA wascarried out on the comprehension data with NARA-R accuracy scores entered as acovariate.

Reading comprehensionMean NARA-R reading comprehension raw scores, and mean scores statistically adjustedfor accuracy using ANCOVA. are presented in Table 2.

Table 2. Reading comprehension scores: Means (and standard deviations) for raw scores and means

adjusted for NARA-R reading accuracy presented by language and gender

Raw scores

Means adjusted

for accuracy

Male

15.20(6.61)

16.66

Monolingual

Female

16.12(8.33)

17.52

Mean

15.72(7.56)

17.09

Male

15.62(6.35)

12.85

EAL

Female

14.40 (7.68)

14.09

EAL

14.96(7.05)

13.47

Co-varying reading accuracy scores in the analysis of comprehension data revealeda significant main effect of language group (/=•(!.87) = 15.860. p < .001. rj^ = .154)with significantly higher comprehension scores from the monolingual children thanfrom children learning EAL. There were no other significant effects.

Listening comprehensionMean raw scores obtained on the NARA-R measure of listening comprehension andthe LCTS (levels B and C scores combined) are reported in Table 3. A monolingual childwas unavailable to complete the NARA-R Form 2; data for this measure is thereforebased on a sample of 45 monolingual children and 46 children learning EAL.

Table 3. Mean raw scores (and standard deviations) fo r listening comprehension as a function of

language and gender group

NARA-R

LCTS

Male

8.42(5.31)

39.45 (10.50)

Monolingual

Female

9.85 (4.36)

45.69 (9.52)

Mean

9.24 (4.78)

42.98(10.33)

Male

6.10(3.28)

38.90(10.31)

EAL

Female

7.32 (4.90)

40.04(12.48)

Mean

6.76 (4.24)

39.52(11.43)

The monolingual children obtained significantly higher listening comprehensionscores on the NARA-R measure than the children learning EAL (^(1.87) = 6.463.p = .013. 1? = .069). Differences between language groups were not significant.

The comprehension sk/Ws of children learning EAL 741

however, on the LCTS (F(l,88) = 1.877, p = .174, 17 = .021). There were no othersignificant effects on either measure.

Expressive and receptive vocabularyTable 4 displays the raw scores obtained on the vocabulary measures.

Table 4. Mean raw scores (and standard deviations) for expressive and receptive vocabulary as a

function of language and gender group

Monolingual EAL

Male Female Mean Male Female Mean

Expressive 71.05(13.64) 72.69(14.40) 71.98(13.95) 57.62(6.81) 58.56(15.48) 58.13(12.19)

vocabularyReceptive 83.05(11.83) 83.81(11.31) 83.48(11.41) 71.67(9.16) 74.72(15.42) 73.33(12.90)

vocabulary

There was a significant main effect of language in favour of monolingual children inthe analysis of both expressive (/?(1,88) = 24.650, p < .001, 17 = .219) and receptivevocabulary scores (77(1,88) = 15.822,/? < .001,77^ = .152). No other significant effectswere found.

From these results, it is clear that children learning EAL experience difficulties withcomprehension relative to their monolingual peers, and that these difficulties are not aconsequence of poor decoding skills but are more likely to be related to weakervocabulary. A series of standard multiple regressions was conducted to further explorethe contribution of expressive and receptive language skills to reading and listeningcomprehension. A standard regression was computed for each of the criterion variables(LCTS listening comprehension, NARA-R listening comprehension, and NARA-R readingcomprehension) separately for the monolingual children and the children learning EAL;results are displayed in Table 5.

Taken together, expressive and receptive vocabulary scores explain a significantamount of variance in comprehension for both monolingual children and childrenlearning EAL. For both groups, only expressive vocabulary makes a significant uniquecontribution to listening comprehension. This relationship is stronger for themonolingual children; though expressive vocabulary is significant for children learningEAL, the measure explains less variance in listening comprehension for this group. Inaddition, expressive vocabulary explains a greater amount of variance in comprehen-sion as measured by the NARA-R, than the LCTS measure.

The two language measures differentially predicted reading comprehension formonolingual children and children learning EAL. For children learning EAL, the patternof results followed that for listening comprehension as only expressive vocabulary madea statistically significant unique contribution to reading comprehension. Thecontribution of receptive vocabulary to reading comprehension was not significantfor this group, but was much greater than that found on either measure of listeningcomprehension. In contrast, expressive vocabulary made no significant uniquecontribution to the prediction of reading comprehension for monolingual children. Forthis group, receptive vocabulary was a significant predictor of NARA-R readingcomprehension.

742 Kelly Burgoyne et al.

Table S. Standard multiple regression predicting listening and reading comprehension

LCTS

Expressive

Receptive

NARA-R listening

Expressive

Receptive

NARA-R reading

Expressive

Receptive

Monolingual

ß t

0.469 2.995

0.242 i.548

R = .43, adjusted R = .41,

F(2,43)= 16.379, p < . 0 0 1

0.653 4.756

0.142 1.036

R = .57, adjusted R^ = .55,

F(2,42) = 2 8 . l 7 3 , p < . 0 0 l

0.121 .748

0.539 3.329

R^ = .39, adjusted R^ = .37,

F(2,43)= 13.933, p < .001

Sig.t

0.005

0.129

< 0.001

0.306

0.458

0.002

ß

0.382

0.246

R = .34,

F{2,43) =

0.511

0.141

EAL

t

2.179

1.404

adjusted R

11.087, p

3.012

.829

R^ = .38, adjusted P

F(2,43) =

0.420

0.286

R = .43,

F(2,43) =

13.339, p

2.573

1.753

adjusted R

16.140, p

^ = .31,

<.OOI

:^ = .35,

<.OOI

^ = .40,

<.OOI

Sig.t

0.035

0.168

0.004

0.412

0.014

0.087

Discussion

The results of this study demonstrate that children learning EAL have difficulties inunderstanding written and spoken text relative to their tnonolingual peers. Readingcomprehension difficulties are not a consequence of poor decoding skills; rather, thelower levels of English vocabulary knowledge experienced by many children learningEAL place significant constraints on comprehension.

Consistent with previous work (Rosowsky, 2001), children learning EAL scoredsignificantly higher than the monolingual children on measures of both single-wordreading and text reading accuracy. Stronger reading accuracy skills enabled the EALlearners to read more passages and thus attempt more comprehension questions thanthe monolingual children. When these differences were controlled statistically, themonolingual children demonstrated significantly higher levels of text comprehension.Thus, children learning EAL were achieving significantly lower levels of readingcomprehension scores than their monolingual peers in spite of good reading accuracy.Potentially, Arabic reading experiences, which emphasize the importance of decodingabove meaning may contribute to this pattern of scores: for children learning EAL,reading comprehension may therefore be partially constrained, not by inefficient wordreading skills, but by a focus on this skill at the expense of meaning.

One factor which had a clear impact on both reading and listening comprehensionscores was vocabulary. Children learning EAL had significantly lower expressive andreceptive vocabulary scores than monolingual children. Vocabulary was a significantpredictor of both reading and listening comprehension for both groups of children. Theweaker vocabulary skills of EAL learners therefore place significant constraints on theircomprehension of written and spoken text.

The regression analyses highlight the significant contribution of expressivevocabulary skills to the listening comprehension of both children learning EAL andmonolingual children. Receptive vocabulary scores made no significant uniquecontribution. A different pattern emerged for reading comprehension. On this measure.

The comprehension skills of children learning EAL 743

receptive vocabulary emerged as the only significant predictor for the monolingualchildren. In contrast, only expressive vocabulary made a significant contribution to thereading comprehension scores of children learning EAL. However, receptive vocabularymade a greater contribution to this measure than to listening comprehension. Thissuggests differences between reading and listening tasks and important differencesbetween monolingual children and children learning EAL. In tasks of readingcomprehension, children are able to return to the text when answering comprehensionquestions. In this way, the text can be used to support comprehension and languageskills, as the itiformation and vocabulary that is needed to answer the questions can belocated in the text. In listening comprehension tasks, the text is not available to referback to. Demands on expressive language skills are therefore greater on tests of listeningcomprehension than on tests of reading comprehension. Lower levels of vocabularyknowledge suggest that, even when the text is available, EAL learners are less able totake advantage of this resource to support the formulation and expression of responsesto comprehension questions. As a consequence, expressive vocabulary remains asignificant predictor of reading comprehension for this group of chudren.

Tests of listening comprehension were included to explore whether readingcomprehension difficulties refiect general language comprehension difficulties, or arespecific to reading. Using a measure that was parallel to that used to assess readingcomprehension revealed significant differences between the groups: the monolingualchildren attained significantly higher listening comprehension scores on the NARA-Rthan the children learning EAL. This suggests that many children learning EALexperience a general language comprehension problem.

This pattern was not, however, reflected by the results of the LCTS on which nosignificant group differences were found. In attempting to understand this, differencesbetween the tests must be considered. Firstly, the LCTS employs multiple-choice andtrue/false questions, where the NARA-R uses open-ended comprehension questions thatrequire children to formulate and verbally express their answers. As confirmed by theregression analysis, this latter method clearly imposes much greater demands onexpressive language skills. The heavy demands on expressive language imposed by theNARA-R may consequently underestimate the comprehension skills of many EALlearners. Work with monolingual, English-speaking children has found that childrenwho score poorly on open-ended comprehension questions perform significantly betteron reading comprehension tests with closed questions (Spooner et al, 2004). Eurtherwork is therefore necessary to explore the impact of question format on thecomprehension performances of children learning EAL.

Though multiple choice and true/false questions may reduce the demands placed onexpressive language skills, Cain and Oakhill (2006a) suggest that assessingcomprehension in this way is less likely to capture specific comprehension skills,such as inferential processing skills, than open-ended questions. A multiple choiceformat, such as that used in the LCTS, must necessarily provide an option in which theinferable information is spelled out. In this case, the child does not need to make theinference to answer the question correctly as they can succeed by recognizing that astatement is consistent with a text representation. As Cain and Oakhill (2006a) argue,this is not the same as generating the inference at text presentation; consequently, testswhich use this type of question format may be a less reliable measure of specificcomprehension skills. Scores on the NARA-R may therefore provide a more accuratereflection of this. It would be of interest for further work to examine comprehensionscores more closely by, for example, comparing performance across literal and

744 Keiiy Burgoyne et ai

inferential question types. Inference generation is a source of difficulty for many poorcomprehenders (e.g. YuUl & Oakhill, 1991); clearly, further work that explores specificcomprehension subskills including inference making in children learning EAL isnecessary.

The NARA-R and the LCTS also differ in the nature of the stimulus they present. TheNARA-R is a test of narrative comprehension which uses formal vocabulary and sentencestructure, and reflects the type of academic, de-contextualized language typical to testsof reading ability. In contrast, the materials used in the LCTS are more comparable toeveryday spoken language communication, including for example, a classroomdiscussion. It is therefore reasonable to suggest that children learning EAL are morefamiliar with, and are therefore better able to comprehend, the materials presented inthe LCTS.

Differences between the two tests suggest that they are not comparable measures oflistening comprehension. Though these results suggest that children learning EAL donot always perform more poorly than monolingual children on tests of listeningcomprehension, it is clear that, when assessed on measures that are parallel to tests ofreading comprehension, children learning EAL attain lower listening comprehensionscores. This analysis suggests that comprehension problems extend beyond the readingdomain, and reflect problems with understanding regardless of the modality ofpresentation. NARA-R comprehension scores attained by children learning EALtherefore appear to reflect a general problem with English language comprehensionrelative to monolingual English-speaking children. As comprehension skill issignificantly related to academic attainment (Cain & Oakhill, 2006b), this hasconsiderable implications for many pupils learning EAL. Further work to explore therelationships between performance on the reading and language assessments used hereand SATS performance in Year 6 would shed more light on the impact of comprehensionand vocabulary on academic attainment for this group of children.

In contrast to previous reports of gender differences in literacy attainment in favourof girls (e.g. DCSF, 2006/07; Sammons et al, 2004; Whitehead, 2006) there was littleevidence of gender differences for monolingual English-speaking children and forchildren learning EAL in the research reported here: girls and boys performed similarlyon the measures of reading accuracy, comprehension, and vocabulary. This finding isconsistent with other literacy research that has failed to find evidence of genderdifferences (e.g. Sen & Blatchford, 2001) and may suggest that gender effects aresurpassed by other variables such as ethnicity as argued elsewhere (e.g. CotinoUy, 2006).Eor the children participating in this project, gender does not appear to be a significantfactor in the development of literacy skills.

The results of this study have significant implications for educational practice. Firstly,children learning EAL are often able to develop strong word recognition and decodingskills. In this study, the children learning EAL were able to progress further through thereading comprehension passages, and therefore attempt more comprehensionquestions, than the monolingual children. As a consequence, there appeared to be nodifferences in reading comprehension when looking solely at the raw scores;differences were only apparent when scores were adjusted to account for readingaccuracy skills. Within the classroom, these children may not therefore be identified ashaving difficulties in reading; well-developed reading accuracy skills are likely to maskunderlying comprehension difficulties.

Secondly, it is clearly incorrect to assume that reading difficulties, and lower levels ofreading attainment, necessarily reflect difficulties with decoding. For many children

The comprehension skills of children learning EAL 745

learning EAL. it is difficulties understanding text that constrain performance. Thissignals the need for direct and targeted support to address the comprehensiondifficulties of children learning EAL. Recent research exploring the multifaceted natureof comprehension has identified several components that underlie comprehension skill(e.g. Nation. 2005; Nation & Angelí. 2006; Perfetti et al, 1996. 2005; Stuart. 2003).Further work is currently underway to assess the contribution of these components tothe comprehension skills of children learning EAL. Despite this developing knowledge-base surrounding the components of comprehension, it has been argued that thisknowledge is unrecognized in the National Literacy Framework (Stuart. 2003). To makefull use of the emergent understanding of the components of comprehension and thedevelopment of comprehension skills, this knowledge must be incorporated intoteaching and intervention strategies. Without this, it is unlikely that appropriate andspecific support would be typically available even where particular difficulties had beenidentified.

This research confirms that vocabulary knowledge has a key role in comprehensionskill. The lower levels of vocabulary knowledge experienced by many children learningEAL place significant constraints on their ability to understand English text. Orallanguage skills are often taken for granted in the school setting such that formalinstruction rarely takes place in British schools (Stuart. 2003). It is clear from thisresearch that age-appropriate English language skills cannot be assumed for the manyEAL pupils that attend English-speaking schools, implicating a clear need to extendvocabulary learning in the classroom. Without instruction, this group of learners willcontinue to experience significantly lower levels of vocabulary in comparison to theirmonolingual. English-speaking peers (Cameron. 2002; Hutchinson et al, 2003. 2004).Greater emphasis on language development is therefore needed in the schoolcurriculum to attempt to address the limited language skills of children learning EAL.Furthermore, this needs to begin earlier than Key Stage 2 in order to establish a level oflanguage skill which may support the development of comprehension in Year 3 andbeyond.

Lastly, this study suggests that for children learning EAL. a focus on decoding whenreading texts contributes to comprehension difficulties. Children learning EAL need agreater emphasis on reading for meaning, brought about through discussion of textwithin classrooms so that children may learn to make sense of text. Furthermore, thestrong reading accuracy skills demonstrated by many children learning EAL suggests thatcomprehension instruction and a focus on meaning could be integrated into readingpractices earlier than Year 3-

ReferencesAugust. D.. Francis. D. J.. Hsu. H. -Y. A.. & Snow. C. E. (2006). Assessing reading comprehension

in bilinguals. The Elementary School Journal, 107(T), 221-238.Beech. J. R.. & Keys. A. (1997). Reading, vocabulary and language preference in 7- to 8-year old

bilingual Asian children. British Journal of Educational Psychology, 67, 405-414.Browneil. R. (Ed.). (2000). Expressive and Receptive One-Word Picture Vocabulary Tests (3rd ed.).

Novato. CA: Academic Therapy Publications.Bryant. P. E.. MacLean. M.. Bradley. L. L.. & Crossland. J. (1990). Rhyme, alliteration, phoneme

detection and learning to read. Developmental Psychology, 26, 429-438.Cain. K.. & Oakhill. J. (2006a). Assessment matters: Issues in the measurement of reading

comprehension. British Journal of Educational Psychology, 76, 697-708.

746 Kelly Burgoyne et o/.

Cain, K., & Oakhill, J. V. (2006b). Profiles of children with specific reading comprehensiondifficulties. British Joumat of Educational Psychology, 76, 683-696.

Cameron, L. (2002). Measuring vocabulary size in English as an additional language. LanguageTeaching Research, 6(1), 145-173.

Carlisle, J. F. (1989). Diagnosing comprehension deficits through listening and reading. Annals ofDyslexia, 39, 159-176.

Connolly, P. (2006). The effects of social class and ethnicity on gender differences in GCSEattainment: A secondary analysis of the Youth Cohort Study of England and Wales 1997-2001.British Educational Research Journal, 32(1), 3-21.

Department for ChUdren, Schools and FamUies [DCSF]. (2006, September). Primary NationalStrategy: Primary Framework for Literacy and Mathematics. Ref: 02011-2006BOK-EN.

DCSF (2006/07). SFR 38/2007, National Curriculum Assessment, GCSE and Equivalent Attainmentand PosM6, 2006/07.

DCSF (2008). Statistical First Release (SFR) 09/2008. PupU Characteristics and Class Sizes inMaintained Schools in England, January 2008 (Provisional).

Frederickson, N., & Frith, U. (1998). Identifying dyslexia in bilingual children: A phonologicalapproach with inner London SyUieti speakers. Dyslexia, 4, 119-131.

Gernsbacher, M. A., Varner, K. R., & Faust, M. E. (1990). Investigating differences in generalcomprehension skill. Journal of Experimental Psychology: Learning, Memory andCognition, 16(5), 430-445.

Groen, M. A., Yasin, 1., Laws, G., Barry, }.G.,& Bishop, D. V M. (2008). Weak hand preference inchildren with Down syndrome is associated with language deficits. DevelopmentalPsychobiology, 50(5), 242-250.

Hagues, N., Siddiqui, R., & Merwood, P (1999). Listening comprehension test series.NFER-Nelson.

Hargrave, A. C, & Senechal, M. (2000). A book reading intervention with preschool children whohave limited vocabularies: The benefits of regular reading and dialogic reading. EarlyChildhood Research Quarterly, 15(1), 75-90.

Hatcher, P J., Hulme, G., & EUis, A. W (1994). Ameliorating early reading faUure by integrating theteaching of reading and phonological skills: The phonological linkage hypothesis. ChildDevelopment, 65, 41-57.

Hutchinson, J. M., Whiteley, H. E., Smith, G. D., & Connors, L. (2003). The developmentalprogression of comprehension-related skills in children learning EAL. Journal of Research inReading, 26(1), 19-32.

Hutchinson, J. M., Whiteley, H. B., Smith, G. D., & Gonnors, L. (2004). The early identification ofdyslexia: Ghildren with English as an additional language. Dyslexia, 10, 179-195.

Jiminez, R. T. (1997). The strategic reading abilities and potential of 5 low-literacy Latina/o readersin middle school. Reading Research Quarterly, 32(5), 224-243.

Muter, V, Hulme, G., Snowling, M., & Taylor, S. (1997). Segmentation, not rhyming, predicts earlyprogress in learning to read. Journal of Experimental Child Psychology, 65, 370-396.

NALDIG (2008). EAL learners read more but achieve less. News item. Retrieved January 24, 2008,from http://www.naldic.org.uk/docs/news/archive/news_item.cftn?newsid=250átPp=l

Nation, K. (2005). Ghildren's reading comprehension difficulties. In M. J. Snowling & G. Hulme(Eds.), We science of reading: A handbook (pp. 248-266). Oxford: Blackwell Publishing.

Nation, K., & Angelí, P. (2006). Learning to read and learning to comprehend. London Review ofEducation, 4(1), 77-87.

Neale, M. (1997). Neale analysis of reading ability (Second Revised British Edition). Windsor:NFER-Nelson.

Oakhill, J., Gain, K., & Bryant, P E. (2003). The dissociation of word reading and textcomprehension: Evidence from component skills. Language and Cognitive Processes 18443-468.

Perfetti, G. A. (1985). Reading ability. New York: Oxford University Press.

The comprehension skills of children learning EAL 747

Perfetti, C. A., Landi, N., & Oakhill, J. (2005). The acquisition of reading comprehension skill. InM. J. Snowling & C. Hulme (Eds.), The science of reading: A handbook (pp. 227-247). Oxford:Blackwell Publishing.

Perfetti, C. A., Marron, M. A., & Foltz, P. W. (1996). Sources of comprehension failure: Theoreticalperspectives and case studies. In C. Cornoldi & J. Oakhill (Eds.), Reading comprehensiondifflcuities: Processes and interventions (pp. 137-167). Mahwah, NJ: Erlbaum.

Prezbindowski, A. K., & Lederberg, A. R. (2003). Vocabulary assessment of deaf and hard-of-hearing children from infancy through the pre-school years. Journal of Deaf Studies andDeaf Education, 8(4'), 383-400.

Robertson, L. H. (2002). Parallel literacy classes and hidden strengths: Learning to read in English,Urdu and classical Arabic. Reading Literacy and Language, 36(3), 119-126.

Rose, J. (2006). Independent review of the teaching ofeariy reading: Einal Report. London: DfES.Rosowsky, A. (2001). Decoding as a cultural practise and its effects on the reading process

of bilingual pupils. Language and Education, 15ÇI), 56-70.Sammons, P, Elliot, K., Sylva, K., Melhuish, E., Siraj-Blatcbford, I., &Taggart, B. (2004). The impact

of pre-school on young children's cognitive attainments at entry to reception. BritishEducationai Research Journal, 30(5), 691-712.

Sen, R., & Blatchford, P. (2001). Reading in a second language: Eactors associated with progress inyoung children. Educational Psychology, 21(2), 189-202.

Spooner, A. L. R., Baddeley, A. D., & Gathercole, S. E. (2004). Can reading and comprehension beseparated in the Neale analysis of reading ability? British Joumai of Educational Psychology,74, 187-204.

Sticht, T. G., & James, J. H. (1984). Listening and reading. In D. Pearson (Ed.), Handbook ofresearch on reading (Vol. 1, pp. 293-317). New York: Longman.

Stuart, M. (1999). Getting ready for reading: Early phoneme awareness and phonics teachingimproves reading and spelling in inner-city second language learners. British Journal ofEducationai Psychoiogy, 69, 587-605.

Stuart, M. (2003). Eine tuning the National Literacy Strategy to ensure continuing progress inimproving standards of reading in the U.K.: Some suggestions for change. Paper written inresponse to the DfES Phonics Seminar, March, 2003.

Stuart, M. (2004). Getting ready for reading: A follow-up study of inner city second languagelearners at the end of Key Stage 1. British Journal of Educational Psychology, 74, 15-36.

Swanson, H. L., Rosston, K., Gerber, M., & Solad, E. (2008). Influence of oral language andphonological awareness on children's bilingual reading. Journal of School Psychology, 46,413-429.

Twist, L., Schagen, I., & Hodgson, C. (2007). Readers and reading: The national report forEngland 2006 (PLRLS: Progress in Intemationai Reading Literacy Study). Slough: NEER.

Whitehead, J. M. (2006). Starting school - why girls are already ahead of boys. TeacherDeveiopment, 10(2), 249-270.

Wilkinson, G. S. (1993). Wide range achievement test 3- London: Psychological CorporationLimited.

YuiU, N., & Oakhill, J. (1991). Chiidren's problems in text comprehension: An experimentalinvestigation. Cambridge: Cambridge University Press.

Received 30 june 2008; revised version received 12 January 2009

Copyright of British Journal of Educational Psychology is the property of British Psychological Society and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.