Elsevier

Research in Developmental Disabilities

Volume 32, Issue 5, September–October 2011, Pages 1782-1791
Research in Developmental Disabilities

The role of vocabulary, working memory and inference making ability in reading comprehension in Down syndrome

https://doi.org/10.1016/j.ridd.2011.03.007Get rights and content

Abstract

Thirteen children and young adults with Down syndrome (DS) completed tests of language and reading and their performance was compared to that of three control groups. Reading comprehension was confirmed to be a specific deficit in DS and found to be strongly correlated with underlying language skills. Although reading comprehension was more strongly related to language ability in the DS group, this was shown to be a function of more advanced word recognition rather than a characteristic of DS per se. Individuals with DS were found to have greater difficulty with inferential comprehension questions than expected given their overall comprehension ability and the reading profile associated with DS was found to be similar to that of children known as poor comprehenders. It is recommended that oral language training programs, similar to those that have been shown to improve reading comprehension in poor comprehenders, be trialed with children who have DS.

Highlights

► Reading comprehension is a specific deficit in DS and strongly related to underlying language skills. ► Making inferences during reading is poorer than expected given overall comprehension ability. ► DS is characterised by a reading profile similar to that found in ‘poor comprehenders’. ► Intervention focusing on oral language skills is recommended to improve reading comprehension in DS.

Introduction

Down syndrome (DS) is the most common genetic cause of learning disability with a prevalence of 1 in 700 (Roizen, 1997). Reading is a critical skill for individuals with learning disabilities as it may open up vocational opportunities, facilitate increased independence and help improve language and communication skills (Buckley, 1985). The cognitive profile observed in DS is typically uneven with language ability being more affected than nonverbal ability (e.g., Laws & Bishop, 2003) and reading accuracy (Cardoso-Martins, Peterson, Olson, & Pennington, 2009). Most children with Down syndrome can learn to read, although levels of attainment vary considerably (Kay-Raining Bird et al., 2008, Laws and Gunn, 2002, Sloper et al., 1990). Published research focusing on reading comprehension in DS is more limited than that on reading accuracy, but the available evidence has identified reading comprehension as an area of significant difficulty (e.g., Carr, 1995).

Studies have shown that reading comprehension is correlated with measures of language (e.g., Boudreau, 2002, Laws and Gunn, 2002), and data from a recent study in Italian (Roch & Levorato, 2009) suggest that language comprehension may in fact be a more important determinant of reading comprehension in this population. On this view, the reading comprehension deficit in DS is the result of underlying language difficulties. Alternatively, or additionally, the reading comprehension deficit could be caused by weaknesses in higher level processes; the process of making inferences during reading was identified as a potential difficulty in DS in a case study by Groen, Laws, Nation, & Bishop (2006). The current study tested the hypotheses that reading comprehension in DS is more strongly related to language skills than in typically developing (TD) children and that inferential comprehension questions pose a particular problem for individuals with DS.

In order to understand the reading comprehension difficulties associated with DS it is useful to consider models of reading comprehension and relationships between reading comprehension and underlying skills in typical development. Theoretical models typically conceptualise reading comprehension as the product of two skills or sets of skills. In the ‘Simple View of Reading’ (SVR) Gough and Tunmer (1986) proposed that reading comprehension is the product of decoding (word reading) and linguistic comprehension. Linguistic comprehension is seen as a complex process, involving the ability to access lexical information and derive sentence and discourse interpretations. In support of the SVR, word recognition skills and components of linguistic comprehension (vocabulary and grammar) have been found to be significant predictors of reading comprehension in typically developing children (e.g., Catts, Adlof, & Weismer, 2005, Kendeou et al., 2009, Muter et al., 2004). According to the convergent skills model proposed by Vellutino, Tunmer, Jaccard, & Chen (2007) the relative contribution of word reading and language skills to reading comprehension are not stable across development; word reading is the best predictor of reading comprehension in the early stages of reading development, while language skills become the dominant predictor once accurate word identification has been achieved.

Recent models of reading comprehension emphasize the role of lower level language skills, such as knowledge of word meanings and syntax. The convergent skills model (Vellutino et al., 2007) views semantic and syntactic knowledge as key skills that feed into language comprehension and hence into reading comprehension. The lexical quality hypothesis (Perfetti & Hart, 2002) also emphasizes the role of language skills in reading comprehension; placing word meanings at the interface between word identification and reading comprehension. While semantic and syntactic knowledge enable the reader to compute the meanings of individual words and sentences, other processes must be brought to bear to integrate the meanings of successive sentences and incorporate background knowledge to create a representation of the text. Inferences are necessary to establish links between events in the text and establish connections between what is presently being read and existing knowledge (Graesser, Singer, & Trabasso, 1994). There are many different forms of inferences, such as pronoun resolution, bridging, elaborative and evaluative inferences. Of these, bridging inferences are considered to be of particular importance, as according to Haviland and Clark (1974) they serve to establish connections between events in the text. For example, to fully understand the statement ‘The man threw his paper on the fire. The ashes rose up the chimney’. The listener/reader must remember the two sections of text and integrate them to understand the relationship between the two events; that the first event caused the second (Singer, 1993). Without this connection coherence would be lost. Inference making takes place within the constraints of working memory, as noted in models such as the Landscape model (van den Broek, Risden, Fletcher, & Thurlow, 1996). Therefore, limitations in working memory may affect a reader's ability to make inferences while reading. In support of this, verbal working memory has been found to correlate with performance on inferencing tasks and directly with reading comprehension (e.g., Oakhill, Cain, & Bryant, 2003).

The most obvious cause of reading comprehension failure is a word reading difficulty. If an individual is unable to read with sufficient accuracy and fluency then the information extracted from the text may not be sufficient for comprehension. According to the reading comprehension models presented above, language difficulties provide another source of reading comprehension failure. Children referred to as ‘poor comprehenders’ are children who experience comprehension difficulties despite having age appropriate word recognition and decoding skills (Cain and Oakhill, 1999b, Nation and Snowling, 1998). Research with this group of children allows for the exploration of the role of factors other than word recognition in reading comprehension. Studies have reported that poor comprehenders exhibit deficits in lower level language processes such as vocabulary knowledge (Catts et al., 2005) and morphosyntactic skills (Nation and Snowling, 2000, Nation et al., 2004). Evidence for a causal role for poor vocabulary knowledge in reading comprehension difficulties comes from a randomised control trial by Clarke et al. (2010), which showed that an oral language program led to greater gains in vocabulary, which led to gains in reading comprehension. There is also evidence for deficits in higher level processes such as verbal working memory (Nation, Adams, Bowyer-Crane, & Snowling, 1999), comprehension monitoring (Oakhill, Hartt, & Samols, 2005) and inference generation (Cain, Oakhill, Barnes, & Bryant, 2001). Cain and Oakhill (1999a) argue that a difficulty making inferences is causally related to the reading comprehension deficit as such difficulties are poorer than expected given overall comprehension level.

Reading comprehension has been found to be an area of particular difficulty for individuals with DS. Carr (1995) found that the average lag between reading accuracy and reading comprehension was 11 months in a group of adults. It has been suggested that underlying language weaknesses contribute to the reading comprehension deficit found in individuals with DS and in support of this, studies have found significant correlations between measures of language and reading comprehension (e.g., Boudreau, 2002, Laws and Gunn, 2002). However, the measures of reading comprehension used in these earlier studies focussed on the comprehension of single written words, phrases or sentences rather than passages of text, and floor effects were evident in the DS groups. Data from a recent study in Italian (Roch & Levorato, 2009) suggest that language comprehension may in fact be a more important predictor of reading comprehension in DS than in TD children. Roch and Levorato (2009) went on to hypothesize that individuals with DS show an uneven reading profile similar to that exhibited by poor comprehenders and propose that an investigation of factors found to influence reading comprehension in poor comprehenders, such as inferential skills, working memory and meta-comprehension skills, could prove fruitful.

The findings of a case study by Groen et al. (2006) suggest that individuals with DS may have particular difficulties making inferences during reading, a difficulty that could be causally related to the reading comprehension deficit as has been proposed in poor comprehenders. An 8 year old girl with DS (KS) was found to have a level of reading comprehension significantly below her reading accuracy level when assessed using The Neale Analysis of Reading Ability – Revised (NARA II, form II, Neale, 1997) but not when assessed using the WORD (Wechsler, 1993). Critically, the NARA II includes questions assessing both memory for information presented in the text (literal questions) and questions that require an inference to be made, while the WORD (Wechsler, 1993) contains mainly literal questions. Inferential questions have been shown to pose greater difficulties than literal questions for typically developing readers (Priya & Wagner, 2009) and it has been reported that poor comprehenders have greater difficulty with inference based questions than literal ones (e.g. Bowyer-Crane and Snowling, 2005, Cain and Oakhill, 1999a, Cain and Oakhill, 1999b).

The present study compared the reading and related cognitive skills of a group of children and adolescents with DS to that of three groups of children, typically developing children matched on single-word reading ability, typically developing children matched on reading comprehension ability and a group of ‘poor comprehenders’ matched on reading accuracy and reading comprehension ability. Matching groups on reading accuracy (reading age, RA) ensures that any resulting difference in reading comprehension is not the product of differences in reading accuracy and must therefore reflect differences in other underlying skills. The use of a comprehension age matched comparison (CAM) group means that differences in underlying skills, such as vocabulary, working memory or inferencing skill, are not a result of differences in reading comprehension and likely to be associated with the cause of difficulties. The poor comprehender group was included to investigate the extent of the similarity between the two developmental disorders as highlighted by Roch and Levorato (2009).

  • 1.

    Do individuals with DS have a difficulty with inferential questions that goes beyond their general comprehension level? Based on the Groen et al. (2006) study we predicted that the DS group would have greater difficulty with the inferential comprehension questions than the RA and CAM control groups.

  • 2.

    What are the contributions of word reading and language skills to reading comprehension in DS? Roch and Levorato (2009) found that underlying language skills made a greater contribution to reading comprehension in Italian children with DS compared to a comprehension-age matched control group. On the basis of this we predicted that this would be the case in our sample, when the DS group was compared to the CAM group. We included measures of vocabulary and verbal working memory as these have been found to be related to reading comprehension in typically developing children (Muter et al., 2004, Oakhill et al., 2003) and found to be impaired in poor comprehenders (Catts et al., 2005, Nation et al., 1999). A verbal working memory measure such as listening recall requires both the processing and storage of verbal information. The processing component of this task requires language comprehension and so this task can be considered to be a measure of higher level language ability.

  • 3.

    Do individuals with DS have a similar reading profile to poor comprehenders? It has been hypothesized that there is overlap in the reading profiles of the two groups, but this has not been tested empirically.

All groups were assessed using a range of standardised tests measuring single word reading, decoding, passage reading accuracy and comprehension, vocabulary, non-verbal ability, and verbal working memory. An experimental passage reading task was created with literal and inferential comprehension questions.

Section snippets

Participants

Seventeen participants with Down syndrome (DS) were initially recruited to this study. However, four were not able to read sufficiently well to complete all the reading tasks and the final sample size for the DS group was thirteen (aged 11 years 4 months–19 years 3 months). Twelve participants had full trisomy 21, while one individual had translocation DS. Furthermore, 12 participants attended mainstream education and only one participant attended a school specialising in learning difficulties.

Group profiles on the standardised measures of language and reading

Descriptive data for the standardised measures are shown in Table 1, presented as mean raw scores. Separate between subject ANOVAS were run for each variable. Significant main effects were followed up with Tukey's HSD post hoc test and significant group differences are indicated. To explore the relative contributions of word reading and language skills to reading comprehension in the DS group, performance will first be compared to that of the RA group and then to the CAM group. Correlational

Discussion

Previous studies have identified reading comprehension as an area of particular difficulty for individuals with DS and shown that it is related to language ability. However, these earlier studies were limited both by the low levels of reading comprehension of their DS participants and by the measures used. The current study sought to extend previous work by including individuals with DS who had measurable levels of passage reading comprehension, measures of both lower and higher level language

Acknowledgements

We would like to thank Alexandra Petrov and Lisa Henderson for their assistance with data collection and of course the children, families and schools for their time. The preparation of this manuscript was supported by Wellcome Programme Grant 082036 and we would like to thank Maggie Snowling and Paula Clarke for their helpful comments.

References (42)

  • K. Nation et al.

    Working memory deficits in poor comprehenders reflect underlying language impairments

    Journal of Experimental Child Psychology

    (1999)
  • K. Nation et al.

    Semantic processing and the development of word recognition skills: Evidence from children with reading comprehension difficulties

    Journal of Memory and Language

    (1998)
  • D. Boudreau

    Literacy skills in children and adolescents with Down syndrome

    Reading and Writing: An Interdisciplinary Journal

    (2002)
  • C. Bowyer-Crane et al.

    Assessing children's inferencing generation: What do tests of reading comprehension measure?

    British Journal of Educational Psychology

    (2005)
  • S. Buckley

    Attaining basic educational skills: Reading, writing and number

  • K. Cain et al.

    Inference making and its relation to comprehension failure

    Reading and Writing

    (1999)
  • K. Cain et al.

    Reading comprehension development from 8 to 14 years: The contribution of component skills and processes

  • K. Cain et al.

    Comprehension skill, inference making ability, and their relation to knowledge

    Memory & Cognition

    (2001)
  • C. Cardoso-Martins et al.

    Component reading skills in Down syndrome

    Reading and Writing

    (2009)
  • J. Carr

    Down's syndrome: Children growing up

    (1995)
  • H.W. Catts et al.

    Language deficits in poor comprehenders: A case for the simple view of reading

    Journal of Speech, Language and Hearing Research

    (2005)
  • P.J. Clarke et al.

    Ameliorating children's reading comprehension difficulties: A randomised controlled trial

    Psychological Science

    (2010)
  • L.M. Dunn et al.

    British Picture Vocabulary Scale II

    (1997)
  • C.D. Elliott et al.

    British Ability Scales II (BAS-II)

    (1996)
  • P.B. Gough et al.

    Decoding, reading and reading disability

    Remedial and Special Education

    (1986)
  • A. Graesser et al.

    Constructing inferences during text comprehension

    Psychological Review

    (1994)
  • M. Groen et al.

    A case of exceptional reading accuracy in a child with Down syndrome: Underlying skills and the relation to reading comprehension

    Cognitive Neuropsychology

    (2006)
  • P. Hatcher

    Predictors of reading recovery book levels

    Journal of Research in Reading

    (2000)
  • S.E. Haviland et al.

    What's new? Acquiring new information as a process in comprehension

    Journal of Verbal Learning & Verbal Behaviour

    (1974)
  • K. Kay-Raining Bird et al.

    Written and oral narratives of children and adolescents with Down syndrome

    Journal of Speech, Language and Hearing Research

    (2008)
  • P. Kendeou et al.

    Revisiting the simple view of reading

    British Journal of Educational Psychology

    (2009)
  • Cited by (0)

    View full text