Cognition and Literacy in English Language Learners at Risk for Reading Disabilities

Cognition and Literacy in English Language Learners at Risk for Reading Disabilities
of 20
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Journal of Educational Psychology Cognition and Literacy in English Language Learners atRisk for Reading Disabilities H. Lee Swanson, Michael J. Orosco, and Cathy M. LussierOnline First Publication, November 28, 2011. doi: 10.1037/a0026225CITATIONSwanson, H. L., Orosco, M. J., & Lussier, C. M. (2011, November 28). Cognition and Literacyin English Language Learners at Risk for Reading Disabilities.  Journal of EducationalPsychology  . Advance online publication. doi: 10.1037/a0026225  Cognition and Literacy in English Language Learners at Risk forReading Disabilities H. Lee Swanson, Michael J. Orosco, and Cathy M. Lussier University of California, Riverside This study explores the cognitive basis of reading disabilities (RDs) in Spanish-speaking children whoare learning English as a second language. Children (  N   393) designated as English language learners(ELLs) or bilingual with and without RDs in Grades 1, 2, and 3 were administered a battery of cognitive(short-term memory, working memory, rapid naming, random letter and number generation), vocabulary,and reading measures in both Spanish and English. Four important findings emerged. First, children atrisk for RD and ELL children share common problems in English phonological processing and namingspeed, as well as on language general measures of working memory and ratings of classroom attention.Second, children at risk for RD in both bilingual and ELL samples share similar cognitive difficulties,but Spanish phonological processing partials out performance differences between RD and non-RDgroups only in the bilingual sample. Third, differences among RD subgroups were isolated to measures of classroom inattention, English naming speed, and phonological processing. Fourth, performance as a functionof language status and RD was related to a language general working memory factor. In general, the resultssupport the notion that Language 1 phonological processing as well as a general working memory systemunderlie second-language acquisition and RD in children whose first language is Spanish. Keywords:  working memory, English language learners, cognitive processing, reading, language acqui-sition Currently, there are approximately 5.5 million students attend-ing U.S. public schools whose native or first language is notEnglish (see McCardle, Mele-McCarthy, Cutting, Leos, &D’Emilio, 2005, for a review). Of these students, 80% speak Spanish (McCardle et al., 2005). School achievement is lower forLatino children than for many other groups, and children withSpanish as their first language are more likely to have readingdifficulties (e.g., English word identification and comprehension)than other minority and Caucasian children (August & Hakuta,1997). Of those children with reading difficulties, several may beat increased risk for reading disabilities (RDs). Unfortunately, thereason behind the prevalence of RDs in English language learners(ELLs) in the public school system is unclear because neither amethod for accurate identification nor a consistent definitionacross states exists (e.g., McCardle, Keller-Allen, & Shuy, 2008).Further, the figures available indicate the complexity of this issue(McCardle et al., 2005). For example, national estimates revealthat ELLs are underrepresented overall in special education, mean-ing that a smaller percentage of these students are receivingservices than would be expected, given the proportion of theoverall population that they represent (e.g., U.S. Department of Education, National Center for Education Statistics, 2004). Theseissues underscore the need for better tools and methods for accu-rately identifying ELL children at risk for RD.The purpose of this study was to explore those cognitive pro-cesses that underlie RD in ELL children. The candidates mostoften referred to in the literature are phonological processes andoral language (e.g., see Lesaux & Geva, 2006, for a review). Thereis considerable evidence that phonological processing and vocab-ulary are major cognitive determinants of word reading skills,especially in the early phases of learning (National Reading Panel,2000). Cross-cultural linguistic studies have demonstrated that inelementary ELL students, phonological awareness (Dickinson,McCabe, Clark-Chiarelli, & Wolf, 2004; Gottardo, Collins, Baciu,& Gebotys, 2008; Lindsey, Manis, & Bailey, 2003) and vocabu-lary development (Ordo´n˜ez, Carlo, Snow, & McLaughlin, 2002)correlate with Language 2 (L2, English) reading outcomes. Inaddition to phonological processing and vocabulary, recent studieshave linked short-term memory (STM) and working memory(WM) to ELL children at risk for RD (e.g., Swanson, Sa´ez, Gerber,& Leafstedt, 2004). Although both STM and WM involve transientmemory, their relationships to reading are different (e.g., Alloway, H. Lee Swanson, Department of Educational Psychology, University of California, Riverside; Michael J. Orosco, Department of Special Educa-tion, University of California, Riverside; Cathy M. Lussier, Department of Educational Psychology, University of California, Riverside.This longitudinal study is funded by Grant R324A090092 awarded tothe first author by the U.S. Department of Education, Cognition andStudent Learning, Institute of Education Sciences. Special thanks are givento Erin Bostick Mason, Valerie Perry, Joseph Rios, Elizabeth Arellano,Nicole Garcia, Alfredo Aviles, and Steve Go´mez for their work with thedata collection. Special appreciation is given to the Colton School Districtand the San Bernardino County Superintendent of Schools, and we thank Jose Espinoza and Yolanda Cabrera for their extensive administrativesupport.Correspondence concerning this article should be addressed to H. LeeSwanson, Graduate School of Education, Area of Educational Psychology,University of California, Riverside, Riverside, CA 92521. Journal of Educational Psychology © 2011 American Psychological Association2011, Vol.  ●● , No.  ● , 000–000 0022-0663/11/$12.00 DOI: 10.1037/a0026225 1  Gathercole, Willis, & Adams, 2004; Swanson, 2008). STM is morelikely to predict word identification, whereas WM is more likely topredict comprehension (e.g., Swanson & Berninger, 1995).The majority of studies that have compared children who vary inlanguage and reading skills assume that STM measures tap aphonological system or what Baddeley (2007) denotes as the  phonological loop  (e.g., Alloway et al., 2004). The phonologicalloop has been referred to as STM because it involves two majorcomponents discussed in the STM literature: a speech-based pho-nological input store and a rehearsal process (see Baddeley, 1986,for a review). Research to date indicates that children with diffi-culties in reading and second-language acquisition have difficul-ties on tasks requiring the short-term retention of ordered infor-mation (e.g., Thorn & Gathercole, 1999; Thorn, Gathercole, &Frankish, 2002), an indication of inefficient phonological rehearsalprocessing.On the other hand, WM involves active manipulation of infor-mation while also mentally storing other information (Baddeley &Logie, 1999; Gathercole, Pickering, Ambirdge, & Wearing, 2004).That is, simultaneous storage and processing are believed to occurwhen one recruits WM processes (Daneman & Carpenter, 1980;see Baddeley, Eysenck, & Anderson, 2009, for a review). Forexample, rehearsing the pronunciation of a newly learned word inprint would involve STM processing, whereas rehearsing pronun-ciation and simultaneously continuing to read text invokes WMprocessing. With respect to RD among ELLs, some studies (Swan-son et al., 2004; Swanson, Sa´ez, & Gerber, 2006) have suggestedthat with ELL first graders, deficits in the Spanish (L1) STMsystem underlie reading problems in the second language (L2). Forexample, Swanson et al. (2004) found that although children at risk for RD manifested severe deficits in a general WM system, whatprimarily differentiated children with and without RD were STMdeficits within a specific language system. This finding was in linewith several studies showing that the phonological component of memory plays a major role in both L2 acquisition (e.g., Service,1992) and reading (see Baddeley, Gathercole, & Papagno, 1998,for a review).Given the literature just discussed, we consider two models toexplain the role of memory in accounting for RD in ELL children:One focuses on processing efficiency at a phonological level, andthe other focuses on processes independent of phonological pro-cessing. What is at issue is not whether children’s risk for RDrelates to the phonological system, but whether differences inreading within or between the two language systems can also beattributed to other processes (e.g., WM) beyond the well-attestedcontribution of phonological processing to second-language acqui-sition and reading.One model tested here is that children’s reading skills areprimarily due to processing efficiency at the phonological level.This model is based on findings showing that phonological mem-ory is an important component of second-language vocabularyacquisition (e.g., Thorn & Gathercole, 1999; Thorn et al., 2002) aswell as reading acquisition (e.g., Harrington & Sawyer, 1992;Lipka & Siegel, 2007; Stanovich & Siegel, 1994). For example,deficits in the phonological system have been attributed to RD inEnglish (e.g., Stanovich & Siegel, 1994) and Spanish (e.g.,Gonza´lez & Valle, 2000). These findings converge on the viewthat children with relatively poor phonological memory are lesssuccessful in learning the sound structure of new words (e.g.,Chiappe & Siegel, 2006; Chiappe, Siegel, & Wade-Woolley, 2002;Lesaux & Siegel, 2003; Palladino & Ferrari, 2008). Poor phono-logical memory is assumed to be related to a weak phonologicalstore in which phonological traces are constantly refreshed byrehearsal (referred to as the phonological loop). Thus, if childrensuffer processing deficits related to the phonological system, thenthey are unable to temporarily hold unfamiliar phonological formsof information to allow more permanent memory presentations tobe constructed (Baddeley et al., 1998).In contrast to the model just described, the second model viewsexecutive processes as providing resources to lower order (phono-logical system) skills as well as a general system of monitoringindependent of those skills (e.g., Baddeley & Logie, 1999). Thus,in contrast to the aforementioned model that assumes phonologicalprocesses play the dominant role in RD, the present model assumesthat executive processes, primarily those related to WM, related assignificantly to L2 reading performance. Previous studies haveshown that the executive component of WM has been significantlyrelated to both L2 word identification and comprehension (e.g.,Swanson et al., 2006). For this study, we narrowly define execu-tive processing of the WM system as reflecting controlled attention(see Kane, Conway, Hambrick, & Engle, 2007, for a review). Weassume that when the effects of WM on reading and languagemeasures partial out the influence of storage systems (e.g., pho-nological STM), the remaining residual variance reflects con-trolled attention (e.g., Engle, Tuholski, Laughlin, & Conway,1999). Parallel activities investigated in this study that we assumedwere reflective of controlled attention were the inhibition of rote orhabitual responses as measured by a Random Generation task (Towse & Cheshire, 2007) and inattention in the classroom (e.g.,Gathercole et al., 2008). Several models of WM also assume thatoperations are related to naming speed (e.g., Henry & Millar,1993). That is, increases in naming speed enhance the effective-ness of subvocal rehearsal processes and hence reduce the decay of memory items prior to output. Therefore, letter and digit namingspeed may underlie the general pattern of cognitive difficultiesnoted in the present study.In general, the two aforementioned models provide competinginterpretations on the primary role of the phonological system. Thefirst model suggests that the phonological system has a bottom-upinfluence on WM, whereas the second suggests that the executivesystem of WM underlies some of difficulties in reading acquisi-tion. This second model has some support in that problems in WMhave been significantly correlated with difficulties in reading com-prehension (e.g., Carretti, Borella, Cornoldi, & De Beni, 2009). Itmay be the case, however, that the validity of either model rests onwhether children primarily experience difficulties in reading orsecond-language acquisition. That is, although both second-language learners and children at risk for RD draw on a phono-logical system, the influence of one process (e.g., phonologicalprocessing) when compared with another (WM) may be greater inone group than in the other. For example, in this study, weinvestigated the possibility that WM processes operate somewhatindependently of the phonological system in second-languagelearners without RD when compared with children at risk for RD.This possibility is based on the assumption that problems in basicphonological processes for children at risk for RD create a bottle-neck in the processing of information within the executive system.This prediction is based on earlier work (Swanson et al., 2004) 2  SWANSON, OROSCO, AND LUSSIER  showing that L2 English word identification was best predicted bya general WM latent measure that combined Spanish and Englishtask performance and a Spanish-specific latent measure related toSTM. This finding suggested that L2 word reading involves bothlanguage general WM processes and an L1 STM system.This study also examined whether there is a bilingual advantageon some cognitive measures for children at risk for RD. Severalrecent studies have explored how bilingualism affects executivefunctioning, flexibility, and intentional control in children (e.g.,Bialystok, 2007; Bialystok & Martin, 2004). Some studies suggestthat children who are bilingual excel on tasks that require con-trolled attention when compared with monolingual peers. Al-though this finding has not been consistently replicated (e.g.,Namazi & Thordardottir, 2010; Sanchez et al., 2010), there areindications that bilingual children excel on tasks requiring con-trolled attention (e.g., WM) when compared with their monolin-gual peers. This bilingual advantage has been attributed to bilin-gual children’s continual practice in ignoring the irrelevantlanguage to communicate effectively in the relevant one. Overtime, this continual practice in a linguistic domain generalizes tocognitive processes. Thus, on the basis of the literature, we as-sessed whether bilingual children at risk for RD may have somecognitive advantages when compared with ELL children at risk forRD whose first language is Spanish.In summary, the purpose of this study was to determine if thecognitive processes than underlie second-language acquisition arethe same as those that underlie RDs in ELL children. The studyaddresses three questions.1. Do ELL children at risk for RD differ from bilingualchildren at risk for RD on cognitive measures? Althoughsome studies have compared ELL children at risk for RDand monolingual children at risk for RD on cognitivemeasures (e.g., Geva, Wade-Woolley, & Shaney, 1997;Jean & Geva, 2009; Lesaux, Rupp, & Siegel, 2007), fewhave compared ELL children at risk for RD with children atriskforRDwhoarerelativelyproficientinbothSpanishandEnglish. Because several studies have found that deficits inphonological processing, oral language, and WM are com-parable among ELL and monolingual children at risk forRD (e.g., Chiappe & Siegel, 2006; Lesaux, Lipka, & Siegel,2006), we predicted that children at risk for RD who areeither ELL or bilingual share similar cognitive deficits. Wealsoinvestigatedwhetherthispatternholdsforchildrenwithreading difficulties only in the area of comprehension. Theliterature has suggested that although ELL children at risk for RD make progress in acquiring word recognition skills(e.g., Kieffer & Lesaux, 2008; Manis, Lindsey, & Bailey,2004; Proctor, Carlo, August, & Snow, 2005), difficultiespersist in reading comprehension (e.g., Lesaux, Crosson,Kieffer, & Pierce, 2010). Thus, the present study also com-pared the cognitive performance of children at risk for RDwho yielded low scores in both word recognition and com-prehension to children with deficits only in comprehension.2. What cognitive measures clearly separate the perfor-mance of ELL children from those at risk for RD? Somestudies have shown that individual differences in second-language acquisition are related to language-specific pho-nological skills (e.g., Palladino & Ferrari, 2008). Thus,we expected that some parallels may exist between ELLchildren without reading difficulties and ELL childrenwith reading difficulties in the phonological domain.However, the literature is unclear as to whether the twogroups (ELL children with and without RD) manifestdifferent profiles on English and Spanish measures otherthan phonological skills. Although some literature hassuggested that the two groups can be separated on Eng-lish measures of WM and naming speed (e.g., Manis etal., 2004; Swanson et al., 2006), it is unclear from theliterature whether group differences on such measuresmanifest themselves in the child’s primary language(Spanish in this case).3. Does a domain general WM system and/or language-specific WM system underlie second-language acquisi-tion and RD? The majority of research on RD and lan-guage acquisition assumes phonological processes play adominant role in predicting second-language acquisitionand reading. A component of WM clearly related tophonological processing (i.e., the phonological loop) hasbeen implicated as one component of the WM systemrelated to RD and second-language learning (e.g., Lind-sey et al., 2003; Thorn & Gathercole, 1999). 1 This studyalso tests whether executive processes, primarily thoserelated to WM, are significantly related to L2 languageand reading performance. There is some support for thenotion that a general executive WM system is related toL2 language and L2 reading performance (e.g., Har-rington & Sawyer, 1992; Jongejan, Verhoeven, & Siegel,2007; Namazi & Thordardottir, 2010; Palladino & Fer-rari, 2008; Swanson et al., 2004). For example, studieshave shown advantages for bilingual children in atten-tional control and WM (Bialystok & Marin, 2004; Na-mazi & Thordardottir, 2010) when compared with mono-lingual children, suggesting that some general system isin operation as children become bilingual. When appliedto ELL children, Swanson et al. (2004) found significantcross-language transfer (Spanish    English) in Grade 1children on WM measures ( r  s    .80). The study also 1 Our framework to capture diverse memory processes as they apply tosecond language acquisition and reading is Baddeley’s multicomponentmodel of WM (e.g., Baddeley & Logie, 1999). This tripartite view char-acterizes WM as comprising a central executive controlling system thatinteracts with a set of two subsidiary storage systems: the speech-basedphonological loop and the visual–spatial sketchpad. The phonological loopis responsible for the temporary storage of verbal information; items areheld within a phonological store of limited duration, and the items aremaintained within a store through the process of subvocal articulation. Thevisual–spatial sketchpad is responsible for the storage of visual–spatialinformation over brief periods and plays a key role in the generation andmanipulation of mental images. The central executive is involved in thecontrol and regulation of the WM system. According to Baddeley (Bad-deley & Logie, 1999), it coordinates the two subordinate systems, focusingand switching attention and activating representations within LTM. Thismodel has been revised to include an episodic buffer (Baddeley, Allen, &Hitch, 2010), but support for the tripartite model has been found acrossvarious age groups of children (Gathercole et al., 2004). 3 COGNITION AND LITERACY  showed that that L2 English word identification was bestpredicted by a general WM latent measure that combinedSpanish and English WM task performance. In contrast,a Spanish-specific latent measure related to Spanish STMwas related to RD. This finding suggested that L2 wordreading involved a language general WM processes,whereas RD among ELL children was related to an L1phonological system. This finding has not been repli-cated; therefore, attributing RD within an ELL sample tothe first-language STM system needs to be consideredwithin both an ELL and a bilingual sample. MethodParticipants Participants in this longitudinal study were 383 students inGrades 1, 2, and 3 from three large school districts in SouthernCalifornia. The first wave of data collection was completed in thespring of 2010. We excluded from data analysis children whosefluid intelligence scores (Raven Progressive Matrices Test) werebelow 80, because the focus of this study was not on generalintellectual abilities. The final sample (  N     383) included 169boys and 214 girls. All children were Hispanic. Interviews indi-cated that the primary language spoken at home was Spanish for80% of the children. Of the sample, 14% primarily spoke Englishand Spanish at home, whereas the remainder of the children (6%)primarily spoke English at home. Of the sample 97% participatedin the federally funded Free/Reduced Meals Program. Classroomreading instruction was in English or was a combination of Englishand Spanish. The reading programs in each school district placeda heavy emphasis on phonics instruction.From this sample, six comparison groups based on norm-referenced scores in reading and vocabulary were formed. Chil-dren were first separated into those at risk for RD in comprehen-sion only, those at risk for RD in both word identification andcomprehension, and those not at risk for RD (non-RD). There is nodoubt that there is controversy over the definition of RD, espe-cially in ELL children (e.g., Wagner, Francis, & Morris, 2005).There is growing consensus among researchers, however, that it ismore appropriate to use an absolute definition of RD (below somecutoff score on reading achievement) rather than a discrepancybetween achievement and IQ; therefore, this study adheres to theabsolute definition (e.g., Fletcher, Epsy, Francis, Davidson,Rourke, & Shaywitz, 1989). The 25th percentile on a standardizednorm-referenced reading measure is a common standard to identifychildren at risk for RD (Fletcher et al., 1989; Siegel & Ryan, 1989;Stanovich & Siegel, 1994). There is also controversy over what theabsolute cutoff should be to determine RD, especially in ELLchildren. Previous studies have examined RD difficulties in ELLgroups below the 25th percentile in word recognition skills. Be-cause the focus of this study is on English reading proficiency, wechose to examine children whose composite scores in Englishword recognition and English word attack were below the 25thpercentile. Because the literature has indicated that there are sub-groups within the group of ELL children at risk for RD (e.g.,Lesaux et al., 2006), we subgrouped children with passage com-prehension scores above and below the 25th percentile. Compre-hension deficits when compared with word recognition difficultiesoccur frequently in ELL children at risk for RD. Thus, to enhancethe generalizability of the findings, three subgroups were com-pared: children at risk for RD in word identification and passagecomprehension (RD-WC), children at risk for RD in passagecomprehension (RD-C), and children without RD (non-RD).Our criteria for identifying participants at risk for RD were (a)all children must score in the average range (  80) on a measure of fluid intelligence (Raven Progressive Matrices Test) and (b) allchildren must score above the 25th percentile in math. Children atrisk for RD in word identification and comprehension (RD-WC)scored below the 25th percentile on the Woodcock–Mun˜oz Lan-guage Survey Test on English measures of word identification andpassage comprehension subtests and English word attack from theWoodcock Reading Mastery Test (Woodcock, 1998). Children atrisk for RD in comprehension (RD-C) scored above the 25thpercentiles in English word identification and word attack butbelow the 25th percentile in English passage comprehension. Chil-dren identified as not at risk for RD (non-RD) scored above the25th percentile in word identification, word attack, and passagecomprehension.As with the definition of RD, the definition of ELL is alsocontroversial (i.e., whether to use expressive vs. receptive lan-guage, frequency of English spoken at home, etc.). The firstlanguage for all children participating in this study was Spanish.As indicated by current school records, children’s performance onthe California English Language Development Test (CELDT) waslow. For this study, however, ELLs were operationally defined asachieving a composite score on English receptive language (on thePeabody Picture Vocabulary Test; Dunn & Dunn, 1997) andEnglish Expressive language (on the Expressive One-Word PictureVocabulary Test Spanish-Bilingual Edition; Brownell, 2001) at orbelow a standard score of 90. That is, we assumed that if childrenwere below the 25th percentile norms in English language (eventhough Spanish receptive vocabulary may be in the average range),their language understanding during English reading instructionwould be strained. The remaining L1 (Spanish) children withcomposite English vocabulary scores above a standard score of 90were designated as non-ELL (for lack of a better term) or bilingual.The correlation between language status on our measures andoverall score on the California English Language DevelopmentTest (low scores reflect lower language proficiency) was signifi-cant,  r  (365)  .38,  p  .0001.Our cutoff score for determining whether a child is bilingual orELL is arbitrary; therefore, our attempt to identify children whoare ELL or bilingual needs justification. We compared two cutoff points (standard scores of 85 and 90, respectively) from ourcomposite English vocabulary scores for identifying ELL childrenversus bilingual students. We used a measure referred to asaffected-status agreement (Cicchetti & Feinstein, 1990; Waesche,Schatschneider, Maner, Ahmed, & Wagner, 2011), which in thiscase, is the proportion of students classified as bilingual by eithera cutoff score of 85 or 90 or both. The number of childrenidentified as bilingual (from a total  N   383) on both cutoff scoreswas 173. The number of additional children identified as bilingualwith a cutoff score of 85 was 66. The affected status agreementwas .72 (173/173    66    0). We computed the standard error(.029; see Waesche et al., 2011, p. 300, for the formula) anddetermined the 95% confidence interval (.029    1.96), whichyielded an affected status range from .67 to .77. Because our status 4  SWANSON, OROSCO, AND LUSSIER
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!