who will develop dyslexia? - uva · who will develop dyslexia? cognitive precursors in parents and...

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Who will develop dyslexia? Cognitive precursors in parents and children

van Bergen, E.

Link to publication

Citation for published version (APA):van Bergen, E. (2013). Who will develop dyslexia? Cognitive precursors in parents and children.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 02 Apr 2020

https://dare.uva.nl/personal/pure/en/publications/who-will-develop-dyslexia-cognitive-precursors-in-parents-and-children(eef38ac0-c599-467f-95df-8852c9275441).html

omslag_Elsje-van-Bergen_v2.indd 1 20-12-12 19:37

Who will develop dyslexia?

Cognitive precursors in parents and children

Elsje van Bergen

23629 Bergen, Elsje van V.indd 1 19-12-12 12:24

The Dutch Dyslexia Programme was supported by the Netherlands Organisation

for Scientific Research (NWO) under project number 200-62-304.

This thesis was sponsored by Stichting Kohnstamm Fonds

Copyright © 2013 Elsje van Bergen

ISBN/EAN: 978-90-6464-625-6

NUR: 773

Cover design by Esther Ris, Proefschriftomslag.nl, Koog a/d Zaan

Layout by Ferdinand van Nispen, Citroenvlinder-dtp.nl, Bilthoven

Printed by GVO | Ponsen & Looijen drukkers & vormgevers BV, Ede


Who will develop dyslexia?

Cognitive precursors in parents and children

ACADEmISCh PROEFSChRIFT

ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam

op gezag van de Rector magnificus

prof. dr. D.C. van den Boom

ten overstaan van een door het college voor promoties

ingestelde commissie,

in het openbaar te verdedigen in de Agnietenkapel

op donderdag 14 februari 2013, te 14:00 uur

door

Elisabeth van Bergen

geboren te Landsmeer


Promotiecommissie

Promotores: Prof. dr. P.F. de Jong

Prof. dr. D.A.V. van der Leij

Co-promotor: Prof. dr. F.J. Oort

Overige Leden: Dr. B. Boets

Dr. J.E. Rispens

Prof. dr. m.J. Snowling

Prof. dr. L.T.W. Verhoeven

Prof. dr. F.N.K. Wijnen

Faculteit der maatschappij- en Gedragswetenschappen


Voor Yves


Contents

Chapter 1General Introduction 11

1.1 Dyslexia 121.2 Theoretical Framework 14

1.2.1 The multiple deficit model 141.2.2 The generalist genes hypothesis 191.2.3 The hybrid model 21

1.3 Precursors in Children 231.3.1 IQ 241.3.2 Preliteracy skills 25

1.4 Precursors in Families 281.4.1 Home literacy environment 281.4.2 Parental skills 29

1.5 The Current Studies 301.5.1 The Dutch Dyslexia Programme 301.5.2 Outline of the thesis 30

1.6 References

Chapter 2Dutch children at family risk of dyslexia: Precursors, reading development, and parental effects

39

2.1 Introduction 412.2 Methods 44

2.2.1 Participants 442.2.2 Measures 462.2.3 Procedure 482.2.4 Analytic approach 48

2.3 Results 482.3.1 Parent characteristics and home-literacy environment 482.3.2 Child characteristics 51

2.4 Discussion 532.5 References 58


Chapter 3Child and parental literacy levels within families with a history of dyslexia

63

3.1 Introduction 653.1.1 FR children without dyslexia 653.1.2 Intergenerational transfer 67

3.2 Methods 683.2.1 Participants 683.2.2 Measures 693.2.3 Procedure 71

3.3 Results 723.3.1 Data screening 723.3.2 Children 723.3.3 Intergenerational transfer 74

3.4 Discussion 793.5 Key Points 833.6 References 84

Chapter 4The effect of parents’ literacy skills and children’s preliteracy skills on the risk of dyslexia

87

4.1 Introduction 894.1.1 Risk factors in families 894.1.2 Risk factors in children 904.1.3 Specificity of precursors 91

4.2 Method 924.2.1 Participants 924.2.2 Measures 94

4.3 Results 964.3.1 Family characteristics in relation to children’s group 1014.3.2 Children’s characteristics in relation to children’s group 1014.3.3 Predictors of children’s reading skills 101

4.4 Discussion 1044.4.1 Intergenerational transfer 1064.4.2 Home literacy environment 1074.4.3 Preliteracy skills and their specificity 107

4.5 References 112


Chapter 5IQ of four-year-olds who go on to develop dyslexia 117

5.1 Introduction 1195.2 Method 121

5.2.1 Participants 1215.2.2 Measures 1225.2.3 Procedure 125

5.3 Results 1255.3.1 Structure of the IQ test 1255.3.2 The relationship of IQ at age 4 and second-grade school

achievement128

5.3.3 Differences among reading groups 1285.4 Discussion 1315.5 References 135

Chapter 6General Discussion 139

6.1 Precursors in Children 1416.1.1 IQ 1416.1.2 Preliteracy skills 143

6.2 Literacy Skills in Children 1466.3 Precursors in Families 147

6.3.1 Home literacy environment 1486.3.2 Parental skills 150

6.4 The Intergenerational Multiple Deficit Model 1536.5 Future Directions 1566.6 Concluding Remarks 1586.7 References 159

Summary 164Samenvatting (Summary in Dutch) 167Acknowledgements 173Dankwoord (Acknowledgements in Dutch) 177List of Publications 182Curriculum Vitae 184


Writing – the art of communicating thoughts to the mind, through the eye – is

the great invention of the world. Great, very great in enabling us to converse

with the dead, the absent, and the unborn, at all distances of time and space;

and great, not only in its direct benefits, but greatest help, to all other inventions.

- Abraham Lincoln -

General Introduction

Chapter

chapter-intros_Elsje-van-Bergen_v2.indd 1 18-12-12 20:4523629 Bergen, Elsje van V.indd 10 23-12-12 12:49

General Introduction

Chapter


Chapter 1

12

At this very moment, precisely controlled eye movements make you process a

few black marks on a white page from the general introduction of this thesis.

The message I convey from my brain to your brain is in fact coded as just a

collection of straight and curved lines. Yet such a collection of tiny lines can

transfer knowledge or bring a story to life. Our ability to read and write is an

impressive feat and life today is impossible to imagine without it.

From an evolutionary perspective, reading and writing are relatively

recent inventions. homo sapiens have existed for 200,000 years, but reading

and writing dates back roughly 5.000 years (Powell, 2009). Nevertheless,

it has been only some 100 years since nearly all people in Western societies

have become literate. It probably does not come as a surprise that the start of

compulsory education, and hence reading education for all, roughly coincides

with the birth of its study, the science of reading.

From a neuroscientific perspective, simply reading a single word is

already a very complex task, let alone reading this thesis. Learning to read is not

something that comes naturally to us, like learning to walk or talk. Instead, it takes

a couple of years of deliberate practice. But once mastered, we recognize words

instantly. A 10-year old skilled reader can read a word that is flashed on a screen

for just 200 milliseconds (Yap & van der Leij, 1993). On seeing the word a cascade

of orchestrated neural firing is triggered, making us recognize the black marks

as a meaningful word. The beauty of it is that this all happens very quickly and

automatically; you cannot help but read.

1.1 Dyslexia

In view of the complexity of word recognition it is perhaps not surprising that

not all children learn to read without difficulty. Although the exact definition

and diagnostic criteria of dyslexia are still hotly debated (e.g., Brown Waesche,

Schatschneider, maner, Ahmed, & Wagner, 2011; Fletcher, Coulter, Reschly, &

Vaughn, 2004; Kleijnen et al., 2008; Snowling & hulme, 2012), there is consensus

among researchers and practitioners that individuals with dyslexia experience

difficulty in mastering word-level reading skills and show persistent problems

in the accuracy or fluency of word-reading. These problems in word-level

reading1 often go hand in hand with problems in word-level spelling.

Throughout this thesis I will use ‘reading’ to refer to word-level reading , rather than reading comprehension.


General IntroductIon

13

1 There are large individual differences in reading fluency, that is,

how quickly and accurately single words are read. Individuals with dyslexia

are at the bottom of the distribution of normal variation in reading fluency.

Slow speed of processing and relatively high error rates indicate their lack of

automatic decoding (van der Leij & van Daal, 1999). It can readily be seen that

the prevalence rate of dyslexia reflects the cut points established as criteria

for identification, as well as the exclusion criteria that are used. According to

the criteria proposed in the Dutch Dyslexia protocol, it is estimated that 4% of

Dutch primary school children have dyslexia (Blomert, 2006).

Children with dyslexia might experience great difficulty in education.

After the initial stage of learning acquisition, an educational transition takes

place from learning to read to reading to learn. Struggling with the task of

reading itself hampers a child with dyslexia to employ reading to learn about

for example science or history. Indeed, dyslexia negatively impacts long-term

academic outcomes (Willcutt et al., 2007). moreover, dyslexia is associated with

increased risks of anxiety disorders, depression, attention-deficit/hyperactivity

disorder (ADhD), conduct disorder (Bosman & Braams, 2005; Carroll, maughan,

Goodman, & meltzer, 2005; Willcutt et al., 2007), and even criminal behaviour

(Kirk & Reid, 2001; macdonald, 2012).

Given the trouble that dyslexia might cause, it is very important to study

in what way individuals with dyslexia differ from their peers without dyslexia.

The ultimate goal is to unravel the causal pathway of aetiological risk factors

leading to dyslexia. Insight into these pathways will inform the development

of effective interventions to ameliorate the impact of dyslexia.

From research comparing individuals with and without dyslexia, we

know that those with dyslexia generally show deficits in specific cognitive areas,

like the processing of speech sounds in spoken words. Finding a cognitive deficit

linked to dyslexia raises the question as to whether the deficit is a consequence

of children’s reading problems or whether it was already present before they

came to the task of learning to read. If so, the deficit is said to be a precursor

of dyslexia. Furthermore, to better understand the developmental pathways we

need to find out whether a precursor of reading ability or disability is specifically

linked to reading development or whether it is shared with for instance arithmetic

development. Beyond studying these cognitive characteristics intrinsic to

the child, the current research will explore a new avenue by investigating the

cognitive risks that parents pass on to their child.


Chapter 1

14

1.2 Theoretical Framework

The issues pursued in the current thesis were guided by previous empirical

work on cognitive deficits associated with dyslexia and by theoretical work,

specifically the multiple deficit model and the generalist genes hypothesis.

This theoretical framework will be discussed in the following two sections,

concluded by a third section about the proposed hybrid model.

1.2.1 The Multiple Deficit ModelResearch into dyslexia was dominated for a long time by the quest for the holy

Grail: the single cognitive deficit that is necessary and sufficient to cause all

behavioural characteristics of the disorder. In the case of dyslexia the dominant

hypothesis about a single causal factor has been the phonological-deficit

hypothesis (e.g., Snowling, 1995; Wagner, 1986); other hypotheses include

core deficits in rapid auditory processing (Tallal, 1980), amplitude-envelope

rise-time discrimination (Goswami et al., 2002), visual attention span (Bosse,

Tainturier, & Valdois, 2007), visuo-spatial attention (Vidyasagar & Pammer, 2010),

magnocellular processing of fast sensory information (Stein & Walsh, 1997), the

ability to perform skills automatically due to cerebellar deficits (Nicolson, Fawcett,

& Dean, 2001), and the forming of stimulus-specific anchors (Ahissar, 2007).

however, single cognitive deficit models have a number of

shortcomings (see Pennington, 2006, for a comprehensive overview). First,

there is no single cognitive deficit found that can explain all behavioural

symptoms of all cases with dyslexia. For example, not all individuals with

dyslexia show a phonological deficit (e.g., Pennington et al., 2012; Valdois et

al., 2011), which is the main candidate for a single cognitive deficit explanation.

Conversely, not all individuals with a phonological deficit have dyslexia (e.g.,

Bekebrede, van der Leij, Schrijf, & Share, 2010; Snowling, 2008). This questions a

one-to-one mapping and points to the possibility that various constellations of

underlying cognitive deficits can lead to the behavioural symptoms of dyslexia.

Second, these models cannot readily explain the phenomenon of

comorbidity. In the case of dyslexia, comorbidity refers to the fact that dyslexia

co-occurs more often than expected by chance with other developmental

disorders, like dyscalculia, specific language impairment, speech-sound

disorder, or ADhD. To illustrate this point, suppose disorder A and B each

have a prevalence of 5% in the general population. If disorder A and B were



15

1independent, then the chance that A and B co-occur would only be 0.05 * 0.05

= 0.0025, or 0.25%. however, comorbidity rates for developmental disorders

are commonly in the order of 30%, for example between dyslexia and speech-

sound disorder or dyslexia and ADhD (Pennington, 2006). The huge discrepancy

between these figures (0.25% vs. 30%) implies that developmental disorders

are not independent.

The single deficit model requires for each comorbidity (pair of disorders)

a distinct account. Pennington (2006) discusses as an example the comorbidity

between dyslexia and speech-sound disorder. Speech-sound disorder is

defined by difficulties in the development of spoken language, especially

problems with the intelligible production of speech sounds. Approximately

30% of children with early language or speech problems go on to develop

dyslexia. A parsimonious single deficit model to explain this comorbidity is the

severity hypothesis. The severity hypothesis states that speech-sound disorder

and dyslexia have the same underlying phonological deficit, with speech-

sound disorder being an earlier developmental manifestation of this deficit

than dyslexia. Comorbid cases will have the most severe phonological deficit.

If the phonological deficit is less severe, speech-sound disorder will not reach

clinical boundaries but dyslexia will. To account for cases with early speech-

sound disorder but without later dyslexia, the model must pose a subtype

of speech-sound disorders that is caused by a phonological deficit distinct

from the phonological deficit as seen in cases with dyslexia. Alternatively, the

phonological deficit in such cases must be resolved by the time they come to

the task of learning to read. however, Snowling, Bishop, and Stothard (2000)

followed a group of former language-impaired children into adolescence.

Those with early speech-sound disorder (isolated phonological impairments

at 4 years of age) had normal reading skills at age 15, but continued to show

phonological deficits. Similar results were obtained by Peterson, Pennington,

Shriberg, and Boada (2009). In their study many children with early speech-

sound disorder went on to learn to read normally despite a lasting phonological

deficit. Thus, in both studies the children with early speech-sound disorder

had a phonological deficit similar to children with dyslexia. This conclusion is

inconsistent with the single cognitive deficit severity hypothesis.

While research at the cognitive level of explanation was still searching

for a single deficit, studies at the genetic level converged on the conclusion

that the aetiology of dyslexia and other developmental disorders is genetically


Chapter 1

16

complex (Pennington, 2006). So instead of a single gene determining dyslexia,

many genes act probabilistically, each having only a small contributory effect

to the aetiology of dyslexia (Bishop, 2009). moreover, behavioural genetic

studies showed for certain developmental disorders that the relation between

two traits (like reading ability and inattention) is larger in monozygotic twin

pairs than in dizygotic twin pairs (Willcutt, Pennington, & DeFries, 2000). Such

a bivariate heritability supports genetic overlap between the conditions, in this

example between dyslexia and ADhD. The partly shared aetiology of dyslexia

and ADhD does not yet rule out the possibility of a distinct single cognitive

deficit for each disorder. however, studies have demonstrated that a processing

speed impairment is not only a characteristic of dyslexia, but also of ADhD (e.g.,

Willcutt, Pennington, Olson, Chhabildas, & hulslander, 2005), suggesting that

processing speed is a shared cognitive risk factor (mcGrath et al., 2011). The

accumulating evidence for aetiological and cognitive overlap between dyslexia

and ADhD speaks against a single deficit model for explaining their frequent

co-occurrence. Also for other dyslexia comorbidities, shared cognitive deficits

are found, for example a phonological deficit in specific language impairment

(e.g., Bishop, mcDonald, Bird, & hayiou-Thomas, 2009) and a processing-speed

deficit in dyscalculia (e.g., van der Sluis, de Jong, & Leij, 2004).

It seems that single deficit models are untenable and must give way to

multiple cognitive deficit models for understanding developmental disorders.

The multiple cognitive deficit model proposed by Pennington (2006) is depicted

schematically in Figure 1.1. In his model, multiple genetic and environmental

risk factors operate probabilistically by increasing the liability to a disorder;

conversely, protective factors decrease the liability. These aetiological factors

produce the behavioural symptoms of developmental disorders by influencing

the development of relevant neural systems and cognitive processes.

Importantly, there is no single aetiological or cognitive factor that is sufficient

to cause a disorder. Instead, multiple cognitive deficits (each due to multiple

aetiological factors) need to be present to produce a disorder at the behavioural

level. Some of the aetiological and cognitive risk factors are shared with other

disorders. Consequently, comorbidity among developmental disorders is to be

expected, rather than something that requires additional explanations. Finally,

from Pennington’s multiple deficit model it follows that the liability distribution

for a given disorder is continuous and quantitative, rather than being discrete

and categorical. Therefore, thresholds to define disorders are rather arbitrary.


general introduction

17

1

Behavioural disorders

Cognitive processes

Neural systems

Aetiological risk and protective factors

Level of analysis

GG1 EE1 GG2

NN1 NN2 NN3

CC1 CC2 CC3

DD1 DD2 DD3

Figure 1.1. Pennington’s multiple de cit model. Double headed arrows indicate

interactions. Causal connections between levels of analysis are omitted. G = genetic risk or protective factor

E = environmental risk or protective factor N = neural system

C = cognitive process D = disorder

Figure 1.1. Pennington’s multiple defi cit model. Double headed arrows indicate interactions. Causal connections between levels of analysis are omitted.

G = genetic risk or protective factorE = environmental risk or protective factorN = neural systemC = cognitive processD = disorder


Chapter 1

18

Pennington (2006) concludes his paper by remarking that it remains

challenging to test the multiple cognitive deficit model. The model is much

more complex than single deficit models, which are attractively parsimonious,

but this complexity is needed to account for the observations at the different

levels of analysis. The model is universally applicable to developmental

disorders, but therefore remains abstract. It is not specified which aetiological

factors, neural systems, and cognitive processes interact to produce a given

disorder.

We would like to argue that a line of inquiry that can contribute to

testing and specifying the multiple deficit model are family-risk studies. In

family-risk studies, children are followed who are at risk of dyslexia by virtue

of having an immediate dyslexic family member (usually a parent). Such

studies have shown that 34 to 66% of them develop dyslexia (Elbro, Borstrøm,

& Petersen, 1998; Pennington & Lefly, 2001; Scarborough, 1990; Snowling,

Gallagher, & Frith, 2003; Torppa, Lyytinen, Erskine, Eklund, & Lyytinen, 2010),

depending on the stringency of the dyslexia criteria. The much higher

prevalence of dyslexia among offspring of parents with dyslexia is consistent

with twin studies showing moderate to strong heritability of dyslexia (e.g.,

Olson, Byrne, & Samuelsson, 2009).

From the multiple deficit model it follows that children at family risk

experience at least some of the aetiological risk factors: they inherit genetic

risk factors and might experience a less rich literacy environment. hence, it is

hypothesized that at-risk children have a higher genetic and environmental

liability than children without a family history of dyslexia (labelled control

children). Furthermore, the at-risk children who go on to develop dyslexia

are expected to show cognitive deficits in several processes. Some of these

cognitive processes are expected to be affected even before the onset of

reading instruction, as a consequence of aetiological risk factors and deficient

neural systems.

A key prediction of the multiple deficit model for family-risk studies

concerns the at-risk children who do not develop dyslexia. If liability to dyslexia

were discrete (as would happen if only one factor, say a gene, were involved), at-

risk non-dyslexic children would not differ from controls. however, according to

the multiple deficit model, liability is continuously distributed. This also follows

from the fact that reading ability is influenced by many genes of small effect,

producing normal distributions of phenotypes (Plomin, DeFries, mcClearn, &



19

1mcGuffin, 2008, p. 33). Consequently, the multiple deficit model predicts that at-

risk children without dyslexia also inherit at least some disadvantageous gene

variants from their dyslexic parents, giving them a higher liability than control

children, although still lower than at-risk dyslexic children. At the behavioural

and cognitive level this should translate into mild deficits in literacy skills and

some of its cognitive underpinnings. When plotting mean performances of the

three groups, a step-wise pattern (i.e., at-risk dyslexic < at-risk non-dyslexic <

controls) would support a continuum of liability, one of the characteristics of

the multiple deficit model.

Comparing the three groups of children on behavioural measures sheds

light on cognitive deficiencies and behavioural symptoms, the bottom two levels

in Figure 1.1. These three groups have also been compared on neural processing

of visual and auditory stimuli (e.g., Leppänen et al., 2010; Plakas, van Zuijen, van

Leeuwen, Thomson, & van der Leij, 2012; Regtvoort, van Leeuwen, Stoel, & van

der Leij, 2006), the second level of the multiple deficit model. Some family-risk

studies have also examined aspects of the home environment, which belong

to the aetiological level. however, specific genetic risk factors remain hidden in

family-risk studies. As genetic screening of children for their dyslexia susceptibility

is still far away, we propose an indicator of their genetic risk. Since reading ability

is moderately to highly heritable and children receive their genetic material from

their parents, we argue that cognitive abilities of parents can partly reveal their

offspring’s liability. The parents of at-risk children will have weaker reading skills

than those of control children, reflecting selection criteria in family-risk studies,

but the key issue is whether parental reading skills differentiate between at-risk

children with and without dyslexia. Based on the multiple deficit model it is

expected that at-risk children who develop dyslexia have inherited more genetic

risk variants than at-risk children without dyslexia and that this difference can be

revealed by lower reading performance of parents of the at-risk dyslexic children.

In Section 1.4.2 I will elaborate upon parental effects.

1.2.2 The Generalist Genes HypothesisIn our family-risk study we seek to identify cognitive processes playing a role in

the developmental pathways that lead to dyslexia. The multiple deficit model

states that some cognitive deficits are shared among disorders. This raises

the question of which cognitive precursors of dyslexia are distinct and which

are shared with other disorders. With regard to learning abilities, like reading


Chapter 1

20

ability, there is a hypothesis that addresses this specificity issue: the generalist

genes hypothesis (Kovas & Plomin, 2007; Plomin & Kovas, 2005).

The generalist genes hypothesis stems from behavioural genetic studies

employing the twin design. The twin design is the major method to quantify

genetic and environmental influences on a trait. It compares the resemblance of

monozygotic twins, who are genetically identical, with dizygotic twins, who are

on average 50% similar genetically. If for a certain trait monozygotic twins are

more similar than dizygotic twins, genetic factors must play a role. If there is no

difference in resemblance heritability is negligible. For dyslexia, the concordance

rate (that is, the likelihood that one twin will be affected if the other twin is affected)

is 75% for monozygotic twins and 43% for dizygotic twins (Stromswold, 2001),

suggesting substantial heritability. Estimates for the heritability of reading ability

are in the range of .47 to .84 (Taylor, Roehrig, hensler, Connor, & Schatschneider,

2010, and Byrne et al., 2009, respectively).

Recently, the field of behavioural genetics has moved beyond

quantifying genetic and environmental influences on a trait to studying genetic

and environmental overlap between traits. For the three learning abilities

reading, arithmetic, and language, empirical data have shown that the genes

important for one learning ability largely overlap with the genes important for

the other learning abilities. The genetic correlation is the measure that quantifies

this: it indexes the extent to which genetic influences on one trait correlate with

the genetic influences on another trait (independently of the heritability of the

traits). The genetic correlation between learning abilities is about .70 (Kovas &

Plomin, 2007; Plomin & Kovas, 2005). This suggests that half (.702) of the genes

associated with reading ability are generalists: they also influence other learning

abilities. hence Plomin and Kovas (2005) named their hypothesis the ‘generalist

genes hypothesis’. As genetic correlations are not 1.0, there are also specialist

genes: genes that contribute to dissociations among learning abilities. however,

most genes associated with reading are generalists, also in another respects:

twin studies have demonstrated that genetic influences on the lower end of the

reading distribution (containing children with dyslexia) are largely the same as

the genetic influences on the entire distribution (Kovas, haworth, Dale, & Plomin,

2007; Plomin & Kovas, 2005).

Observed differences in learning abilities among individuals are also

partly due to differences between the environments in which individuals

were born, were brought up and live. Behavioural genetics subdivides



21

1environmental influences into those that make family members similar (called

shared environmental effects) and those that do not contribute to resemblance

among family members (called non-shared environmental effects). Shared

environmental effects include shared family experiences or sharing the same

school or community. Examples of non-shared environmental effects are

accidents, illnesses, hobbies, and influences of individuals such as peers and

teachers. Also for these environmental components statistics exist analogous

to genetic correlation. Shared environmental correlations among learning

abilities are as high as genetic correlations, so shared environmental effects

are also largely general effects (Kovas & Plomin, 2007). In contrast, non-shared

environmental correlations are low. This indicates that these effects primarily

act as specialists, contributing to performance differences in learning abilities

within a child (Kovas & Plomin, 2007).

1.2.3 The Hybrid ModelThe generalist genes hypothesis and the multiple cognitive deficit model

complement each other well. The multiple deficit model is more general

because it holds for all common developmental disorders, while the generalist

genes hypothesis specifically pertains to learning abilities and disabilities.

Furthermore, the multiple deficit model includes four levels of explanation,

whereas the generalist genes hypothesis only concerns the aetiological level.

Nevertheless, the generalist genes hypothesis details for learning abilities the

degree of overlapping and unique influences in each of the three aetiological

components (genetical, shared environmental, and non-shared environmental

influences). I have visualised the generalist genes hypothesis and incorporated

it into the multiple deficit model, yielding the hybrid model depicted in Figure

1.2. In this model only the first and the fourth layer are further specified

because the generalist genes hypothesis only deals with these two levels. The

aetiological factors of the first level influence the behavioural manifestations at

the fourth level by acting through the second and third level.

The hybrid model quantifies the overlap in aetiological factors between

learning abilities: genetic and shared environmental effects are largely shared

by the three learning domains, whereas the non-shared environmental effects

are largely distinct. These differential overlaps are visualised in the hybrid


cHaPter 1

22


Cognitive processes

Neural systems


Level of analysis

N1 N2 N3

C1 C2 C3

RD AD LD

Figure 1.2. Hybrid model for learning (dis)abilities, consisting of the generalist

genes hypothesis (Kovas & Plomin, 2007) and the multiple de cit model (Pennington, 2006). Double headed arrows indicate interactions. Causal

connections between levels of analyses are omitted. N = neural system

C = cognitive process RD = reading disorder; AD = arithmetic disorder; LD = language disorder

Genetic Shared environment

Non-shared environment

Figure 1.2. hybrid model for learning (dis)abilities, consisting of the generalist genes hypothesis (Kovas & Plomin, 2007) and the multiple de!cit model (Pennington, 2006). Double headed arrows indicate interactions. Causal connections between levels of analyses are omitted.

N = neural systemC = cognitive processRD = reading disorder; AD = arithmetic disorder; LD = language disorder



23

1model as the degree of overlap between the circles. Despite this quantification

of aetiological overlap, the hybrid model does not specify which aetiological

factors are relevant. Regarding genetic factors, molecular genetic studies will

ultimately inform us which genes are implicated in dyslexia. Knowledge of

specific genes contributing to dyslexia susceptibility promises to help bridge

the gap from genes to neural systems, cognitive processes, and behavioural

outcomes (Fisher & Francks, 2006). Until now, only a handful of candidate

genes are identified, each of small effect. As an example of the small effects, the

most common form of the candidate gene KIAA0319 is found equally frequent

in people with and without dyslexia. The high-risk version is found in 35% of

those with dyslexia and 27% of those without, whereas the figures for the low-

risk version are 24% versus 36% (Bishop, 2009).

Insight into which specific neural systems, cognitive skills, and

behavioural symptoms are implicated in dyslexia can be gained from family-

risk studies. The hybrid model points to the opportunity to study reading in

combination with arithmetic or language to increase insight into shared and

distinct factors. We chose to focus on reading and arithmetic, both basic

school skills central during early primary school. As the model suggests, its

disorders, dyslexia and dyscalculia, indeed often co-occur (Landerl & moll,

2010). moreover, this pair of comorbidity is under-researched compared to the

comorbidity of dyslexia with ADhD or language disorders. We aimed to study

the comorbidity issue at the cognitive level of explanation. Which cognitive

factors are specific and which are shared between the development of reading

and arithmetic?

1.3 Precursors in Children

This thesis will deal with two types of cognitive skills: general cognitive skills

and preliteracy skills. The key questions regarding cognitive precursors in

children are: Do the children who go on to develop dyslexia show cognitive

deficits before they come to the task of learning to read? And do the at-risk

children who do not develop dyslexia perform at the same level as the control

children?

Reading performance is the outcome of the combined action of

multiple factors, as depicted in Figure 1.2. Knowledge about these factors can


Chapter 1

24

be obtained by studying predictors of reading outcome. In addition, knowledge

about predictors of dyslexia is of clinical importance because it enables early

identification and therefore remedial teaching of children at heightened risk of

dyslexia.

1.3.1 IQReading ability and IQ are correlated. Therefore, when selecting a group of

children with dyslexia based on low reading achievement, it can be expected

that their group mean on IQ falls below that of a group of children who do

not show reading problems. however, scores on IQ tests might be adversely

affected by reading problems and, as a result, less print exposure or print

exposure to less complex texts (Cunningham & Stanovich, 1997). Consider

for instance conceivably poorer performance on the Verbal Comprehension

Index (including vocabulary, similarities, comprehension, information, and

word-reasoning subtests) of the WISC (Wechsler, 2004), a world-wide used

IQ battery. Thus, the correlation between reading ability and IQ might partly

be a consequence of reading ability. Therefore, longitudinal data from before

and after reading onset are required. The only two previous family-risk studies

that report IQ before Grade 1 (Snowling et al., 2003; Torppa et al., 2010) report

that the at-risk group with later dyslexia performed lower on verbal IQ, but the

studies are inconsistent regarding nonverbal IQ. moreover, IQ scores are given

but the papers are silent about their interpretation.

A partly causal relation between early verbal IQ and later reading

development is plausible, since the cognitive system for written language

builds on the cognitive system for oral language (hulme, Snowling, Caravolas,

& Carroll, 2005). Regarding nonverbal IQ, if it is associated with later reading, a

causal interpretation is less self-evident. It might be that nonverbal IQ indicates

a child’s ability to acquire cognitive skills in general. It cannot be readily seen

how this general ability translates into a specific problem in learning to decode

words. Alternatively, a possible relation between early nonverbal IQ and later

reading might be explained by a third common underlying variable. It might be

that both abilities share aetiological factors, as do different domains of learning

abilities according to the generalist genes hypothesis (see Figure 1.2). Chapter

5 and Chapter 6 (General Discussion) will further discuss the interpretation of

longitudinal relations between IQ and reading.

Chapter 5 aims to investigate the longitudinal relations between IQ



25

1and reading (dis)ability by examining IQ performance at age 4 and reading

outcome at age 8 years. Another goal of this chapter is to investigate the

specificity of the potential predictive relation between early IQ and later reading.

Are early verbal and nonverbal IQ uniquely related to later fluency in word

reading or also to later fluency in mental arithmetic? As mentioned previously,

word reading and mental arithmetic are arguably the most important basic

school skills acquired during the early school years. moreover, impairments in

acquiring these skills (i.e., dyslexia and dyscalculia, respectively) often co-occur

(Landerl & moll, 2010), making them a good pair of skills to study shared and

distinct links with early IQ. As can be seen in the hybrid model (Figure 1.2),

the aetiological and cognitive factors influencing reading and arithmetic are

assumed to be partly shared and partly unique. Investigating the relationship

of reading and arithmetic with early IQ offers insight into whether verbal and

nonverbal ability are cognitive factors related to these school skills and if so,

whether they are shared between the school skills.

1.3.2 Preliteracy SkillsBesides studying general cognitive abilities we investigated specific cognitive

abilities before the start of reading education. The cognitive skills that are

thought of as causally linked to subsequent development of decoding skills and

later development of fluent reading are phonological skills and knowledge of

grapheme-phoneme –or letter-sound– correspondence (e.g., hulme, Bowyer-

Crane, Carroll, Duff, & Snowling, in press). A third ability found to be a powerful

longitudinal (and concurrent) predictor of reading outcome is the ability to

rapidly name a matrix of items such as colours, objects, or once mastered,

digits or letters (e.g., de Jong & van der Leij, 1999; de Jong & van der Leij, 2003;

Lervåg, Bråten, & hulme, 2009).

The main indicator of phonological skills is phonological awareness,

which refers to the ability to detect and manipulate sounds in spoken words,

like phonemes or rimes. The first task to measure phonological awareness,

phoneme deletion, was developed in 1964 (Bruce) and is still widely used (see

e.g., Chapter 3). In phoneme deletion the participant is required to identify

how a spoken word would sound if one sound were omitted. For example,

“What is smile without /s/?” Other phonological-awareness tasks include

phoneme blending (e.g., “What does /c/ /a/ /t/ say?”) and segmentation (e.g.,

“What sounds do you hear in the word /s/ /u/ /n/?”), both of which are used in


Chapter 1

26

Chapter 2 and 4. Rapid naming is sometimes subsumed under the umbrella

of phonological skills (Boets et al., 2010; Wagner & Torgesen, 1987), since fast

retrieval of phonological codes is one of its task components, alongside the

intermodal coupling of visual input to phonological output (Warmington &

hulme, 2012). Others see rapid naming as a separate cognitive underpinning

of reading ability (Wolf & Bowers, 1999).

The mentioned trio of preliteracy skills –phonological awareness, rapid

naming, and letter knowledge– will be investigated in this thesis. These skills

were measured in kindergarten in both the Amsterdam sample (Chapter 2)

and the national sample (Chapter 4; see for the sample descriptions Section

1.5.1). Furthermore, Chapter 3 (national sample) will present findings on the

concurrent relations of phonological awareness and rapid naming with reading

skills, all assessed at the end of second grade. As literacy is acquired, letter

knowledge loses its correlations with reading ability (halfway through Grade 1

all letters are taught, producing ceiling effects on letter-knowledge measures),

but phonological awareness and rapid naming continue to show concurrent

correlations with reading ability (e.g., de Jong, 2011; Vaessen & Blomert, 2010),

even into adulthood (e.g., Bekebrede, van der Leij, Plakas, Share, & morfidi,

2010). The three groups of children in our family-risk studies, at-risk dyslexia,

at-risk non-dyslexia, and controls, will be compared on these preliteracy or

underlying cognitive skills of reading.

A couple of previous family-risk studies compared the groups of

children at kindergarten age. In general, the at-risk children who later become

dyslexic are found to have difficulty with phonological-awareness tasks, are

slower on rapid-naming tasks, and are less familiar with the letters of the

alphabet compared to their typically developing peers (Elbro et al., 1998;

Pennington & Lefly, 2001; Scarborough, 1990; Snowling et al., 2003; Torppa et

al., 2010). moreover, after kindergarten at-risk children with dyslexia perform

poorly on tests of phonological awareness and rapid naming, in addition to the

known difficulties in reading and spelling (Boets et al., 2010; de Bree, Wijnen, &

Gerrits, 2010; Pennington & Lefly, 2001; Snowling, muter, & Carroll, 2007).

The at-risk children who do not meet dyslexia criteria nevertheless

typically perform less well than control children on reading and spelling tasks

(Boets et al., 2010; Pennington & Lefly, 2001; Snowling et al., 2003; but see Torppa

et al., 2010 for an exception). When group means are visualized in a graph this

yields the mentioned stepwise pattern on literacy tasks: the at-risk dyslexics



27

1perform weakest, the at-risk non-dyslexics take up an intermediate position,

and the controls perform best. This pattern of results fits with the continuity

of family risk (Pennington & Lefly, 2001; Snowling et al., 2003) and with a

multifactorial aetiology of dyslexia (Pennington, 2006, Figure 1.1). Previous

research has shown mixed results, however, regarding the performance of the

at-risk non-dyslexic group on the cognitive underpinnings of reading. During

the early school years this group either performs somewhat lower or equally

well as the controls (Boets et al., 2010; Pennington & Lefly, 2001; Snowling et

al., 2003), but no significant differences have been reported on phonological

awareness and rapid naming during kindergarten (Boets et al., 2010; Elbro et

al., 1998; Pennington & Lefly, 2001; Snowling et al., 2003; Torppa et al., 2010).

Our national sample is larger, and hence, analyses have more statistical power

than those in previous studies. Therefore we have the opportunity to detect

also subtle impairments.

Returning to the multiple deficit model (Figure 1.1), this model predicts

a normally distributed liability continuum. Relating this liability continuum to

our three groups, we hypothesize the at-risk dyslexics to be in the high tail

of the distribution, the at-risk non-dyslexics to be somewhat above average

(inheriting some disadvantageous gene variants from their dyslexic parents),

and the controls to have a somewhat below average liability profile (as their

parents are average to good readers). This outlined model is in line with

the reported phenotypic findings of a step-wise pattern for literacy skills.

Combining this model with the fact that literacy, phonological-awareness, and

rapid-naming skills are normally distributed and interrelated, it follows that a

step-wise pattern for phonological awareness and rapid naming is also most

likely.

As for the relation between general cognitive ability and later reading,

we also investigated the specificity question for the relation between the

preliteracy skills and later reading. Chapter 4 examines whether the predictive

value of the trio of preliteracy skills –phonological awareness, rapid naming,

and letter knowledge– is specific for later word-reading fluency or whether

these skills are also predictive for later mental-arithmetic fluency. De Smedt

and colleagues (Boets & De Smedt, 2010; De Smedt, Taylor, Archibald, & Ansari,

2010) have proposed an hypothesis to explain the correlation between word

reading and mental arithmetic, which states that performance on these

tasks are related because both depend upon the quality of phonological


Chapter 1

28

representations in long-term memory. mental arithmetic involves fast

retrieval of phonological-based arithmetic facts. mobilizing those facts

inefficiently limits memory resources available for selecting and carrying out

suitable arithmetic procedures (hecht, Torgesen, Wagner, & Rashotte, 2001).

Therefore, we expected that both phonological awareness and rapid naming

in kindergarten would be significant predictors of third-grade arithmetic,

although not as strong as for reading, since these skills are the core precursors

of literacy, while the core precursor of arithmetic is early number competence

(e.g., Jordan, Kaplan, Ramineni, & Locuniak, 2009).

1.4 Precursors in Families

The previous section focused on an individual child (with or without family risk)

and his or her cognitive characteristics or risk factors for dyslexia. In addition,

the current thesis encompasses a broader domain of plausible risk factors

present before formal reading instruction starts. We will investigate whether

characteristics of children’s home literacy environment are related to children’s

reading outcome. Last but not least, we will study parent-child resemblance,

which signifies intergenerational transfer of cognitive skills of parents to their

offspring.

1.4.1 Home Literacy EnvironmentDoes the literacy environment that parents provide at home before school

entry influence children’s subsequent reading development? This question

will be addressed for both the Amsterdam (Chapter 2) and the national

sample (Chapter 4). We tested whether individual differences in home literacy

environment before first grade were related to individual differences in reading

performance some years later. home literacy environment is usually quantified

as the availability of books and other reading material in the home and how

often parents read to their children. Parents are often told that they should

read to their child to promote reading development. however, the effect of

shared-reading interventions on eventual reading performance is surprisingly

under-researched and those studies addressing this issue failed to demonstrate

that parent reading enhances reading ability (Sénéchal & Young, 2008).

Correspondingly, the family-risk studies that looked into literacy characteristics



29

1of the family environment did not show effects on later reading achievement

(Elbro et al., 1998; Snowling et al., 2007; Torppa et al., 2007). These findings are

in agreement with findings from behavioural genetic studies showing strong

genetic and small shared-environmental influences on reading ability (Byrne et

al., 2009; haworth et al., 2009).

1.4.2 Parental SkillsAs mentioned earlier, in family-risk studies two samples are followed, differing in

family risk by virtue of having or not having a parent with dyslexia. Accordingly,

these studies are implicitly based on a dichotomy of family risk. however, it

might well be that within the at-risk sample, the groups of children who do and

do not go on to develop dyslexia differ in their degree of family risk for dyslexia,

despite the fact that they all have an affected parents.

As briefly touched upon before, we propose that the assumption that

the at-risk children with and without dyslexia have equal liability to dyslexia is

testable by considering the cognitive skills of their parents. multiple genetic

and environmental factors act probabilistically, leading to a continuum (rather

than a dichotomy) of liability to dyslexia. Since parents provide both genes and

environments for their child, variability among children must be associated

at least partly with variability among their parents. Therefore, we studied the

relation between parents and offspring in their reading and reading-related

skills.

To return to the two groups of at-risk children, we investigated

whether the reading outcome of children is related to the severity of affected

parent’s dyslexia. This would indicate that within the at-risk sample the

affected children have a higher liability than the unaffected children, as the

multiple deficit model predicts: affected children inherit and experience more

of the numerous aetiological risks. This is tested for the Amsterdam sample

in Chapter 2 and for the national sample in Chapter 3 and 4. Replication of

findings across two independent samples is important as we are the first to

investigate this issue. Furthermore, Chapter 4 also investigates the importance

of the literacy levels of the parent without dyslexia. Chapter 6 concludes by

proposing an extended version of the multiple deficit model of Pennington

(2006; Figure 1.1): the intergenerational multiple deficit model, which includes

cognitive risks that parents pass on to their offspring.


Chapter 1

30

1.5 The Current Studies

The current set of studies are based on data of the Dutch Dyslexia Programme

(DDP). In what follows I will introduce the DDP and I will conclude this chapter

with a brief overview of the issues addressed in the chapters to come.

1.5.1 The Dutch Dyslexia ProgrammeThe Dutch Dyslexia Programme (DDP) is a multidisciplinary and multicentre

research programme, funded by The Netherlands Organisation for Scientific

Research (NWO), and initiated and executed by the University of Amsterdam,

the University of Groningen, and the Radboud University Nijmegen. It brings

researchers together from a variety of disciplines, including educational

sciences, psychology, linguistics, neurology, and genetics. The DDP consists

of three components, namely a prospective longitudinal study, intervention

studies, and genetic studies.

The present thesis presents findings of two family-risk samples, which

I will refer to as the Amsterdam and the national sample. Chapter 2 presents

the Amsterdam family-risk study. This sample of children originates from the

two intervention studies of the DDP which were conducted in Amsterdam

(Regtvoort & van der Leij, 2007; van Otterloo & van der Leij, 2009). Children

in this sample were followed from 5 (kindergarten) till 11 years of age (Grade

5). The other three studies of my thesis employ data of the ongoing national

longitudinal study of the three DDP universities. In this larger-scale family-risk

study, couples with and without a family history of dyslexia who expected a

baby were recruited. In short, children were considered at family risk if (at least)

one of their parents and one other family member had dyslexia. The children

have been followed from infancy. Chapter 3, 4, and 5 belong to this prospective

longitudinal study, covering the development of 3½ to 9 years of age (Grade 3).

1.5.2 Outline of the Thesismy thesis aims to shed light on the constellation of risks that ultimately leads

to dyslexia. In Section 1.2 and 1.3 I have introduced the topics that will be

covered. As there is no one-to-one relation between the four topics and the

four following chapters, I will provide an overview of the topics per chapter.

Chapter 2 (van Bergen et al., 2011) concerns the Amsterdam sample, which is

followed from kindergarten to Grade 5. It examines home literacy environment,



31

1preliteracy skills, reading development, and parental skills. Chapter 3 to 5 are

part of the ongoing national study. The focus of Chapters 3, 4, and 5 will be

children’s cognitive skills at age 8 (Grade 2), 6 (kindergarten), and 4 years,

respectively. Chapter 3 (van Bergen, de Jong, Plakas, maassen, & van der Leij,

2012) discusses group differences on literacy and its cognitive underpinnings

at the end of Grade 2. moreover, the effect of the literacy skills of the dyslexic

parent on children’s reading outcome are examined. Chapter 4 (van Bergen,

de Jong, maassen, & van der Leij, submitted) the intergenerational transfer

from both parents to their offpspring is investigated. moreover, reading and

arithmetic skills in Grade 3 will be predicted from home literacy environment

and preliteracy skills measured before learning to read. In Chapter 5 (van

Bergen et al., in revision) focusses on the relation between general cognitive

ability at 4 years of age and reading and arithmetic four years later, at the end

of Grade 2. Finally, in Chapter 6 the key findings and theoretical implications of

the present thesis will be discussed.


Chapter 1

32

1.6 References

Ahissar, m. (2007). Dyslexia and the ancho-ring-deficit hypothesis. Trends in Cognitive Sciences, 11(11), 458-465. doi: 10.1016/j.tics.2007.08.015

Bekebrede, J., van der Leij, A., Plakas, A., Sha-re, D., & morfidi, E. (2010). Dutch dyslexia in adulthood: Core features and variety. Scientific Studies of Reading, 14(2), 183-210.

Bekebrede, J., van der Leij, A., Schrijf, G., & Share, D. (2010). Cognitive profiles of poor readers compared to typical readers in middle childhood, early and late ado-lescence, and adulthood.. In: The role of orthographic and phonological processing in dyslexia and reading (Doctoral disserta-tion). University of Amsterdam.

Bishop, D. V. m. (2009). Genes, cognition, and communication. Annals of the New York Academy of Sciences, 1156 (The Year in Cognitive Neuroscience 2009), 1-18. doi: DOI: 10.1111/j.1749-6632.2009.04419.x

Bishop, D. V. m., mcDonald, D., Bird, S., & hayiou-Thomas, m. E. (2009). Children who read words accurately despite lan-guage impairment: Who are they and how do they do it? Child Development, 80(2), 593-605. doi: 10.1111/j.1467-8624.2009.01281.x

Blomert, L. (2006). Protocol dyslexie diagnos-tiek en behandeling [Protocol for diagno-sing and treating dyslexia]

Boets, B., & De Smedt, B. (2010). Single-digit arithmetic in children with dyslexia. Dys-lexia, 16(2), 183-191. doi: 10.1002/dys.403

Boets, B., De Smedt, B., Cleuren, L., Van-dewalle, E., Wouters, J., & Ghesquière, P. (2010). Towards a further characterization of phonological and literacy problems in Dutch-speaking children with dyslexia. British Journal of Developmental Psycho-logy, 28, 5-31.

Bosman, A. m. T., & Braams, T. (2005). Depres-sie en angst bij basisschoolleerlingen met dyslexie [Depression and anxiety in pri-mary school children with dyslexia]. Tijd-schrift Voor Orthopedagogiek, 44, 213-223.

Bosse, m., Tainturier, m. J., & Valdois, S. (2007). Developmental dyslexia: The vi-sual attention span deficit hypothesis. Cognition, 104(2), 198-230. doi: 10.1016/j.cognition.2006.05.009

Brown Waesche, J. S., Schatschneider, C., ma-ner, J. K., Ahmed, Y., & Wagner, R. K. (2011). Examining agreement and longitudinal stability among traditional and RTI-based definitions of reading disability using the affected-status agreement statistic. Jour-nal of Learning Disabilities, 44(3), 296-307. doi: 10.1177/0022219410392048

Bruce, D. (1964). The analysis of word sounds by young children. British Journal of Edu-cational Psychology, 34(2), 158-170.

Byrne, B., Coventry, W. L., Olson, R. K., Sa-muelsson, S., Corley, R., Willcutt, E. G., . . . DeFries, J. C. (2009). Genetic and en-vironmental influences on aspects of li-teracy and language in early childhood: Continuity and change from preschool to Grade 2. Journal of Neurolinguistics, 22(3), 219-236.

Carroll, J. m., maughan, B., Goodman, R., & meltzer, h. (2005). Literacy difficulties and psychiatric disorders: Evidence for comor-bidity. Journal of Child Psychology and Psy-chiatry, 46(5), 524-532.

Cunningham, A. E., & Stanovich, K.,E. (1997). Early reading acquisition and its relation to reading experience and ability 10 ye-ars later. Developmental Psychology, 33(6), 934-945.

de Bree, E., Wijnen, F., & Gerrits, E. (2010). Non-word repetition and literacy in dutch children at-risk of dyslexia and children



33

1

with SLI: Results of the follow-up study. Dyslexia, 16(1), 36-44.

de Jong, P. F. (2011). What discrete and se-rial rapid automatized naming can reveal about reading. Scientific Studies of Rea-ding, 15(4), 314-337.

de Jong, P. F., & van der Leij, A. (1999). Speci-fic contributions of phonological abilities to early reading acquisition: Results from a dutch latent variable longitudinal study. Journal of Educational Psychology, 91(3), 450-476.

de Jong, P. F., & van der Leij, A. (2003). Deve-lopmental changes in the manifestation of a phonological deficit in dyslexic children learning to read a regular orthography. Journal of Educational Psychology, 95(1), 22-40. doi: 10.1037/0022-0663.95.1.22

De Smedt, B., Taylor, J., Archibald, L., & Ansari, D. (2010). how is phonological processing related to individual diffe-rences in children’s arithmetic skills? De-velopmental Science, 13(3), 508-520. doi: 10.1111/j.1467-7687.2009.00897.x

Elbro, C., Borstrøm, I., & Petersen, D. K. (1998). Predicting dyslexia from kindergarten: The importance of distinctness of pho-nological representations of lexical items. Reading Research Quarterly, 33(1), 36-60.

Fisher, S. E., & Francks, C. (2006). Genes, cog-nition and dyslexia: Learning to read the genome. Trends in Cognitive Sciences, 10(6), 250-257. doi: 10.1016/j.tics.2006.04.003

Fletcher, J., Coulter, W., Reschly, D., & Vaug-hn, S. (2004). Alternative approaches to the definition and identification of learn-ing disabilities: Some questions and ans-wers. Annals of Dyslexia, 54(2), 304-331. doi: 10.1007/s11881-004-0015-y

Goswami, U., Thomson, J., Richardson, U., Stainthorp, R., hughes, D., Rosen, S., & Scott, S. K. (2002). Amplitude envelope onsets and developmental dyslexia: A new hypothesis. Proceedings of the Natio-nal Academy of Sciences, 99(16), 10911.

haworth, C. m. A., Kovas, Y., harlaar, N., hayiou-Thomas, m. E., Petrill, S. A., Dale, P. S., & Plomin, R. (2009). Generalist genes and learning disabilities: A multivariate genetic analysis of low performance in reading, mathematics, language and ge-neral cognitive ability in a sample of 8000 12-year-old twins. Journal of Child Psycho-logy and Psychiatry, 50(10), 1318-1325. doi: 10.1111/j.1469-7610.2009.02114.x

hecht, S. A., Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (2001). The relations bet-ween phonological processing abilities and emerging individual differences in mathematical computation skills: A longi-tudinal study from second to fifth grades. Journal of Experimental Child Psychology, 79(2), 192-227.

hulme, C., Bowyer-Crane, C., Carroll, J. m., Duff, F. J., & Snowling, m. J. (in press). The causal role of phoneme awareness and letter-sound knowledge in learning to read combining intervention studies with mediation analyses. Psychological Science, doi: 10.1177/0956797611435921

hulme, C., Snowling, m., Caravolas, m., & Car-roll, J. (2005). Phonological skills are (pro-bably) one cause of success in learning to read: A comment on castles and coltheart. Scientific Studies of Reading, 9(4), 351-365.

Jordan, N. C., Kaplan, D., Ramineni, C., & Lo-cuniak, m. N. (2009). Early math matters: Kindergarten number competence and later mathematics outcomes. Develop-mental Psychology, 45(3), 850.

Kirk, J., & Reid, G. (2001). An examination of the relationship between dyslexia and offending in young people and the impli-cations for the training system. Dyslexia, 7(2), 77-84.

Kleijnen, R., Bosman, A, de Jong, P.F., henne-man, K., Pasman, J., Paternotte, A., Ruijs-senaars, A., . . . Wijnen, F. (2008). Diagnose en behandeling van dyslexie. Brochure van stichting dyslexie Nederland (SDN). [Diag-nosis and treatment of dyslexia. Booklet of


Chapter 1

34

the Dutch foundation of dyslexia]. Biltho-ven: SDN.

Kovas, Y., haworth, C. m. A., Dale, P. S., & Plomin, R. (2007). The genetic and en-vironmental origins of learning abilities and disabilities in the early school years. Monographs of the Society for Research in Child Development, 72(3), 1-160. doi: 10.1111/j.1540-5834.2007.00453.x

Kovas, Y., & Plomin, R. (2007). Learning abili-ties and disabilities: Generalist genes, spe-cialist environments. Current Directions in Psychological Science, 16(5), 284-288.

Landerl, K., & moll, K. (2010). Comorbidity of learning disorders: Prevalence and fami-lial transmission. Journal of Child Psycho-logy and Psychiatry, 51(3), 287-294.

Leppänen, P. h. T., hämäläinen, J. A., Salmi-nen, h. K., Eklund, K. m., Guttorm, T. K., Lohvansuu, K., . . . Lyytinen, h. (2010). Ne-wborn brain event-related potentials re-vealing atypical processing of sound fre-quency and the subsequent association with later literacy skills in children with fa-milial dyslexia. Cortex, 46(10), 1362-1376.

Lervåg, A., Bråten, I., & hulme, C. (2009). The cognitive and linguistic foundations of early reading development: A Norwegian latent variable longitudinal study. Deve-lopmental Psychology, 45(3), 764–781.

macdonald, S. J. (2012). Biographical path-ways into criminality: Understanding the relationship between dyslexia and educa-tional disengagement. Disability & Society, 27(3), 427-440.

mcGrath, L. m., Pennington, B. F., Shanahan, m. A., Santerre Lemmon, L. E., Barnard, h. D., Willcutt, E. G., . . . Olson, R. K. (2011). A multiple deficit model of reading disabi-lity and attention deficit/hyperactivity disorder: Searching for shared cognitive deficits. Journal of Child Psychology and Psychiatry, 52(5), 547-557.

Nicolson, R. I., Fawcett, A. J., & Dean, P. (2001). Developmental dyslexia: The cerebellar

deficit hypothesis. Trends in Neurosciences, 24(9), 508-511.

Olson, R. K., Byrne, B., & Samuelsson, S. (2009). Reconciling strong genetic and strong environmental influences on indi-vidual differences and deficits in reading ability. In K. Pugh, & P. mcCardle (Eds.), How children learn to read. New York: Psy-chology Press.

Pennington, B. F. (2006). From single to mul-tiple deficit models of developmental disorders. Cognition, 101(2), 385-413.

Pennington, B. F., & Lefly, D. L. (2001). Early reading development in children at family risk for dyslexia. Child Development, 72(3), 816-833.

Pennington, B. F., Santerre–Lemmon, L., Rosenberg, J., macDonald, B., Boada, R., Friend, A., . . . Willcutt, E. G. (2012). Indi-vidual prediction of dyslexia by single versus multiple deficit models. Journal of Abnormal Psychology, 121(1), 212.

Peterson, R. L., Pennington, B. F., Shriberg, L. D., & Boada, R. (2009). What influences literacy outcome in children with speech sound disorder? Journal of Speech, Lan-guage, and Hearing Research, 52(5), 1175.

Plakas, A., van Zuijen, T., van Leeuwen, T., Thomson, J. m., & van der Leij, A. (2012). Impaired non-speech auditory processing at a pre-reading age is a risk-factor for dyslexia but not a predictor: An ERP study. Cortex,

Plomin, R., DeFries, J. C., mcClearn, G. E., & mcGuffin, P. (2008). Behavioral genetics (5th ed.). New York: Worth Publishers.

Plomin, R., & Kovas, Y. (2005). Generalist ge-nes and learning disabilities. Psychological Bulletin, 131(4), 592-617.

Powell, B. B. (2009). Writing: Theory and his-tory of the technology of civilization. Ox-ford: Blackwell.

Regtvoort, A. G. F. m., & van der Leij, A. (2007). Early intervention with children of



35

1

dyslexic parents: Effects of computer-ba-sed reading instruction at home on liter-acy acquisition. Learning and Individual Differences, 17(1), 35-53.

Regtvoort, A. G. F. m., van Leeuwen, T. h., Stoel, R. D., & van der Leij, A. (2006). Effi-ciency of visual information processing in children at-risk for dyslexia: habituation of single-trial ERPs. Brain and Language, 98(3), 319-331.

Scarborough, h. S. (1990). Very early langua-ge deficits in dyslexic children. Child Deve-lopment, 61(6), 1728-1743.

Sénéchal, m., & Young, L. (2008). The effect of family literacy interventions on children’s acquisition of reading from kindergarten to Grade 3: A meta-analytic review. Review of Educational Research, 78(4), 880-907. doi: 10.3102/0034654308320319

Snowling, m. J. (1995). Phonological proces-sing and developmental dyslexia. Journal of Research in Reading, 18(2), 132-138.

Snowling, m. J. (2008). Specific disorders and broader phenotypes: The case of dyslexia. The Quarterly Journal of Experimental Psy-chology, 61(1), 142-156.

Snowling, m. J., Bishop, D. V. m., & Stothard, S. E. (2000). Is preschool language impair-ment a risk factor for dyslexia in adoles-cence? Journal of Child Psychology and Psy-chiatry, 41(5), 587-600. doi: 10.1111/1469-7610.00651

Snowling, m. J., Gallagher, A. m., & Frith, U. (2003). Family risk of dyslexia is continuo-us: Individual differences in the precurs-ors of reading skill. Child Development, 74(2), 358-373.

Snowling, m. J., & hulme, C. (2012). Annual research review: The nature and classifica-tion of reading disorders – A commentary on proposals for DSm-5. Journal of Child Psychology and Psychiatry, 53(5), 593-607. doi: 10.1111/j.1469-7610.2011.02495.x

Snowling, m. J., muter, V., & Carroll, J. (2007). Children at family risk of dys-

lexia: A follow-up in early adolescence. Journal of Child Psychology and Psychia-try, 48(6), 609-618. doi: 10.1111/j.1469-7610.2006.01725.x

Stein, J., & Walsh, V. (1997). To see but not to read; the magnocellular theory of dys-lexia. Trends in Neurosciences, 20(4), 147-152.

Stromswold, K. (2001). The heritability of language: A review and metaanalysis of twin, adoption, and linkage studies. Lan-guage, , 647-723.

Tallal, P. (1980). Auditory temporal percep-tion, phonics, and reading disabilities in children. Brain and Language, 9(2), 182-198.

Taylor, J., Roehrig, A. D., hensler, B. S., Connor, C. m., & Schatschneider, C. (2010). Teacher quality moderates the genetic effects on early reading. Science, 328(5977), 512-514. doi: 10.1126/science.1186149

Torppa, m., Lyytinen, P., Erskine, J., Eklund, K., & Lyytinen, h. (2010). Language de-velopment, literacy skills, and predictive connections to reading in finnish children with and without familial risk for dyslexia. Journal of Learning Disabilities, 43(4), 308-321. doi: 10.1177/0022219410369096

Torppa, m., Poikkeus, A., Laakso, m., Tol-vanen, A., Leskinen, E., Leppänen, P. h. T., . . . Lyytinen, h. (2007). modeling the early paths of phonological awareness and factors supporting its development in children with and without familial risk of dyslexia. Scientific Studies of Reading, 11(2), 73-103.

Vaessen, A., & Blomert, L. (2010). Long-term cognitive dynamics of fluent reading de-velopment. Journal of Experimental Child Psychology, 105(3), 213-231.

Valdois, S., Bidet-Ildei, C., Lassus-Sangosse, D., Reilhac, C., N’guyen-morel, m. A., Gui-net, E., & Orliaguet, J. P. (2011). A visual processing but no phonological disorder in a child with mixed dyslexia. Cortex,


Chapter 1

36

van Bergen, E., de Jong, P. F., maassen, B., & van der Leij, A. (submitted). The effect of parents’ literacy skills and children’s preli-teracy skills on the risk of dyslexia.

van Bergen, E., de Jong, P. F., maassen, B., Krikhaar, E., Plakas, A., & van der Leij, A. (in revision). IQ of four-year-olds who go on to develop dyslexia.

van Bergen, E., de Jong, P. F., Plakas, A., maas-sen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of Child Psychology and Psychiatry, 53(1), 28-36. doi: 10.1111/j.1469-7610.2011.02418.x

van Bergen, E., de Jong, P. F., Regtvoort, A., Oort, F., van Otterloo, S., & van der Leij, A. (2011). Dutch children at family risk of dyslexia: Precursors, reading develop-ment, and parental effects. Dyslexia, 17(1), 2-18. doi: 10.1002/dys.423

van der Leij, A., & van Daal, V. h. P. (1999). Automatization aspects of dyslexia speed limitations in word identification, sen-sitivity to increasing task demands, and orthographic compensation. Journal of Learning Disabilities, 32(5), 417-428.

van der Sluis, S., de Jong, P. F., & Leij, A. v. d. (2004). Inhibition and shifting in children with learning deficits in arithmetic and reading. Journal of Experimental Child Psychology, 87(3), 239-266. doi: 10.1016/j.jecp.2003.12.002

van Otterloo, S. G., & van der Leij, A. (2009). Dutch home-based pre-reading interven-tion with children at familial risk of dys-lexia. Annals of Dyslexia, 59(2), 169-195.

Vidyasagar, T. R., & Pammer, K. (2010). Dys-lexia: A deficit in visuo-spatial attention, not in phonological processing. Trends in Cognitive Sciences, 14(2), 57-63.

Wagner, R. K. (1986). Phonological proces-sing abilities and reading. Journal of Learn-ing Disabilities, 19(10), 623-629.

Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological processing and its causal role in the acquisition of reading skills. Psychological Bulletin, 101(2), 192-212.

Warmington, m., & hulme, C. (2012). Pho-neme awareness, visual-verbal paired-as-sociate learning, and rapid automati-zed naming as predictors of individual differences in reading ability. Scientific Studies of Reading, 16(1), 45-62. doi: 10.1080/10888438.2010.534832

Wechsler, D. (2004). The wechsler intelligence scale for children - Fourth edition. London: Pearson Assessment.

Willcutt, E. G., Betjemann, R. S., Pennington, B. F., Olson, R. K., DeFries, J. C., & Wads-worth, S. J. (2007). Longitudinal study of reading disability and attention deficit/hyperactivity disorder: Implications for education. Mind, Brain, and Education, 1(4), 181-192.

Willcutt, E. G., Pennington, B. F., & DeFries, J. C. (2000). Twin study of the etiology of comorbidity between reading disability and attention deficit/hyperactivity disor-der. American Journal of Medical Genetics, 96(3), 293-301.

Willcutt, E. G., Pennington, B. F., Olson, R. K., Chhabildas, N., & hulslander, J. (2005). Neuropsychological analyses of comor-bidity between reading disability and attention deficit hyperactivity disorder: In search of the common deficit. Develop-mental Neuropsychology, 27(1), 35-78. doi: 10.1207/s15326942dn2701_3

Wolf, m., & Bowers, P. G. (1999). The dou-ble-deficit hypothesis for the develop-mental dyslexias. Journal of Educational Psychology, 91(3), 415-438.

Yap, R., & van der Leij, A. (1993). Word pro-cessing in dyslexics. Reading and Writing, 5(3), 261-279.


The great tragedy of science – the slaying of a beautiful hypothesis by an ugly

fact.

- Thomas Huxley -

Dutch children at family risk of dyslexia: Precursors, reading development, and

parental effects

van Bergen, E., de Jong, P.F., Regtvoort, A., Oort, F., van Otterloo, S., & van der Leij, A. (2011). Dutch children at family risk of dyslexia: Precursors, reading development, and

parental effects. Dyslexia, 17(1), 2-18.

Chapter


Dutch children at family risk of dyslexia: Precursors, reading development, and

parental effects

van Bergen, E., de Jong, P.F., Regtvoort, A., Oort, F., van Otterloo, S., & van der Leij, A. (2011). Dutch children at family risk of dyslexia: Precursors, reading development, and

parental effects. Dyslexia, 17(1), 2-18.

Chapter


Chapter 2

40

Abstract

The study concerns reading development and its precursors in a

transparent orthography. Dutch children differing in family risk for

dyslexia were followed from kindergarten through 5th grade. In 5th grade,

at-risk dyslexic (n = 22), at-risk non-dyslexic (n = 45), and control children

(n = 12) were distinguished. In kindergarten, the at-risk non-dyslexics

performed better than the at-risk dyslexics, but worse than the controls

on letter knowledge and rapid naming. The groups did not differ on

phonological awareness. At-risk dyslexics read less fluently from 1st

grade onwards than the other groups. At-risk non-dyslexics’ reading

fluency was at an intermediate position between the other groups at

the start of reading. By 5th grade they had reached a similar level as

the controls on word reading, but still lagged behind on pseudoword

reading. Results further showed that the parents of the groups of at-

risk children differed in educational level and reading skills. Overall,

the groups of at-risk children differed on pre-reading skills as well as

on reading development. These differences do not seem to stem from

differences in intellectual abilities or literacy environment. Instead, the

better reading skills of parents of at-risk non-dyslexics suggest that

these children might have a lower genetic liability.


Dutch chilDren at family risk of Dyslexia

41

2

2.1 Introduction

It is well established that reading and spelling problems of dyslexics are

accompanied by deficits in phonological processing. Dyslexics’ impairments

have been found in phonological awareness, verbal short-term memory,

and rapid naming (see for reviews Elbro, 1996; Vellutino, Fletcher, Snowling,

& Scanlon, 2004). most of the evidence is based on studies in which the

phonological deficits were examined after reading problems had become

manifest. There is, however, a small number of studies that started before

reading onset and continued to follow children until an age that reading

problems could be established. Such longitudinal studies can provide stronger

evidence for a causal relationship between deficits in phonological processing

and the development of dyslexia.

most prospective studies involved children who have a dyslexic parent

and, hence, have a family risk of becoming dyslexic. Such studies provide the

opportunity to examine the differential development of at-risk children who

do and do not become dyslexic (hereafter called dyslexics and at-risk non-

dyslexics, respectively), and controls (i.e., not at-risk non-dyslexics). The majority

of studies concerned children learning to read in an opaque orthography. The

first issue of importance in the present study is that it involved children learning

to read in a relatively transparent orthography: Dutch. Several studies suggest

that the differential importance of precursors of later reading success or failure

is moderated by orthographies’ transparencies (de Jong & van der Leij, 2003;

Puolakanaho et al., 2008; Wimmer, mayringer, & Landerl, 2000). The second

issue that is addressed in this study is the impact of parental reading status

on their offspring’s reading success or failure. Although differences between

the three groups on child characteristics are well documented, differences on

parental characteristics have largely been neglected.

Several earlier prospective studies showed that phonological deficits

were already present before reading started (Boets, Wouters, van Wieringen, &

Ghesquière, 2007; Elbro, Borstrøm, & Petersen, 1998; Pennington & Lefly, 2001;

Scarborough, 1990; Snowling, Gallagher, & Frith, 2003). For example, Elbro,

Borstrøm, and Petersen (1998) followed the progress of Danish children with

and without family risk from kindergarten through the beginning of second

grade, when dyslexia was determined. In kindergarten, the at-risk dyslexics

showed deficits as compared to the control group in phoneme awareness,


Chapter 2

42

verbal short-term memory, and distinctness of phonological representations,

whereas the at-risk non-dyslexics did not. In contrast, the at-risk non-dyslexics

did perform as poorly as the at-risk dyslexics on morpheme awareness and

articulation, whereas they performed better than the at-risk dyslexics but less

well than the controls on letter knowledge.

Snowling and colleagues carried out a similar study with children

learning to read in English (Gallagher, Frith, & Snowling, 2000; Snowling,

Gallagher, & Frith, 2003; Snowling, muter, & Carroll, 2007) . Children’s reading

status was determined at the age of 8. At 3 years and 9 months, the at-risk

non-dyslexics performed better than the at-risk dyslexics but poorer than the

controls on non-word repetition and letter knowledge. The dyslexic, but not

the non-dyslexic at-risk readers, exhibited early deficits in vocabulary, narrative

skills, and verbal short-term memory. Snowling et al. (2003; 2007) concluded

that the children with a family risk that did not become dyslexic had probably

compensated for their deficits in grapheme-phoneme knowledge with strong

oral language skills.

Pennington and Lefly (2001) also conducted an at-risk study with

English-speaking children. Dyslexia was assessed in second grade. At the start

of first grade, the at-risk dyslexics had, compared to both groups of future

normal readers, impairments on emergent literacy skills and verbal short-term

memory. The at-risk dyslexics also performed poorer on phoneme awareness

and speech perception than at-risk non-dyslexics, who performed poorer than

controls. Interestingly, Pennington and Lefly also assessed rapid naming skills,

on which the at-risk dyslexics also scored weakest and the controls highest.

Generally, learning to read in an opaque orthography like Danish or

English poses extra challenges to the beginning reader and might affect the

various manifestations of dyslexia (Seymour, Aro, & Erskine, 2003; Ziegler et al.,

2005). In a study of dyslexics (with and without risk combined), at-risk non-

dyslexics, and controls in a relatively transparent orthography (Dutch), Boets

et al. (2007; 2010) found that the dyslexic readers performed as kindergartners

worse than the not at-risk non-dyslexic readers on letter knowledge,

verbal short-term memory, phonological awareness and rapid naming. The

performance of the at-risk non-dyslexic group was at an intermediate position

between that of the other two groups, but did not differ significantly from

either of them. These results are roughly in accordance with studies in opaque

orthographies, but the results on phonological awareness are at odds with



43

2

findings that kindergarten phonological awareness predicts reading skills less

well in transparent orthographies (de Jong & van der Leij, 2003; Puolakanaho

et al., 2008; Wimmer, mayringer, & Landerl, 2000).

Regardless of the subtle differences between orthographies in the

cognitive profile of dyslexic readers, the question remains in what respect

at-risk children that do not develop dyslexia differ from those that do, and

especially, where these differences stem from. One possibility is that at-risk

non-dyslexic children are able to compensate their impaired reading potential

with other skills, as suggested by Snowling et al. (2003). Another possibility

is that this group is at lower genetic risk. Finally, it could be that this group

experiences more advantageous environmental factors.

Torppa et al. (2007) assessed the relation between home literacy,

reading interest, phonological awareness, vocabulary, and emergent literacy

variables. They found a direct effect of home-literacy environment (shared

reading) on vocabulary growth, but not on emergent literacy. Similarly,

Elbro et al. (1998) did not find differences between dyslexic and non-dyslexic

parents, nor between parents of dyslexic and non-dyslexic children in the

amount of shared reading in kindergarten. Finally, Snowling et al. (2007) did

not find differences between the two at-risk groups in parent reading behavior,

family literacy behavior, and socioeconomic background. Overall, there is no

strong evidence that children who develop dyslexia grow up in a relatively

disadvantageous literacy environment.

Dyslexia is commonly seen as a complex multifactorial disorder

with numerous genes involved that interact with one another and with the

environment (Bishop, 2009). Each gene affects the probability of the disorder,

but is in itself not necessary and sufficient for causing it. The involvement of

multiple genes may function as a normally distributed genetic liability (Rutter,

2006; van den Oord, Pickles, & Waldman, 2003): as more genes are affected the

probability increases that dyslexia becomes manifest. A multifactorial etiological

conceptualization resulting from a normally distributed genetic liability is in

accordance with the view from behavioral studies (Snowling, Gallagher, & Frith,

2003) that the family risk of dyslexia is continuous. This model may be useful

to speculate about the question why some at-risk children develop dyslexia

whereas others do not.

Surprisingly, possible differences in reading ability between the

dyslexic parents of the two at-risk groups have been largely neglected. Such


Chapter 2

44

differences might suggest that at-risk dyslexic children have a higher genetic

liability than at-risk non-dyslexics. As an exception, Snowling et al. (2007)

reported that the parents of at-risk dyslexics did not differ from the parents of

at-risk non-dyslexics on spelling and word-reading accuracy. however, it should

be noted that in these analyses the pairs of parents were divided according to

gender instead of according to reading status. Furthermore, within the at-risk

group the literacy levels of the children and their parents were not associated.

It is of interest whether similar results will be found for word-reading fluency

or rate. Word-reading fluency is seen as the most important feature of reading

ability in transparent orthographies (Wimmer, 2006).

In conclusion, studies following children at-risk of dyslexia indicate

that pre-readers who become dyslexic show early phonological deficits

and broader language difficulties. In addition, at-risk non-dyslexics are not

completely unaffected but show a milder phenotype. What is missing, though,

are long-term studies in orthographically transparent languages. In addition,

attention should be devoted to the reading skills of the parents.

In our study, the literacy development of Dutch children with and

without family risk for dyslexia was followed from kindergarten through fifth

grade. In fifth grade groups of at-risk dyslexics, at-risk non-dyslexics, and controls

were distinguished. Next we considered, retrospectively, differences in the

development of (pre-)reading skills among these groups. In addition, differences

in parental characteristics and home-literacy environment were examined.

2.2 Method

2.2.1 ParticipantsTwenty-two children were included in the at-risk dyslexic group and 45 in

the at-risk non-dyslexic group. The control group included 12 not at-risk non-

dyslexic children (see Table 2.1).

The children were recruited earlier (Regtvoort & van der Leij, 2007; van

Otterloo & van der Leij, 2009) . The original sample consisted of 98 children

(82 at-risk). None of the parents reported (suspicion of ) developmental

disorders. In addition, the children were recruited from regular education. In

the Netherlands children with developmental disorders are usually referred

to special education. Fifty-nine at-risk children received intervention before



45

2

they received formal reading instruction. The intervention group showed a

beneficial effect on phonological awareness and letter knowledge at the end

of kindergarten, but this head start did not lead to better literacy levels in first

grade1. The only measurement occasion with an intervention effect present

(i.e., end of kindergarten) was not included in the present study.

Table 2.1 Characteristics of the Children from the Three Groups

Characteristic At-risk dyslexic

At-risk non-dyslexic

Control F (2, 76) p η2p

Sample size 22 45 12No (%) of boys 16

a (73%) 28

a (62%) 6

a (50%)

No (%) of children in intervention

14a (64%) 32

a (71%) 0 (0%)

Child characteristicsAge in months (KG) 70.43

a (4.56) 70.61

a (3.80) 69.95

a (4.12) 0.126 .882 .003

Vocabulary (KG) 72.00a (9.90) 72.22

a (8.95) 73.00

a (7.89) 0.049 .952 .001

Nonverbal IQ (Gr1) 25.05a (5.19) 26.00

a (4.30) 27.00

a (4.97) 0.714 .493 .018

Note. Characteristics are given in raw scores, standard deviations in parentheses. Numbers and means in the same row that do not share subscripts differ at p < .05 on the χ2-test and the Tukey honestly significant difference comparison. No = number; KG = kindergarten; Gr = Grade.

Children were considered at-risk if at least one of the parents reported

to be dyslexic and if this was confirmed by tests measuring word and

pseudoword-reading fluency (Brus & Voeten, 1972; van den Bos, lutje Spelberg,

Scheepstra, & de Vries, 1994) . Also, the subtest similarities of the Wechsler Adult

Intelligence Scale was administered to measure verbal reasoning (Wechsler,

1997). For inclusion of a child in the at-risk group, at least one parent had to

score below or at the 20th percentile on both reading tests, or below or at the

10th percentile on one of them (with the other ≤ 40th percentile). Children of

parents who showed a large discrepancy (≥ 60 percentiles) between verbal

reasoning and one of the reading-tests were also included, with the restriction

that neither reading-test percentile score exceeded 402. Of not at-risk children

both parents had to score ≥ 40th percentile on both reading tests. 1 T-tests (comparing at-risk children with and without kindergarten intervention) on

January Grade 1 measures showed no significant differences: for letter-naming fluency, mean difference = 0.008, t(65) = 0.13, p = .896; for word-reading fluency, mean difference = -1.56, t(65) = -0.67, p = .504.

2 There were 14 at-risk children for whose parents only the discrepancy criterion applied. Excluding this group of children did not alter the pattern of results.


Chapter 2

46

One of the 16 not at-risk and 15 of the 82 at-risk children dropped out

of the study. Within the at-risk group, children who dropped out had a smaller

vocabulary than the others, t (80) = 2.21, p = .030, but they did not differ in

age, t (80) = 1.59, p = .146. Their parents did not differ on word-reading fluency,

t (80) = -0.71, p = .481, pseudoword- reading fluency, t (80) = -0.95, p = .346,

and verbal reasoning, t (80) = -1.05, p = .298. The attrition rate in the not at-

risk group was too small to do group comparisons. One not at-risk child was

excluded from the study because an elder sibling was diagnosed as dyslexic.

The child itself also became dyslexic.

Reading status was determined in fifth grade (at a mean age of 10 years

and 10 months). Children with a reading score on a word-reading fluency task

(WRF2, see below) corresponding to the weakest ten percent in the population

were considered dyslexic. The not at-risk dyslexic group consisted of only two

children and was excluded in the study. The remaining three groups did not differ

on sex, age, vocabulary, and nonverbal IQ (Table 2.1). The proportion of children

in the at-risk groups who had received intervention in kindergarten did not differ.

2.2.2 Measuresmeasures were selected to investigate intelligence, phonological awareness,

letter knowledge, rapid naming, and (pseudo)word-reading skills. Test

reliabilities ranged from .73 to .97.

Intelligence

Vocabulary. In the receptive vocabulary test (Verhoeven & Vermeer,

1996) the child had to choose among four alternatives the picture that best

matched a given word. The 98 items were of increasing difficulty. Administration

was stopped when the child failed six out of the last eight items.

Nonverbal IQ. Coloured Progressive matrices (J. C. Raven, Court, &

Raven, 1984) were used to measure nonverbal IQ.

Home-literacy environment

Parents were given a short questionnaire when their child attended

kindergarten. They were asked to indicate whether their child was a library

member, to estimate their shared reading frequency on a scale from 1 (never)

to 5 (more than five times a week) and the number of books they had at home

scaled from 1 (less than 50) to 5 (over 200). Next, the level of education of

both parents was ranked on a scale ranging from 1 (primary school only) to 7

(university degree).



47

2

Phonological Awareness

Alliteration (Irausquin, 2001). The word with a different onset sound

from three words had to be named (‘odd one out’). All words had a consonant-

vowel-consonant (CVC) structure. The experimenter named the three words

and repeated them when necessary. There were 3 practice items (with feedback)

and 10 test items, all of approximately equal difficulty.

Phoneme blending and segmentation. In phoneme-blending

(Verhoeven, 1993a) the child was required to blend aurally presented phonemes

into a word. For example, children listened to the successive phonemes /r/ /u/

/p/ /s/, after which they had to merge this into /rups/ (caterpillar). Phoneme-

segmentation (Verhoeven, 1993b) was the reverse of phoneme-blending. Now

the child had to segment a given word into its constituent phonemes. Both

tasks began with three practice trials (with feedback). Test items consisted of 20

monosyllabic words per test, increasing from two to five phonemes, with four

to six items per number of phonemes. The tests were stopped when all items

with the same number of phonemes were blended or segmented incorrectly.

Rapid Naming

Serial rapid naming (van den Bos, 2003) was measured for colors (black, yellow,

red, green, and blue) and objects (bike, tree, duck, scissors, and chair). Each of

the tasks consisted of 50 randomly ordered symbols arranged in five columns

of 10 symbols each. Before test administration, the child practiced by naming

the last column of symbols. Children were instructed to name the symbols

column-wise as quickly as possible. For both tasks, the time to completion

was transformed to number of symbols per second to normalize the score

distribution.

Letter Knowledge

Receptive letter-knowledge. The test (Verhoeven, 2002) required the child

to point from six alternative lowercase letters to the letter that matched a given

sound. For instance, the child was asked “Where do you see the /m/ of mooi

(beautiful)?” The knowledge of 32 graphemes (including digraphs) was tested.

Letter-naming fluency. Children were asked to provide the sound of 34

graphemes (including digraphs) correctly as quickly as possible (Verhoeven,

1993a). When the child gave letter names instead, the child was corrected

and the test was started once again. The randomly ordered graphemes were

printed in lowercase in two columns of 17 items each. Scores were transformed

to number of correctly-read letters per second.


Chapter 2

48

Reading Skills

Word-reading fluency (WRF). This was measured as how many

disconnected words one can read correctly within a certain time allowed. Three

lists were used (Verhoeven, 1995; Verhoeven & van Leeuwe, 2009). The first

and second list (WRF1 and WRF2, respectively) consisted of 150 monosyllabic

words each. WRF1 included words with a CV, VC, or CVC pattern. All words

in WRF2 contained at least one consonant cluster. WRF3 comprised 120

polysyllabic words with various orthographic complexities. The words had to

be read correctly as quickly as possible. The score per list was the number of

words read correctly within one minute. WRF2 was used to assess dyslexia in

Grade 5, because it differentiates best between normal and dyslexic readers.

Pseudoword-reading fluency. The test (van den Bos, lutje Spelberg,

Scheepstra, & de Vries, 1994) consisted of a list of 116 pseudowords of increasing

difficulty. The child was asked to correctly read aloud as many pseudowords as

possible within two minutes.

2.2.3 ProcedureTests were administered individually in a fixed order by trained graduate

students in a separate room at school, in January of the second kindergarten

year, in January and June of Grade 1, June of Grade 2, and January of Grade 5.

2.2.4 Analytic ApproachOne-way analyses of variance (ANOVAs) and χ2-tests were used to compare

groups. Contrasts were evaluated post-hoc with Tukey’s procedure to correct

for multiple testing. When the same ability was measured on several occasions,

repeated-measures ANOVAs were performed with Group as between subject-

factor and Time as within-subject factor.

2.3 Results

2.3.1 Parent Characteristics and Home-Literacy EnvironmentThe characteristics of the children’s parents and home environment are

presented in Table 2.2. In both at-risk groups, almost 75% of the children had

a dyslexic father. Regarding the level of education for mothers and fathers,

control families had the highest level whereas the lowest level was found in



49

2

Tab

le 2

.2 C

hara

cter

istic

s of

the

Par

ents

and

hom

e En

viro

nmen

t fr

om t

he T

hree

Gro

ups,

with

χ2 a

nd A

NO

VA

Resu

lts

Cha

ract

eris

ticA

t-ris

k dy

slex

icA

t-ris

k no

n-dy

slex

icC

ontr

olF

dfp

η2 p

Pare

nt c

hara

cter

istic

s

No

(%) o

f dys

lexi

c fa

ther

s16

a (73%

)32

a (71%

)

Leve

l of e

duca

tion

of

mot

her

4.32

a (1.4

3)5.

07ab

(1.5

7)5.

92b (1

.31)

4.58

(2, 7

6).0

13.1

08

Fath

er3.

86a (1

.96)

4.91

ab (1

.85)

6.17

b (1

.53)

6.27

(2, 7

6).0

03.1

42

Dys

lexi

c p

aren

t4.

00a (1

.98)

4.73

a (1.8

3)2.

24(1

, 65)

.139

.033

Non

-dys

lexi

c p

aren

t4.

18a (1

.44)

5.24

b (1

.55)

7.25

(1, 6

5).0

09.1

00

Verb

al re

ason

ing

17.5

0 a (5.2

0)16

.22 a (4

.78)

18.4

2 a (4.1

4)1.

13(2

, 76)

.329

.029

Wor

d-re

adin

g flu

ency

60.4

1 a (11.

96)

69.0

0 b (1

3.61

)97

.83 c (1

1.08

)34

.33

(2, 7

6)<

.001

.475

Pseu

dow

ord-

read

ing

fluen

cy38

.77 a (1

5.32

)53

.22 b

(17.

85)

101.

75c (8

.45)

61.5

5(2

, 76)

< .0

01.6

18

hom

e-lit

erac

y en

viro

nmen

t

No

(%) o

f lib

rary

mem

ber

s16

a (73%

)32

a (71%

)6 a (5

0%)

Shar

ed re

adin

g4.

05a (1

.29)

4.39

a (0.9

1)4.

83a (0

.39)

2.55

(2, 7

6).0

85.0

63

Book

s at

hom

e3.

05a (1

.53)

3.31

a (1.4

9)4.

00a (1

.65)

1.54

(2, 7

6).2

20.0

39

Not

e. P

aren

tal r

eadi

ng fl

uenc

y an

d re

ason

ing

are

scor

es o

f the

wea

kest

-rea

ding

par

ent.

Num

ber

s an

d m

eans

in th

e sa

me

row

th

at d

o no

t sh

are

sub

scrip

ts d

iffer

at

p <

.05

on t

he χ

2 -tes

t an

d th

e Tu

key

hone

stly

sig

nific

ant

diff

eren

ce c

omp

aris

on. N

o =

nu

mb

er.


Chapter 2

50

Table 2.3 Descriptive Statistics and ANOVA Results for the measures of Phonological Awareness, Rapid Naming, and Letter Knowledge

At-risk dyslexic

At-risk non-

dyslexic Control

measure M SD M SD M SD F (2, 76) p η2p

Phonological awareness (Jan. KG)

Alliteration 4.15a 2.99 4.53

a 3.08 4.83

a 3.10 0.21 .808 .006

Phoneme blending 6.18a 6.63 7.27

a 6.15 7.67

a 6.51 0.29 .748 .008

Phoneme segmentation 4.86a 4.84 5.22

a 5.35 4.58

a 5.55 0.09 .918 .002

Rapid naming (Jan. KG)

Colors 0.48a 0.15 0.63

b 0.18 0.74

b 0.21 9.85 < .001 .206

Objects 0.53a 0.14 0.62

ab 0.17 0.74

b 0.17 6.72 .002 .150

Letter knowledge

Receptive (Jan. KG) 10.36a 4.86 14.42

b 6.70 18.00

b 5.94 6.51 .002 .146

Naming fluency (Jan. Gr.1) 0.82a 0.21 0.96

b 0.24 1.18

c 0.23 9.52 < .001 .200

Naming fluency (June Gr.1) 1.01a 0.26 1.24

b 0.28 1.24

b 0.29 5.72 .005 .131

Note. maximum scores are 10, 20, and 20 for the phonological-awareness tasks, and 32 for the receptive letter-knowledge task. Jan. = January. Numbers and means in the same row that do not share subscripts differ at p < .05 on the Tukey honestly significant difference comparison. KG = kindergarten; Gr = Grade.

the families of the at-risk dyslexics. Within at-risk families, we also investigated

level of education for dyslexic and non-dyslexic parents. An ANOVA (with

Group as between-subject and Parent as within-subject factor) revealed that

the parents of the at-risk non-dyslexics had a higher educational level than

those of the at-risk dyslexics (main effect of Group: F (1, 65) = 5.26, p = .025, η2p

= .075). The main effect of Parent (F (1, 65) = 2.73, p = .103, η2p = .040) and the

interaction effect (F < 1) were not significant. In contrast to level of education,

no differences were found for parental verbal reasoning. Interestingly, reading

fluency of the parents of at-risk dyslexics was significantly lower than of at-

risk non-dyslexics for both words and pseudowords. This finding was further

endorsed by a correlation of .27 (p = .026) between reading fluency scores (a

composite of words and pseudowords) of the dyslexic parents and their offspring.

With respect to home-literacy environment in kindergarten (Table

2.2), the proportion of children who were member of the library did not differ

significantly among the groups, nor did the groups differ significantly in the

amount of shared reading. Finally, the Group effect was not significant for the

number of books at home.



51

2

2.3.2 Child CharacteristicsPre-Reading Skills

The scores on the measures of phonological awareness, rapid naming, and

receptive letter knowledge (Table 2.3) were subjected to a multivariate analysis

of variance (mANOVA). The three groups differed significantly: Wilks’ λ = .668,

F (12, 142) = 2.65, p = .003, η2p = .183. The univariate results revealed that the

groups did not differ on phonological-awareness tasks. Conversely, differences

were found on rapid naming of colors and objects, and on letter knowledge.

The at-risk dyslexic children were slower on rapid naming and knew fewer

letters than the controls. The at-risk non-dyslexics also seemed to perform more

poorly than controls, but this did not reach significance at the (conservative)

Tukey post-hoc comparisons.

Reading Development

Letter-naming fluency (Table 2.3) was measured at the middle and end of

first grade. There was a main effect of both Time, F (1, 76) = 21.23, p < .001,

Table 2.4 Descriptive Statistics and ANOVA Results for the Reading Fluency measures

measure

At-risk dyslexic At-risk non-dyslexic Control

F (2, 76) η2p

M SD M SD M SD

Grade 1, January

Words-WRF1 11.14a 4.52 20.76

b 8.62 37.75

c 15.02 33.81 .471

Grade 1, June

Words-WRF1 22.68a 11.12 41.42

b 15.48 58.33

c 15.20 25.70 .403

Words-WRF2 10.68a 7.10 26.78

b 13.17 43.92

c 17.18 28.49 .428

Words-WRF3 6.77a 7.10 16.58

b 9.44 29.67

c 11.91 23.99 .387

Pseudowords 11.32a 8.91 20.82

b 9.71 36.58

c 11.91 25.58 .402

Grade 2, June

Words-WRF1 42.09a 17.10 66.29

b 15.30 77.58

b 11.81 26.15 .409

Words-WRF2 30.18a 16.57 56.53

b 18.48 72.50

c 12.52 27.72 .422

Words-WRF3 19.05b 13.34 41.33

b 16.40 52.17

b 12.81 23.38 .381


b 12.86 47.50

c 8.93 24.24 .389

Grade 5, January

Words-WRF1 63.45a 15.78 93.04

b 12.44 95.67

b 13.21 39.45 .509

Words-WRF2 55.00a 16.20 89.76

b 11.48 93.25

b 12.68 58.44 .606

Words-WRF3 43.82a 16.70 75.71

b 12.05 80.75

b 11.15 48.81 .562


b 14.35 78.50

c 11.71 37.77 .498

Note. The maximum scores for the word-reading fluency (WRF)-lists are 150, 150, and 120, respectively; for pseudoword-reading 116. Numbers and means in the same row that do not share subscripts differ at p < .05 on the Tukey honestly significant difference comparison.


Chapter 2

52

January Gr5June Gr2June Gr1January Gr1

Num

ber o

f wor

ds

100

80

60

40

20

0

At-risk DyslexicAt-risk Non-dyslexicControl

January Gr5June Gr2June Gr1

Num

ber o

f pse

udow

ords

80

60

40

20

At-risk DyslexicAt-risk Non-dyslexicControl

Figure 2.1. Development of reading. First panel: words (i.e., WRF1); second panel: pseudowords.



53

2

η2p = .218, and Group, F (2, 76) = 9.71, p < .001, η2

p = .203. The Time by Group

interaction approached significance, F (2, 76) = 2.59, p = .082, η2p = .064. Pairwise

comparisons revealed that halfway through Grade 1 the at-risk non-dyslexics

were significantly slower than the controls, but significantly faster than the at-

risk dyslexics. By the end of the school year, this group had reached the same

level as the controls.

The statistics for reading fluency are presented in Table 2.4. In addition,

the results for WRF1 and pseudowords are presented graphically in Figure

2.1. On all word and pseudoword reading tests, the main effects of both

Group and Time, as well as their interaction were significant (ps < .001, results

available on request). On all word and pseudoword reading tests and on all

occasions, the at-risk dyslexic group lagged significantly behind the other

two. moreover, the performance on both word and pseudoword reading

of the at-risk non-dyslexics was significantly below that of controls in first

and second grade. In Grade 5, however, the at-risk non-dyslexics still read

pseudowords significantly less fluently than the controls, but performed at

the same level on word reading.

2.4 Discussion

The reading development of children born to families with or without a history

of dyslexia was followed from kindergarten to fifth grade. In fifth grade reading

status (dyslexic or not dyslexic) was determined. Then, the development of

early reading-related abilities and reading in the groups differing in family-risk

status and reading status were compared.

In kindergarten, we found that the amount of letter-knowledge and

rapid naming differed among the groups of at-risk dyslexic, at-risk non-dyslexic,

and not at-risk normal-reading children (controls). however, differences among

the groups in the development of phonological awareness were not found.

The finding that the level of phonological awareness in kindergarten

was similar in dyslexic and normal reading children is in accordance with results

reported by de Jong and van der Leij (2003). They also did not find differences

between groups of kindergartners that later did or did not develop dyslexia.

De Jong and van der Leij attributed this null-result to Dutch kindergartners’

low level of phonological awareness (Wesseling & Reitsma, 1998). Apparently,


Chapter 2

54

as long as instruction pertaining to phonological awareness is sparse, there are

no group differences (de Jong & van der Leij, 1999).

however, in a recent prospective study, Boets et al. (2007; 2010) did find

deficits in a Dutch speaking dyslexic group (with and without risk combined)

relative to controls on phonological awareness assessed in kindergarten.

There are a number of reasons that might account for the different findings

on phonological awareness between the current study and the study of

Boets et al. (2010). To start with, it should be acknowledged that most of the

phonological-awareness performance in our study was at floor level, which

makes it difficult to detect differences. Another possibility concerns the

moment of assessing reading status. Boets et al. assessed reading status at the

start of third grade, whereas we assessed it mid-fifth grade. It is plausible that,

in general, groups will differ more on a correlate of reading when the time span

between the assessment of the correlate and the categorization into reading

groups is shorter. Indeed, when we repeated the analyses with the classification

based on reading scores at the end of second grade, the group differences on

phonological awareness were larger, although still not significant. In addition,

the relationship of phonological awareness and reading might change

over time. Vaessen and Blomert (2010) showed that the relation between

phonological awareness and word-reading fluency decreases as a function of

reading experience, so assessing reading status earlier might lead to groups

that differ more in kindergarten phonological awareness.

In contrast to phonological awareness, risk status and reading

outcome were related to the kindergartners’ knowledge of letters, as well as

their ability to rapidly name colors and objects. In line with previous studies

(Pennington & Lefly, 2001; Snowling, Gallagher, & Frith, 2003), rapid naming

and letter knowledge were impaired in the at-risk children who were later

classified as dyslexic, whereas the at-risk children who were later classified

as normal readers seemed to show milder impairments. These outcomes

support the notion that it is rapid naming rather than phonological awareness

that discriminates between reading groups in transparent orthographies,

as dyslexics in transparent orthographies are characterized by sufficient

reading accuracy but deficient reading fluency (de Jong & van der Leij, 2003;

Georgiou, Parrila, & Papadopoulos, 2008; Landerl & Wimmer, 2008). Since the

genes contributing to the manifestation of phonological awareness and rapid

naming seem to be only partially shared (Naples, Chang, Katz, & Grigorenko,



55

2

2009), a consequence of this notion is that the set of genes involved in the

etiology of dyslexia might differ somewhat between language environments

that differ in their orthography’s transparency. At present, this intriguing idea

remains speculative and warrants further investigation.

After half a year of reading instruction, the at-risk dyslexics were slower

in the naming of letters compared to the at-risk non-dyslexics, who were slower

than the controls. By the end of first grade the future dyslexics still lagged

behind, but the at-risk non-dyslexics performed at the level of controls. The

development of word-reading fluency showed a similar pattern, although over

a longer time period. Throughout the primary-school years, at-risk children

later classified as dyslexic read less fluently than the other two groups. The

group of at-risk non-dyslexics performed better than the at-risk dyslexics but

poorer than the controls in first and second grade, but in fifth grade they had

closed the gap and reached the same level of word reading as the controls.

meanwhile, the group differences in pseudoword-reading fluency were stable

over the five years studied, with the at-risk dyslexics performing less well and

the controls better than the at-risk non-dyslexics at all times.

Taken together, the results of the development of letter knowledge

and reading suggest that the at-risk non-dyslexics are genetically affected by

their dyslexic parent. Although they become normal readers in the long run,

they show subclinical deficits: first in pre-reading skills and later in reading

itself. These findings at the behavioral level are in line with an earlier report on

the present sample, showing atypical brain functioning during the processing

of simple visual stimuli at the age of five in the at-risk non-dyslexics (Regtvoort,

van Leeuwen, Stoel, & van der Leij, 2006). Recently, a study of Finnish researchers

revealed atypical processing of simple auditory stimuli of at-risk non-dyslexics

at birth (Leppänen et al., 2010). Despite these subtle deficits of at-risk non-

dyslexics, they catch-up with the controls on letter-naming and word-reading

fluency. Together with the finding that this group remains to lag behind in

pseudoword reading, these results might be taken to suggest that at-risk non-

dyslexics need more exposure to letters and words to reach the same level as

the controls. As a related possibility, it could be that this group has a larger

orthographic competence, the ability to store associations between written

and spoken word forms at the word and subword level (Bekebrede, van der Leij,

& Share, 2009), than their at-risk dyslexic peers. At the beginning of learning

to read differences between the at-risk dyslexic and non-dyslexic groups are


Chapter 2

56

relatively small, as word reading is highly dependent on phonological recoding.

however, as reading development proceeds and orthographic competence

becomes more important, at-risk non-dyslexic children outperform dyslexic

children and reach, though somewhat later, a similar level of word reading as

the controls.

The main focus of our study is to elucidate why some of the at-risk

children, about one-third, develop dyslexia whereas others with a similar

family background do not. Snowling et al. (2003) suggest that this latter group

compensates their phonological deficits with a higher general cognitive ability

or with good oral language skills. The current study does not provide support

for this explanation, as the groups neither differed in nonverbal IQ nor in

vocabulary. Likewise, there was no evidence of differences in nonverbal IQ and

vocabulary between the two at-risk groups in the studies of Elbro et al. (1998)

and Pennington and Lefly (2001). Still, we do not rule out that mild impairments

could be observed when broader oral language skills are assessed.

Similar to previous studies (Elbro, Borstrøm, & Petersen, 1998;

Snowling, muter, & Carroll, 2007), we also did not find evidence for the second

possible explanation that the at-risk non-dyslexic children experience a

more advantageous literacy environment compared to the at-risk dyslexic

children, though admittedly none of the studies (including ours) measured the

environment thoroughly (Torppa et al., 2007). however, in the present study

the parents of the non-dyslexic at-risk children had a higher level of education

compared to those of the dyslexic children, which has not been found before.

It is ambiguous whether this indicates a difference in the quality of their

offspring’s home literacy environment and/or a difference in their offspring’s

genetic potential.

A major finding of the current study is that within the group of dyslexic

parents, those whose child developed dyslexia performed worse in both word

and pseudoword reading. This might suggest that parents with more severe

dyslexia are genetically more affected and have offspring at higher genetic

risk, which fits with a continuous view of the family risk of dyslexia. Although

individual differences in parental reading skills cannot be attributed to genetic

influences alone, we suggest that it could be seen as an indication of their

offspring’s genetic liability for reading failure. The present finding needs

replication, but it shows that examining the relation between children’s reading

outcome and parents’ literacy levels is intriguing in itself.



57

2

One of the limitations of our study is that our control group is rather

small. however, the focus of this study is the difference between the two at-

risk groups which are of substantial size. moreover, our finding that the control

group outperforms the at-risk groups in letter knowledge, rapid naming, and

reading skills is consistent with previous longitudinal at-risk studies (Boets,

Wouters, van Wieringen, & Ghesquière, 2007; Elbro, Borstrøm, & Petersen,

1998; Pennington & Lefly, 2001; Scarborough, 1990; Snowling, Gallagher, &

Frith, 2003). Furthermore, the small effect sizes on differences in phonological-

awareness suggest that the non-significant results can hardly be attributed to

a lack of power.

It should be noted that about two-thirds of the at-risk children received

intervention to promote literacy development during kindergarten, which

might hamper interpretation of the data for the current aim. But although

there were intervention effects observable directly after the intervention had

finished, the differences between the trained and untrained at-risk children

vanished when all children started formal reading instruction in first grade

(Regtvoort & van der Leij, 2007; van Otterloo & van der Leij, 2009). The absence

of long-term intervention effects is also shown in a meta-analysis of Bus and van

IJzendoorn (1999). Finally, it is reassuring that the children who had received

intervention spread equally over the dyslexic and non-dyslexic groups.

In sum, our data show that about two-third of the children born to

families with a history of dyslexia acquire in the end adequate word-reading

skills, despite mild impairments in rapid naming and letter knowledge in

kindergarten and mild difficulties in pseudoword reading throughout middle

childhood. The profile of reading and reading-related skills of the at-risk dyslexic

group indicates stable and persistent impairments. Overall, our findings

suggest that it is not likely that at-risk children who do and do not develop

dyslexia differ in their intellectual abilities or literacy environment. Rather, they

might differ in genetic liability.


Chapter 2

58

2.5 References

Bekebrede, J., van der Leij, A., & Share, D. L. (2009). Dutch dyslexic adolescents: Pho-nological-core variable-orthographic dif-ferences. Reading and Writing, 22(2), 133-165. doi:10.1007/s11145-007-9105-7

Bishop, D. V. m. (2009). Genes, cognition, and communication. Annals of the New York Academy of Sciences, 1156 (The Year in Cognitive Neuroscience 2009), 1-18. doi:10.1111/j.1749-6632.2009.04419.x

Boets, B., De Smedt, B., Cleuren, L., Vandewalle, E., Wouters, J., & Ghesquière, P. (2010). To-wards a further characterization of phono-logical and literacy problems in Dutch-spea-king children with dyslexia. British Journal of Developmental Psychology, 28, 5-31.

Boets, B., Wouters, J., van Wieringen, A., & Ghesquière, P. (2007). Auditory proces-sing, speech perception and phonologi-cal ability in pre-school children at high-risk for dyslexia: A longitudinal study of the auditory temporal processing theory. Neuropsychologia, 45(8), 1608-1620.

Brus, B. T., & Voeten, m. J. m. (1972). Eén-mi-nuut-test [One-minute-test]. Lisse: Swets & Zeitlinger.

Bus, A.,G., & van IJzendoorn, m.,h. (1999). Phonological awareness and early rea-ding : A meta-analysis of experimental training studies. Journal of Educational Psychology, 91(3), 403-414.

de Jong, P. F., & van der Leij, A. (1999). Speci-fic contributions of phonological abilities to early reading acquisition: Results from a Dutch latent variable longitudinal study. Journal of Educational Psychology, 91(3), 450-476.

de Jong, P. F., & van der Leij, A. (2003). Deve-lopmental changes in the manifestation of a phonological deficit in dyslexic children learning to read a regular orthography. Journal of Educational Psychology, 95(1), 22-40. doi:10.1037/0022-0663.95.1.22

Elbro, C. (1996). Early linguistic abilities and reading development: A review and a hy-pothesis. Reading and Writing, 8(6), 453-485.


Gallagher, A. m., Frith, U., & Snowling, m. J. (2000). Precursors of literacy delay among children at genetic risk of dyslexia. Journal of Child Psychology and Psychiatry, 41(2), 203-213.

Georgiou, G. K., Parrila, R., & Papadopoulos, T. C. (2008). Predictors of word decoding and reading fluency across languages va-rying in orthographic consistency. Journal of Educational Psychology, 100(3), 566-580. doi:10.1037/0022-0663.100.3.566

Irausquin, R. S. (2001). Alliteratie test [allitera-tion test]. Unpublished manuscript.

Landerl, K., & Wimmer, h. (2008). Deve-lopment of word reading fluency and spelling in a consistent orthography: An 8-year follow-up. Journal of Educational Psychology, 100(1), 150-161.

Leppänen, P. h. T., hämäläinen, J. A., Salmi-nen, h. K., Eklund, K. m., Guttorm, T. K., Lohvansuu, K., Puolakanaho, A., Lyytinen, h. (2010). Newborn brain event-related potentials reveal atypical processing of sound frequency and the subsequent association with later literacy skills in children with familial dyslexia. Cortex, 46(10), 1362-1376. doi:10.1016/j.cor-tex.2010.06.003

Naples, A. J., Chang, J. T., Katz, L., & Grigoren-ko, E. L. (2009). Same or different? Insights into the etiology of phonological aware-ness and rapid naming. Biological Psycho-logy, 80(2), 226-239.



59

2


Puolakanaho, A., Ahonen, T., Aro, m., Eklund, K., Leppänen, P. h. T., Poikkeus, A., . . . Lyyti-nen, h. (2008). Developmental links of very early phonological and language skills to second grade reading outcomes: Strong to accuracy but only minor to fluency. Journal of Learning Disabilities, 41(4), 353-370. doi:10.1177/0022219407311747

Raven, J. C., Court, J. h., & Raven, J. (1984). Co-loured progressive matrices. London: Lewis.

Regtvoort, A. G. F. m., & van der Leij, A. (2007). Early intervention with children of dyslexic parents: Effects of computer-ba-sed reading instruction at home on liter-acy acquisition. Learning and Individual Differences, 17(1), 35-53.

Regtvoort, A. G. F. m., van Leeuwen, T. h., Stoel, R. D., & van der Leij, A. (2006). Effi-ciency of visual information processing in children at-risk for dyslexia: habituation of single-trial ERPs. Brain and Language, 98(3), 319-331.

Rutter, m. (2006). Genes and behavior. Natu-re-nurture interplay explained. Oxford, UK: Blackwell Publishing.


Seymour, P. h. K., Aro, m., & Erskine, J. m. (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94, 143-174.


Snowling, m. J., muter, V., & Carroll, J. (2007). Children at family risk of dys-lexia: A follow-up in early adolescence. Journal of Child Psychology and Psychia-

try, 48(6), 609-618. doi:10.1111/j.1469-7610.2006.01725.x



van den Bos, K. P. (2003). Serieel benoemen en woorden lezen [Serial naming and word reading]. Groningen: Rijksunversiteit Gro-ningen.

van den Bos, K. P., lutje Spelberg, h. C., Scheepstra, A. J. m., & de Vries, J. R. (1994). De klepel: Een test voor de leesvaardigheid van pseudowoorden [The klepel: A test for the reading skills of pseudowords]. Lisse: Swets & Zeitlinger.

van den Oord, E. J. C. G., Pickles, A., & Wald-man, I. D. (2003). Normal variation and abnormality: An empirical study of the lia-bility distributions underlying depression and delinquency. Journal of Child Psycho-logy and Psychiatry, 44(2), 180-192.

van Otterloo, S. G., & van der Leij, A. (2009). Dutch home-based pre-reading interven-tion with children at familial risk of dys-lexia. Annals of Dyslexia, 59(2), 169-195.

Vellutino, F. R., Fletcher, J. m., Snowling, m. J., & Scanlon, D. m. (2004). Specific reading disability (dyslexia): What have we learned in the past four decades? Journal of Child Psychology and Psychiatry, 45(1), 2-40.

Verhoeven, L. (1993a). Grafementoets. Toets voor auditieve synthese. Handleiding [Grap-heme test. Test for phoneme blending. Ma-nual]. Arnhem, the Netherlands: Cito.

Verhoeven, L. (1993b). Toets voor auditieve analyse. Fonemendictee. Handleiding [Test


Chapter 2

60

for phoneme segmentation. Dictation of phonemes. Manual]. Arnhem, the Nether-lands: Cito.

Verhoeven, L. (1995). Drie-minuten-toets. Handleiding [Three-minutes-test. Manual]. Arnhem: Cito.

Verhoeven, L. (2002). Passieve letterkennis [Receptive letter knowledge]. Arnhem: Cito.

Verhoeven, L., & van Leeuwe, J. (2009). modeling the growth of word-deco-ding skills: Evidence from Dutch. Scien-tific Studies of Reading, 13(3), 205-223. doi:10.1080/10888430902851356

Verhoeven, L., & Vermeer, A. (1996). Taaltoets allochtone kinderen [Language test for children of ethnic minorities]. Tilburg, The Netherlands: Zwijsen.

Wechsler, D. (1997). Wechsler adult intelligen-ce scale-III (WAIS-III). San Antonio: Psycho-logical Corporation.

Wesseling, R., & Reitsma, P. (1998). Phonemi-cally aware in a hop, skip, and a jump. In P. Reitsma, & L. Verhoeven (Eds.), Problems and interventions in literacy development (pp. 81-94). Dordrecht, the Netherlands: Kluwer Academic.

Wimmer, h. (2006). Don’t neglect reading fluency! Developmental Science, 9(5), 447-448.

Wimmer, h., mayringer, h., & Landerl, K. (2000). The double-deficit hypothesis and difficulties in learning to read a regular or-thography. Journal of Educational Psycho-logy, 92(4), 668-680.

Ziegler, A., König, I. R., Deimel, W., Plume, E., Nöthen, m. m., Propping, P., . . . Schul-te-Körne, G. (2005). Developmental dys-lexia - Recurrence risk estimates from a German bi-center study using the single proband sib pair design. Human Heredity, 59(3), 136-143.


I wish you to gasp not only at what you read

But at the miracle of its being readable.

- Vladimir Nabokov -

Child and parental literacy levels within families with a history of

dyslexia

van Bergen, E., de Jong, P.F., Plakas, A., Maassen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of

Child Psychology and Psychiatry, 53(1), 28-36.

Chapter


Child and parental literacy levels within families with a history of

dyslexia

van Bergen, E., de Jong, P.F., Plakas, A., Maassen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of

Child Psychology and Psychiatry, 53(1), 28-36.

Chapter


Chapter 3

64

Abstract

Background: The present study concerns literacy and its underlying

cognitive skills in Dutch children who differ in familial risk (FR) for

dyslexia. Previous studies with FR-children were inconclusive regarding

the performance of FR-children without dyslexia as compared to the

controls. moreover, van Bergen et al. (2011) recently showed that FR-

children with and without dyslexia differed in parental reading skills,

suggesting that those who go on to develop dyslexia have a higher

liability. The current study concerned 1) the comparison of three groups

of children at the end of 2nd grade and 2) the intergenerational transfer

of reading and its underlying cognitive skills from parent to child.

Method: Three groups of children were studied at the end of 2nd grade:

FR-dyslexia (n = 42), FR-no-dyslexia (n = 99), and control children (n = 66).

Parents and children were measured on naming, phonology, spelling,

and word and pseudoword reading.

Results: The FR-dyslexia children were severely impaired across all tasks.

The FR-no-dyslexia children performed better than the FR-dyslexia

children, but still below the level of the controls on all tasks; the only

exception was RAN, on which they were as fast as the controls. Focusing

on the FR sub-sample, parental reading and RAN were related to their

offspring’s reading status.

Conclusions: We replicated and extended van Bergen et al.’s study in

showing that the FR-children who develop dyslexia are likely to have

a higher liability. Both the group comparisons and the parent-child

relations highlight the importance of good RAN skills for reading

acquisition.


Literacy LeveLs within famiLies with a history of dysLexia

65

3

3.1 Introduction

There is a fair amount of evidence to support the observation that dyslexia

tends to run in families. In studies with children at familial risk (FR) of dyslexia,

i.e. children with at least one dyslexic parent, approximately 33% to 66% have

been reported to become dyslexic (Boets et al., 2010; Pennington & Lefly,

2001; Scarborough, 1990; Snowling, Gallagher, & Frith, 2003; Torppa, Lyytinen,

Erskine, Eklund, & Lyytinen, 2010; van Bergen et al., 2011). In contrast, the

percentage of dyslexic children of parents without dyslexia has been found

to range from 6% to 16%. In combination with the results from quantitative

genetic studies (e.g., Olson, Byrne, & Samuelsson, 2009; Pennington & Olson,

2005), such results suggest a strong hereditary basis of dyslexia.

Reading ability and disability is the end product of the effects of many

genes in interaction with the environment. From this multifactorial etiology

it follows that the underlying liability (see Falconer, 1965) for reading failure

is continuously distributed. Plomin, DeFries, mcClearn and mcGuffin (2008, p.

33) show that the involvement of only a few genes already lead to continuous

variation at the phenotypic level. From a liability continuum two propositions

can be deduced. Firstly, children with a FR of dyslexia but without dyslexia

themselves (FR non-dyslexics) have a higher position on the liability continuum

than control children. That is, they inherit and/or experience more risk factors

and therefore are expected to show weaknesses in literacy and its underlying

cognitive skills. Secondly, within the group of FR children there might be a

difference in liability to dyslexia between children with and without dyslexia.

Insight into this liability might be gained by the skills of their parents. The

current study explores both of these propositions.

3.1.1 FR Children Without DyslexiaSeveral FR-studies demonstrated that not only FR children who went on to

become dyslexic were impaired on precursors of reading. Those who became

normal readers performed in general better than the FR-dyslexia children, yet

poorer than the controls. This is in accordance with a continuum of liability to

dyslexia, in which FR-no-dyslexia children have a higher liability than controls.

however, there are only a few studies in which groups of FR-dyslexia, FR-no-

dyslexia, and control children were compared on reading and its underlying

cognitive skills after a few years of reading instruction. The first focus of the


Chapter 3

66

current study was on the differences among these three groups when the

children were 8 years of age, that is, after two years of formal reading instruction.

Specifically, the emphasis was on the status of the FR-no-dyslexia children as

compared to the controls.

Previous studies have provided a fairly clear picture of the performance

of the groups regarding literacy skills, but nevertheless, they have been

inconclusive regarding the cognitive skills underlying reading. Underlying

skills are of interest because they represent endophenotypes, that is, the

layer between the genotype and the phenotype of the reading behavior itself

(Snowling, 2008). The most important cognitive skills underlying reading are

assumed to be rapid naming (RAN) and phonological awareness (PA) (Snowling

& hulme, 2000). RAN is the rapid naming of highly familiar visual symbols and

PA refers to the ability to identify and manipulate the sounds in spoken words.

Verbal short-term memory is sometimes seen as a third underlying cognitive

skill (Boets et al., 2010), but in literate individuals it tends to be closely tied

to PA (de Jong & van der Leij, 1999), which suggests that both measures tap

phonological processing.

Comparing at age 8 (i.e., after two years of reading instruction)

FR-children classified as dyslexic with controls, studies report pronounced

difficulties in reading, spelling, PA, verbal short-term memory, and RAN (Boets

et al., 2010; de Bree, Wijnen, & Gerrits, 2010; Pennington & Lefly, 2001; Snowling,

muter, & Carroll, 2007). These studies did not agree, however, on whether the

two groups of normal readers differ from each other. It is particularly intriguing

that the FR-children classified as non-dyslexic showed mild impairments

relative to the controls on reading and spelling, while the findings regarding the

difference between these two groups on the underlying skills are inconsistent.

In the study of Pennington and Lefly (2001) the FR-no-dyslexia children were

comparable to the controls on PA, but showed difficulties in verbal short-term

memory, RAN, and all literacy measures. Snowling et al.(2003) also found that

FR-no-dyslexia children and controls differed on the literacy tasks. In addition,

there was a trend for the FR-no-dyslexia children to perform somewhat weaker

on PA. however, they performed equally well on the repetition of nonwords,

a phonological task that measures verbal short-term memory. more recently,

Boets et al.(2010) reported mild impairments in the FR-no-dyslexia group

relative to controls on two out of the six literacy measures, although it seemed

that larger groups would yield a significant difference on the other literacy



67

3

measures as well. The FR-no-dyslexia children were comparable to the controls

on RAN and verbal short-term memory, whereas there was a trend, as in the

study of Snowling et al., for PA to be lower. Taken together, it seems that

difficulties of FR-children without dyslexia are subtle and may not be detected

in studies with small samples.

3.1.2 Intergenerational TransferSince the etiology of dyslexia is multifactorial, involving both genetic and

environmental influences, individual differences between children in dyslexic

families must be associated at least in part with variations in their parents’

reading and cognitive skills. As an extension to earlier FR-studies, the second

focus of our study was the relation between the skills of the children and those

of their parents. In particular, we examined the relation between child and

parental literacy achievement and their underlying cognitive skills.

FR-studies start from the premise that children with and without a

FR of dyslexia differ in their liability of dyslexia. Implicitly, the assumption has

been that within the FR-subsample, children do not differ in their FR. Since

the reading skills of dyslexic parents all fall in the lower tail of the distribution,

differences in parental reading skills between FR-children with and without

dyslexia are not expected. To our knowledge, only two studies have tested

this assumption. Snowling et al.(2007) looked at the relation between

parents’ and children’s reading skills when the children were 12-13 year old.

Although reading skills of parents and children were related within the control

subsample, they were not within the FR-subsample. Based on these findings,

Snowling et al. conclude that “it does not seem to be the case that more severely

affected parents have more severely affected offspring”. This conclusion was

contradicted by the study of van Bergen et al. (2011). Despite the fact that

the dyslexic parents performed on average close to the fifth percentile, they

found that the dyslexic parents whose child manifested dyslexia at age 11 were

more severely impaired on word and nonword reading than those whose child

acquired reading skills within the normal range. This finding suggests that the

two groups of FR-children are not equal in their FR, i.e. the FR-children who

become dyslexic have a higher liability of developing dyslexia. Differences

in parental reading skills raise the question as to whether the FR-groups also

differ in parents’ cognitive skills underlying reading.

In the present study we compared groups of children differing in FR-


Chapter 3

68

status and reading outcome on literacy and its underlying skills. In addition,

we assessed literacy and its underlying skills in their parents, which allowed

us to investigate intergenerational transfer of skills from parent to child.

Given a liability continuum and the high heritability of both reading and its

underpinnings, we expected the FR-dyslexia children to show deficits across all

tasks and the FR-no-dyslexia children to display mild weaknesses. Based on the

findings of van Bergen et al. (2011), we predicted the FR-dyslexia children to

have more severely affected dyslexic parents than the FR-no-dyslexia children.

3.2 Method

3.2.1 ParticipantsThree groups of children were involved in this study. The FR-dyslexia group

consisted of 42 dyslexic children with a FR for dyslexia. The FR-no-dyslexia

group comprised 99 children with a FR for dyslexia but without dyslexia. Finally,

the control group included 66 children without a FR for dyslexia and without

dyslexia.

All children participated in the Dutch Dyslexia Programme (DDP):

a longitudinal multi-center study including the universities of Amsterdam,

Nijmegen, and Groningen. Couples with and without a history of dyslexia who

expected a baby were recruited between 1999 and 2002. They were recruited

via advertisements in newspapers and leaflets at GPs and midwives. Only

healthy, full-term babies were admitted to the program. The program was

approved by the ethical commission and all parents gave informed consent

for participation. About two-thirds of the sample consisted of families with a

history of dyslexia.

Children were considered to have a FR for dyslexia if at least one of

the parents and a close relative reported to have dyslexia. The dyslexia of the

parent had to be confirmed by tests measuring word and nonword-reading

fluency (see below). moreover, the subtest Similarities of the Wechsler Adult

Intelligence Scale was administered to measure verbal reasoning (Wechsler,

1997). For inclusion of a child in the FR-group, the parent with dyslexia had to

fulfill at least one out of the following criteria: 1) scores on both reading tests

≤ 20th percentile, 2) one reading score ≤ 10th percentile and the other reading

score ≤ 40th percentile, or 3) a discrepancy of ≥ 60 percentiles between the



69

3

verbal-reasoning score and one of the reading scores, with the restriction that

neither reading-test percentile score exceeded 40. On average, the dyslexic

parents of the FR-children scored on the reading tests close to the 5th percentile.

For inclusion in the no-FR-group both parents had to score ≥ 40th percentile on

both reading tests.

Dyslexia in children was assessed at the end of second grade. Children

were considered dyslexic when their reading score on a word-reading fluency

task (WRF2, see below) corresponded to the weakest ten percent in the

population (see Boets et al., 2010 for a similar criterion). In the no-FR-group

only two children turned out to be dyslexic. These children were excluded from

the study. The characteristics of the three groups are presented in Table 3.1.

Percentages of boys were similar across groups. however, compared to the

other two groups, the FR-dyslexia group had a lower nonverbal IQ (SON-R,

Tellegen, Winkel, Wijnberg-Williams, & Laros, 1998) at 48 months of age.

Furthermore, the FR-dyslexia group was slightly older than the controls.

Table 3.1 Children’s Characteristics by Group

FR

Characteristic Dyslexia No-dyslexia Control F(2, 204)

Sample size 42 99 66

No (%) of boys 25a (60%) 56

a (57%) 39

a (59%)

Nonverbal IQ (at 48 months) 105.90a (15.27) 113.70

b (14.31) 114.14

b (15.38) 4.75

At assessment end Grade 2

Age in months 98.60a (5.52) 97.67

ab (4.47) 96.41

b (4.34) 3.02

No of months reading instruction 18.69a (0.68) 18.87

a (0.74) 18.80

a (0.66) 0.96

Note. The group means are given with standard deviations in parentheses. Numbers and means in the same row that do not share subscripts differ at p < .05 on the χ2-test or Tukey’s test. No = number.

3.2.2 MeasuresChildren

The children were measured at the end of second grade on reading accuracy

and fluency, spelling, PA, and RAN.

Word and nonword-reading accuracy. In these tests (de Jong & Wolters,

2002) the child was required to read a list of (non)words without time pressure.

Each test consisted of 40 items increasing in difficulty from one to four


Chapter 3

70

syllables. Testing was discontinued when six out of the last eight items were

read incorrectly.

Word-reading fluency (WRF). This was assessed using three lists

(Three-minutes-Test, Verhoeven, 1995). The first and second (WRF1 and WRF2,

respectively) consisted of 150 monosyllabic words each. WRF1 included

words with a CV, VC, or CVC pattern. All words in WRF2 contained at least

one consonant cluster. WRF3 comprised 120 polysyllabic words with various

orthographic complexities. The score per list was the number of words read

correctly within one minute. WRF2 was used to assess dyslexia, as it has a high

reliability (.96 at the end of Grade 2, moelands, Kamphuis, & Verhoeven, 2003)

and differentiates well between children at this age.

Nonword-reading fluency. This test (Klepel: van den Bos, lutje Spelberg,

Scheepstra, & de Vries, 1994) consisted of a list of 116 pseudowords of increasing

difficulty. The child was asked to read as many pseudowords as possible within

two minutes, with the score being the number correctly read.

Spelling. The child had to write down (without time pressure) mono-

and disyllabic words presented in a sentence (van den Bosch, Gillijns, Krom, &

moelands, 1993). The test had two versions, an easier and a more difficult one,

consisting of 38 and 36 items. The easier version was given at two centers, the

difficult version at the third. According to the test manual the scores of the tests

could be transformed to one scale (based on Rasch analyses). The scores on this

scale ranged from 80 to 144.

PA. PA was measured using a phoneme-deletion test (de Jong & van

der Leij, 2003). On each item of the test a phoneme, always a consonant, had

to be deleted from a nonword resulting in another nonword. The test consisted

of two parts. The first part comprised nine monosyllabic and nine disyllabic

nonwords. The second part consisted of nine disyllabic nonwords, with the

phoneme that had to be deleted occurring twice. Testing was discontinued

when six consecutive items were answered incorrectly in the first part or three

in the second part. Each attempt was scored as either correct or incorrect. Both

parts of the test started with two items for practice.

RAN. Serial RAN (RAN: van den Bos, 2003) was measured for numbers (2, 4,

5, 8, and 9) and colors (black, yellow, red, green, blue). Each of the tasks consisted of

fifty randomly ordered symbols arranged in five columns of 10 symbols (numbers

or colors) each. Before test administration, the child practiced by naming the last

column of symbols. Children were instructed to name the symbols column-wise



71

3

as quickly as possible. The time to completion was transformed to number of

symbols per second to normalize the score distribution.

Parents

The dyslexic parent of the FR-children and both parents of the control children

were assessed on reading fluency, spelling, nonword repetition, and RAN.

The data presented are the data of the weakest-reading parent. In addition,

the level of education of both parents was ranked on a scale ranging from 1

(primary school only) to 5 (university degree).

Word-reading fluency. The test (One-minute-Test: Brus & Voeten, 1972)

consisted of a list of 116 words of increasing difficulty. The parent was asked to

correctly read as many words as possible within one minute.

Nonword-reading fluency. The same test as for the children was used

(see above).

Spelling. The ability to apply phonological analysis and spelling rules

was tested with a test that required writing nonwords to dictation. The 26

nonwords ranged from one to three syllables.

Nonword repetition. The test (de Jong & van der Leij, 1999) consisted

of 48 randomly ordered nonwords ranging from one to four syllables. The

nonwords were presented twice on audiotape. After the nonword was

presented the parent had to repeat back the nonword. The task started with

three practice items. Each attempt was scored as either correct or incorrect.

RAN. Three serial RAN tasks were developed for the present study:

one for upper case letters (all, except I, Q, and Y), one for numbers (2 up to

and including 11), and one for colors (green, red, yellow, and blue). Each of the

tasks consisted of 50 randomly ordered symbols arranged in five columns of

10 symbols each. Before test administration, the parent practiced by naming

the last column of symbols. Parents were instructed to name the symbols

column-wise as quickly as possible. For all three tasks, the time to completion

was transformed to number of symbols per second to normalize the score

distribution.

3.2.3. ProcedureParents and children were tested individually by trained graduate students.

The parents were tested in a quiet room at home or at the hospital around the

birth of their child. The children were tested in a quiet room at the university

between may and July of Grade 2.


Chapter 3

72

3.3 Results

3.3.1 Data ScreeningOne dyslexic parent and their dyslexic child were removed from the analyses

because of missing scores and scores on literacy measures below 3.3 SD’s of the

mean of their group. Data were missing for just a couple of the parents’ tests

(1-2%) and educational levels (4%). The analyses of these variables therefore

included somewhat fewer cases. All score distributions were approximately

normal. The one exception was for word-reading accuracy, which was too

easy for the two groups of normal readers. These scores were logarithmically

transformed before analysis.

3.3.2 ChildrenTable 3.2 displays the performance of the three groups of children (FR-dyslexia,

FR-no-dyslexia, and controls) on the literacy and cognitive measures. One-

way ANOVAs followed up with Tukey’s test were used to investigate group

differences. Effect sizes (Cohen’s d’s) were computed using the SD’s of the

control group.

All literacy measures showed the same pattern: the FR-no-dyslexia

group scored considerably higher than the FR-dyslexia group (with effect sizes

between 1.22 and 2.41), but was still impaired relative to the control group

(with effect sizes between 0.37 and 0.74). On average, the children with dyslexia

could only read the mono- and disyllabic words and monosyllabic nonwords of

the accuracy tests.

Similar group differences were observed for PA, but not for RAN. The

FR-dyslexia group was significantly impaired on both RAN tasks; however, the

FR-no-dyslexia group performed similarly to the control group.

Finally, some extra analyses were performed to investigate whether the

observed group differences on literacy(-related) skills were genuine rather

than due to group differences on children’s non-verbal IQ (Table 3.1) or

parental educational level (Table 3.5). Nonverbal IQ correlated between .24 and

.32 with the reading measures, .37 with spelling, .28 with PA and .07 and .20

with RAN digits and colors. Controlling for IQ in a series of one-way ANCOVAs

yielded virtually the same results as the reported ANOVAs (Table 3.2), as did

ANOVAs on a subset of data with groups matched for IQ. The level of education

of the weakest-reading parent only correlated significantly with the outcome



73

3

Tab

le 3

.2 C

hild

ren’

s Pe

rfor

man

ce o

n Li

tera

cy a

nd C

ogni

tive

mea

sure

s by

Gro

up

FR

Con

trol

Effec

t siz

e (C

ohen

’s d)

Dys

lexi

aN

o-dy

slex

ia

mea

sure

max

.M

SDM

SDM

SDF

(2, 2

04)

FRD

vs.

FRN

DFR

D v

s. C

FRN

D v

s. C

Read

ing

accu

racy

Wor

ds40

20.8

6 a10

.52

34.6

0 b5.

7337

.64 c

2.69

71.5

31.

732.

460.

74N

onw

ords

408.

19a

6.22

19.3

2 b10

.16

24.8

3 c8.

9943

.07

1.24

1.85

0.61

Read

ing

fluen

cyW

ords

- W

RF1

150

32.5

2 a11

.83

68.6

4 b16

.85

75.0

6 c15

.01

109.

422.

412.

830.

43W

ords

- W

RF2

150

19.0

7 a7.

1957

.51 b

19.4

867

.70 c

18.9

510

4.77

2.03

2.57

0.54

Wor

ds -

WRF

312

012

.95 a

7.02

42.6

8 b18

.17

53.5

6 c18

.21

79.4

91.

632.

230.

60N

onw

ords

116

13.0

0 a4.

5530

.37 b

12.6

037

.48 c

14.0

654

.73

1.24

1.74

0.51

Spel

ling

144

108.

50a

6.38

119.

04b

8.57

122.

21c

8.61

37.7

61.

221.

590.

37

PA27

9.26

a4.

8014

.47 b

6.00

18.1

7 c4.

8034

.16

1.09

1.86

0.77

RAN

Dig

its1.

22a

0.29

1.56

b0.

351.

61b

0.34

19.8

11.

001.

150.

15C

olor

s0.

83a

0.20

0.94

b0.

190.

96b

0.21

5.63

0.52

0.62

0.10

Not

e. C

hild

ren’

s sk

ills

wer

e as

sess

ed a

t th

e en

d of

Gra

de 2

. max

. = m

axim

um s

c ore

; WRF

= w

ord-

read

ing

fluen

cy. N

umb

ers

and

mea

ns in

the

sam

e ro

w t

hat

do n

ot

shar

e su

bsc

ripts

diff

er a

t p <

.05

on T

ukey

’s p

ost-

hoc

test

. Coh

en’s

d is

cal

cula

ted

usin

g th

e SD

s of

the

cont

rols

. max

. = m

axim

um s

c ore

, FRD

= F

R-dy

slex

ia, F

RND

= F

R no

-dys

lexi

a, C

= c

ontr

ol.


Chapter 3

74

variable PA: r = .16. Adjusting for parental education in ANCOVAs did not alter

the pattern of results. Although equating for group differences can be disputed

(miller & Chapman, 2001), the fact that the ANCOVAs yield the same picture as

the ANOVAs strengthen our findings.

3.3.3 Intergenerational TransferTables 3.3 and 3.4 display correlations between children’s and parents’ measures

of literacy and their underlying skills. As an illustration, the relation between

nonword-reading fluency in children and parents is shown graphically in

Figure 3.1. This measure was chosen since it was assessed with the same test in

children and parents. Due to our strict selection criteria (i.e., percentile scores

of the dyslexic parents ≤ 40 and those of the control parents ≥ 40), correlations

within the FR and control subsamples suffered from restriction of range in the

parental measures, which suppressed correlations. The correlations presented

are therefore of the FR and control subsamples combined. In the control

sample only the data of the weakest-reading control parent was selected.

Correlations are given separately for the weakest-reading mothers and their

children and the weakest-reading fathers and their children. The correlations

show that reading skills of parents and children were moderately correlated. As

expected, mothers’ nonword repetition correlated strongest with children’s PA.

mothers’ RAN, however, correlated stronger with children’s reading than with

children’s RAN. No such clear pattern was observed for father-child correlations.

Children’s reading fluency seemed to correlate higher with the performance of

mothers than of fathers. The significance of this difference was tested using

mx (Neale, 2004). It appeared that an unconstrained two-group model fitted

the data better than one with the set of eight correlations (i.e., two reading-

fluency measures times four parental measures) constrained to be equal across

mothers and fathers, Δχ2(8, N=205) = 15.55, p = .049. Children’s PA and RAN,

on the other hand, did not correlate differently with test scores of mother and

fathers, Δχ2(8, N=205) = 7.86, p = .447. The proportion of dyslexic parents was

virtually equal in the two groups, as was the variability in test scores. hence,

these factors could not explain the differences between maternal and paternal

correlations.



75

3

Table 3.3 Correlations among skills of mothers and their children

mother

Reading fluency

Words NonwordsNonwordrepetition RAN digits

ChildrenReading fluency

Words .43 .47 .23 .38Nonwords .51 .56 .21 .49

Phonological awareness .54 .50 .34 .31RAN digits .13 .12 .10 .19

Note. Word-reading fluency of children is WRF2. n’s [83, 87]. For n = 85: p < .05 for r ≥ .21 and p < .001 for r ≥ .35.

Table 3.4 Correlations among skills of fathers and their children

Father

Reading fluency

Words NonwordsNonword repetition RAN digits

ChildrenReading fluency

Words .39 .38 .22 .27Nonwords .30 .32 .15 .15

Phonological awareness .22 .29 .22 .16RAN digits .28 .26 .17 .19

Note. Word-reading fluency of children is WRF2. n’s [119, 120]. For n = 119: p < .05 for r ≥ .18 and p < .001 for r ≥ .30.


Chapter 3

76

Figure 3.1. Scatterplots (with regression lines) illustrating the relation for nonword-reading fluency between children and mothers (top panel, r = .56) and children and fathers (lower panel, r = .32).



77

3

Differences among the FR-dyslexia, FR-no-dyslexia, and control children

were also investigated for parental characteristics and skills. There was no difference

between the two FR-groups in gender of the dyslexic parent: The dyslexic parent

was the father in 55% of the cases in the FR-dyslexia group compared to 62%

in the FR-no-dyslexia group (χ2 < 1). The other parental characteristics and test

scores by group are shown in Table 3.5. The effect sizes of the comparisons with

the control group were large; the comparisons of the two FR-groups yielded

small effect sizes, except for word-reading fluency and alphanumeric RAN with

medium effect sizes. Regarding background characteristics, the mothers and

fathers of both FR-groups had a lower educational level compared to the control

group. The two FR-groups differed in educational level of the non-dyslexic

parent, but not in that of the dyslexic parent. The general Dutch population

has an average level of education of 3.20 (with SD = 0.94, Statline, 2010), which

indicates that the parents in the FR and control samples were slightly (Cohen’s d

= .31) and considerably (Cohen’s d = 1.01) above average, respectively.

The parental skills investigated included verbal reasoning, literacy, and

literacy correlates. The parents in the control group obtained the highest scores

on verbal reasoning, followed by the FR-dyslexia and FR-no-dyslexia groups,

respectively. By definition, the dyslexic parents in the FR-groups were weaker

readers than the parents in the control group, but they scored also significantly

lower than the controls on spelling. Since we hypothesized that the parents of

the FR-dyslexia children would score lower than those of the FR-no-dyslexia

children on literacy(-related) tasks, we performed planned t-tests. As predicted,

the dyslexic parents of the FR-dyslexia group scored significantly lower than

those of the FR-no-dyslexia group on word-reading fluency (t(139) = 1.69, p =

.047, one-tailed). however, the FR-groups did not differ in nonword-reading

fluency and spelling. The reading measures correlated similarly with educational

level (within FR-sample: word reading: r = .38; nonword reading: r = .42).

Concerning the correlates of reading, the parents of the controls were

better than those of the FR-children on nonword repetition and RAN. The two

FR-groups did not differ significantly on nonword repetition and RAN of colors,

but they did on alpha-numeric RAN: the dyslexic parents of the FR-no-dyslexia

children were significantly better on both letters (t(137) = 2.16, p = .016, one-

tailed) and digits (t(137) = 2.52, p = .006, one-tailed). Parental RAN hardly

correlated with their educational level (within FR-sample: .14, .19, and .10 for

letters, digits, and colors, respectively).


Chapter 3

78

Tab

le 3

.5 P

aren

ts’ C

hara

cter

istic

s an

d Pe

rfor

man

ce o

n Li

tera

cy a

nd C

ogni

tive

mea

sure

s by

Gro

up

FR

Dys

lexi

aN

o-dy

slex

iaC

ontr

olEff

ect s

ize

(Coh

en’s

d)

Cha

ract

eris

ticm

ax.

MSD

MSD

MSD

Fdf

FRD

vs.

FRN

DFR

D v

s. C

FRN

D v

s. C

Leve

l of e

duca

tion

ofm

othe

r5

3.40

a0.

873.

56a

0.89

4.14

b0.

7212

.69

(2, 1

96)

0.22

1.03

0.81

Fath

er5

3.34

a0.

883.

51a

0.91

4.16

b0.

8813

.66

(2, 1

96)

0.19

0.93

0.74

Dys

lexi

c p

aren

t5

3.41

a0.

873.

41a

0.90

< 1

(1, 1

35)

0.00

Non

-dys

lexi

c p

aren

t5

3.33

a0.

893.

67b

0.89

4.26

(1, 1

32)

0.38

Verb

al re

ason

ing

2616

.98 ab

3.04

16.4

3 a3.

7918

.43 b

3.16

6.63

(2, 2

02)

-0.1

70.

460.

63

Read

ing

fluen

cyW

ords

116

61.8

1 a12

.19

65.8

6 a13

.35

96.6

1 b8.

4816

7.62

(2, 2

04)

0.48

4.10

3.63

Non

wor

ds11

646

.17 a

17.9

447

.73 a

15.0

010

0.95

b9.

8431

8.42

(2, 2

04)

0.16

5.57

5.41

Spel

ling

2612

.98 a

4.79

13.4

7 a3.

9618

.66 b

3.12

41.5

4(2

, 200

)0.

161.

821.

66

Non

wor

d re

pet

ition

4836

.50 a

5.99

37.2

4 a5.

6742

.83 b

3.50

28.0

7(2

, 200

)0.

211.

811.

60

RAN Le

tter

s1.

91a

0.45

2.10

a0.

502.

79b

0.41

60.3

7(2

, 200

)0.

462.

151.

68

Dig

its1.

91a

0.47

2.14

b0.

532.

72c

0.44

41.2

3(2

, 200

)0.

521.

841.

32

Col

ors

1.41

a0.

271.

45a

0.29

1.87

b0.

2950

.48

(2, 1

99)

0.14

1.59

1.45

Not

e. T

est

scor

es b

elon

g to

the

wea

kest

-rea

ding

par

ent.

Num

ber

s an

d m

eans

in t

he s

ame

row

tha

t do

not

sha

re s

ubsc

ripts

diff

er a

t p

< .0

5 on

Tuk

ey’s

pos

t-ho

c te

st.

Coh

en’s

d is

cal

cula

ted

usin

g th

e SD

s of

the

cont

rols

. max

. = m

axim

um s

c ore

, FRD

= F

R-dy

slex

ia, F

RND

= F

R no

-dys

lexi

a, C

= c

ontr

ol.



79

3

Finally, it was investigated whether the effect of parental characteristics

on children’s literacy outcome was fully mediated by children’s characteristics.

Therefore, a set of hierarchical regression analyses focusing on the FR-group

was performed. In predicting children’s literacy outcome (i.e., WRF2), first child

predictors were entered (nonverbal IQ, PA, and RAN) and then one of the

parental characteristic. It was found that parental word-reading fluency (β = .136,

p = .031) and alphanumeric RAN (letters: β = .149, p = .018; digits: β = .123, p =

.054) added further variance to the model, indicating that they affected children’s

literacy outcome (partly) independent from children’s cognitive skills. Parental

characteristics that did not improve the model were nonword-reading fluency

(β = .096, p = .128), RAN colors, nonword repetition, and level of education of the

dyslexic and non-dyslexic parent (all ps > .40). The predictive value of parental

RAN on children’s literacy outcome is nicely illustrated when comparing groups

of dyslexic parents with and without a RAN deficit (using the scores marking the

bottom 10% in the control group as cut-offs). In the group without a RAN deficit

in letters or digits 14% and 16%, respectively, manifested dyslexia, while these

proportions where 39% and 41%, respectively, for the parents with a RAN deficit.

3.4 Discussion

After two years of reading instruction, 30% of the children with a family history

of dyslexia had developed dyslexia, compared to only 3% of the children

without such a history. These figures agree with other studies using a similarly

(strict) criterion to identify dyslexia.

The FR-dyslexia children were significantly impaired relative to the FR-

no-dyslexia children on PA, RAN, spelling, and reading accuracy and fluency of

both words and nonwords. This is in line with previous longitudinal FR-studies

contrasting groups of 8-year-olds (Boets et al., 2010; Pennington & Lefly, 2001;

Snowling et al., 2003; Torppa et al., 2010; van Bergen et al., 2011). In contrast

to several earlier studies in transparent orthographies (e.g., de Jong & van der

Leij, 2003; Wimmer, 1993), but in accordance with Boets et al. (2010), reading

accuracy differentiated dyslexic from normal readers and even FR-no-dyslexia

children from controls. This shows that differences in accuracy are observable

until at least second grade, provided that the task is not restricted to one- or

two-syllable items.


Chapter 3

80

In accordance with previous studies (Boets et al., 2010; Snowling et al.,

2003), the FR-no-dyslexia group performed better than the FR-dyslexia group,

but weaker than the controls on PA and literacy, although their performance

was well within the normal range. In fact, compared to the national norm

scores for the reading-fluency and spelling tests, the performance of the FR-no-

dyslexia group was just below average, while the performance of the controls

was just above.

Regarding PA, this study is the first study that shows reliable differences

between FR-no-dyslexia children and controls. This confirms the earlier

reported trends observed by Snowling et al. (2003) and Boets et al. (2010).

Strikingly, the FR-no-dyslexia children were unimpaired on RAN. Of the two

other FR-studies that also included RAN, our result resembles that of Boets et

al. (2010), but contrasts with that of Pennington and Lefly (2001), who found

the FR-no-dyslexia group to perform slower. Unlike Boets et al. and our study,

Pennington and Lefly used reading accuracy to diagnose dyslexia. however,

reanalyzing RAN differences with reading status based on word-reading

accuracy still yielded equivalent performances of our two non-dyslexic groups.

The poor performance on RAN of the FR-children with dyslexia does

not seem to be a consequence of their poor reading. The current study showed

that the FR-children without dyslexia were faster on RAN than their relatively

weak reading fluency suggested. Furthermore, longitudinal studies have

demonstrated that FR-children who later manifested dyslexia were already

slow at RAN before they came to the task of learning to read (Boets, Wouters,

van Wieringen, & Ghesquière, 2007; Pennington & Lefly, 2001; van Bergen et al.,

2011). Finally, Lervåg and hulme (2009) found RAN to be a predictor of reading

development, but failed to find support for a reciprocal influence of reading

on RAN. These findings argue against RAN performance as a consequence of

reading performance.

Besides the skills of the children in the three groups, we aimed to

investigate the relation between child and parental skills. Child and parental

reading skills were moderately correlated: ~.35 for fathers and ~.50 for

mothers. These correlations are higher than the correlation of .28 reported by

Bynner and Parsons (2006), although this may well be explained by the fact

that parental literacy skills in their cohort study were self-reported instead

of tested. Interestingly, children’s reading fluency, but not their PA and RAN,

correlated higher with mothers’ than with fathers’ skills. This is suggestive of an



81

3

environmental influence: it could be that mothers are in general more involved

in parenting and that this has an effect on their offspring’s reading skill but not

on its underlying cognitive skills. On the other hand, we cannot rule out the

possibility that some of the fathers in our sample are not the biological fathers.

This would suppress father-child correlations. moreover, the finding should be

interpreted with caution since it involves only data from the weakest-reading

parent. Replication with complete parent data is therefore warranted.

Importantly, we found that the dyslexic parents whose child manifested

dyslexia were even more impaired on word-reading fluency than those whose

child had an age-appropriate reading level. however, the difference was

smaller (i.e., a medium effect size of d = .48) than the difference observed by

van Bergen et al. (i.e., a large effect size of d = .78). moreover, the current study

failed to replicate a difference in nonword-reading fluency, as reported by van

Bergen et al.. Since reading ability is not yet fully stabilized after two years of

reading instruction, the moment of assessing reading status could partially

explain the differences between the studies. Indeed, when we reanalyzed van

Bergen et al.’s data with reading status based on children’s reading skills at

the end of second grade (instead of fifth grade, as in the original study), the

difference in parental reading ability between the two FR-groups decreased,

although it remained larger than in the current sample. Furthermore, we

wondered whether the presence of a word-reading difference and the absence

of a nonword-reading difference in the current sample could be subscribed to

a difference in the amount the two groups of dyslexic parents read. however,

the two reading measures correlated similarly with educational level, which

argues against this notion.

As expected, the dyslexic parents were impaired on all underlying

cognitive skills. Although nonword repetition (as a proxy for PA) discriminated

well between parents with and without dyslexia, within the FR-subsample

parental nonword repetition appeared to be unrelated to children’s reading

outcome. Interestingly, the dyslexic parents of the FR-dyslexia children were

somewhat slower on RAN than those of the FR-no-dyslexia children. Parental

differences in reading and RAN might be related to differences in liability of

their offspring. Compared with parental reading, parental RAN might be a

cleaner indicator of their genetic potential, because it is less influenced by

reading experience (as shown by the correlations with educational level).

Another indicator of children’s liability is that of the parents’ level of


Chapter 3

82

education. As also observed by van Bergen et al. (2011), the FR-no-dyslexia

children have non-dyslexic parents with a higher educational level compared

to FR-dyslexia children. This might be an additional factor which affects

the outcome of children, operating via inheritance and the home-literacy

environment.

Subsequently, we examined possible intergenerational transfer by

examining whether cognitive skills of the dyslexic parents were related to

children’s reading outcome. It appeared that differences in parental word-

reading fluency and alphanumeric RAN were predictive of differences in

children’s reading ability, beyond the effect of children’s cognitive skills. The

predictive value of parental RAN is also shown by the fact that of the children of

a dyslexic parent without a RAN deficit just 15% developed dyslexia, compared

to as many as 40% of those with a RAN deficit. This agrees with the finding that

the two groups of FR-children differed in parental RAN. Both underline that

RAN deficits in dyslexic parents added to their offspring’s risk for dyslexia, over

and above their reading status. Furthermore, parental RAN correlated stronger

with children’s reading than with children’s RAN. Apparently, RAN of parents

and RAN of children somehow measure different aspects of children’s reading

ability.

The mild deficits in PA and literacy displayed by the FR no-dyslexia

group corroborate a liability continuum. Additionally, both children’s data and

parent-child relations highlight that – at least within transparent orthographies

– RAN plays a more important role in reading acquisition than PA. A parallel can

be drawn between the present findings and those of Bishop, mcDonald, Bird,

and hayiou-Thomas (2009), who studied another group at-risk for dyslexia:

children with language impairment. They also found that those who achieved

adequate reading skills despite their risk showed remarkable good RAN skills.

They suggest that what RAN taps might protect against reading failure. Our

data support this hypothesis, although longitudinal studies are required to

elucidate whether and how the skills underlying RAN performance might

protect children at FR for dyslexia.



83

3

3.5 Key Points

• Prospective studies with children at FR of dyslexia demonstrate the

multifactorial etiology of dyslexia.

• The current study shows that FR-children who do not meet criteria for

dyslexia are mildly impaired on literacy and PA, but not on RAN.

• Dyslexic parents of dyslexic children tended to be more severely impaired

on word-reading and RAN skills than dyslexic parents of non-dyslexic

children.

• Parental characteristics which affect the reading outcome of their offspring

are reading, RAN, and level of education. These factors can be used to

improve a child’s risk assessment for developing dyslexia.


Chapter 3

84

3.6 References

Bishop, D. V. m., mcDonald, D., Bird, S., & hayiou-Thomas, m. E. (2009). Children who read words accurately despite lan-guage impairment: Who are they and how do they do it? Child Development, 80(2), 593-605. doi: 10.1111/j.1467-8624.2009.01281.x


Boets, B., Wouters, J., van Wieringen, A., & Ghesquière, P. (2007). Auditory proces-sing, speech perception and phonologi-cal ability in pre-school children at high-risk for dyslexia: A longitudinal study of the auditory temporal processing theory. Neuropsychologia, 45(8), 1608-1620.


Bynner, J., & Parsons, S. (2006). New light on literacy and numeracy. London: NRDC.

de Bree, E., Wijnen, F., & Gerrits, E. (2010). Non-word repetition and literacy in dutch children at-risk of dyslexia and children with SLI: Results of the follow-up study. Dyslexia, 16(1), 36-44.

de Jong, P. F., & van der Leij, A. (1999). Speci-fic contributions of phonological abilities to early reading acquisition: Results from a dutch latent variable longitudinal study. Journal of Educational Psychology, 91(3), 450-476.


de Jong, P. F., & Wolters, G. (2002). Fonemisch bewustzijn, benoemsnelheid en leren le-zen [Phonemic awareness, rapid naming, and learning to read]. Pedagogische Studi-en [Pedagogical Studies], 79(1), 53-63.

Falconer, D. S. (1965). The inheritance of liability to certain diseases, estimated from the incidence among relatives. An-nals of Human Genetics, 29(1), 51-76. doi: 10.1111/j.1469-1809.1965.tb00500.x

Lervåg, A., & hulme, C. (2009). Rapid auto-matized naming (RAN) taps a mechanism that places constraints on the develop-ment of early reading fluency. Psycho-logical Science, 20(8), 1040-1048. doi: 10.1111/j.1467-9280.2009.02405.x

miller, G. A., & Chapman, J. P. (2001). misun-derstanding analysis of covariance. Jour-nal of Abnormal Psychology, 110(1), 40-48.

moelands, F., Kamphuis, F., & Verhoeven, L. (2003). Verantwoording drie-minuten-toets (DMT) [Account three-minutes-test]. Arn-hem: Cito.

Neale, m. C. (2004). Mx: Statistical modeling



Pennington, B. F., & Olson, R. K. (2005). Gene-tics of dyslexia. In m. J. Snowling, & C. hulme (Eds.), The science of reading: A handbook (pp. 453-472). malden: Blackwell Publishing.

Plomin, R., DeFries, J. C., mcClearn, G. E., & mcGuffin, P. (2008). Behavioral genetics (5th ed.). New York: Worth Publishers.



85

3


Snowling, m. J. (2008). Specific disorders and broader phenotypes: The case of dyslexia. The Quarterly Journal of Experimental Psy-chology, 61(1), 142-156.


Snowling, m. J., & hulme, C. (Eds.). (2000). The science of reading: A handbook. Blackwell Publishing.

Snowling, m. J., muter, V., & Carroll, J. (2007). Children at family risk of dys-lexia: A follow-up in early adolescence. Journal of Child Psychology and Psychia-try, 48(6), 609-618. doi: 10.1111/j.1469-7610.2006.01725.x

Statline. (2010). Labour force: Level of educati-on (August 2010 ed.) Statistics Netherlands.

Tellegen, P. J., Winkel, m., Wijnberg-Williams, B. J., & Laros, J. (1998). Snijders-Oomen nonverbal intelligence test, SON-R 2.5–7, manual and research report. Lisse: Swets & Zeitlinger.

Torppa, m., Lyytinen, P., Erskine, J., Eklund, K., & Lyytinen, h. (2010). Language de-velopment, literacy skills, and predictive connections to reading in Finnish children with and without familial risk for dyslexia. Journal of Learning Disabilities, 43(4), 308-321. doi: 10.1177/0022219410369096




van den Bosch, L., Gillijns, P., Krom, R., & moe-lands, F. (1993). Schaal vorderingen in spel-lingvaardigheid [Scale progress in spelling skills]. Arnhem: Cito.

Verhoeven, L. (1995). Drie-minuten-toets. Handleiding [Three-minutes-test.Mmanu-al]. Arnhem: Cito.

Wechsler, D. (1997). Wechsler adult intelligen-ce scale-III (WAIS-III). San Antonio: Psycho-logical Corporation.

Wimmer, h. (1993). Characteristics of developmental dyslexia in a regu-lar writing system. Applied Psycholin-guistics, 14(01), 1-33. doi: doi:10.1017/S0142716400010122


The whole is more than the sum of its parts.

- Aristotle -

The effect of parents’ literacy skills and children’s preliteracy skills on the

risk of dyslexia

van Bergen, E., de Jong, P.F., Maassen, B., & van der Leij, A. (submitted). The effect of parents’ literacy skills and children’s preliteracy skills on the risk of dyslexia.

Chapter


The effect of parents’ literacy skills and children’s preliteracy skills on the

risk of dyslexia

van Bergen, E., de Jong, P.F., Maassen, B., & van der Leij, A. (submitted). The effect of parents’ literacy skills and children’s preliteracy skills on the risk of dyslexia.

Chapter


Chapter 4

88

Abstract

This familial-risk study examines plausible preschool risk factors for

dyslexia. Participants included at-risk children with (n = 50) and without

(n = 82) dyslexia in Grade 3 and controls (n = 64). Firstly, children from

families in which both parents experienced literacy difficulties were 2½

times more likely to develop dyslexia than children from families with only

one affected parent. Secondly, we found impairments in phonological

awareness, rapid naming, and letter knowledge in at-risk kindergartners

who later developed dyslexia, and mild phonological-awareness deficits

in at-risk kindergartners without subsequent dyslexia. We also examined

the specificity of these preliteracy skills. Although these skills were also

related to later arithmetic, associations with later reading were stronger.

Finally, we propose an intergenerational multiple deficit model in which

both parents pass on cognitive risks.


Preschool risk factors for dyslexia

89

4

4.1 Introduction

Prospective studies in which children are followed from the preschool years

can provide valuable insights into the characteristics of children who go on to

develop dyslexia. however, it is still an open question whether the precursors of

dyslexia specifically predict dyslexia and reading development or whether they

additionally predict dyscalculia and arithmetic development. Furthermore, less

is known about the characteristics of the families in which children who go on

to develop dyslexia are born. Knowledge about risk factors of dyslexia sheds

light on the causal pathways that ultimately lead to dyslexia. These insights

might also be useful for tracing children who are at heightened risk of reading

failure and need timely support. The current study investigates plausible risk

factors for dyslexia present before reading instruction, both at the level of

cognitive skills of the child, as well as at the level of family characteristics.

4.1.1 Risk Factors in FamiliesPart of the prospective studies included children with an increased risk

of dyslexia because they are born into families with a history of dyslexia.

Depending on the definition of dyslexia, 33 to 66% of the children at familial

risk have been found to develop dyslexia (Elbro, Borstrøm, & Petersen, 1998;

Pennington & Lefly, 2001; Scarborough, 1990; Snowling, Gallagher, & Frith, 2003;

Torppa, Lyytinen, Erskine, Eklund, & Lyytinen, 2010; van Bergen et al., 2011).

The fact that the prevalence of dyslexia is consistently found to be higher in

children with a familial history of dyslexia compared to children without such

a history is in accordance with a difference between these samples in liability

or risk for dyslexia. Nevertheless, in such a design liability to dyslexia is in fact

dichotomized, which implies that implicitly the assumption has been made

that at-risk children with and without dyslexia have equal liabilities. Given

that dyslexia is influenced by a wide range of environmental and genetic risk

factors (Pennington, 2006), its underlying liability distribution, however, must

be continuously distributed.

In three recent papers (Torppa, Eklund, van Bergen, & Lyytinen, 2011;

van Bergen et al., 2011; van Bergen, de Jong, Plakas, maassen, & van der Leij,

2012) the equal-liability assumption has been tested by comparing the two

at-risk groups on the reading and reading-related skills of the parent with

dyslexia. Evidence was found against the equal-liability assumption, because


Chapter 4

90

the parents of the affected children were more severely dyslexic than those

of the unaffected children. On the other hand, information regarding possible

differences between the groups in reading skills of the spouse, the parent

without dyslexia, is as yet lacking. We will present data on parents’ self-reported

reading and spelling skills, a valid indicator of tested skills (Snowling, Dawes,

Nash, & hulme, 2012). We expected the unaffected parents of the affected

children to report more literacy difficulties than those of the unaffected

children, mirroring the findings in affected parents.

Differences between the two risk groups in parental reading skills

could suggest that these two groups of children differ in genetic predisposition.

however, the alternative explanation is that weaker reading parents offer their

child a less advantageous literacy environment. Such differences in literacy

stimulation might have been the primary cause of children’s differences in later

reading success. We will investigate several characteristics of the home literacy

environment when children were 3½ years old, and test whether differences

are related to children’s reading status five years later. The combination of

findings on parental literacy skills and home literacy environment sheds light

on whether the intergenerational transmission of risks is predominantly via

genetic or environmental pathways.

4.1.2 Risk Factors in ChildrenAlongside characteristics of families that might indicate heightened risk for

dyslexia, we investigated characteristics of kindergartners that could signify

increased risk. more specifically, we examined whether differences between

at-risk dyslexic, at-risk non-dyslexic, and control children are only present after

some years of reading instruction, or whether the groups already demonstrate

differences before the start of reading instruction. By comparing at-risk children

who go on to develop dyslexia with controls retrospectively, it has been shown

that phonological awareness, rapid naming, and letter knowledge are among

the most important precursors of dyslexia. Familial-risk studies consistently

found that children with dyslexia were impaired across these skills during

the preschool years (Elbro, Borstrøm, & Petersen, 1998; Pennington & Lefly,

2001; Scarborough, 1990; Snowling, Gallagher, & Frith, 2003; Torppa, Lyytinen,

Erskine, Eklund, & Lyytinen, 2010; van Bergen et al., 2011).

In addition, familial-risk studies offer the interesting possibility to

compare the at-risk children classified as non-dyslexic with their peers without



91

4

familial risk (controls). Although these at-risk children do not meet dyslexia

criteria, they typically perform less well than controls on reading and spelling

tasks after a few years of reading instruction (Boets et al., 2010; Pennington &

Lefly, 2001; Snowling et al., 2003; van Bergen et al., 2011; van Bergen et al., 2012)

(but see Torppa et al., 2010 for an exception). These findings provide further

support for the continuity of familial risk (Pennington & Lefly, 2001; Snowling et

al., 2003; van Bergen et al., 2012), and are consistent with a multifactorial model

of the aetiology of dyslexia (Pennington, 2006).

The somewhat lower literacy skills of the at-risk children without

dyslexia raise the question whether, before the start of reading instruction, they

also exhibit mild deficiencies in the cognitive skills underpinning reading, or

whether they start off first grade performing as well as controls but experience

slower development in reading skills. In previous familial-risk studies the

performance of the at-risk non-dyslexics was not significantly weaker on

phonological awareness, rapid naming, and letter knowledge in kindergarten

(Boets et al., 2010; Elbro et al., 1998; Pennington & Lefly, 2001; Snowling et al.,

2003; Torppa et al., 2010; van Bergen et al., 2011); the only exception being

letter knowledge in the study of Elbro and colleagues. Based on the mentioned

studies, with no clear differences between at-risk children without dyslexia

and controls, we would expect to find no differences either. Yet we might find

differences for two reasons. First, in previous studies the at-risk non-dyslexics

performed similar or even worse, though not significantly, than controls.

As we have a larger sample in the present study, subtle differences might

reach significance. Second, in an earlier paper about the same sample as the

current paper we reported mild difficulties at the end of Grade 2 of the at-risk

children without dyslexia on literacy and phonological awareness, but not on

rapid naming (van Bergen et al., 2012). Accordingly, we expected to find mild

difficulties in kindergarten on letter knowledge and phonological awareness,

but not on rapid naming.

4.1.3 Specificity of PrecursorsAn additional aim of the current study was to consider the specificity of known

precursors for dyslexia. Comorbidity rates of dyslexia and dyscalculia are higher

than expected by chance (Landerl & moll, 2010). Accordingly, it is important

to examine whether the above mentioned trio of preliteracy skills are also

predictive of arithmetic skills. Put differently, are the precursors of reading


Chapter 4

92

specifically related to basic reading skills or are they also precursors of basic

arithmetic skills? According to De Smedt and colleagues (Boets & De Smedt,

2010; De Smedt, Taylor, Archibald, & Ansari, 2010), reading and arithmetic

ability are concurrently associated because word reading and arithmetic

fact retrieval both depend upon the quality of phonological representations

in long-term memory. If so, we should find a longitudinal relation between

phonological awareness and arithmetic. Building on this hypothesis, the speed

of retrieval of phonological representations form long-term memory (as tapped

by rapid naming) should be related to subsequent mental arithmetic skills,

since the latter requires quick retrieval of arithmetic facts. Inefficient retrieval

of phonologically coded arithmetic facts leaves limited memory resources for

selecting and carrying out appropriate procedures (hecht, Torgesen, Wagner,

& Rashotte, 2001).

The current paper will report new data of an ongoing longitudinal

study (see e.g. van Bergen et al., 2012). Parental literacy and home literacy

environment when children were 3½ years old and children’s cognitive skills

in kindergarten will be related to their reading and arithmetic skills in Grade 3.

Adding to previous studies, in the current study we also examined literacy skills

of the non-dyslexic parent and the specificity of known precursors of reading.

4.2 Method

4.2.1 ParticipantsThree groups of children of the Dutch Dyslexia Programme (see van Bergen et

al., 2012) were involved in this study. Children who had data present at ages

3½ (literacy questionnaire) and/or 6 (kindergarten), as well as at age 9 (third

grade) were eligible for inclusion in this study (N = 196). The at-risk dyslexic

group consisted of 50 dyslexic children at high familial risk for dyslexia. The

at-risk non-dyslexic group comprised 82 children at high familial risk but

without dyslexia. Finally, the control group included 64 children at low familial

risk and without dyslexia. Six children at low familial risk were categorized as

dyslexic and were omitted from the study. All at-risk children had at least one

parent and one close relative with dyslexia. In 13 cases (9.8%) both parents had

dyslexia. On average, the scores of the dyslexic parents (i.e. weakest reading

parent) belonged to the bottom five percent on reading fluency. A detailed



93

4

description of the assessment of familial risk is given in van Bergen et al. (2012).

Assessment of dyslexia in the children was done in Grade 3. Children were

considered to have dyslexia when their score on the word-reading fluency task

(described below) corresponded to the weakest ten percent in the population

(equivalent to a Wechsler score of ≤6.2 norm scores taken from van den Bos,

lutje Spelberg, Scheepstra, & de Vries, 1994). The Grade 3 assessment took

place between January and may. Cut-off scores were adjusted according to the

month of assessment.

Table 4.1 shows the characteristics of the three groups. Percentages of

boys were similar across groups. The at-risk dyslexic children had lower IQs than

the non-dyslexic children. The groups did not differ in age at the questionnaire

administration. The control children were two months younger at the

kindergarten and Grade 3 sessions, but the groups did not differ in the number

of months of reading instruction when seen in third grade. Kindergarten

encompasses two years in the Netherlands. The kindergarten assessment took

place between April and August of the second year, when the children were

between 67 and 79 months old. Four children were kept down in Grade 1 or 2.

Like the other children, they were seen for the Grade 3 assessment after almost

three years of reading instruction, when these four children attended Grade 2.

Finally, parental education (averaged over both parents; scale 1-5) was lower in

the at-risk groups than in the control group.

Table 4.1 Group Characteristics

At-risk

Dyslexic Non-dyslexic Control p

Sample size 50 82 64No. (%) of boys 30

a (60%) 45

a (55%) 39

a (61%) .728

Full-scale IQ (at age 4) 105.37a (9.90) 111.08

b (10.75) 113.24

b (10.53) <.001

Age in months at assessments

Questionnaire 41.75a (2.03) 42.52

a (3.75) 41.08

a (2.73) .214

Kindergarten 73.02ab

(3.31) 73.29a (3.08) 71.85

b (3.11) .023

Grade 3 107.96a (4.65) 107.46

ab (3.99) 105.83

b (3.97) .014

No of months reading at assessment

Grade 3* 26.18a (1.00) 26.23

a (0.89) 25.92

a (1.12) .158

Parental education 3.41a (0.72) 3.42

a (0.77) 4.16

b (0.74) <.001

Note. The group means are given with standard deviations in parentheses. Numbers and means in the same row that do not share subscripts differ at p < .05 on the χ2-test or Tukey’s test. No. = number; * = 10 months instruction per year.


Chapter 4

94

4.2.2 MeasuresWhen children were 3½ years old, parents were asked to fill out a questionnaire

about their own literacy and the home literacy environment. Subsequently, the

children were tested on preliteracy skills at the end of kindergarten (at age 6)

and on school achievement in Grade 3 (at age 9).

Literacy Questionnaire at Age 3½

Parental literacy. Both fathers and mothers were asked about their

print exposure and literacy difficulties. Three questions were related to print

exposure: how many hours per week do you spend on average on reading for

1) work/study and 2) leisure? and 3) how many hours per week do you spend

on average on writing (e-mails, letters, postcards, diary etc.)? Each was scored

on a scale ranging from 1 (less than one hour a week) to 5 (more than 10 hours

a week). The print-exposure measure (range 3-15, Chronbach’s α .62) was the

sum of the scores for these items. Furthermore, three questions (on a scale of

1 to 3) concerned literacy difficulties: 1) Do you think you are a fast, average or

slow reader? 2) Do you have trouble following the subtitles on TV? and 3) Do

you think you have more, average or less difficulties with spelling than other

people? The sum of the scores for these items formed the literacy-difficulties

measure (range 3-9).

For 78 fathers and 70 mothers we had scores on both the literacy-

difficulties measure and reading-fluency tests of words and nonwords, which

allowed us to investigate criterion validity. The reading-fluency tests (see van

Bergen et al., 2012, for descriptions) were assessed around the time of their child’s

birth. The correlation between the literacy-difficulties measure and a composite

of word- and nonword-reading fluency was -.84 for fathers and -.85 for mothers.

Home Literacy Environment. The literacy questionnaire also contained

questions regarding home literacy environment. Both fathers and mothers were

asked to indicate the frequency of storybook reading in a typical week on a scale

from 1 (never) to 5 (more than five times a week) and whether they had a magazine

or newspaper subscription. Furthermore, parents were asked to estimate the

number of books available in the home (1 = fewer than 20 to 5 = more than

150). Finally, part of an existing questionnaire was used to measure cognitive

stimulation (Leseman, 1994; Sigel, 1982) with statements like ‘I ask my child

questions during storybook reading’. There were six statements (Chronbach’s α

.54), rated on a scale ranging from 1 (strongly disagree) to 6 (strongly agree).



95

4

Preliteracy at Age 6

Rapid naming. Serial rapid naming (van den Bos, 2003) consisted of

50 randomly ordered patches of colours (black, yellow, red, green, and blue)

arranged in five columns of 10 symbols each. Before test administration,

children practiced by naming the last column. Children were instructed to

name the colours column-wise as quickly as possible. The time to completion

was transformed to number of colours per second to normalize the score

distribution. The reliability for 6-year-olds is .80.

Phonological awareness. Two tests measured phonological awareness:

phoneme blending and phoneme segmentation. In phoneme blending

(Verhoeven, 1993a) the child was required to blend aurally presented phonemes

into a word. For example, children listened to the successive phonemes /r/ /u/

/p/ /s/, after which they had to merge this into rups [caterpillar]. Phoneme

segmentation (Verhoeven, 1993b) was the reverse of phoneme-blending. Now

the child had to segment a given word into its constituent phonemes. Both

phoneme blending and phoneme segmentation began with three practice

trials (with feedback). Test items consisted of 20 monosyllabic words per test,

increasing from two to five phonemes, with four to six items for each specific

number of phonemes. The tests were stopped when all items with the same

number of phonemes were failed. Chronbach’s α for both tests is above .85

(Verhoeven, 2000).

Letter knowledge. Both receptive and productive letter knowledge

were assessed. The receptive test (Verhoeven, 2002) required the child to

point from six alternative lowercase letters to the letter that matched a given

sound. For instance, the child was asked “Where do you see the /m/ of mooi

(beautiful)?” The knowledge of 32 graphemes (including digraphs) was tested.

Chronbach’s α is .88 (Eleveld, 2005). In the productive knowledge test, children

were asked to provide the sound of 34 graphemes (including digraphs), but

letter names were also considered correct (Verhoeven, 1993a). The randomly

ordered graphemes were printed in lowercase in two columns of 17 items

each. Chronbach’s α is above .85 (Verhoeven, 2000).

School Achievement at Age 9

Reading. To assess word-reading fluency, children were given the

One-minute-Test (Brus & Voeten, 1972), which consists of a list of 116 words

of increasing difficulty. They were asked to read as many words as possible

correctly within one minute. The reliability is .90 for Grade 3.


Chapter 4

96

Arithmetic. Two subtests of the arithmetic tempo test (de Vos, 1992)

were administered: addition and subtraction. Each paper-and-pencil subtest

includes 40 problems of increasing difficulty. Per problem two operands have

to be added or subtracted (e.g. 13 + 4 = …). All operands and outcomes are

below 100. The number of correctly solved problems within one minute forms

the raw score. A total score was computed as the sum of the standard scores

over both subtests. The reliability of each subtest is .93 (Stock, Desoete, &

Roeyers, 2010).

4.3 Results

About half of the participants returned the questionnaire that was sent by mail.

however, percentage of data present was nicely distributed among the groups

(at-risk dyslexic: 56%, at-risk non-dyslexic: 51%, and controls: 50%). missing-

value analyses indicated that the subsamples with and without missing

questionnaire data did not differ significantly on parental education (t(189.9) =

-0.57, p = .572), word-reading fluency of the weakest-reading parent (t(192.2)

= -0.17, p = .865), or word-reading fluency of the child (t(193.0) = -1.52, p

= .131). hence, further analyses of the questionnaire data were deemed

appropriate. The preliteracy data were present for 96% of the children. Grade

3 data were complete. One outlier on rapid naming in the control group was

removed because the score was 3.4 standard deviations above the control

group’s mean.

Differences among the groups on continuous measures were

evaluated using one-way ANOVAs, followed by pairwise comparisons with

Tukey’s correction for multiple testing. Group means, group effects, and

effects sizes are presented in Tables 4.2, 4.3, and 4.4. Results are described

below for family and child characteristics in relation to the three groups, and

subsequently, results concerning the prediction of children’s reading skills are

given.

4.3.1 Family Characteristics in Relation to Children’s GroupParental print exposure and literacy difficulties in relation to children’s group

can be found in Table 4.2. The fathers of the control children spent significantly

more time on reading and writing than those of the at-risk children.



97

4

Tab

le 4

.2 P

aren

tal L

itera

cy in

Rel

atio

n to

Chi

ldre

n’s

Gro

up

At-

risk

Effec

t siz

e(C

ohen

’s d)

Dys

lexi

cN

on-d

ysle

xic

Con

trol

Mea

sure

MSD

MSD

MSD

Fdf

FRD

vs.

FRN

DFR

D v

s. C

FRN

D v

s. C

Prin

t exp

osur

e

Fath

er7.

81a

3.36

7.53

a 3.

089.

81b

3.06

4.99

(2, 9

2)-0

.09

0.65

0.75

Mot

her

7.96

a 2.

798.

10a

3.00

9.03

a 2.

871.

24(2

, 95)

0.05

0.37

0.32

Dys

lexi

c p

aren

t7.

56a

2.98

7.08

a 3.

08<

1(1

, 64)

-0.1

6

Non

-dys

lexi

c p

aren

t8.

23a

3.15

8.56

a 2.

84<

1(1

, 63)

0.11

Lite

racy

diffi

cult

ies

Fath

er6.

68a

1.52

6.20

a 1.

833.

91b

0.93

29.9

1(2

, 94)

-0.5

2-2

.98

-2.4

6

Mot

her

6.00

a 1.

945.

10a

2.07

3.84

b

1.17

10.8

3(2

, 98)

-0.7

7-1

.85

-1.0

8

Dys

lexi

c p

aren

t7.

56a

0.96

7.24

a 1.

111.

39(1

, 64)

-0.3

0

Non

-dys

lexi

c p

aren

t5.

19a

1.57

4.02

b

1.31

10.8

8(1

, 66)

-1.1

1

Not

e. N

umb

ers

and

mea

ns in

the

sam

e ro

w t

hat

do n

ot s

hare

sub

scrip

ts d

iffer

at

p <

.05

on T

ukey

’s te

st. C

ohen

’s d

is c

alcu

late

d us

ing

the

SDs

of t

he c

ontr

ols.

FRD

=

fam

ilial

-ris

k dy

slex

ic; F

RND

= fa

mili

al-r

isk

non-

dysl

exic

; C =

con

trol

.


Chapter 4

98

Tab

le 4

.3 H

ome

Lite

racy

Env

ironm

ent i

n Re

latio

n to

Chi

ldre

n’s

Gro

up

At-

risk

Effec

t siz

e(C

ohen

’s d)

Dys

lexi

cN

on-d

ysle

xic

Con

trol

Mea

sure

MSD

MSD

MSD

Fdf

FRD

vs.

FRN

DFR

D v

s. C

FRN

D v

s. C

Book

s at

hom

e*4.

04a

1.14

4.07

a1.

114.

62a

0.75

Shar

ed re

adin

g

Fath

er3.

20a

1.26

3.29

a1.

293.

50a

1.91

< 1

(2, 9

5)0.

050.

160.

11

Mot

her

4.23

a0.

994.

48a

0.77

4.19

a1.

001.

09(2

, 97)

0.25

-0.0

4-0

.29

Dys

lexi

c p

aren

t3.

70a

1.27

3.67

a1.

32<

1(1

, 67)

-0.0

2

Non

-dys

lexi

c p

aren

t3.

75a

1.23

4.12

a1.

051.

67(1

, 63)

0.25

Cog

nitiv

e st

imul

atio

n29

.01 a

4.48

29.7

9 a3.

3529

.28 a

3.45

< 1

(2, 9

9)0.

230.

08-0

.15

Not

e. N

umb

ers

and

mea

ns in

the

sam

e ro

w t

hat

do n

ot s

hare

sub

scrip

ts d

iffer

at

p <

.05

on T

ukey

’s te

st. C

ohen

’s d

is c

alcu

late

d us

ing

the

SDs

of t

he c

ontr

ols.

FRD

=

fam

ilial

-ris

k dy

slex

ic; F

RND

= fa

mili

al-r

isk

non-

dysl

exic

; C =

con

trol

; * =

test

ed n

on-p

aram

etric

ally

and

eff

ect s

izes

not

giv

en, s

ee te

xt.



99

4

Tab

le 4

.4 C

hild

ren’

s Pr

elite

racy

Ski

lls a

t 6 Y

ears

and

Sch

ool A

chie

vem

ent a

t 9 Y

ears

At-

risk

Effec

t siz

e(C

ohen

’s d)

Dys

lexi

cN

on-d

ysle

xic

Con

trol

Mea

sure

MSD

MSD

MSD

Fdf

FRD

vs.

FRN

DFR

D v

s. C

FRN

D v

s. C

Prel

itera

cy s

kills

(6yr

)

Rap

id n

amin

g co

lour

s0.

58a

0.16

0.72

b0.

180.

75b

0.18

14.1

9(2

, 182

)0.

790.

960.

17

Phon

olog

ical

aw

aren

ess

Blen

ding

7.57

a5.

7011

.56 b

6.25

14.4

5 c5.

1419

.51

(2, 1

86)

0.78

1.34

0.56

Segm

enta

tion

5.08

a4.

979.

46b

6.67

12.3

2 c6.

0919

.49

(2, 1

86)

0.72

1.19

0.47

Lett

er k

now

ledg

e

Rece

ptiv

e13

.61 a

6.58

21.2

7 b6.

5922

.73 b

7.20

27.9

1(2

, 186

)1.

061.

270.

20

Prod

uctiv

e9.

02a

6.28

17.6

6 b8.

1318

.50 b

8.22

24.8

9(2

, 185

)1.

051.

150.

10

Scho

ol a

chie

vem

ent (

9yr)

Read

ing

28.6

4 a8.

2154

.72 b

12.1

061

.53 c

12.2

412

9.98

(2, 1

93)

2.13

2.69

0.56

Arit

hmet

ic-1

.35 a

1.61

0.15

b1.

790.

90c

1.75

24.2

2(2

, 193

)0.

861.

290.

43

Not

e. N

umb

ers

and

mea

ns in

the

sam

e ro

w t

hat

do n

ot s

hare

sub

scrip

ts d

iffer

at

p <

.05

on T

ukey

’s te

st. C

ohen

’s d

is c

alcu

late

d us

ing

the

SDs

of t

he c

ontr

ols.

FRD

=

fam

ilial

-ris

k dy

slex

ic; F

RND

= fa

mili

al-r

isk

non-

dysl

exic

; C =

con

trol

.


Chapter 4

100

Although maternal print exposure seemed not to be related to children’s group,

a multivariate analyses (with data of both parents analysed simultaneously) did

not show a significant parent by group interaction, F(2, 90) = 1.63, p = .202, but

only an effect of group, F(2, 90) = 3.73, p = .028. Next, in the at-risk sample we

subdivided parents according to their reading status (dyslexic vs. non-dyslexic1).

In this sample, parental print exposure was unrelated to reading outcome of at-

risk children. Non-dyslexic parents (M = 8.43, SD = 2.95) appeared to read and

write approximately equally frequent as the control parents (M = 9.42, SD =

2.62; t(94) = 1.59, p = .115).

Concerning literacy difficulties, parents in the control group reported

less problems with reading and spelling, as was expected given the selection

criteria. Interestingly, within the at-risk group parental literacy difficulties

seemed to differentiate children with and without dyslexia. This difference was

large and significant (Cohen’s d = -1.11, p = .002) for the non-dyslexic parent.

The interaction between parental status (dyslexic vs. non-dyslexic) and child

outcome (dyslexic vs. non-dyslexic) approached significance, F(1, 63) = 3.85,

p = .054. The non-dyslexic parents of the non-dyslexic children (M = 4.02, SD =

1.31) reported similar levels of literacy as the control parents (M = 3.88, SD =

0.67; t(71) = 0.59, p = .560).

home literacy environment was measured at 3½ years by whether

parents had a subscription to a newspaper or magazine, how many books

they had at home, and how often they read to their child. The percentages

of fathers/mothers with a subscription were 68%/68% in the at-risk dyslexic

group, 73%/76% in the at-risk non-dyslexic group, and 100%/91% in the

control group. The differences were significant for fathers, suggesting more

subscriptions in control families, χ2(2, N=101) = 11.86, p = .003, but not for

mothers, χ2(2, N=101) = 4.82, p = .090. The variable about number of books

in the home (Table 4.3) was strongly skewed and therefore analysed with the

nonparametric Kruskal-Wallis test. Overall, the difference between groups was

significant, K(2, N=102) = 7.01, p = .030, suggesting more books in the homes

of control families, but pairwise comparisons with adjusted p-values missed

significance. The frequency of storybook reading was virtually the same over

groups; the only difference being that mothers read more than fathers. Parents

provided similar levels of cognitive stimulation.

1 In 9 out of the 70 at-risk children with questionnaire data present, both parents had dyslexia. ‘Dyslexic parent’ refers to the weakest reading parent and ‘non-dyslexic’ to the other parent.



101

4

4.3.2 Children’s Characteristics in Relation to Children’s GroupGroup means on precursors of reading at the end of kindergarten are shown

in Table 4.4. All group effects on the preliteracy tasks were highly significant

(ps < .001). The at-risk dyslexic group was slower on rapid naming and knew

fewer letters compared to the two non-dyslexic groups, which were statistically

indistinguishable. however, the group means on phonological awareness

showed a stepwise pattern, with the at-risk dyslexic children performing

lowest, followed by the at-risk non-dyslexic, and thereafter by the control

children. Group differences were not attributable to the group differences on

parental education (Table 4.1), as analyses with parental education as covariate

(ANCOVA for rapid naming and mANCOVAs for phonological awareness and

letter knowledge, with each two indicators) also yielded highly significant group

effects (ps < .001) and insignificant covariate effects (ps < .333). Controlling for

IQ differences (Table 4.1) in (m)ANCOVAs as above showed significant effects

of IQ (ps < .001), but all group effects remained highly significant (ps < .001).

As can be seen in Table 4.4, groups did not only differ on reading but

also on arithmetic ability. When applying a similar criterion for dyscalculia

(≤10, or Wechsler scale score ≤6.2, norm scores taken from melis, 2002) as for

dyslexia, the percentages of children identified with dyscalculia were found

to differ significantly: 42% (21/50) in the at-risk dyslexic group, 20% (16/82)

in the at-risk non-dyslexic group, and 8% (5/64) in the control group. These

differences were confirmed by a chi-square test, χ2(2, N=196) = 19.79, p < .001.

4.3.3 Prediction of Children’s Reading Skills After confirming comorbidity, we examined in the full sample how much of the

variance in school achievement can be explained by the preliteracy skills, and

whether these skills are specifically related to reading or arithmetic (see for the

correlations Table 4.5). Results of regression analyses with all predictors included

are presented in Table 4.6. The preliteracy skills together explained 43.3% of

the variance in reading; letter knowledge (akin to an autoregressor) made the

strongest contribution. After controlling for arithmetic ability, the preliteracy

skills explained an additional 21.4% of the variance, confirming the specific

relation between preliteracy and literacy skills. With respect to arithmetic, the

preliteracy skills explained 22.5% of the variance, and even made a small but

significant unique contribution to arithmetic (ΔR2 = 3.2%, p = .039) after reading

was controlled. Phonological awareness did not explain variance in arithmetic


Chapter 4

102

above that accounted for by rapid naming and letter knowledge, as shown by

its insignificant β. however, when letter knowledge (which correlated .73 with

phonological awareness) was excluded, rapid naming (β = .33, p < .001) as well

as phonological awareness (β = .19, p = .008) were significant.

Table 4.5 Correlations between Children’s Preliteracy Skills at 6 Years and School Achievement at 9 Years

School achievement (9yr)

Reading Arithmetic

Preliteracy skills (6yr)

Rapid naming colours .49 .41

Phonological awareness

Blending .49 .31

Segmentation .51 .34

Letter knowledge

Receptive .57 .41

Productive .57 .41

Note. All ps < .001.

Table 4.6 β-weights and Total R2 in Regression models Predicting School Achievement at 9 Years from Preliteracy Skills at 6 Years

Reading Arithmetic

Predictor Without arithmetic With arithmetic Without reading With reading

Controlling variable

Arithmetic - .30*** - -

Reading - - - .42***

Preliteracy skills (6yr)

Rapid naming colours .26*** .17** .29*** .18*

Phonological awareness a .17* .16* .03 -.04

Letter knowledge a .36*** .29*** .24* .09

Total R2 .43*** .51*** .23*** .32***

Note. a Summed standardized scores of the two measures. *p<.05; **p<.01; ***p<.001.



103

4

Lastly, we examined the utility of parental literacy difficulties in

predicting children’s reading skills. Therefore, three regression analyses were

conducted on the complete sample, followed by three in the at-risk sample.

In the complete sample (n = 97, Table 4.7) the literacy difficulties of fathers

and mothers together explained 21.3% of the variance in children’s word-

reading fluency. After controlling for parental education, literacy difficulties

of fathers and mothers explained an additional 15.3% of the variance. Even

after accounting for parental education and all preliteracy skills of the children,

the literacy difficulties of fathers and mothers made a unique contribution,

explaining 4.3% additional variance in children’s reading fluency.

Table 4.7 Regression models Predicting Children’s Reading at 9 Years from Parental Literacy difficulties in the Complete Sample

At-risk + Control Sample

Predictor ΔR2 β

model 1

Step 1 .213***

Father’s literacy -.32***

mother’s literacy -.37***

model 2

Step 1 .065*

Parental education .26*

Step 2 .153***

Father’s literacy -.30**

mother’s literacy -.34***

model 3

Step 1 .540***

Parental education .25**

Children’s preliteracy skills

Rapid naming .34***

Phonological awareness a .21

Letter knowledge a .30*

Step 2 .043*

Father’s literacy -.20**

mother’s literacy -.16



Chapter 4

104

In the at-risk sample (n = 65, Table 4.8) we again subdivided parents

according to reading status. We anticipated that knowledge of the self-

reported literacy difficulties of the dyslexic parent would not add much

over and above the fact that they have dyslexia. This is supported by the

significantly lower SDs of the dyslexic parents (Table 4.2; Levene’s test: F(1, 132)

= 9.15, p = .003). Therefore, in subsequent analyses literacy difficulties of the

dyslexic parent were entered before those of the non-dyslexic parent, allowing

to test whether literacy difficulties of the non-dyslexic parent are related to

their offspring’s reading skills, controlling for literacy difficulties of the dyslexic

parent. It appeared that literacy difficulties of the dyslexic parent indeed

did not explain a significant amount of variance in at-risk children’s reading

fluency, but differences in literacy difficulties of the non-dyslexic parent did

explain additional variance. Over and above parental education, and literacy

difficulties of the dyslexic parent, the literacy difficulties of the non-dyslexic

parent accounted for an additional 11.5% in children’s reading fluency. In a

final analysis, differences in parental education and children’s preliteracy skills

together explained an impressive 62.8% of the variance in the reading fluency

of at-risk children, but parental literacy difficulties did not significantly add to

this prediction.

To illustrate the effect of the non-dyslexic parent on their offspring’s

risk for dyslexia, we dichotomised the literacy-difficulty measure: ‘non-

dyslexic’ parents scoring ≥62 were categorized as having literacy difficulty. It

appeared that within the group of at-risk children with a non-dyslexic parent

without literacy difficulties 30% (16/54) developed dyslexia, compared to 79%

(11/14) in the group of at-risk children with a ‘non-dyslexic’ parent with literacy

difficulties.

4.4 Discussion

The current paper studied putative risk factors for dyslexia at the level of

literacy behaviours and skills of parents, and at the level of cognitive skills of

kindergartners. This was done utilizing data of an ongoing longitudinal study

in which children with and without familial risk are followed from infancy.

2 This cut-off score was chosen because parents scoring ≥6 indicated to have difficulties on at least one of the three questions. moreover, the score distribution showed a jump between score 5 and 6.



105

4

Table 4.8 Regression models Predicting Children’s Reading at 9 Years from Parental Literacy difficulties in the At-risk Sample

At-risk Sample

Predictor ΔR2 β

model 1

Step 1 .023

Dyslexic parent’s literacy -.15

Step 2 .135**

Non-dyslexic parent’s literacy -.37**

model 2

Step 1 .026

Parental education .16

Step 2 .017


Step 3 .115**

Non-dyslexic parent’s literacy -.37**

model 3

Step 1 .628***

Parental education .18*

Children’s preliteracy skills

Rapid naming .29**

Phonological awareness a .13

Letter knowledge a .50***

Step 2 .020


Step 3 .003

Non-dyslexic parent’s literacy -.06



Chapter 4

106

4.4.1 Intergenerational TransferAn important issue addressed was the intergenerational transfer of reading

skills. We measured parental reading and spelling difficulties using a rating

scale (assessed when children were 3½), which demonstrated high validity:

self-reported and tested reading performance in a subsample correlated

.84-.85. First, we investigated within the at-risk sample the effect of the non-

dyslexic parent, whose contribution has been neglected in previous work.

We found that self-reported literacy difficulties (i.e., reading and spelling

difficulties) of the non-dyslexic parent differentiated at-risk children with

and without dyslexia. moreover, in regression analyses parental literacy

difficulties predicted at-risk children’s reading fluency, even after accounting

for differences in literacy difficulties of the dyslexic parent and the educational

level of both parents. This supports the view that children who go on to

develop dyslexia have a higher genetic predisposition towards dyslexia and

that parental skills are indicative of their children’s liability (van Bergen et al.,

2012). Second, the literacy difficulties of the dyslexic parent, conversely, did

not differentiate at-risk children with and without dyslexia and did not explain

variance in children’s reading. Put differently, their self-reported difficulties did

not add to the prediction of children’s reading beyond the fact that they have

dyslexia. however, in an earlier paper about this sample (van Bergen et al.,

2012) we showed that individual differences in an objective measure of word-

reading fluency did differentiate at-risk children with and without dyslexia and

did explain variance in children’s word-reading fluency. The group difference

on this objective measure was somewhat larger than on the self-reported

measure (Cohen’s d = 0.48 vs. 0.30, respectively), presumably reflecting that

measuring skills is still more reliable than asking about skills (Snowling et al.,

2012). Third, intergenerational transfer of literacy skills was also observed in

the at-risk and control samples combined: both maternal and paternal self-

reported literacy difficulties predicted children’s reading, even after accounting

for parental education and children’s preliteracy skills.

Only recently an interest has emerged in the predictive value of

parental skills for the development of children’s reading skills. Previous familial-

risk studies have found associations between reading accuracy, reading fluency,

spelling, phonological skills, and vocabulary of the dyslexic parent and their

offspring’s reading outcome (Torppa et al., 2011; van Bergen et al., 2011; van

Bergen et al., 2012). most importantly, our findings are the first to demonstrate



107

4

the importance of the literacy skills of the parent without dyslexia. They

predicted children’s reading beyond the effect of the dyslexic parent. Within

the children with a dyslexic parent, the risk for dyslexia was 2½ times higher if

the other parent had literacy difficulties as well.

4.4.2 Home Literacy EnvironmentInvestigation of environmental influences did not indicate links between

children’s home literacy environment and children’s reading outcome. Before

school entry, the home environment of at-risk children with and without

dyslexia did not differ in terms of shared reading, cognitive stimulation, how

much parents read and write, newspaper subscriptions, and number of books,

although control families tended to have more newspapers and books. Likewise,

in previous familial risk studies no group differences have been found in shared

reading, neither in children’s access to print, like library membership or number

of books at home (Elbro et al., 1998; Torppa et al., 2007; van Bergen et al., 2011).

Although many educators and parents have strong beliefs about the impact

of story book reading on later reading success, research repeatedly fails to find

such an association (see for a review Sénéchal & Young, 2008), which is in line

with the present findings. The absence of a relation between preschool home

literacy environment and subsequent reading attainment found in the current

and previous familial-risk studies is in agreement with behavioural genetic

studies showing low levels of shared environment influence, whereas up to

80% is due to genetic influences (e.g., Byrne et al., 2009; haworth et al., 2009).

however, the possibility remains that environmental influences in familial-

risk studies are downplayed because they are not measured thoroughly, and

because familial-risk studies are based on voluntary samples and therefore

might include a somewhat restricted range of environments.

4.4.3 Preliteracy Skills and their SpecificityIn line with the literature, in kindergarten the children who went on to develop

dyslexia (assessed in Grade 3) were impaired on rapid naming, phonological

awareness, and letter knowledge. Interestingly, the at-risk children without

later dyslexia had age-adequate rapid naming and letter knowledge, but

were mildly impaired on phonological awareness. Despite adequate letter

knowledge and rapid naming before the start of formal reading, these children

read less fluently in Grade 2 (van Bergen et al., 2012) and Grade 3 (Table 4.4).


Chapter 4

108

The different results for letter knowledge and phonological awareness were

unexpected given the often observed reciprocal development between these

skills (e.g., de Jong, 2007; Wagner, Torgesen, & Rashotte, 1994). however,

teaching of letters by parents affects children’s letter knowledge (Torppa,

Poikkeus, Laakso, Eklund, & Lyytinen, 2006). The dyslexic families in our study

are possibly aware of their children’s risk for dyslexia and might have paid extra

attention to teaching letters, which might explain the good letter knowledge

of the at-risk non-dyslexic children. Alternatively, the learning capacity of these

children is less affected and more similar to that of control children, making

them more likely to pick up letter-sound knowledge in their environment.

The finding of weak phonological awareness in combination with

good rapid naming of the at-risk non-dyslexics in kindergarten mirrors the

pattern found in this group at the end of second grade (van Bergen et al., 2012),

indicating longitudinal stability. Our finding highlights that phonological

awareness and rapid naming have partially unique influences on reading

ability and disability, as proposed by the double-deficit theory (Wolf & Bowers,

1999). Although their correlation with later reading was remarkably similar

(.5), group differences showed different patterns for the two precursors and

the regression analyses revealed partly unique contributions to reading. In

line with this, van den Boer, de Jong, and haentjens-van meeteren (in press)

recently demonstrated differential effects of phonological awareness and

rapid naming on the developing reading system. In an experiment with words

of different lengths, they modeled the amount of serial (or letter-by-letter)

processing as indexed by the time needed to read each additional letter (the

word length effect). They disentangled the word length effect from overall

reading speed and showed that phonological awareness is associated with

the degree of serial processing, whereas rapid naming is related to overall

reading speed, irrespective of the degree of serial processing. Van den Boer

et al. argue that “poor phonological awareness results in poor build-up of

orthographic representations, and therefore in continued reliance on a serial

processing strategy”. Translating their conclusions to the current study, the at-

risk non-dyslexic group is hypothesized to have some difficulty with building

up orthographic knowledge (a requisite for sight-word reading), but not with

fast processing of orthographic knowledge and fast retrieval of phonological

codes. The at-risk dyslexic group seems to be impaired in all these elements.

Our study once more confirmed that poor rapid naming, phonological



109

4

awareness, and letter knowledge in kindergarten are cognitive risk factors for

dyslexia. With regard to arithmetic, our study showed a clear stepwise pattern

of (co)morbidity with dyscalculia. Rates of dyscalculia in the at-risk dyslexic,

at-risk non-dyslexic, and control groups were 42%, 20%, and 8%, respectively.

The demonstrated association between reading (dyslexia) and arithmetic

(dyscalculia) raises the question as to whether the three cognitive risk factors

for dyslexia are specific for reading. Our analysis controlling for arithmetic

showed that all three precursors uniquely captured variance in later reading

skills (21% in all). Visa versa, only rapid naming had a small (3%) but significant

influence on arithmetic, while holding reading constant (in agreement with de

Jong & van der Leij, 1999; and hecht et al., 2001), despite the clear associations

between the three preliteracy skills and subsequent arithmetic achievement.

These results suggest that the preliteracy trio is rather specifically related to the

academic domain of reading. Note that the measurement methods of reading

(word decoding fluency) and arithmetic (mental arithmetic fluency) closely

resemble each other. As lower overlap between word reading and broader

mathematical skills are to be expected, this strengthens the finding that the

preliteracy skills, especially letter knowledge and phonological awareness, are

better predictors of reading.

Consistent with the hypothesis that the ease of arithmetic fact retrieval is

affected by the quality of phonological representations (Boets & De Smedt, 2010;

De Smedt et al., 2010) we found longitudinal relations between phonological

processing and arithmetic. It should be noted that this hypothesis, that explains

why reading and arithmetic are associated, is not directly testable with the

current arithmetic task, since this graded task does not differentiate between

problems solved by fact retrieval and by procedural strategies (although fact

retrieval might be an element in procedural strategies as well). Rather, children

likely started the task using fact retrieval and shifted somewhere along to

using procedures. Based on the findings of De Smedt and colleagues higher

correlations with arithmetic might be expected when arithmetic was tested

using only problems with a high probability of being solved by retrieval. Since

our findings showed higher relations for rapid naming than for phonological

awareness in predicting arithmetic, future arithmetic studies should include

rapid naming as well. The importance of efficient retrieval of phonological

codes (as tapped by rapid naming) for arithmetic efficiency has been shown

before (e.g., van der Sluis, de Jong, & van der Leij, 2007) and supports the view


Chapter 4

110

that efficient retrieval of arithmetic facts leaves sufficient memory resources

for the selection and implementation of appropriate procedures (hecht et al.,

2001). Although our data showed relations between phonological processing

and subsequent arithmetic, relations were stronger for subsequent reading.

Our findings fit well in Pennington’s multiple deficit model (2006).

Pennington’s model is built on the multifactorial and probabilistic aetiology of

developmental disorders. It postulates that a cognitive developmental disorder

is the behavioural outcome of multiple interacting risk and protective factors.

Some of these factors influence several disorders (causing comorbidity) and

some are specific to one particular disorder. The investigated preliteracy skills

were good predictors of primarily subsequent reading, although they also

predicted subsequent arithmetic to some extent, especially rapid naming.

hence, deficits in phonological awareness and rapid naming can be viewed as

two of the multiple cognitive deficits that increase the probability of dyslexia.

Poor letter knowledge can be seen as a third underlying cognitive deficit, or as

an earlier developmental manifestation of dyslexia (akin to an autoregressor of

reading). Because rapid naming was predictive of later reading and arithmetic

in the current and previous studies, slow rapid naming might well be a cognitive

deficit in common to dyslexia and dyscalculia. Shared deficits can account for

the frequent co-occurrence of these developmental disorders.

The parent-child resemblance reported in the current and other

recent literature seems to imply that the multiple cognitive deficit model of

Pennington (2006) can be extended to an intergenerational multiple cognitive

deficit model. As we argued in an earlier paper (van Bergen et al., 2012), literacy

abilities of parents might be viewed as indicators of their offspring’s risk or

liability for literacy difficulties, since parents provide their offspring with their

genetic and family endowment. In Pennington’s model the focus is on a specific

individual, whereas we propose to include cognitive factors of parents. In the

intergenerational multifactorial deficit model cognitive skills of both parents

can be seen as risk and protective factors for their offspring’s predisposition

towards developmental disorders. These cognitive factors may include skills

like reading, spelling, arithmetic, and language, or may include their cognitive

underpinnings like phonological awareness, rapid-naming ability, verbal short-

term memory, number sense, and executive functioning. On each of these

continua parents occupy a position in multivariate space. The constellation of

these factors of both parents gives an indication of their offspring’s liability to



111

4

develop certain cognitive disorders. The findings from our studies suggest that

parents confer risk factors for dyslexia predominantly via genetic rather than

environmental pathways. more research investigating the cognitive risks that

parents pass on to their children is needed to further develop the proposed

intergenerational multiple deficit model. Such studies will help to contribute

to unravelling the aetiology of developmental disorders. Additionally, such

research could lead to better identifying the children at-risk for cognitive

developmental disorders and provide them timely with appropriate support.


Chapter 4

112

4.5 References


Boets, B., De Smedt, B., Cleuren, L., Vandewalle, E., Wouters, J., & Ghesquière, P. (2010). To-wards a further characterization of phono-logical and literacy problems in Dutch-spea-king children with dyslexia. British Journal of Developmental Psychology, 28, 5-31.


Byrne, B., Coventry, W. L., Olson, R. K., Samu-elsson, S., Corley, R., Willcutt, E. G., . . . De-Fries, J. C. (2009). Genetic and environmen-tal influences on aspects of literacy and language in early childhood: Continuity and change from preschool to Grade 2. Journal of Neurolinguistics, 22(3), 219-236.

de Jong, P. F. (2007). Phonological awareness and the use of phonological similarity in letter-sound learning. Journal of Experi-mental Child Psychology, 98(3), 131-152.


De Smedt, B., Taylor, J., Archibald, L., & Ansari, D. (2010). how is phonological processing related to individual diffe-rences in children’s arithmetic skills? De-velopmental Science, 13(3), 508-520. doi: 10.1111/j.1467-7687.2009.00897.x

de Vos, T. (1992). Tempo test rekenen [Arith-metic speed test]. Lisse: Swets & Zeitlinger.


Eleveld, m. A. (2005). At risk for dyslexia: The role of phonological abilities, letter know-ledge, and speed of serial naming in early intervention and diagnosis. Antwerpen/Apeldoorn: Garant.

haworth, C. m. A., Kovas, Y., harlaar, N., hayiou-Thomas, m. E., Petrill, S. A., Dale, P. S., & Plomin, R. (2009). Generalist genes and learning disabilities: A multivariate genetic analysis of low performance in reading, mathematics, language and ge-neral cognitive ability in a sample of 8000 12-year-old twins. Journal of Child Psycho-logy and Psychiatry, 50(10), 1318-1325. doi: 10.1111/j.1469-7610.2009.02114.x

hecht, S. A., Torgesen, J. K., Wagner, R. K., & Ras-hotte, C. A. (2001). The relations between phonological processing abilities and emer-ging individual differences in mathematical computation skills: A longitudinal study from second to fifth grades. Journal of Expe-rimental Child Psychology, 79(2), 192-227.


Leseman, P. P. m. (1994). Preventie van on-derwijsachterstand [Prevention of edu-cational backwardness]. In J. Rispens, P. Goudena & J. Groenendaal (Eds.), Van kindmodel naar modelkind. [From child model to model child.] (pp. 63-92). Gronin-gen: Stichting Kinderstudies.

melis, G. (2002). DLE boek [DLE book] Edufor-ce/Swets & Zeitlinger B.V.





113

4


Sénéchal, m., & Young, L. (2008). The effect of family literacy interventions on children’s acquisition of reading from kindergarten to Grade 3: A meta-analytic review. Review of Educational Research, 78(4), 880-907. doi: 10.3102/0034654308320319

Sigel, I. E. (1982). The relation between pa-rental distancing strategies and the child’s cognitive behavior. In L. m. Laosa, & I. E. Si-gel (Eds.), Families as learning environments for children. New York: Plenum Press.

Snowling, m. J., Dawes, P., Nash, h., & hulme, C. (2012). Validity of a protocol for adult self-report of dyslexia and related diffi-culties. Dyslexia, 18(1), 1-15. doi: 10.1002/dys.1432


Stock, P., Desoete, A., & Roeyers, h. (2010). Detecting children with arithmetic disa-bilities from kindergarten: Evidence from a 3-year longitudinal study on the role of preparatory arithmetic abilities. Journal of Learning Disabilities, 43(3), 250-268. doi: 10.1177/0022219409345011

Torppa, m., Eklund, K., van Bergen, E., & Lyytinen, h. (2011). Parental literacy pre-dicts children’s literacy: A longitudinal family-risk study. Dyslexia, 17(4), 339-355. doi: 10.1002/dys.437


Torppa, m., Poikkeus, A., Laakso, m., Eklund, K., & Lyytinen, h. (2006). Predicting delay-

ed letter knowledge development and its relation to Grade 1 reading achievement among children with and without familial risk for dyslexia. Developmental Psycho-logy, 42(6), 1128-1142.


van Bergen, E., de Jong, P. F., Plakas, A., maas-sen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of Child Psychology and Psychiatry, 53(1), 28-36. doi: 10.1111/j.1469-7610.2011.02418.x


van den Boer, m., de Jong, P. F., & haentj-ens-van meeteren, m. (in press). modeling the length effect: Specifying the relation with visual and phonological correlates of reading. Scientific Studies of Reading, doi: 10.1080/10888438.2012.683222



van der Sluis, S., de Jong, P. F., & van der Leij, A. (2007). Executive functioning in children, and its relations with reasoning, reading, and arithmetic. Intelligence, 35(5), 427-449.


Chapter 4

114

Verhoeven, L. (1993a). Grafementoets. Toets voor auditieve synthese. Handleiding [Grap-heme test. Test for phoneme blending. Ma-nual]. Arnhem, the Netherlands: Cito.

Verhoeven, L. (1993b). Toets voor auditieve analyse. Fonemendictee. Handleiding [Test for phoneme segmentation. Dictation of phonemes. Manual]. Arnhem, the Nether-lands: Cito.

Verhoeven, L. (2000). Components in early second language reading and spelling. Scientific Studies of Reading, 4(4), 313-330.

Verhoeven, L. (2002). Passieve letterkennis [Receptive letter knowledge]. Arnhem: Cito.

Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). Development of reading-re-lated phonological processing abilities: New evidence of bidirectional causality from a latent variable longitudinal study. Developmental Psychology, 30(1), 73-87.

Wolf, m., & Bowers, P. G. (1999). The dou-ble-deficit hypothesis for the develop-mental dyslexias. Journal of Educational Psychology, 91(3), 415-438.


Essentially, all models are wrong, but some models are useful.

- George E.P. Box -

IQ of four-year-olds who go on to develop dyslexia

van Bergen, E., de Jong, P.F., Maassen, B., Krikhaar, E., Plakas, A., & van der Leij, A. (in revision). IQ of four-year-olds who go on to develop dyslexia.

Chapter


IQ of four-year-olds who go on to develop dyslexia

van Bergen, E., de Jong, P.F., Maassen, B., Krikhaar, E., Plakas, A., & van der Leij, A. (in revision). IQ of four-year-olds who go on to develop dyslexia.

Chapter


Chapter 5

118

Abstract

Do children who go on to develop dyslexia show normal verbal and

nonverbal development before reading onset? According to the

aptitude-achievement discrepancy model, dyslexia is defined as a

discrepancy between intelligence and reading achievement. One of the

underlying assumptions is that the general cognitive development of

children who fail to learn to read has been normal. The current study

tests this assumption. Additionally, we investigated whether possible

IQ deficits are uniquely related to later reading or are also related to

arithmetic. Four-year-olds (N = 212) with and without familial risk for

dyslexia were assessed on 10 IQ subtests. Reading and arithmetic skills

were measured four years later, at the end of Grade 2. Relative to the

controls, the at-risk group without dyslexia had subtle impairments

only in the verbal domain, while the at-risk group with dyslexia lagged

behind across IQ-tasks. Nonverbal IQ was associated with both reading

and arithmetic, whereas verbal IQ was uniquely related to later reading.

The children who went on to develop dyslexia performed relatively

poorly in both verbal and nonverbal abilities at age 4, which challenges

the discrepancy model. Furthermore, we discuss possible causal and

epiphenomenal models explaining the links between early IQ and later

reading.


IQ of four-year-olds who go on to develop dyslexIa

119

5

5.1 Introduction

The first case reports of developmental dyslexia describe children who fail

to master literacy despite normal development in other cognitive domains

(e.g., Kerr, 1897; morgan, 1896). Therefore, the reading problems they faced

were unexpected within the context of their otherwise normal cognitive

development. This ‘unexpectedness’ is translated into diagnostic criteria that

take intelligence measures into account. According to the so-called aptitude-

achievement discrepancy model, dyslexia is defined as a discrepancy between

intelligence and reading achievement (Fletcher, Lyon, Fuchs, & Barnes, 2007).

Although several researchers have advised against the use of IQ-scores in the

diagnosis of learning disorders (e.g., Fletcher, Coulter, Reschly, & Vaughn, 2004;

Siegel, 1989), it is still commonly used.

One of the assumptions underlying a discrepancy model is that the

general cognitive development of children who fail to learn to read has been

normal. This assumption is rarely tested, however. After the start of reading,

the assumption cannot be adequately examined because the development

of verbal and nonverbal IQ of children with reading problems might have

been adversely affected by less print exposure as a consequence of their poor

reading (Cunningham & Stanovich, 1997). hence, longitudinal studies that start

before reading onset are required to investigate the issue of normal cognitive

development.

There are several prospective studies assessing cognitive precursors of

reading ability and disability, but these studies usually match groups of reading-

disabled and typically-reading children on IQ (e.g., de Jong & van der Leij, 2003)

and are therefore not suitable for the present purpose. A more appropriate

design is to follow children at-familial risk of dyslexia, that is, children with

dyslexic relatives. The current study is a familial-risk study that reports verbal and

nonverbal IQ-data of four-year-olds who went on to become dyslexic. Do these

children indeed show normal cognitive development? To pursue this question

these children were compared to children with and without a familial risk who

did not develop dyslexia on nonverbal and verbal IQ (i.e., language skills).

In a landmark study, Scarborough (1990) compared at-risk children who

later became dyslexic with control children. She found that the at-risk dyslexic

children were deficient in several language skills from age 2½ onwards. Slow

language development in the at-risk children with dyslexia was also shown by


Chapter 5

120

a prospective study of Snowling, Gallagher, and Frith (2003). When the children

were about 4 years old, they had deficiencies in vocabulary, narrative skills,

and verbal short-term memory. At this age, nonverbal IQ was not significantly

lower, despite the observed medium effect size of .5. Yet at 6 years of age both

verbal and nonverbal IQ of the dyslexic children were significantly lower than

that of their peers without familial risk by more than one standard deviation.

Furthermore, the at-risk children who did not become dyslexic showed at both

measurement occasions normal nonverbal IQ, but they had a slightly lower

verbal IQ than the children without familial risk. This latter difference had a

medium effect size, but missed significance possibly due to a lack of power.

Recently, Torppa, Lyytinen, Erskine, Eklund, and Lyytinen (2010)

examined very early language markers of dyslexia in a familial-risk study. Various

receptive and expressive language abilities were examined in this group between

1½ and 5½ years of age. From age 2 onwards, the at-risk and non at-risk children

who later became dyslexic showed impairments relative to the controls. At-risk

children who became normal readers tended to show weak expressive syntax

and vocabulary at age 5, although this difference (d = .4) was not significant.

Torppa et al. also found that at age 2, the future dyslexic children had significantly

lower full-scale IQ than the other two groups, with a medium effect size. At age

5 they performed significantly lower (d = 1) than the control group on verbal IQ.

however, at this age nonverbal IQ was similar in all groups.

Taken together, these studies show that children with dyslexia typically

had problems with language acquisition in the preschool years. however, the

findings in the few studies in which lags in nonverbal abilities were examined

are less consistent. Snowling et al. (2003) found that children who later faced

dyslexia performed slightly lower at age 4 and evidently lower at age 6. In

contrast, the children with dyslexia in the study of Torppa et al. (2010) had

normal nonverbal IQ before reading onset.

In the current study we compared at-risk dyslexic, at-risk non-dyslexic,

and non-dyslexic control children without familial risk, on several subtests

of verbal and nonverbal IQ when they were 4 years of age. In addition, we

addressed two more issues. Firstly, whereas various abilities within the verbal

domain have been considered, a differentiation among nonverbal abilities has

not been made. In the current study different aspects of nonverbal IQ were

examined. As a second issue, we considered whether the possibly lower IQ of

children who go on to develop dyslexia specifically relates to their later reading



121

5

difficulties, or whether their lower IQ is due to (subclinical) problems that are

comorbid to dyslexia. We hypothesized that this could especially be true for

a possible lower nonverbal IQ: in models of decoding, verbal abilities (e.g.,

the phonological lexicon in dual-route models or the semantic component in

connectionist models) do explicitly play a part, whereas nonverbal abilities do

not. Because dyslexia and dyscalculia often co-occur (Landerl & moll, 2010),

we chose to include arithmetic ability in the present study. In addition, basic

arithmetic skills are, besides basic reading skills, arguably the most important

foundational skills acquired during the early school years.

5.2 Method

5.2.1 ParticipantsThree groups of children of the Dutch Dyslexia Programme (see van Bergen,

de Jong, Plakas, maassen, & van der Leij, 2012) were involved in this study. The

at-risk dyslexic group consisted of 44 dyslexic children with a familial risk for

dyslexia. The at-risk non-dyslexic group comprised 100 children with a familial

risk but without dyslexia. Finally, the control group included 68 children without

a familial risk and without dyslexia. Two children without a familial risk were

categorized as dyslexic and were omitted from the study. All at-risk children

had at least one parent and one close relative with dyslexia. On average, the

scores of the dyslexic parents belonged to the bottom 5% on reading fluency.

A detailed description of the assessment of familial risk is given in van Bergen

et al. (2012). Informed consent was obtained from all parents. All but one child

attended mainstream education.

At the end of Grade 2, children were considered dyslexic when their

score on the word-reading fluency task described below corresponded to the

weakest 10% in the population. In an earlier paper we reported that these dyslexic

children were impaired in reading accuracy and fluency, spelling, phonological

awareness, and rapid naming (van Bergen et al., 2012). In the present paper we

report new data of the same sample. Table 5.1 presents characteristics of the three

groups. Percentages of boys were similar across groups and age differences were

not significant. however, parental education (averaged over both parents) was

lower in the at-risk than in the control group. Furthermore, the at-risk dyslexic

group lagged behind on arithmetic (bottom Table 5.2).


Chapter 5

122

Table 5.1 Children’s Characteristics by Group

At-risk

Dyslexic Non-dyslexic Control p

Sample size 44 100 68

No (%) of boys 27a (61%) 57

a (57%) 41

a (60%) .855

Age in months

At nonverbal ability assessment 48.20a (2.74) 47.68

a (3.02) 47.07

a (2.25) .095

At verbal ability assessment 53.19a (1.60) 53.01

a (1.45) 52.69

a (1.07) .144

Parental education 3.33a (0.86) 3.53

a (0.78) 4.17

b (0.72) <.001

Note. The group means are given with standard deviations in parentheses. Numbers and means in the same row that do not share subscripts differ at p < .05 on the χ2-test or Tukey’s test. No = number.

5.2.2 Measuresmeasures were selected for nonverbal IQ (at age 4), verbal IQ (at age 4½),

and school achievement (at age 8; end second grade). The reliabilities of the

measures are listed in Table 5.2.

Nonverbal IQ

The nonverbal-IQ test battery of Tellegen, Winkel, Wijnberg-Williams, and

Laros (1998) is widely used in the Netherlands and Flanders because of its high

psychometric quality (Evers, van Vliet-mulder, & Groot, 2000). It was originally

developed for deaf children; hence, measures are genuinely nonverbal. For

hearing children spoken language is used during assessment to create a natural

situation, but the verbal instructions do not provide additional information. The

test battery is composed of six subtests. The performance tasks (block design,

patterns, and object assembly) tap visual-spatial processing, visual-motor

coordination, and persistence. Solutions can be obtained by trial-and-error. The

reasoning tasks (picture completion, categories, and analogies) measure fluid

reasoning about concrete and abstract categories. From the presented material

a principle has to be detected and subsequently applied to new material.

Each subtest consists of two parts, with in total 14 to 17 items

of increasing difficulty. In Part 1 of each subtest, the first few items are first

demonstrated by the tester to show the requirements of the items. Part

2 of each subtest starts with an item for practice. The items in Part 2 of the

Performance tests are stopped after 2’30 minutes for practical reasons. All test

items are scored as either correct or incorrect. Following an incorrect response,



123

5

the tester shows the correct solution so the child understands the purpose of

the test without verbal explanation. Each subtest contains a discontinuation

rule. Scaled subtest scores were derived from the manual. The scores combine

to form a nonverbal-IQ score with a reliability of .90. The six subtests are

described below.

Performance tests.

Block design. The child arranges coloured squares according to a

design depicted in a booklet. The difficulty increases from designs with only

red coloured squares to designs with red, yellow, and red/yellow squares.

Patterns. The test involves copying abstract line patterns from a

booklet. The child has to draw the patterns by connecting dots. It starts with

simple patterns (e.g., connecting 2 dots) and increases to patterns that require

connecting 4, 9, or 16 dots. Drawings are scored according to whether the right

dots are connected, rather than manual dexterity.

Object assembly. The test requires the child to do jigsaw puzzles of

three to six pieces. The puzzle images include familiar objects or animals. In

Part 1 the puzzles come with the help of a picture and a frame to fit the puzzle

in. In later items no frame and picture are provided.

Reasoning tests.

Picture completion. The child completes a set of pictures of which parts

are missing. In the items in Part 1 four pictures are shown. Of each picture a part

is missing (e.g., four different animals are depicted without legs). The missing

parts are depicted on four cards. Part 2 consists of items depicting a scene (e.g.,

a boy in a kitchen) with one or two elements missing. The child completes the

picture with the appropriate cards.

Categories. In this test the child has to detect the underlying concept

of a set of pictures. The first part requires the child to sort cards in two semantic

categories, like dolls and teddy bears. In the second part three pictures are

shown that belong to the same semantic category (e.g., fruit). Out of five cards

two have to be picked that also belong to this category.

Analogies. In Part 1, pieces have to be sorted in two small trays. The

pieces differ on three dimensions: form (circles, squares, and triangles), colour

(blue and red), and size (small and large). The booklet shows the sorting

principle above the trays. In Part 2 the same changing principle as in the

example analogy has to be applied. For instance, the example analogy depicts

a small blue triangle that is transformed into a big blue triangle, meaning a


Chapter 5

124

change in the dimension size.

Verbal IQ

We selected the adapted and standardized Dutch version of the Reynell

Developmental Language Scales (van Eldik, Schlichting, lutje Spelberg, van der

meulen, & van der meulen, 2001) and three tests from a language-skill battery

(Schlichting, van Eldik, lutje Spelberg, van der meulen, & van der meulen,

2003). All items are scored as either correct or incorrect. Each subtest contains

a discontinuation rule. Standard scores were derived from the manuals.

Language comprehension. The test assesses sentence comprehension.

The sentences concern relations between objects and increase in complexity.

The child has to show the meaning of the sentence using the objects provided,

for example “Put the knife on the plate”. There are 87 items. The test’s reliability

is .90.

Expressive vocabulary. The child is required to name pictured objects

or to complete the description of a picture (e.g., “That are trains. This one is

long and this one is…” [short]). In this way the knowledge of 62 nouns, verbs,

adjectives, adverbs, and prepositions was assessed. The reliability is .90.

Expressive syntax. In this test syntactic structures are elicited using

pictures and objects. In the first 12 items the child has to repeat a sentence

literally. In later items a sentence needs to be repeated with a variation, or a

sentence needs to be completed. The sentences are embedded in a short story.

For example, the experimenter shows a picture of a man and a woman and

two apples, and says “They are both going to eat an apple. This one is for him

and …” after which the child says: “this one is for her”. There are 40 items. The

reliability is .86.

Verbal short-term memory. Children repeat series of one to six words.

Words are one-syllabic concrete nouns. The test starts with a practice and a

test item of one word. Subsequently, blocks of items increasing from two to

five words are given. Each block consists of one practice and three test items.

Finally, there are two 6-word items, adding up to 15 items. The reliability is .79.

School Achievement

Reading. Word-reading fluency was assessed using a list comprising 150

monosyllabic words (Verhoeven, 1995). All words contain at least one consonant

cluster. The score is the number of words read correctly within one minute. This

test is widely used in Dutch education. Its reliability for this age is .96.

Arithmetic. Two subtests of the arithmetic tempo test (de Vos, 1992)



125

5

were administered: addition and subtraction. Each subtest includes 40

problems of increasing difficulty. Per problem two operands have to be added

or subtracted. All operands and outcomes are below 100. The score was the

number of correctly solved problems within one minute. The reliability of each

subtest is .93 (Stock, Desoete, & Roeyers, 2010). A total score was computed as

the sum of the standard scores.

5.2.3 ProcedureTesting lasted one hour at age 4, two hours at age 4½, and two hours at age

8. The tests at age 8 reported in this paper are part of a larger test battery (see

van Bergen et al., 2012, for an overview and results). Tests were administered

individually in a quiet room at university.

5.3 Results

Group means on the IQ subtests are depicted in Figure 5.1 and 5.2. Table 5.2 gives

the descriptive statistics of the reading and arithmetic measures. Correlations

among all measures can be found in the online appendix. Distributions were

approximately normal and missingness varied from 0 to 4%. The results are

presented in three sections. First, we consider the factor structure of the

various verbal and nonverbal IQ tests. Next, we relate IQ at the age of 4 to

reading and arithmetic achievement at the end of second grade. Finally, we

examined differences in verbal and nonverbal IQ among at-risk dyslexic, at-risk

non-dyslexic, and non-dyslexic control children without familial risk.

5.3.1 Structure of the IQ TestsConfirmatory factor analysis was used to test the hypothesised structure of the 10

IQ-subtests in the full sample. The analyses were conducted with mplus (version

2.21, muthén & muthén, 2007) using maximum likelihood estimation. One latent

variable was specified for verbal IQ (four indicators) and two for nonverbal IQ: a

performance factor and a reasoning factor, each with three indicators. Allowing

a correlated error between object assembly and categories, this model had an

excellent fit: χ2 (31, N=212) = 34.24, p = .315, RmSEA = .022 [90% CI: .000-.058],

SRmR = .037, CFI = .995. A two-group model with an at-risk and a non at-risk

group also had an excellent fit and yielded virtually identical parameter estimates.


Chapter 5

126

9

10

11

12

13

Blockdesign

Patterns Objectassembly

Picturecompletion

Categories Analogies

At-risk dyslexicAt-risk non-dyslexicControl

100

105

110

115

Languagecomprehension

Expressivevocabulary

Expressive syntax Verbal short-termmemory


Figure 5.1. Children’s Performance on Nonverbal Ability measures by Group9

10

11

12

13

Blockdesign

Patterns Objectassembly

Picturecompletion

Categories Analogies


100

105

110

115

Languagecomprehension

Expressivevocabulary

Expressive syntax Verbal short-termmemory


Figure 5.2. Children’s Performance on Verbal Ability measures by Group



127

5

Tab

le 5

.2 C

hild

ren’

s Pe

rfor

man

ce b

y G

roup

on

IQ F

acto

r Sco

res

and

mea

sure

s of

Sch

ool A

chie

vem

ent

At-

risk

Dys

lexi

cN

on-d

ysle

xic

Con

trol

Effec

t siz

e (C

ohen

’s d)

MSD

MSD

MSD

RD v

s. R

ND

RD v

s. C

RND

vs.

C

Non

verb

al a

bili

ty (4

yr)

Perf

orm

ance

-0.4

2 a0.

820.

12b

0.86

0.10

b0.

930.

580.

57-0

.01

Reas

onin

g-0

.38 a

0.91

0.04

b0.

840.

18b

0.82

0.52

0.69

0.17

Verb

al a

bili

ty (4

½ y

r)-0

.45 a

0.93

-0.0

1 b0.

930.

30b

0.80

0.55

0.94

0.39

Scho

ol a

chie

vem

ent (

end

Gra

de 2

)

Read

ing

19.2

1 a7.

1657

.01 b

19.4

967

.50 c

18.9

02.

002.

560.

56

Arit

hmet

ic

Add

ition

12.1

8 a4.

2816

.12 b

4.36

17.2

1b4.

250.

931.

180.

26

Sub

trac

tion

9.91

a4.

6513

.85 b

4.91

14.8

5 b3.

901.

011.

270.

26

Not

e. N

umb

ers

and

mea

ns in

the

sam

e ro

w th

at d

o no

t sha

re s

ubsc

ripts

diff

er a

t p <

.05

on T

ukey

’s p

ost-

hoc

test

. Coh

en’s

d is

cal

cula

ted

usin

g th

e SD

s of

the

cont

rols

. RD

= a

t-ris

k dy

slex

ic, R

ND

= a

t-ris

k no

n-dy

slex

ic, C

= c

ontr

ol.


Chapter 5

128

5.3.2 The Relationship of IQ at Age 4 and Second-Grade School Achievement Structural equation modelling was used to examine a model of the relationship

of the three IQ factors with later reading and arithmetic achievement. In this

model, the three IQ factors were specified to load on one second-order factor,

labelled IQ. Note that with three factors this second-order model is equivalent

to the three-factor model that was examined in the previous section. Reading

and arithmetic achievement were also specified to load (with equal factor

loadings) on a common factor, labelled School Achievement. Finally, School

Achievement was regressed on IQ.

The fit of this model was adequate: χ2 (50, N=212) = 69.57, p = .035,

RmSEA = .043 [.012-.066], SRmR = .047, CFI = .974. however, the model fit was

improved significantly by adding a direct effect from verbal IQ to reading (Δχ2(1)

= 7.20, p = .007). The final model (presented in Figure 5.3) provided a good fit to

the data: χ2 (49, N=212) = 62.37, p = .095, RmSEA = .036 [.000-.060], SRmR = .043,

CFI = .982. Note that the factor School Achievement reflects the variance that

arithmetic and reading have in common. The significant extra path from verbal

IQ to reading indicates that verbal abilities predict unique variance in reading

after differences in arithmetic have been controlled.

5.3.3 Differences Among Reading GroupsThe at-risk dyslexic, at-risk non-dyslexic, and control children were compared

on the three IQ factors (the verbal, performance and reasoning factors). To

this end factor scores were derived on the basis of the three-factor model that

adequately described the 10 IQ subtests (see above). Descriptive statistics for

the scores on the three factors are presented in Table 5.2. Differences among

the groups on the three IQ factors were evaluated using one-way ANOVAs

followed by two planned contrasts: at-risk dyslexic versus at-risk non-dyslexic,

and at-risk non-dyslexic versus controls. Subsequently, the analyses were

performed with arithmetic or parental education as covariate.

Nonverbal Abilities

Nonverbal abilities were subdivided into a performance and a reasoning factor,

but the results were very similar. For both, the group effect was significant

(performance factor: F(2, 209) = 6.54, p = .002; reasoning factor: F(2, 209) =

6.22, p = .002). The at-risk non-dyslexic group outperformed the at-risk dyslexic

group (performance: F(1, 209) = 11.64, p = .001; reasoning: F(1, 209) = 7.67,


iQ oF Four-year-oldS wHo go on to develoP dySleXia

129

5nvIQ

re

ason

ing

.66

.68

.50

.72 .6

4

.82

.69

.14

.20 .4

2

.69 .9

3 .73

.82

Scho

ol

nvIQ

p

erfo

rman

ce

vIQ

IQ

.14

.93.73

.20

.69

.66

.68 66

.46

.52

.64

Patt

erns

Ob

ject

ass

emb

ly

Cat

egor

ies

Pict

ure

com

pl.

Bloc

k de

sign

Ana

logi

es

Lang

uage

com

pr.

Exp

r. vo

cab

Exp

r. sy

ntax

verb

al S

TM

68

Arit

hmet

ic

66

Read

ing

.82

69.69

.69 .5

5 .5

5

.50

.72

.72

.72

Lang

uage

com

pr.

.81

.74 .8

0

.60

Fig

ure

5.3.

Str

uctu

ral E

quat

ion

mod

el o

f the

Eff

ects

of I

Q a

t Age

4 o

n Ba

sic

Scho

ol A

chie

vem

ent a

t Age

8.


Chapter 5

130

p = .006), but did not significantly differ from the control group, (performance:

F(1, 209) < 1; reasoning: F(1, 209) = 1.15, p = .285).

Next, we examined whether the lower nonverbal ability of the at-risk

dyslexics was related to their later arithmetic difficulties (see bottom Table 5.2).

Arithmetic as covariate controls for the variance in IQ that is related to arithmetic

and that arithmetic and reading have in common (similar to the model in Figure

5.3). In ANCOVAs, the group differences on both performance (F(2, 208) = 2.11,

p = .124) and reasoning (F(2, 208) = 2.18, p = .115) disappeared, after controlling

for arithmetic (performance: F(1, 208) = 13.46, p < .001; reasoning: F(1, 208) =

8.75, p = .003).

We also conducted ANCOVAs with parental education as covariate

to test whether the lower performance of the at-risk dyslexic group could be

due to a lower SES. The results for the two ANCOVAs were similar: Although

the effect of parental education was significant (performance: F(1, 203) = 5.57,

p = .019; reasoning: F(1, 203) = 11.38, p = .001), the group effect remained

significant (performance: F(2, 203) = 5.39, p = .005; reasoning: F(2, 203) = 3.64,

p = .028). moreover, the follow-up contrast comparing the at-risk children with

and without dyslexia remained significant (performance: F(1, 203) = 10.78, p =

.001; reasoning: F(1, 203) = 7.14, p = .008).

Verbal Abilities

The ANOVA on the verbal-IQ factor revealed a group difference, F(2, 209) = 9.56,

p < .001. The contrasts confirmed a stepwise pattern: the at-risk dyslexics scored

lower than the at-risk non-dyslexics, F(1, 209) = 7.57, p = .006, who scored lower

than the controls F(1, 209) = 4.90, p = .028.

In contrast to the nonverbal-IQ findings, the ANCOVA on verbal IQ with

arithmetic as covariate showed that the effect of arithmetic was not significant,

F(1, 208) = 1.85, p = .175, and that the effect of group (F(2, 208) = 6.17, p = .003)

and the two contrasts remained significant.

Finally, in an ANCOVA the effect of parental education was significant,

F(1, 203) = 19.47, p < .001. Nevertheless, the overall group effect remained

significant, F(2, 195) = 4.10, p = .018, as did the difference between the two

at-risk groups, F(1, 203) = 6.45, p = .012. however, after controlling for parental

education the significant difference between the at-risk non-dyslexic and the

control group disappeared, F(1, 195) < 1.



131

5

5.4 Discussion

The current study examined the relationship between intelligence in four-year-

olds and their reading status four years later in a sample of children with and

without a familial background of dyslexia. We found that at-risk children who

went on to be classified as dyslexic were impaired relative to control children

(i.e., not at-risk and non-dyslexic) on both verbal and nonverbal IQ. The at-risk

children who became normal readers showed slight but significantly poorer

performance in the verbal, but not in the nonverbal domain. All reported

deficits of the at-risk dyslexic children are relative to the other two groups rather

than relative to norm scores. Our sample scored in general above average on

IQ, probably due to the overrepresentation of parents with a high educational

level (see van Bergen et al., 2012, p.33).

Given that reading and IQ are correlated, it might be regarded as

obvious that a group of poor readers has a somewhat lower IQ than a group

of normal readers. however, within an IQ-matched subsample of children with

below-average full-scale IQ and with (n = 75) and without (n = 26) familial risk,

37% versus 8% of the children developed dyslexia. This difference demonstrates

once more that low reading cannot be explained by low IQ alone.

The at-risk dyslexic children lagged behind across tests of verbal IQ. At 4

years old, the at-risk dyslexia group was characterized by relatively poor sentence

comprehension, poor expressive syntax, weak vocabulary, and poor verbal short-

term memory (d = .94 on the factor score). This confirms previous reports of early

oral-language deficits in dyslexic children (Scarborough, 1990; Snowling et al.,

2003; Torppa et al., 2010). These studies as well as the current one reported large

effect sizes, suggesting that these findings are robust across orthographies.

The verbal IQ of the at-risk children who did not develop dyslexia were

somewhat lower (d = .39) relative to the controls. Previous studies (Snowling

et al., 2003; Torppa et al., 2010) found similar effect sizes, although none of

these effects reached significance. Generally, the at-risk non-dyslexic children

perform slightly lower in the verbal domain, which may only be detectable in

large samples.

Controlling for differences in parental education suggests that, relative

to their peers without familial risk, the lower language skills of children at-risk

might be attributable to their lower socioeconomic background. It could well be

that the at-risk families offer a less rich language environment. however, Neiss


Chapter 5

132

and Rowe (2000) disentangled the environmental and genetic components

underlying the correlation between parental education and children’s verbal

IQ. They estimated that the environmental effect of parental education explains

no more than 4% of children’s verbal IQ. As both children’s IQ (Walker, Petrill, &

Plomin, 2005) and parental education (Rowe, Vesterdal, & Rodgers, 1998) are

for about two-thirds genetic in origin, the disappearance of the group effect

in the ANCOVA might well be through ‘controlling for’ genetic rather than

environmental differences between the groups. Indeed, the parents of the

control children were somewhat better in verbal reasoning (van Bergen et al.,

2012), a proxy for verbal IQ. however, an extra ANCOVA with parental verbal-

reasoning as covariate1 showed no significant effect of verbal reasoning, which

favours the genetic explanation of SES’ effect.

Interestingly, the at-risk dyslexic group also lagged behind on

nonverbal IQ, although somewhat less than in the language domain. In line

with Snowling et al. (2003), we found that around 4 years of age these children

performed on average over half a standard deviation lower on nonverbal IQ

than the controls. however, this difference reached significance only in the

present study due to the larger sample size. In contrast, Torppa et al. (2010)

did not find differences on nonverbal IQ at age 5. Nevertheless, the present

findings agree with both studies in that at-risk children who did not develop

dyslexia had age-adequate nonverbal IQ.

At a more general level, we were interested in the nature of the link

between IQ at age 4 and reading at age 8. Are the lower IQ’s shown by the future

dyslexics specifically related to reading or are they also related to another basic

academic skill, arithmetic? Our results from structural equation modelling as

well as ANCOVAs indicate that verbal IQ is specifically related to reading, not

to arithmetic, which is expected in view of the linkage between the language

and reading systems (hayiou-Thomas, harlaar, Dale, & Plomin, 2010). Yet, our

observed unique relation between verbal IQ and reading does not elucidate

whether they are independent consequences of shared aetiological factors

or, alternatively, whether there is a causal link, with the developing reading

system building on the language system.

In contrast to verbal IQ, we found that nonverbal IQ is related to both

reading and arithmetic. This is in line with the generalist-genes hypothesis put

1 Only scores of the weakest-reading parent were available. Details of this analysis can be obtained from the first author.



133

5

forward by Kovas and Plomin (2007). Kovas, haworth, Dale, and Plomin (2007)

found that one-third of the genes that influence one academic domain are

specific to that domain, one-third overlaps with other academic domains, and

one-third overlaps with genes associated with IQ. For dyslexia, the identified

candidate susceptibility genes are associated with anomalies in neuronal

migration and axon growth (Galaburda, LoTurco, Ramus, Fitch, & Rosen, 2006).

In light of the deficiency in these fundamental processes, pleiotropic cognitive

deficits are to be expected. This suggests that the poor nonverbal IQ of the

future dyslexics might well be epiphenomenal to poor reading, rather than

causally linked.

In the current study we only assessed fairly basic school skills. Even

higher predictive relations are expected between preschool IQ and later

reading comprehension and mathematics. A limitation of the study is the lack of

measures of attention or motor problems for example, since the groups might

differ in this respect. Our study did, however, include measures of arithmetic

problems, of which the findings highlighted the value of studying comorbidity.

Finally, it should be noted that verbal IQ batteries (e.g., the WISC, Wechsler,

2004) usually also include verbal reasoning, besides vocabulary and verbal

short-term memory measures. It is not entirely clear if and how the absence of

a measure of verbal reasoning has affected our findings. If differences among

the groups in verbal reasoning are relatively smaller, then inclusion of verbal

reasoning would have resulted in somewhat smaller differences in verbal IQ.

however, given the robust differences on the other verbal tests it is unlikely that

this would have changed the overall pattern of differences among the groups.

Nevertheless, it is important that our results can be fully compared with those

of other familial risk studies (see Introduction) in which verbal reasoning was

not included either.

Our study is of interest for the debate on the diagnosis of dyslexia. The

children with dyslexia already had a lower full-scale IQ before reading onset,

arguably as part of their disorder. According to the aptitude-achievement

discrepancy model these children need to have even more severe reading

difficulties in order to be diagnosed as dyslexic. Since it is now more or less

accepted that children with reading difficulties might have additional language

difficulties, some scholars have proposed leaving verbal IQ out of the definition

and instead using a discrepancy with nonverbal IQ. however, the future dyslexic

readers in our study also had lower nonverbal IQ. Our findings, together with


Chapter 5

134

evidence from earlier mentioned behavioural and molecular genetic studies,

raise the possibility that dyslexia in itself affects brain circuits not related to

reading. If so, then it seems undesirable to tap these circuits by IQ-tests and to

take these IQ-scores into account in the diagnosis of dyslexia.

Our findings show that at the group level the children who went on to

develop dyslexia performed relatively poorly in both verbal and nonverbal IQ

at 4 years of age. Even though the mechanisms that might link these abilities

to reading acquisition are not fully understood, our findings demonstrate that

the cognitive development of the group of children who will fail to learn to

read is atypical.



135

5

5.5 References

Cunningham, A.,E., & Stanovich, K.,E. (1997). Early reading acquisition and its relation to reading experience and ability 10 ye-ars later. Developmental Psychology, 33(6), 934-945.

de Jong, P. F., & van der Leij, A. (2003). Deve-lopmental changes in the manifestation of a phonological deficit in dyslexic children learning to read a regular orthography. Journal of Educational Psychology, 95(1), 22-40. doi:10.1037/0022-0663.95.1.22

de Vos, T. (1992). Tempo test rekenen [Arith-metic speed test]. Lisse: Swets & Zeitlinger.

Evers, A., van Vliet-mulder, J. C., & Groot, C. J. (2000). Documentatie van tests en testre-search in Nederland [Documentation on tests and test research in the Netherlands]. Assen: Van Gorcum.

Fletcher, J., Coulter, W., Reschly, D., & Vaug-hn, S. (2004). Alternative approaches to the definition and identification of learn-ing disabilities: Some questions and ans-wers. Annals of Dyslexia, 54(2), 304-331. doi:10.1007/s11881-004-0015-y

Fletcher, J., Lyon, G. R., Fuchs, L., & Barnes, m. (2007). Learning disabilities: From identifi-cation to intervention. New York: Guilford Press.

Galaburda, A. m., LoTurco, J., Ramus, F., Fitch, R. h., & Rosen, G. D. (2006). From genes to behavior in developmental dyslexia. Na-ture Neuroscience, 9(10), 1213-1217.

hayiou-Thomas, m. E., harlaar, N., Dale, P. S., & Plomin, R. (2010). Preschool speech, language skills, and reading at 7, 9, and 10 years: Etiology of the relationship. Journal of Speech, Language, and Hearing Rese-arch, 53(2), 311-332. doi:10.1044/1092-4388(2009/07-0145)

Kerr, J. (1897). School hygiene, in its mental, moral, and physical aspects. Journal of the Royal Statistical Society, 60(3), 613-680.

Kovas, Y., haworth, C. m. A., Dale, P. S., & Plomin, R. (2007). The genetic and en-vironmental origins of learning abilities and disabilities in the early school ye-ars. Monographs of the Society for Rese-arch in Child Development, 72(3), 1-160. doi:10.1111/j.1540-5834.2007.00453.x



morgan, W. P. (1896). A case of congenital word blindness. British Medical Journal, 2(1871), 1378.

muthén, L. K., & muthén, B. O. (2007). Mplus user’s guide. Los Angelos: muthén & muthén.

Neiss, m., & Rowe, D. C. (2000). Parental education and child’s verbal IQ in adop-tive and biological families in the na-tional longitudinal study of adolescent health. Behavior Genetics, 30(6), 487-495. doi:10.1023/a:1010254918997

Rowe, D. C., Vesterdal, W. J., & Rodgers, J. L. (1998). herrnstein’s syllogism: Genetic and shared environmental influences on IQ, education, and income. Intelligen-ce, 26(4), 405-423. doi:10.1016/s0160-2896(99)00008-2


Schlichting, J., van Eldik, m., lutje Spelberg, h., van der meulen, S., & van der meulen, B. (2003). Schlichting test voor taalproduc-tie [Schlichting test for language producti-on]. Lisse: Swets & Zeitlinger B.V.


Chapter 5

136

Siegel, L. S. (1989). IQ is irrelevant to the de-finition of learning disabilities. Journal of Learning Disabilities, 22(8), 469-478.


Stock, P., Desoete, A., & Roeyers, h. (2010). Detecting children with arithmetic disa-bilities from kindergarten: Evidence from a 3-year longitudinal study on the role of preparatory arithmetic abilities. Journal of Learning Disabilities, 43(3), 250-268. doi:10.1177/0022219409345011

Tellegen, P. J., Winkel, m., Wijnberg-Williams, B. J., & Laros, J. (1998). Snijders-Oomen nonverbal intelligence test, SON-R 2.5–7, Manual and research report. Lisse: Swets & Zeitlinger.

Torppa, m., Lyytinen, P., Erskine, J., Eklund, K., & Lyytinen, h. (2010). Language de-velopment, literacy skills, and predictive connections to reading in Finnish children with and without familial risk for dyslexia. Journal of Learning Disabilities, 43(4), 308-321. doi:10.1177/0022219410369096

van Bergen, E., de Jong, P. F., Plakas, A., maas-sen, B., & van der Leij, A. (2012). Child and parental literacy levels within families with a history of dyslexia. Journal of Child Psychology and Psychiatry, 53(1), 28-36. doi:10.1111/j.1469-7610.2011.02418.x

van Eldik, m., Schlichting, J., lutje Spelberg, h., van der meulen, B., & van der meulen, S. (2001). Reynell test voor taalbegrip [Rey-nell test for language comprehension]. Lis-se: Swets & Zeitlinger B.V.

Verhoeven, L. (1995). Drie-minuten-toets. handleiding [Three-minutes-test. manual]. Arnhem: Cito.

Walker, S. O., Petrill, S. A., & Plomin, R. (2005). A genetically sensitive investigation of the effects of the school environment and so-cio-economic status on academic achie-vement in seven-year-olds. Educational Psychology, 25(1), 55-73. doi:10.1080/0144341042000294895

Wechsler, D. (2004). The Wechsler intelligence scale for children - Fourth edition. London: Pearson Assessment.


Don’t become a mere recorder of facts, but try to penetrate the mystery of

their origin.

- Ivan Pavlov -

Chapter

General Discussion


Chapter

General Discussion


Chapter 6

140

Who will develop dyslexia? The series of studies reported in this thesis study

cognitive precursors of dyslexia in parents and children. We investigated, firstly,

the cognitive profile characteristic of children who go on to develop dyslexia.

As a related issue we examined whether children’s skills assessed prior to

Grade 1 are specifically predictive of later reading development or are also

involved in later arithmetic development. The second goal was to examine

the impact of the cognitive profile of parents and the literacy environment

parents create on children’s reading outcome. The studies employed a

prospective design, in which the progress of children at high and low family

risk was followed. Children at high risk had a family history of dyslexia. After

some years of reading instruction they were categorized as either dyslexic

or non-dyslexic (below or above the 10th percentile cut-off on word-reading

fluency, respectively). Subsequently, they were compared concurrently and

retrospectively with each other and with typically developing children without

such a family background. The studies reported in the current thesis include

two independent samples of the Dutch Dyslexia Programme (DDP), labelled

the Amsterdam and the national sample (see also the General Introduction,

Section 1.5.1). The Amsterdam sample (N = 79) covers children’s development

from kindergarten to Grade 5 (11 years old). Findings are described in Chapter

2. Chapters 3, 4, and 5 report findings on the national study (N = 212), including

the development from 3½ to 9 years of age (Grade 3).

In what follows I will review the findings regarding precursors and

literacy skills in children and precursors in families. Each topic will be concluded

by linking the findings to Pennington’s (2006) multiple cognitive deficit model,

as outlined in the General Introduction (Section 1.2.1 and Figure 1.1). Family-

risk studies provide the opportunity to test implications of the multiple deficit

model and to add more details for the model of dyslexia. Do the findings of the

current set of studies lend support for this model and can the findings allow

us to start on specifying the model for the case of dyslexia? Thereafter I will

extend Pennington’s model to the intergenerational multiple deficit model,

which includes an extra level to account for parental effects. This chapter will

be closed by suggestions for future research and concluding remarks.


General Discussion

141

6

6.1 Precursors in Children

Two categories of child precursors of dyslexia were studied: general cognitive

ability (IQ) and preliteracy skills. The first were studied at kindergarten entry

(age 4) in Chapter 5 (the national sample). Preliteracy skills were assessed

during the last year in kindergarten (age 5 or 6) and reported in Chapter 2 for

the Amsterdam sample and Chapter 4 for the national sample.

6.1.1 IQThe relation between verbal and nonverbal IQ around the age of four, and

reading outcome at the end of Grade 2 was the focus of Chapter 5. It was found

that at-risk children who go on to become dyslexic were impaired relative to

controls on both verbal and nonverbal IQ, with the gap being larger for verbal

IQ. The at-risk children who do not become dyslexic showed good nonverbal

abilities, but their verbal IQ was slightly but significantly lower than that of

controls. Furthermore, it appeared that nonverbal IQ was equally strongly

related to later reading achievement (e.g., word-reading fluency) as to later

arithmetic achievement (e.g., mental-arithmetic fluency), while verbal IQ was

specifically predictive of reading.

An unresolved issue is the nature of the link between early IQ

and subsequent reading ability. Three possibilities are 1) the inclusion of

reading-related tasks in the IQ battery, 2) IQ and reading being independent

consequences of the same aetiological factors, and 3) a causal link from IQ to

reading. Each of these possibilities will be discussed below for verbal and for

nonverbal IQ.

According to the first possibility, verbal IQ and reading are related

because verbal IQ tests include subtests that assess component skills that are

known to be related to reading ability and known to be deficient in individuals

with dyslexia (cf. Siegel, 1999). Note that this first account is compatible with

both a shared aetiology account and a causal account. That is, the component

skill that is supposedly involved in both verbal IQ and reading could be

associated with reading through partly shared aetiology or could causally

influence reading ability. A component skill shared by verbal IQ and reading

might be verbal short-term memory (deficient in dyslexia, see e.g. de Jong,

1998), which was indeed part of the verbal IQ battery of the study presented

in Chapter 5. Nevertheless, group differences on this subtest did not exceed


Chapter 6

142

group differences on the other verbal IQ subtests (see Chapter 5, Figure 2), a

finding that is not in agreement with this first account. Concerning nonverbal

IQ, this first account cannot explain the relation between early nonverbal IQ

and subsequent reading as there is no common component skill known.

Second, the literature suggests that early language and later reading

ability indeed share aetiological factors, which causes the phenotypic

correlation between them. hayiou-Thomas, harlaar, Dale, and Plomin (2010),

for example, reported that language (or verbal IQ) at age 4 and subsequent

reading performance are largely influenced by common genetic and shared

environmental influences. Knowing that the genetic and shared environmental

correlations are strong does not, however, reveal whether early verbal IQ

and later reading are causally linked, with language supporting reading

development, or whether they independently draw upon common aetiological

influences.

Regarding nonverbal IQ and a shared-aetiology account, we found that

nonverbal IQ was related to later reading as well as later arithmetic skills. This

finding fits with the generalist genes hypothesis (Kovas & Plomin, 2007; Plomin

& Kovas, 2005 - see also the General Introduction, Section 1.2.2). The generalist

genes hypothesis states that genes influencing different academic domains

and IQ partly overlap (Kovas, haworth, Dale, & Plomin, 2007). Evidence from

molecular genetic studies more clearly points to the possibility that reading,

arithmetic, and nonverbal IQ are related because of partly shared aetiological

factors. The candidate genes that have been linked to dyslexia all play a role in

early brain development: they can be responsible for anomalies in neuronal

migration and axon growth (Galaburda, LoTurco, Ramus, Fitch, & Rosen, 2006).

The resulting subtle cortical malformations are probably more widespread than

just cortical areas involved in reading. Therefore, it is expected that cognitive

deficiencies are not solely restricted to the reading process.

Lastly, I will consider a causal account. Verbal IQ was found to be

specific for reading development, which is consistent with a partly causal link,

with the developing reading system building on the language system. Indeed,

models for word decoding generally include aspects of language skills, like the

phonological lexicon in the dual-route cascaded model (Coltheart, Rastle, Perry,

Langdon, & Ziegler, 2001) and phonological and semantic units in the triangle

connectionist model (Seidenberg & mcClelland, 1989). moreover, a causal

link for the phonological aspects of language is often proposed (e.g., hulme,


General Discussion

143

6

Bowyer-Crane, Carroll, Duff, & Snowling, in press; Vellutino, Fletcher, Snowling,

& Scanlon, 2004), although the relation between phonological skills and

reading could also be reciprocal (e.g., Burgess & Lonigan, 1998). Interventions

to promote broader language skills in kindergarten children that would result

in a subsequently faster reading development, would provide evidence for a

causal interpretation. In the absence of such evidence, explaining the relation

between verbal IQ and reading in terms of shared underlying factors is just as

plausible. In sum, our study showed that early verbal IQ is a specific predictor of

word-reading fluency, but the nature of this association remains elusive.

moving on to nonverbal IQ, one could defend a causal link by arguing

that nonverbal IQ indicates an individual’s ability to acquire new insights and

skills. however, this is at best an explanation for early literacy development,

like grasping the alphabetic principle. It is unclear how an individual’s ability

to acquire new insights and skills would translate into developing such a

specific skill as word-level reading fluency, and hence, such an unspecified

explanation is difficult to defend. Therefore, explaining the relation between

early nonverbal IQ and subsequent reading in terms of common aetiological

factors seems most plausible.

Although our data do not reveal the nature of the link between IQ

and reading, they do show that children who go on to develop dyslexia have

a lower IQ than typically developing children. Because at the time their IQ was

assessed they could not yet have failed to learn letters or word reading, their

lower IQ at this age cannot be a consequence of reading failure and reduced

print exposure. In previous prospective studies IQ differences were either

neglected or were not possible to investigate (e.g., due to matching on IQ as in

de Jong & van der Leij, 2003). For that reason, the relative low IQ of four-year-

olds who go on to be dyslexic is in itself a unique finding.

6.1.2 Preliteracy SkillsAs mentioned above, preliteracy skills were assessed during the last year in

kindergarten (age 5 or 6) for both samples. Chapter 2 reports the results for

the Amsterdam sample and Chapter 4 for the national sample. The studies

included the most important preliteracy skills: phonological awareness, rapid

naming, and letter knowledge. Again, the longitudinal character ascertains

that these cognitive deficiencies are due to aetiological factors rather than due

to poor reading and less print exposure.


Chapter 6

144

In the Amsterdam sample, children were categorised as dyslexic or

non-dyslexic in Grade 5. The at-risk dyslexic group read less fluently than the

at-risk non-dyslexics and the controls across reading tasks and throughout

primary school. Retrospectively, in kindergarten the at-risk group with dyslexia

was slower on rapid naming and knew fewer letters than the other two groups.

however, group differences on phonological-awareness measures did not

reach significance. The absence of a difference could be due to low levels of

phonological awareness in Dutch children of this age (see also de Jong & van

der Leij, 1999). For many children the tasks were too difficult, which might have

obscured existing group differences in the latent ability. Another factor that

might play a role is the timing of identifying dyslexia. Reading outcome was

not assessed until fifth grade, when word-reading fluency in Dutch is more

strongly related to rapid naming than to phonological awareness (Vaessen &

Blomert, 2010). hence, a group categorized as dyslexic at this age is expected

to show larger deficiencies in rapid naming than phonological awareness.

In line with this, the children with dyslexia were impaired on rapidly naming

familiar pictures and colours during kindergarten.

moving on to the national sample, Chapter 4 presents the findings

regarding preliteracy skills in kindergarten. In Chapter 4 reading status was

assessed in third grade and preliteracy skills at the end of kindergarten. The

at-risk dyslexic group was impaired on all preliteracy skills. The at-risk children

without later dyslexia showed normal letter knowledge and rapid naming,

but performed below controls on phonological awareness. Note that the two

independent samples converge on the conclusion that at-risk dyslexic children

are impaired, compared to the other two groups, on rapid naming and letter

knowledge in kindergarten. The findings across the samples differ, however,

for phonological awareness: no group differences in the Amsterdam sample

and a step-wise pattern (at-risk dyslexia < at-risk no-dyslexia < controls) in

the national sample. Although this may seem contradictory, it should be

mentioned that in the national sample the phonological-awareness tasks were

more appropriate for the children’s ability level. For this reason, we attach more

importance to the group differences in the national sample.

With regard to the multiple deficit model (Pennington, 2006), the

current set of studies shows that the cognitive processes associated with

dyslexia are phonological awareness, rapid naming, and verbal and nonverbal

IQ. Note that these associations are not necessarily causal (see the discussion


General Discussion

145

6

on IQ in Section 6.1.1). Letter knowledge might be added to the list of cognitive

processes as an underlying skill of reading. however, it could also be regarded

as belonging to the symptom level, being a forerunner or autoregressor of

reading.

The multiple deficit model predicts that normal reading children with

a family risk do slightly poorer on reading-related skills than normal reading

children without such risk. This is because there is evidence that the underlying

cognitive processes of reading are also complex traits, influenced by multiple

genetic and environmental factors (Naples, Chang, Katz, & Grigorenko, 2009;

Petrill, Deater-Deckard, Thompson, De Thorne, & Schatschneider, 2006), so

the at-risk no-dyslexia children are expected to receive some of the involved

risk factors. A difference between the two normal reading groups was indeed

found for verbal IQ and phonological awareness, but not for rapid naming. how

their surprisingly good rapid naming can be explained aetiologically remains

puzzling. It seems that verbal IQ and phonological awareness are related to both

reading status and family-risk status, whereas rapid naming only differentiates

between children differing in reading status but not between children differing

in risk status. The at-risk children without dyslexia are good at rapid naming

and, in anticipation of Section 6.3.2, their affected parents are relatively good at

rapid naming. This suggests that good rapid naming somehow decreases the

probability of becoming dyslexic, but how remains unclear.

The findings regarding preliteracy skills of the at-risk children who

went on to become dyslexic confirmed that phonological awareness, rapid

naming, and letter knowledge are cognitive risk factors for dyslexia. Chapter

4 additionally investigated comorbidity with dyscalculia and whether the

investigated preliteracy skills are also cognitive risk factors for dyscalculia.

Dyslexia and dyscalculia (defined as scoring in the bottom 10% on word-

reading fluency or arithmetic fluency, respectively) were indeed comorbid,

as evidenced by the 42% rate of dyscalculia in the at-risk dyslexic group. In

comparison, these figures were 20% and 8% in the at-risk and not at-risk

normal readers, respectively. The trio of preliteracy skills was shown to be

predictive of later arithmetic achievement as well. Rapid naming was equally

strongly related to reading and arithmetic, but phonological awareness and

letter knowledge were more specific precursors of reading. After controlling

for rapid naming and letter knowledge in a regression analysis, phonological

awareness dropped out as a predictor of arithmetic. Therefore, our data did not


Chapter 6

146

provide unequivocal support for the hypothesis of De Smedt and colleagues

(Boets & De Smedt, 2010; De Smedt & Boets, 2010; De Smedt, Taylor, Archibald,

& Ansari, 2010) that individuals with dyslexia have difficulties with retrieving

arithmetic facts from long-term memory because these facts are phonological

in nature. Our data suggest that the speed of processing or speed of retrieving

phonological codes might be more important in predicting later achievement

in arithmetic fluency.

The conclusion that some of the cognitive processes of importance

to reading are also important for arithmetic, whereas others are distinct to

reading, is in line with the multiple deficit model and the generalist genes

hypothesis (Kovas & Plomin, 2007; Plomin & Kovas, 2005 - see also the hybrid

model in the General Introduction, Section 1.2.3). In sum, nonverbal IQ and

rapid naming are shared and therefore contribute to the correlation between

arithmetic and reading. Likewise, at the lower end of the distribution, they

contribute to the comorbidity between dyscalculia and dyslexia. Verbal IQ,

phonological awareness, and letter knowledge were found to be skill-specific

cognitive processes, contributing to the dissociation between arithmetic and

reading. The inclusion of arithmetic and its associated disorder dyscalculia

in the study of the specificity of dyslexia precursors underscore the value of

studying comorbidity. Other comorbidities are also of interest, but were not

dealt with in the current studies.

6.2 Literacy Skills in Children

Chapter 3 compares for the national sample the three groups of children at

the end of Grade 2 on literacy skills and their main cognitive underpinnings:

phonological awareness and rapid naming. On all five reading tasks, which

measure accuracy and fluency of word and pseudoword reading, the at-

risk children with dyslexia were severely impaired compared to the control

children, as indicated by large effect sizes (1.74 to 2.83). In addition, they made

many errors in spelling words and in phoneme deletion, and were slow in serial

naming of familiar items.

Although the at-risk group without dyslexia had literacy skills within

the normal range for their age, they read significantly less accurately and

fluently than controls on all of these reading measures (with effect sizes in the


General Discussion

147

6

range of 0.43 to 0.74). The same step-wise pattern was found for spelling skills

and phonological awareness. A salient finding was the good performance of

the at-risk children without dyslexia on rapid naming. Despite the fact that they

fared significantly worse than controls on literacy and phonological awareness,

they were equally good on naming digits and colours rapidly.

The group differences for phonological awareness and rapid naming

in second grade mirror those for kindergarten, so the picture did not change.

Apparently, phonological awareness is associated with both reading and risk

status, while rapid naming is only related to reading status. The at-risk children

who go on to develop normal reading skills might do well despite their family

risk because the efficiency of the processes that rapid naming tap might protect

them against dysfluent reading. On the other hand, one could argue that they

still have mild literacy problems due to their mild phonological awareness

deficit.

The multiple deficit model states that developmental disorders,

like dyslexia, are influenced by many genetic and environmental risk factors

that each increase the liability for the disorder without being necessary or

sufficient for causing it. From this it follows that the multiple deficit model

predicts a liability or risk function that is continuously distributed in the normal

population. For our three groups this would mean a high liability for the at-

risk dyslexia group, moderate liability for the at-risk no-dyslexia group, and

low liability for the control group. At the behavioural level these differences

in liability should be observable as differences in literacy skills, with the at-

risk no-dyslexia group taking up an intermediate position. This hypothesis

was confirmed: on all literacy tasks we found that the at-risk children who

did not meet the dyslexia criterion still performed significantly weaker than

controls. Nonetheless, as described in the previous section, this pattern was

not systematically found at the cognitive level.

6.3 Precursors in Families

In addition to predictors of dyslexia residing in children we examined possible

predictors in their families. more specifically, we studied effects of home literacy

environment and parental literacy skills on children’s reading outcome.


Chapter 6

148

6.3.1 Home Literacy EnvironmentChildren’s socio-economic status, as indexed by parental level of education, can

be seen as a distal family background characteristic. The literacy environment

that parents provide at home might act as a mediating proximal process

between socio-economic status and children’s reading development (cf.

Leseman, Scheele, mayo, & messer, 2007). Both these distal and proximal family

characteristics were investigated.

In the Amsterdam as well as in the national sample the families with

a parent with dyslexia (i.e., at-risk families) had lower levels of education

compared to the families with average to good reading parents (i.e., control

families). Within the at-risk sample, the parents of the affected children had

slightly lower educational levels than those of the unaffected children.

Parental level of education could assert an influence on children’s

reading development via the proximal family environment. hence, it was

investigated whether there is a relation between the literacy environment that

parents provide at home and children’s family history of dyslexia and reading

outcome. In the Amsterdam sample (Chapter 2) home literacy environment

was assessed during the second (and final) year of kindergarten. The three

groups were compared on library memberships, frequency of shared reading,

and on the number of books in the home, but they did not differ significantly

on any of these measures. In the national sample (Chapter 4) the groups did

not differ either on cognitive stimulation by parents, but there was a tendency

for parents of control children to own more magazines, newspapers and books.

The two at-risk groups did not differ in any of the measures of home literacy

environment.

Although behavioural genetic studies point to substantial heritability

of reading, they also estimate that roughly 30% of individual differences is due

to environmental factors (Byrne et al., 2009; Petrill et al., 2006; Taylor, Roehrig,

hensler, Connor, & Schatschneider, 2010). The moderate environmental

influences do not leave much room to find effects of home literacy environment,

especially because our volunteer sample probably does not represent the

full environmental range. Indeed, our data about children’s home literacy

environment did not reveal significant group differences. It should be admitted

however, that home literacy environment was only measured superficially and

subjective. When the home literacy environment is assessed more thoroughly,

preferably with objective observations, the question whether children’s literacy


General Discussion

149

6

environment at home influences their reading development can be answered

more reliably. Also other environmental factors warrant further investigation,

such as pre- and perinatal factors and school and classroom characteristics.

Despite the mentioned limitations, our findings are in agreement with findings

from other family-risk studies, which also failed to show effects of home literacy

environment on children’s reading outcome (Elbro, Borstrøm, & Petersen, 1998;

Snowling, muter, & Carroll, 2007; Torppa et al., 2007). Thus, no environmental

risk factors of substantial effect have been identified that would have been

easy targets for intervention.

To summarize, we did find an effect of parent’s educational level, but

this did not translate into detectable effects of literacy practices at home. This

conclusion might seem paradoxical. however, the lower educational level of

the parents of the at-risk sample might be partly a consequence of their reading

difficulties. This could be the main reason for a relation between parental

education and offspring’s reading outcome, rather than via environmental

differences in literacy stimulation. moreover, environmental effects of parental

education are smaller than what might be expected intuitively. Genetic factors

account for about two-thirds of individual differences in parental education

(Rowe, Vesterdal, & Rodgers, 1998) as well as in reading attainment (Byrne et

al., 2009; Petrill et al., 2006; Taylor et al., 2010). Therefore, the intergenerational

relation between parental education and their offspring’s reading ability might

well be largely genetically mediated. In line with the given explanations, Walker,

Petrill, and Plomin (2005) found that school achievement shows substantial

genetic influence (i.e., 69%). Shared environmental influences accounted for

only 12% of the variance. While controlling for genetics, most of these shared

environmental influences were due to effects of social-economic status

(parental education and occupation) rather than school characteristics.

It should be borne in mind, though, that the high heritability of

reading performance does not imply at all that educational improvements

are pointless. Instead, they positively impact on almost all children’s reading

achievement and raise the average of standardized scores of a class receiving

effective reading intervention. Nonetheless, it is likely that individual differences

among children remain largely genetically driven (Olson, Byrne, & Samuelsson,

2009). This suggests that children with a genetic constraint on their reading

development need increased reading instruction, a notion that I will come

back to in Section 6.5.


Chapter 6

150

6.3.2 Parental SkillsThe key innovating factor of the present family-risk studies is probably the

inclusion of cognitive abilities of the parents. We went beyond using parental

literacy for the sole purpose of dichotomizing children into high and low

family-risk samples by examining the relation between reading and reading-

related skills of the parents and reading skills of their children. We had

objective measures of the parents with dyslexia. Although all children in the at-

risk sample have a parent with dyslexia, they might still vary in their degree of

family risk for dyslexia. We tested this in both the Amsterdam and the national

sample by comparing the groups of at-risk children with and without dyslexia

on the reading skills of their affected parent. Since parents pass on their genes

to their offspring and shape their environment, parental reading skills might be

taken as an indicator of the offspring’s liability to dyslexia.

In the Amsterdam sample (Chapter 2) the dyslexia of the parents of

the affected children was more severe than the dyslexia of the parents of the

unaffected children. This was indicated by group differences on reading fluency

of words and pseudowords, with effect sizes of 0.78 and 1.71, respectively. This

is an intriguing finding, because the affected parents read on average at the

fifth percentile compared to national norms. Yet in this restricted range group

large differences were observable.

Employing data of the national sample (Chapter 3), we had the

opportunity to replicate the above findings, and in addition, to extend them by

investigating parental spelling, nonword repetition, and rapid naming. As we

lacked a direct measure of phonological awareness, nonword repetition was

used as a proxy. The difference between the at-risk children with and without

dyslexia was replicated for the affected parent’s word-reading fluency, although

with a smaller effect size (Cohen’s d was 0.48). The two at-risk groups did not

differ in parental pseudoword reading, spelling, and nonword repetition, though

both groups were impaired compared to controls. Interestingly, however, the

parents of the at-risk dyslexia children were slower on alpha-numeric rapid

naming than those of the at-risk no-dyslexia children. This underscores the

special role of rapid naming, at least in transparent orthographies. In regression

analyses parental word-reading fluency and rapid naming also predicted at-risk

children’s word-reading fluency, even after controlling for children’s concurrent

phonological awareness and rapid naming.

In the two above mentioned studies data were reported of the


General Discussion

151

6

parent with dyslexia. Chapter 4 (about the national sample) completes this

by examining the influence of the parent without dyslexia for the first time.

It was unfortunate that we did not have objective measures of the literacy

skills of the parent without dyslexia; we only had their report on whether they

experienced literacy difficulties. Nevertheless, this measure appeared to be

a very good indicator of reading fluency: in a subsample we found a strong

correlation (.85) between assessed and self-reported literacy skills. In addition,

this measure yielded interesting findings. As hypothesised, also for the non-

dyslexic parents there was a difference between the two at-risk groups: the

parents of the affected children reported more literacy difficulties compared to

those of the unaffected children. moreover, the literacy difficulties of the non-

dyslexic parent explained additional variance in children’s reading fluency after

controlling for literacy difficulties of the dyslexic parents.

Thus, the findings in the Amsterdam sample regarding the affected

parent were only partly replicated in the national sample. however, the results

concerning the unaffected parent further support the conclusion that children

at family risk for dyslexia differ in their liability, as indicated by differences in

parental reading skills between at-risk children with and without dyslexia.

moreover, differences between the two family-risk groups in the severity

of the dyslexia of the affected parent have now been replicated in Finnish,

in the Jyväskylä Longitudinal Study of Dyslexia. Torppa, Eklund, van Bergen,

and Lyytinen (2011) showed differences in parental reading fluency, accuracy,

and spelling. Regression analyses showed that these measures also predicted

children’s literacy skills in Grade 3; parental literacy skills even predicted

children’s reading accuracy over and above children’s preliteracy skills.

Do the findings regarding precursors in families lend support for the

multiple deficit model? According to this model, the aetiology of dyslexia (and

other developmental disorders) is multifactorial and probabilistic. multiple

genetic risk variants interact with each other and with multiple environmental

risks to ultimately produce the disorder at the behavioural level. Regarding

environmental risks, we investigated aspects of one environmental factor:

home literacy environment. however, we did not find effects on children’s

reading outcome. Children’s social-economic status (as indexed by parental

level of education) did differ among groups, arguably asserting influence on

children’s reading development via environmental as well as genetic pathways

(see Section 6.3.1).


Chapter 6

152

Genetic risk factors were not measured directly. Although there is

now a huge body of evidence indicating that genes contribute importantly to

individual differences in reading ability (e.g., Byrne et al., 2009; hayiou-Thomas

et al., 2010), the specific gene variants found thus far only explain a tiny

part of these differences (see Bishop, 2009, for the example of the KIAA0319

gene), despite substantive work in the field of molecular genetics. This

phenomenon also applies to other common traits and is called the mystery

of the missing heritability (see e.g., manolio et al., 2009). Genetic screening

is therefore not (yet) informative about a child’s genetic liability to dyslexia

(Bishop, 2009). Nevertheless, we proposed (see General Introduction, Section

1.4.2) that something can be said about children’s liability. Since parents pass

on their genetic material to their offspring and shape their environment,

cognitive abilities of parents could be used as an indicator of the genetic and

environmental risk and protective factors in the multiple deficit model.

The current studies provide two kinds of support for parental skills

being an indicator for children’s liability. First, as in other family-risk studies,

two samples of children were recruited based on having or not having a parent

with dyslexia. The current and previous family-risk studies found a large effect

of having a family history on children’s risk of becoming dyslexic. For example,

in Chapter 3 it was found that the rate of dyslexia was 30% in the high-risk

group and only 3% in the low-risk groups, whereas by definition 10% of an

unselected sample would meet the employed dyslexia criterion. Thus, having

a parent with dyslexia increases the risk (and having a second parent with

literacy difficulties increases the risk even more, see Chapter 4). Conversely,

having average to good reading parents decreases the risk.

Secondly, within the at-risk sample it was found that affected and

unaffected parents of the affected children had more literacy problems than

those of the unaffected children. moreover, when considering at-risk children’s

reading fluency on a continuous scale (rather than having or not having

dyslexia), parental reading skills were significant predictors of children’s reading

skills. The current investigation of the impact of parental skills on children’s skills

are more fine-grained than in previous family-risk studies: Previous family-risk

studies only assessed reading skills of the parents (or even relied on self-report

of dyslexia) for the purpose of dichotomising their sample into children with

and without family risk (Boets et al., 2010; de Bree, Wijnen, & Gerrits, 2010;

Pennington & Lefly, 2001; Scarborough, 1990; Snowling, Gallagher, & Frith,


General Discussion

153

6

2003; Torppa, Lyytinen, Erskine, Eklund, & Lyytinen, 2010 - but see Snowling,

muter, & Carroll, 2007, for an exception). In contrast, we studied parent-child

resemblance for both parents for reading and reading-related skills.

Our findings support the view that skills of parents indicate their

offspring’s liability, which in itself is the combination of all genetic and

environmental factors that affect reading development. Thus, parental skills

might shed light on the aetiological level in the multiple deficit model. It is

not possible to disentangle the genetic and environmental contribution to the

intergenerational transmission of skills. however, according to our data the

transmission of risk seems to be mainly via genes. It is important to note that

genes are inherited, not phenotypic traits. Thus, although our data reinforce

the view that parental skills are indicative of their offspring’s liability, parental

skills will never completely specify it.

6.4 The Intergenerational Multiple Deficit Model

In the discussion of Chapter 4 an extension of the multiple deficit model was

proposed: the intergenerational multiple deficit model. Below I will elaborate

on this model. The model is depicted in Figure 6.1. In the figure it can be seen

that a top layer is added to Pennington’s multiple deficit model (see Chapter

1, Section 1.2.1), which represents characteristics of parents. Note that again

influences between layers are omitted from the figure.

Cognitive abilities of parents, for instance reading ability, form part

of their phenotype (PT in Figure 6.1). Their phenotype is the result of their

genotype (GT) in interaction with their environment. Genes do not code

for cognitive and behavioural traits but for the structure of proteins and the

regulation of gene expression, which in highly complex ways and in interaction

with the environment guides the building and maintenance of the brain (Fisher

& Francks, 2006; Fisher, 2006). Despite this gap between genes and cognition,

for traits that show genetic influences in behavioural genetic studies there

must be a relationship between genotypic and phenotypic variation. In other

words, for heritable traits parental phenotype is a proxy for their genotype.

As both parents pass on half of their genes to their offspring, the genotype

of both of them determines the genotype of their offspring. It follows that

the phenotype of parents must be related to some extent to the genotype


Chapter 6

154

of children, which includes children’s genetic risk and protective factors for a

particular developmental disorder.

In addition to transmission of parental skills via genetic pathways,

parental skills could be passed on via environmental pathways. Parents largely

shape their children’s childhood environment, which creates a relation between

parents’ characteristics and children’s environment. For example, good reading

parents are more likely to spend a lot of time reading (Chapter 4), thereby

providing a role model to their children. moreover, they appear to be better

educated (Chapter 2 and 3), and as a result, might live in better neighbourhoods

and might send their children to better achieving schools. hence, the cognitive

phenotype of parents could be one of the factors that determines children’s

environmental risk and protective factors. however, other aspects of the

phenotype of parents might also influence children’s reading development.

For instance, the behaviour of parents and the interaction between them

determines how structured or chaotic the household is, something that has

been shown to be related to children’s school performance (hanscombe,

haworth, Davis, Jaffee, & Plomin, 2011). In conclusion, the phenotype of

parents must also be related to a certain degree to children’s environmental

risk and protective factors.

Given the above two lines of reasoning, the phenotype of parents is

informative about children’s genetic and environmental factors. Focussing on

developmental disorders, this suggests that certain aspects of the phenotype

of parents can inform us about a child’s liability to a particular developmental

disorder. Regarding dyslexia, the aspects of the phenotype of parents that are

expected to shed light on children’s liability to dyslexia are skills in accurate and

fluent reading, spelling, and their cognitive underpinnings like phonological

awareness and rapid naming. Related skills (such as language and arithmetic)

and their underlying cognitive abilities might also play a role. The ability of

parents on each of the relevant continua can be conceptualised as a position

in multivariate space. The position of father and mother in multivariate space

is proposed to be indicative of a child’s predisposition towards dyslexia. The

intergenerational transfer of skills from parents to children goes through

genes, environment, and their interaction. Family studies can point to the

skills that show intergenerational correlations; genetically sensitive studies can

ultimately disentangle the contributions of genetic and environmental effects

to such an intergenerational correlation.


general diScuSSion

155

6


Cognitive processes

Neural systems


Level of analysis

N1 N

2 N

3

C1 C

2 C

3

Figure 6.1. The intergenerational multiple deficit model (an extension of Pennington’s multiple deficit model). Double headed arrows indicate interactions. Causal connections between levels of analyses are omitted. GT

p1 = genotype of parent 1; PT

p1 = phenotype of parent 1

G = genetic risk or protective factor; E = environmental risk or protective factor N = neural system C = cognitive process

D = disorder

G1 E

1 G

2

D1 D

2 D

3

Parental influences GTp1

PTp1

PTp2

GTp2

Ch

ild

Figure 6.1. The intergenerational multiple defi cit model (an extension of Pennington’s multiple defi cit model). Double headed arrows indicate interactions. Causal connections between levels of analyses are omitted.

GTp1 = genotype of parent 1; PTp1 = phenotype of parent 1G = genetic risk or protective factor; E = environmental risk or protective factorN = neural systemC = cognitive processD = disorder


Chapter 6

156

The intergenerational multiple deficit model is inspired by the

current studies on dyslexia, but is generally applicable to other multifactorial

developmental disorders with a genetic component. Examples of such disorders

include attention-deficit/hyperactivity disorder, developmental coordination

disorder, dyscalculia, specific language disorder, and autism spectrum disorder.

With respect to autism spectrum disorder, a number of studies (e.g., Bölte &

Poustka, 2006; happé, Briskman, & Frith, 2001; Losh et al., 2009) have studied

the cognitive phenotype of parents of probands (as opposed to children of

probands, as in family-risk studies of dyslexia) and found in parents similar but

milder impairments as in their children, indicating parent-child resemblance.

A second example concerns specific language impairment. In a recent study

(Bishop, hardiman, & Barry, 2012), language skills of probands and their parents

were found to correlate. In another recent study (Bishop et al., 2012) it was

found that a parent’s non-word repetition ability was a predictor of whether the

child would develop specific language impairment. These examples provide

evidence of intergenerational transfer of cognitive skills other than reading.

6.5 Future Directions

The present collection of studies has yielded some answers, but meanwhile has

raised new questions. To start with, our and previous family-risk studies did not

show significant effects of the home literacy environment on children’s reading

outcome, but behavioural genetic studies do find shared and non-shared

environmental influences. Is the home literacy environment important but did

we fail to find this because of other reasons (see Section 6.3.1)? Or are other

environmental factors more important, like pre- and perinatal factors, peer

influences, or characteristics of the school and classroom? Genetic sensitive

studies can estimate how ‘environmental’ these factors actually are.

A second interesting area for future studies regards comorbidity. From

the (intergenerational) multiple deficit model it follows that comorbidities

between developmental disorders are to be expected. This points to the

importance of studying developmental disorders that commonly co-occur.

By doing so, one can uncover shared and distinct risk factors at each of the

levels of explanation. This not only helps to understand the origin of the

comorbidity, but also the developmental paths leading to each of the disorders.

N2

N3

E1

G2

D2

D3


General Discussion

157

6

The intergenerational multiple deficit model is inspired by the

current studies on dyslexia, but is generally applicable to other multifactorial

developmental disorders with a genetic component. Examples of such disorders

include attention-deficit/hyperactivity disorder, developmental coordination

disorder, dyscalculia, specific language disorder, and autism spectrum disorder.

With respect to autism spectrum disorder, a number of studies (e.g., Bölte &

Poustka, 2006; happé, Briskman, & Frith, 2001; Losh et al., 2009) have studied

the cognitive phenotype of parents of probands (as opposed to children of

probands, as in family-risk studies of dyslexia) and found in parents similar but

milder impairments as in their children, indicating parent-child resemblance.

A second example concerns specific language impairment. In a recent study

(Bishop, hardiman, & Barry, 2012), language skills of probands and their parents

were found to correlate. In another recent study (Bishop et al., 2012) it was

found that a parent’s non-word repetition ability was a predictor of whether the

child would develop specific language impairment. These examples provide

evidence of intergenerational transfer of cognitive skills other than reading.

6.5 Future Directions

The present collection of studies has yielded some answers, but meanwhile has

raised new questions. To start with, our and previous family-risk studies did not

show significant effects of the home literacy environment on children’s reading

outcome, but behavioural genetic studies do find shared and non-shared

environmental influences. Is the home literacy environment important but did

we fail to find this because of other reasons (see Section 6.3.1)? Or are other

environmental factors more important, like pre- and perinatal factors, peer

influences, or characteristics of the school and classroom? Genetic sensitive

studies can estimate how ‘environmental’ these factors actually are.

A second interesting area for future studies regards comorbidity. From

the (intergenerational) multiple deficit model it follows that comorbidities

between developmental disorders are to be expected. This points to the

importance of studying developmental disorders that commonly co-occur.

By doing so, one can uncover shared and distinct risk factors at each of the

levels of explanation. This not only helps to understand the origin of the

comorbidity, but also the developmental paths leading to each of the disorders.

N2

N3

E1

G2

D2

D3

In examining the specificity of precursors for dyslexia we included arithmetic

and dyscalculia, but comorbidities with other developmental disorders were

not considered. Including more than just a single (dis)ability in future work will

enhance our understanding. The same applies for studying intergenerational

transfer: parents might confer risks in different domains, like reading, language,

and attention. Therefore, future studies are needed to test whether a more

complete picture of parents’ cognitive and behavioural profile gives a more

reliable indication of their offspring’s liability to a particular disorder.

The third future direction concerns the relation between parental

and children’s skills in an unselected sample. The findings reported in this

thesis on parent-child resemblance might not be generalizable to the normal

population since we had selected samples of high and low family risk. Building

on the reported studies, we are currently conducting a study that aims to

investigate relations between parents’ and children’s skills across the entire

range of reading abilities. For that purpose we test whole families in NEmO

– the science museum in Amsterdam – on reading, arithmetic, rapid naming

and phonological awareness, using the same tests for all family members. In

addition, environmental factors are assessed via questionnaires and genetic

factors via saliva samples. Over 600 participants have been tested so far. Their

data allows us to examine intergenerational transfer of skills and to study the

cognitive, environmental, and genetic factors that influence reading ability

and disability.

Finally, our studies have practical implications which warrant further

investigation. The current and previous family-risk studies suggest that children

who have a combination of weak-reading parents and weak preliteracy skills in

kindergarten run the highest risk of developing dyslexia. This group of children

is likely to have a high genetic endowment for dyslexia. This suggests that for

this group classroom instruction in reading is not sufficient. We are currently

investigating whether these children can be assisted by offering a reading

intervention, consisting of regular extra practice from mid-kindergarten till

mid-Grade 2. Preliminary results (Zijlstra, van Bergen, Koomen, & van der Leij,

2012) confirm that these children are most susceptible for dyslexia, but that

early intervention can indeed ameliorate the impact of their high liability,

provided that the intervention is intensive and individually tailored.


Chapter 6

158

6.6 Concluding Remarks

The present thesis has characterized the cognitive profile of children with a

family background of dyslexia that do and do not become dyslexic readers.

This has revealed precursors of dyslexia, and by including arithmetic outcome,

the specificity of these precursors. In addition, it has shed light on the

cognitive profile of their parents, which provided a start in understanding

the intergenerational transfer of reading ability and disability. Together, the

findings on child and parental level have advanced our understanding of the

factors influencing the liability and the development of dyslexia. however,

more research is needed to move the science of reading towards elucidating

the causal chain that leads to the behavioural manifestation of dyslexia.


General Discussion

159

6

6.7 References

Bishop, D. V. m. (2009). Genes, cognition, and communication. Annals of the New York Academy of Sciences, 1156 (The Year in Cognitive Neuroscience 2009), 1-18. doi: DOI: 10.1111/j.1749-6632.2009.04419.x

Bishop, D. V. m., hardiman, m. J., & Barry, J. G. (2012). Auditory deficit as a consequence rather than endophenotype of specific language impairment: Electrophysiologi-cal evidence. PloS One, 7(5), e35851.

Bishop, D. V. m., holt, G., Line, E., mcDonald, D., mcDonald, S., & Watt, h. (2012). Paren-tal phonological memory contributes to prediction of outcome of late talkers from 20 months to 4 years: A longitudinal stu-dy of precursors of specific language im-pairment. Journal of Neurodevelopmental Disorders, 4(1), 3.



Bölte, S., & Poustka, F. (2006). The broader cognitive phenotype of autism in parents: how specific is the tendency for local processing and executive dysfunction? Journal of Child Psychology and Psychiatry, 47(6), 639-645.

Burgess, S. R., & Lonigan, C. J. (1998). Bidirec-tional relations of phonological sensitivity and prereading abilities: Evidence from a preschool sample. Journal of Experimen-tal Child Psychology, 70(2), 117-141. doi: 10.1006/jecp.1998.2450

Byrne, B., Coventry, W. L., Olson, R. K., Sa-muelsson, S., Corley, R., Willcutt, E. G., . . . DeFries, J. C. (2009). Genetic and en-vironmental influences on aspects of li-teracy and language in early childhood: Continuity and change from preschool to Grade 2. Journal of Neurolinguistics, 22(3), 219-236.

Coltheart, m., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: A dual route cascaded model of visual word recogni-tion and reading aloud. Psychological Re-view, 108(1), 204.

de Bree, E., Wijnen, F., & Gerrits, E. (2010). Non-word repetition and literacy in Dutch children at-risk of dyslexia and children with SLI: Results of the follow-up study. Dyslexia, 16(1), 36-44.

de Jong, P. F. (1998). Working memory defi-cits of reading disabled children. Journal of Experimental Child Psychology, 70(2), 75-96. doi: 10.1006/jecp.1998.2451



De Smedt, B., & Boets, B. (2010). Phono-logical processing and arithmetic fact retrieval: Evidence from developmental dyslexia. Neuropsychologia, 48(14), 3973-3981.

De Smedt, B., Taylor, J., Archibald, L., & Ansari, D. (2010). how is phonological processing related to individual diffe-


Chapter 6

160

rences in children’s arithmetic skills? De-velopmental Science, 13(3), 508-520. doi: 10.1111/j.1467-7687.2009.00897.x


Fisher, S. E. (2006). Tangled webs: Tracing the connections between genes and cognition. Cognition, 101(2), 270-297. doi: 10.1016/j.cognition.2006.04.004

Fisher, S. E., & Francks, C. (2006). Genes, cog-nition and dyslexia: Learning to read the genome. Trends in Cognitive Sciences, 10(6), 250-257. doi: 10.1016/j.tics.2006.04.003

Galaburda, A. m., LoTurco, J., Ramus, F., Fitch, R. h., & Rosen, G. D. (2006). From genes to behavior in developmental dyslexia. Na-ture Neuroscience, 9(10), 1213-1217.

hanscombe, K. B., haworth, C., Davis, O. S. P., Jaffee, S. R., & Plomin, R. (2011). Chaotic homes and school achievement: A twin study. Journal of Child Psychology and Psy-chiatry,

happé, F., Briskman, J., & Frith, U. (2001). Exploring the cognitive phenotype of au-tism: Weak “central coherence” in parents and siblings of children with autism: I. Ex-perimental tests. Journal of Child Psycho-logy and Psychiatry, 42(3), 299-307.

hayiou-Thomas, m. E., harlaar, N., Dale, P. S., & Plomin, R. (2010). Preschool speech, language skills, and reading at 7, 9, and 10 years: Etiology of the relationship. Journal of Speech, Language, and Hearing Rese-arch, 53(2), 311-332. doi: 10.1044/1092-4388(2009/07-0145)

hulme, C., Bowyer-Crane, C., Carroll, J. m., Duff, F. J., & Snowling, m. J. (in press). The causal role of phoneme awareness and letter-sound knowledge in learning to read combining intervention studies with mediation analyses. Psychological Science,

Kovas, Y., haworth, C. m. A., Dale, P. S., & Plomin, R. (2007). The genetic and en-vironmental origins of learning abilities and disabilities in the early school years. Monographs of the Society for Research in Child Development, 72(3), 1-160. doi: 10.1111/j.1540-5834.2007.00453.x


Leseman, P. P. m., Scheele, A. F., mayo, A. Y., & messer, m. h. (2007). home literacy as a special language environment to prepare children for school. Zeitschrift Für Erzie-hungswissenschaft, 10(3), 334-355.

Losh, m., Adolphs, R., Poe, m. D., Couture, S., Penn, D., Baranek, G. T., & Piven, J. (2009). Neuropsychological profile of autism and the broad autism phenotype. Archives of General Psychiatry, 66(5), 518.

manolio, T. A., Collins, F. S., Cox, N. J., Gold-stein, D. B., hindorff, L. A., hunter, D. J., . . . Chakravarti, A. (2009). Finding the missing heritability of complex diseases. Nature, 461(7265), 747-753.

Naples, A. J., Chang, J. T., Katz, L., & Grigoren-ko, E. L. (2009). Same or different? insights into the etiology of phonological aware-ness and rapid naming. Biological Psycho-logy, 80(2), 226-239.





General Discussion

161

6

Petrill, S. A., Deater-Deckard, K., Thompson, L. A., De Thorne, L. S., & Schatschnei-der, C. (2006). Reading skills in ear-ly readers: Genetic and shared en-vironmental influences. Journal of Learning Disabilities, 39(1), 48-55. doi: 10.1177/00222194060390010501

Plomin, R., & Kovas, Y. (2005). Generalist ge-nes and learning disabilities. Psychological Bulletin, 131(4), 592-617.

Rowe, D. C., Vesterdal, W. J., & Rodgers, J. L. (1998). herrnstein’s syllogism: Genetic and shared environmental influences on IQ, education, and income. Intelligen-ce, 26(4), 405-423. doi: 10.1016/s0160-2896(99)00008-2


Seidenberg, m. S., & mcClelland, J. L. (1989). A distributed, developmental model of word recognition and naming. Psycholo-gical Review; Psychological Review, 96(4), 523.

Siegel, L. S. (1999). Issues in the definition and diagnosis of learning disabilities. Journal of Learning Disabilities, 32(4), 304-319. doi: 10.1177/002221949903200405


Snowling, m. J., muter, V., & Carroll, J. (2007). Children at family risk of dys-lexia: A follow-up in early adolescence. Journal of Child Psychology and Psychia-try, 48(6), 609-618. doi: 10.1111/j.1469-7610.2006.01725.x

Taylor, J., Roehrig, A. D., hensler, B. S., Connor, C. m., & Schatschneider, C. (2010). Teacher quality moderates the genetic effects on early reading. Science, 328(5977), 512-514. doi: 10.1126/science.1186149

Torppa, m., Eklund, K., van Bergen, E., & Lyytinen, h. (2011). Parental literacy pre-dicts children’s literacy: A longitudinal family-risk study. Dyslexia, 17(4), 339-355. doi: 10.1002/dys.437




Vellutino, F. R., Fletcher, J. m., Snowling, m. J., & Scanlon, D. m. (2004). Specific reading disability (dyslexia): What have we learned in the past four decades? Journal of Child Psychology and Psychiatry, 45(1), 2-40.

Walker, S. O., Petrill, S. A., & Plomin, R. (2005). A genetically sensitive investigation of the effects of the school environment and socio-economic status on acade-mic achievement in seven-year-olds. Educational Psychology, 25(1), 55-73. doi: 10.1080/0144341042000294895

Zijlstra, A. h., van Bergen, E., Koomen, h., & van der Leij, A. (2012). Response-to-inter-vention in low preliterate children with and without familial risk for dyslexia. Six-th European Graduate School on Literacy Acquisition. Egmond aan Zee, the Nether-lands.


The greater becomes the volume of our sphere of knowledge,

the greater also becomes its surface of contact with the unknown.

- Jules Sageret -

Summary / Samenvatting


Summary / Samenvatting


164

Who will develop dyslexia? Cognitive precursors in parents and children

Children with dyslexia are characterized by slow and laborious reading. Dyslexia

negatively affects educational and occupational attainment. Therefore, it is

important to increase our understanding of the causes of dyslexia.

Studies comparing mono- and dizygotic twins have revealed that

dyslexia is moderately to highly heritable. This also fits with the observation

that dyslexia tends to run in families. Family-risk studies employ this fact. Such

studies follow the development of at-risk children: children with a family history

of dyslexia. After a few years of reading instruction, the at-risk children who

have developed dyslexia can be identified. Subsequently, these children can

be compared in retrospect to at-risk children without dyslexia and to typically

developing children without family risk. Did the children with later dyslexia

already demonstrate cognitive deficits before they started to learn to read? In

this way family-risk studies enable the investigation of precursors of dyslexia.

Several models have been formulated in which dyslexia is attributed

to a single cognitive deficit, but none of these models can readily explain

the range of deficits associated with dyslexia. Pennington has developed

a model in which developmental disorders (like dyslexia) are caused by a

combination of underlying cognitive factors, which can be affected to different

degrees. his model –the multiple deficit model– is depicted in Figure 1.1 in

Chapter 1. Some of these cognitive deficits are specific and some are shared

with other developmental disorders. The current thesis studied the specificity

of precursors of reading ability by investigating whether they also predict

arithmetic ability.

Inspired by the results of aetiological studies, the multiple deficit model

includes genetic and environmental factors that influence the probability of

developing a disorder; in this case, dyslexia. From this multifactorial aetiology

if follows that dyslexia is not an all-or-none condition, but instead that the

liability or risk of developing dyslexia is continuously distributed. In the present

thesis it is argued that a family-risk study offers the opportunity to test this

implication. Given that parents pass on their genes to their offspring and create

their childhood environment, it is expected that reading skills of parents can be

seen as indicators for children’s liability to dyslexia.


165

Summary

In short, the aims of the current thesis are threefold:

1. To study children’s cognitive precursors of dyslexia (and their specificity)

before reading instruction begins;

2. To test whether the cognitive profile of parents is indicative of a child’s

liability or predisposition to dyslexia;

3. To contribute to testing, specifying, and extending the multiple deficit

model.

As mentioned above, we employed a family-risk design to pursue these aims.

The reported studies are part of the Dutch Dyslexia Programme. Children

were considered at high risk if they had at least one parent and another family

member with dyslexia. The parents of the low-risk children were both average

to good readers. Two independent samples were followed, referred to as the

Amsterdam (N = 79) and the national (N = 212) sample. In both samples the

analyses consisted of, among others, comparing the following three groups: 1)

at-risk children with dyslexia, 2) at-risk children without dyslexia, and 3) control

children (i.e., low-risk children without dyslexia). Roughly one-third of the at-

risk children developed dyslexia.

In the Amsterdam sample (Chapter 2) children were followed from

kindergarten through Grade 5. The at-risk children who went on to have

dyslexia had poor letter knowledge in kindergarten. In addition, they were

impaired on a cognitive skill called rapid naming. Rapid naming is the ability

to rapidly retrieve the name of a symbol from long-term memory, measured

by naming a list of familiar items, like colours or pictures. The at-risk children

without dyslexia performed better on letter knowledge and rapid naming,

but still slightly below the level of controls. The same pattern was found

for subsequent reading development. A key finding related to the reading

difficulties of the parents with dyslexia: those of the affected children were

more severely impaired than those of the unaffected children.

Chapters 3 to 5 are devoted to the national sample, reporting the

development of 4 to 9 years of age. IQ was assessed at the age of 4 and

preliteracy skills at the end of kindergarten. The investigated preliteracy

skills included letter knowledge, rapid naming, and phonological awareness.

Phonological awareness is the ability to detect and manipulate sounds in

spoken words. In kindergarten this can be measured with items like “What are

the sounds in ‘cat’?” Answer: “/k/ /a/ /t/”. Compared to controls, at-risk children


166

with later dyslexia were impaired across IQ and preliteracy tasks. The other at-

risk children showed mild deficits in verbal IQ and phonological awareness, but

not in nonverbal IQ, letter knowledge, and rapid naming. Relations between

these precursors and later reading and arithmetic revealed that nonverbal IQ

and rapid naming were related to both outcome measures, whereas the other

precursors were more specific for reading. At-risk dyslexia children continued

to show deficits in phonological awareness and rapid naming in second and

third grade, alongside impairments in literacy and arithmetic. At this age, the

at-risk no-dyslexia children had weaknesses in phonological awareness and

literacy, although they scored within normal limits. Importantly, the finding

in the Amsterdam sample regarding differences between the two at-risk

groups in parental reading could be replicated and extended: also parental

rapid naming differentiated the groups, as did literacy difficulties of the parent

without dyslexia. In both the Amsterdam and the national sample there was

no clear evidence of effects of home literacy environment on children’s reading

outcome.

Chapter 6 gives an extensive summary of the findings and discusses

theoretical implications. To return to the multiple deficit model, one of

the predictions of this model is that the liability distribution for a given

developmental disorder is continuous. Both the findings at the child-level and

parent-level support this prediction. It is argued that cognitive skills of parents

are an indicator of their offspring’s position on the liability continuum. Chapter

6 works towards specifying the multiple deficit model for the case of dyslexia

and concludes by presenting an extended model: the intergenerational

multiple deficit model (see Figure 6.1), which includes the effect of parental

characteristics. Finally, recommendations for future research are given.


167

Samenvatting

Wie wordt dyslectisch? Cognitieve voorlopers in ouders en kinderen

Bij kinderen met dyslexie gaat het lezen moeizaam en traag. Dyslexie heeft

een negatieve invloed op zowel schoolprestaties als het beroepsniveau. het is

daarom van belang om de oorzaken van dyslexie beter te begrijpen.

Onderzoek met mono- en dizygote tweelingen heeft uitgewezen dat

dyslexie in sterkte mate erfelijk is. Dit past ook bij de waarneming dat dyslexie

binnen bepaalde families veel voorkomt. Familiaire risico studies maken gebruik van

dit feit. Zulke studies volgen de ontwikkeling van risicokinderen: kinderen bij wie

dyslexie in de familie voorkomt. Wie van deze risicokinderen dyslexie ontwikkeld

hebben kan worden vastgesteld nadat ze een paar jaar leesonderwijs hebben

gehad. Vervolgens kunnen deze kinderen retrospectief worden vergeleken met

risicokinderen zonder dyslexie en met zich normaal ontwikkelende kinderen die

geen familiair risico dragen. Toonden de kinderen met latere dyslexie al cognitieve

tekorten voordat ze leerden lezen? Op deze manier maken familiaire risico studies

het mogelijk voorlopers van dyslexie op te sporen.

Er zijn verschillende modellen geformuleerd waarin dyslexie

toegeschreven wordt aan een enkelvoudig cognitief tekort, maar geen

van deze modellen kan het brede scala van met dyslexie geassocieerde

kenmerken verklaren. Pennington heeft een model ontwikkeld waarin

ontwikkelingsstoornissen (zoals dyslexie) verklaard worden uit een combinatie

van onderliggende cognitieve factoren waarop een individu in meer of

mindere mate uitvalt. Zijn model –het multiple deficit model– is afgebeeld

in Figuur 1.1 in hoofdstuk 1. Sommige van die cognitieve tekorten zijn

specifiek voor één bepaalde stoornis en sommige worden gedeeld met

andere ontwikkelingsstoornissen. In dit proefschrift wordt de specificiteit van

voorlopers van leesvaardigheid bestudeerd door te onderzoeken in hoeverre

zij ook rekenvaardigheid voorspellen.

Gebaseerd op bevindingen van etiologische studies bevat het multiple

deficit model genetische en omgevingsfactoren die de kans op het ontwikkelen

van een stoornis beïnvloeden, in dit geval dyslexie. Uit deze multifactoriële

etiologie volgt dat dyslexie niet een alles-of-niets conditie is, maar dat de

vatbaarheid voor, of het risico op het ontwikkelen van dyslexie continu verdeeld

is. In het huidige proefschrift wordt beargumenteerd dat een familiaire risico

studie de mogelijkheid biedt om deze implicatie te toetsen. Aangezien ouders


168

hun genen doorgeven aan hun kinderen en de omgeving van hun kinderen

creëren, valt het te verwachten dat de leesvaardigheid van ouders kan worden

gezien als een indicator voor het risico op dyslexie dat kinderen lopen.

In het kort zijn de doelen van dit proefschrift:

1. het onderzoeken van cognitieve voorlopers van dyslexie (en hun

specificiteit) voordat het leesonderwijs begint.

2. Toetsen of het cognitieve profiel van ouders indicatief is voor het risico

dat een kind loopt op dyslexie.

3. Bijdragen aan het toetsen, specificeren en uitbreiden van het multiple

deficit model.

Zoals hierboven genoemd maakten we gebruik van een familiair risico design

om deze doelen na te streven. De gerapporteerde studies zijn onderdeel van het

Dutch Dyslexia Programme1. Als kinderen ten minste één dyslectische ouder en

een ander dyslectisch familielid hadden werden ze als hoog risico beschouwd.

De ouders van de laag-risico kinderen waren beiden gemiddelde tot goede

lezers. Twee onafhankelijke steekproeven werden gevolgd: de Amsterdamse (N

= 79) en de nationale steekproef (N = 212). In beide steekproeven werden drie

groepen vergeleken: 1) risicokinderen met dyslexie, 2) risicokinderen zonder

dyslexie, en 3) controle kinderen (d.w.z., kinderen met een laag risico en zonder

dyslexie). Ongeveer een derde van de risicokinderen ontwikkelde dyslexie.

In de Amsterdamse steekproef (hoofdstuk 2) werden de kinderen van

groep 2 tot en met groep 7 gevolgd. De risicokinderen die later dyslectisch

werden hadden als kleuter een zwakke letterkennis. Daarnaast scoorden ze

zwak op een cognitieve maat genaamd snelbenoemen. Snelbenoemen is de

vaardigheid om snel namen van symbolen uit het lange termijn geheugen op

te halen. Dit wordt gemeten met het hardop benoemen van een lijst bekende

items, zoals kleuren of plaatjes. De risicokinderen zonder dyslexie presteerden

beter op letterkennis en snelbenoemen, maar wel net iets onder het niveau van

de controle kinderen. Voor de latere leesontwikkeling werd hetzelfde patroon

gevonden. Een belangrijke bevinding had betrekking op de dyslectische

ouders: de ouders van wie de kinderen dyslectisch werden waren zwaarder

dyslectisch dan de ouders van wie de kinderen niet dyslectisch werden.

1 Zie http://www.nwo.nl/nwohome.nsf/pages/NWOA_6ZJGAJ


169

Samenvatting

De hoofdstukken 3, 4 en 5 zijn gericht op de nationale steekproef

en rapporteren de ontwikkeling van 4 tot 9 jaar. Op 4-jarige leeftijd werd het

IQ bepaald en aan het eind van de kleuterschool werden de voorschoolse

vaardigheden in kaart gebracht. De onderzochte voorschoolse vaardigheden

bestonden uit letterkennis, snelbenoemen en fonologisch bewustzijn.

Fonologisch bewustzijn is de vaardigheid om klanken in gesproken woorden

te herkennen en te manipuleren. In de kleuterklas kun je dit meten met vragen

als: “Welke klanken zitten er in ‘kat’?” met als antwoord: “/k/ /a/ /t/”. Vergeleken

met de controle kinderen scoorden de risicokinderen bij wie later dyslexie

werd vastgesteld zwak op alle IQ taken en voorschoolse vaardigheden. De

andere risicokinderen lieten milde tekorten zien op verbaal IQ en fonologisch

bewustzijn, maar niet op nonverbaal IQ, letterkennis en snelbenoemen.

Analyse van de relaties tussen deze voorlopers en het latere lezen en rekenen

wees uit dat nonverbaal IQ en snelbenoemen gerelateerd zijn aan beide

schoolvaardigheden, terwijl de andere voorlopers meer specifiek zijn voor het

lezen. Risicokinderen met dyslexie bleven in groep 4 en 5 problemen houden

met fonologisch bewustzijn en snelbenoemen, en liepen daarnaast achter

met lezen, spellen en rekenen. De risicokinderen zonder dyslexie waren op

die leeftijd matig (maar niet onvoldoende) in fonologisch bewustzijn, lezen

en spellen. Een belangrijk resultaat was dat de bevinding in de Amsterdamse

steekproef van verschillen in leesvaardigheid tussen de dyslectische ouders

van de twee risicogroepen kon worden gerepliceerd en uitgebreid: ook

snelbenoemen van de ouders maakte onderscheid tussen deze groepen.

Daarnaast was een opvallend resultaat dat de ouder zonder dyslexie ook

zwakker las wanneer het kind dyslectisch werd. Wat betreft de invloed van de

omgeving was er echter in zowel de Amsterdamse als de nationale steekproef

geen duidelijk bewijs voor effecten van geletterdheid van de thuisomgeving

op de leesuitkomst van kinderen.

In hoofdstuk 6 worden de resultaten van de verschillende studies

samengevat en worden theoretische implicaties besproken. Om terug te

komen op het multiple deficit model, één van de voorspellingen van dit

model is dat de risicoverdeling voor een gegeven stoornis continu is. Zowel

de bevindingen op kind- als ouderniveau ondersteunen deze voorspelling.

Beargumenteerd wordt dat de cognitieve vaardigheden van ouders indicatief

zijn voor de positie van hun kinderen op het risicocontinuüm. In hoofdstuk

6 wordt toegewerkt naar de specificatie van het multiple deficit model voor


170

dyslexie. Afgesloten wordt met de introductie van een uitgebreider model: het

intergenerationele multiple deficit model (zie Figuur 6.1), waarin de effecten van

ouderkenmerken zijn toegevoegd. Tot slot worden enige aanbevelingen voor

verder onderzoek gegeven.


Wetenschap vergt offers.

- Meneer Hoop, mijn leraar scheikunde –

[Science requires sacrifices. Mr Hoop, my chemistry teacher]

Acknowledgements / Dankwoord


Acknowledgements / Dankwoord


174

Acknowledgements

The most read section of my thesis; I hope it doesn’t disappoint you. There is but

one name on the cover, but many contributed scientifically or non-scientifically

to its culmination.

First Peter. I couldn’t have wished for a better mentor. I appreciate

your approachability, involvement, and critical approach. Thank you for the

stimulating scientific conversations and for your enthusiasm for your area of

research, that has become mine as well. Aryan, thank you for setting up this

unique and large-scale study, for having faith in me, and for assigning me to

your beautiful project. Frans, my thesis would have never come about without

all your statistics teaching. All three of you, thank you for your supervision and

the freedom you gave me to work from Zürich, to follow courses, and to do extra

research projects. maggie, Bart, Ludo, Frank, and Judith, thank you very much

for taking the time out of your busy schedules to review my thesis and be my

opponents. To all parents and children who participate in the Dutch Dyslexia

Programme (DDP): thank you so much for your dedication. You taught us a

great deal about dyslexia. I have recognized your participation by representing

you and your child on the cover.

The research group Learning Problems (OLP) changes in composition,

but remains characterized by diverse personalities that together form

a warm nest and a close team. I miss you guys. I reminisce about the good

time in G0.05, the Coffee Company (our other office), the conferences and

accompanying holidays, and the writing retreats. Eva and Judith acted as my

big OLP-sisters. Judith, thank you for taking care of me. Eva, thanks for being a

role model. We grew up in neighbouring villages and shared as kids the same

recorder teacher, but only really got to know each other in OLP. Unfortunately

we don’t live in neighbouring villages any more, but nowadays you Skype me

weekly from Sydney, no matter whether I am in Amsterdam, York, or Oxford.

marleen and Debora, once we formed the new OLP trio and shared our joys

and sorrows. Debora, many thanks for your support, smile, and working out-

of-hours company. haytske, you are open, full of energy and special, and I like

that. Titia, so much that we share: passion for science, the DDP, and of course

our self-developed NEmO-project. madelon and marloes, my fellow Peter-PhD-

students, it’s great to have you by my side during this nerve wrecking hour and

the preparations. madelon, I enjoy discussing analyses and reading research


175

acknowledgementS

with you. During one of our Coffee Company sessions even a new project was

born, the Em-project, which will shine in your thesis. marloes: great energy.

Although we only had a short period together in OLP, I got to know you well

because of the writing retreat and the Canada holiday. Britt: great chef, fresh

DDP-blood, and NEmO help; super! Anna, Ben, and Evelien, many thanks for

all your DDP work and for answering my questions. Other colleagues that I like

to mention: Dominik and Julia (we started together in the C-building), maaike

(adoptive sister), Anne and Cristina (fellow reading researchers), Suzanne

(juggling mate), and helma, Jantine, Francine and Bettina (social-emotional

bulwark).

moving on to the researchers outside the University of Amsterdam

that contributed to my scientific development. Faculty of human movement

Sciences, thank you for providing good education and the challenges beyond,

and for fuelling scientific passion. mark, Lisa, and Andrew, thanks to you I had

an amazing internship in Aberdeen. mark, thank you for your confidence in

me as a researcher, especially when I had lost mine. minna and JLD, thank you

for initiating our fruitful international collaboration; kiitos. maggie and CRL, I

had a nice and valuable study visit in York. many thanks for your help with my

research proposal. Excellent that we are now both based in Oxford. Robin, we

never run out of things to talk about, research and stats in particular. I have

good memories of your stay with us in Amsterdam and I love being together

in Oxford. Simon, thank you for the exciting multidisciplinary collaboration.

Dorothy and OSCCI, thanks a lot for welcoming me in the research group and

for offering me this excellent opportunity to expand my (scientific) horizon.

Kate and LCD, thank you for the ‘honorary membership’. Oriel College and

Orielenses, I feel privileged to be part of you. It is a sometimes slightly bizarre

but brilliant experience.

For the highly needed mental distraction and physical outlet I was

lucky to have session for 10 hours/week at gymnastics club Olympia and

sport acrobatics club CCO. It was good fun! Thanks to, among others, map,

Guido, our flyers Simone, Tamara and miranda, and of course my base and

sister Anke. Anke, we have been through a lot together and I am proud of you.

Oxford Sirens, so cool to be part of the cheerleading competitive squad. Yet

again somersaults and throwing girls… Professor Van der Werf and his team:

amazing. my life is a present. my family in law, Zoia, Adi, Pierre and Bernadette,

thanks for your support and interest. Brothers Aart and Geert, and sisters Anke


176

and Janneke, I am so happy to have you! Rita, Bart, and Suzanne, it’s great to

have you in the family. mommy and daddy, you always being there for me gives

me such a secure feeling. I love coming home. Yves, you are my soul mate, my

rock; even though it’s again at distance. I can’t live without you.


177

DankwoorD

Dankwoord

Het meest gelezen onderdeel van mijn proefschrift; Ik hoop dat het niet

tegenvalt… Eén naam op de kaft, maar velen die wetenschappelijk – of juist

niet wetenschappelijk – hebben bijgedragen.

Allereerst Peter. Een betere leermeester had ik me niet kunnen wensen.

Ik waardeer je toegankelijkheid, betrokkenheid en kritische benadering.

Bedankt voor de stimulerende wetenschappelijke gesprekken en voor je

enthousiasme voor je onderzoeksgebied, dat nu ook het mijne geworden is.

Aryan, bedankt voor het opzetten van dit unieke en gigantische onderzoek en

voor het vertrouwen dat je in me stelde door mij op je mooie project te zetten.

Frans, zonder al het statistiekonderwijs van jou was dit proefschrift er nooit

gekomen. Alledrie bedankt voor jullie begeleiding en de ruimte en vrijheid

die jullie me hebben gegeven voor het werken vanuit Zürich, het volgen van

cursussen, en het doen van extra onderzoeksprojecten. Maggie, Bart, Ludo,

Frank en Judith, hartelijk bedankt voor de tijd en energie die jullie hebben

gestoken in het beoordelen van mijn proefschrift en in het opponeren. Alle

ouders en kinderen die meedoen aan het Dutch Dyslexia Programme (DDP):

hartelijk bedankt voor jullie inzet. Zonder jullie waren we geen moer wijzer

geworden over dyslexie. Als dank heb ik u en uw kind op de kaft gezet.

De onderzoeksgroep Onderwijsleerproblemen (OLP) wisselt van

samenstelling, maar blijft gekenmerkt door uiteenlopende karakters die samen

een warm nest en een hecht team vormen. Ik mis jullie. Met weemoed denk ik

terug aan de leuke tijd in G0.05, de Coffee Company (ons andere kantoor), de

conferenties met de vakanties eromheen, en de schrijfweken. Eva en Judith

fungeerden als mijn grote OLP-zussen. Judith, bedankt voor je goede zorgen.

Eva, bedankt voor je voorbeeldfunctie. We groeiden op in buurdorpen en hadden

als kind dezelfde blokfluitjuf, maar leerden elkaar bij OLP pas echt kennen.

Helaas wonen we niet meer in buurdorpen, maar nu Skype je me wekelijks

vanuit Sydney, of ik nu in Amsterdam, York, of Oxford ben. Marleen en Debora,

wij vormden ooit het nieuwe trio bij OLP en deelden lief en leed. Debora, dank

je wel voor je steun, glimlach en overwerkgezelligheid. Haytske, je bent open,

energiek en bijzonder en dat mag ik graag. Titia, zoveel dat we delen: passie

voor wetenschap, het DDP, en natuurlijk ons zelf opgezette NEMO-project.

Madelon en Marloes, mijn mede Peter-promovenda’s, heel fijn dat jullie mij

terzijde willen staan tijdens dat zenuwslopende uur en de aanloop ernaartoe.


178

madelon, heerlijk om me jouw analyses en leesonderzoek te bediscussiëren.

Tijdens één van onze gesprekken in de Coffee Company werd zelfs een nieuw

onderzoek geboren, het Em-project, dat straks in jouw proefschrift schittert.

marloes: heerlijk energiek. In de korte tijd samen in OLP heb ik je toch goed

leren kennen dankzij de schrijfweek en Canadavakantie. Britt: topkok, vers

DDP-bloed en NEmO-hulp; super! Anna, Ben en Evelien, bedankt voor al jullie

werk aan het DDP en het beantwoorden van al mijn vragen. Andere collega’s

die ik graag wil noemen: Dominik en Julia (gedrie gestart in het C-gebouw),

maaike (adoptiezusje), Anne en Cristina (medeleesonderzoekers), Suzanne

(jongleermaatje), en helma, Jantine, Francine en Bettina (sociaal-emotioneel

bolwerk).

Dan de onderzoekers buiten de UvA die hebben bijgedragen aan mijn

wetenschappelijke ontwikkeling. Faculteit der Bewegingswetenschappen

aan de VU, bedankt voor de gedegen opleiding, de uitdagingen daarbuiten,

en jullie aanstekelijke enthousiasme voor onderzoek. mark, Lisa en Andrew,

dankzij jullie had ik een geweldige stage in Aberdeen. mark, bedankt voor je

vertrouwen in mij als onderzoeker, met name toen ik het mijne kwijt was. minna

en JLD, dank je wel voor het opstarten van onze vruchtbare internationale

samenwerking; kiitos. maggie en CRL, ik had een leuke en waardevolle periode

in York. Bedankt voor je hulp met mijn onderzoeksvoorstel. Geweldig dat we

nu allebei in Oxford zitten. Robin, over onderzoek en vooral statistiek raken

wij nooit uitgepraat. Ik heb goede herinneringen aan je verblijf bij ons in

Amsterdam en vind het super dat ik nu bij jou in Oxford ben. Simon, bedankt

voor de spannende multidisciplinaire samenwerking. Dorothy en OSCCI,

bedankt voor het opnemen van mij in de onderzoeksgroep en het bieden van

deze geweldige kans om mijn (wetenschappelijke) horizon te verruimen. Kate

en LCD, bedankt voor het ‘erelidmaatschap’. Oriel College en Orielenses, ik voel

me bevoorrecht dat jullie me hebben opgenomen. het is een soms ietwat

bizarre maar prachtige ervaring.

Voor de broodnodige geestelijke afleiding en fysieke uitlaatklep had ik

gelukkig 10 uur/week trainen bij turnvereniging Olympia en acrogymvereniging

CCO. het was super leuk! O.a. met dank aan juf map, Guido, onze bovenpartners

Simone, Tamara en miranda en natuurlijk mijn onder en zusje Anke. Anke, we

hebben samen heel wat meegemaakt en ik ben trots op je. Oxford Sirens, heel

tof om deel uit te maken van het cheerleading wedstrijdteam. Toch weer salto’s

en meisjes gooien... Prof. Van der Werf en zijn team: geweldig. mijn leven is een


179

dankwoord

cadeautje. mijn schoonfamilie, Zoia, Adi, Pierre en Bernadette, bedankt voor

jullie steun en interesse. Broers Aart en Geert, en zusjes Anke en Janneke, wat

ben ik blij met jullie! Rita, Bart en Suzanne, fijn dat jullie erbij zijn gekomen.

Papa en mama, wat een veilig gevoel dat jullie altijd voor me klaar staan. het is

altijd heerlijk thuis te komen. Yves, mijn rots in de branding, ook al is het weer

op afstand. Ik kan niet zonder je.


Learn from yesterday, live for today, hope for tomorrow. The important thing is

not to stop questioning.

- Albert Einstein -

Publications & Curriculum Vitae


Publications & Curriculum Vitae


182

List of Publications

Litt, R.A., de Jong, P.F., van Bergen, E., & Nation, K. (in press). Dissociating

crossmodal and verbal demands in paired associate learning (PAL):

What drives the PAL-reading relationship? Journal of Experimental Child

Psychology.

van der Leij, A., van Bergen, E., & de Jong, P.F (2013). Longitudinal family-risk

studies of dyslexia: Why some children develop dyslexia and others

don’t. In A.J. Fawcett (Ed.), British Dyslexia Association Handbook 2013.

Bracknell: British Dyslexia Association.

van Bergen, E., de Jong, P.F., Plakas, A., maassen, B., & van der Leij, A. (2012). Child

and parental literacy levels within families with a history of dyslexia.

Journal of Child Psychology and Psychiatry, 53(1), 28-36. Chapter 3 of

the present thesis

Torppa, m., Eklund, K., van Bergen, E., & Lyytinen, h. (2011). Parental literacy

predicts children’s literacy: A longitudinal family-risk study. Dyslexia,

17(4), 339-355.

van Bergen, E., de Jong, P.F., Regtvoort, A., Oort, F., S. van Otterloo, & van der Leij,

A. (2011). Dutch children at family risk of dyslexia: Precursors, reading

development, and parental effects. Dyslexia, 17(1), 2-18. Chapter 2 of

the present thesis

van Swieten, L.m., van Bergen, E., Williams, J.h.G., Wilson, A.D., Plumb, m.S.,

Kent, S.W., & mon-Williams, m.A. (2010). A test of motor (not executive)

planning in developmental coordination disorder and autism. Journal

of Experimental Psychology: Human Perception and Performance, 36(2),

493-499.

van Bergen, E., van Swieten, L.m., Williams, J.h.G., & mon-Williams, m. (2007).

The effect of orientation on prehension movement time. Experimental

Brain Research, 178(2), 180-193.


183

Publications & curriculum Vitae

Submitted or in Revision

van Bergen, E., de Jong, P.F., maassen, B., Krikhaar, E., Plakas, A., & van der Leij,

A. (in revision). IQ of four-year-olds who go on to develop dyslexia.

Chapter 5 of the present thesis

Torppa, m., Eklund, K., van Bergen, E., & Lyytinen, h. (submitted). Late-emerging

and resolving dyslexia: A follow-up study from kindergarten to Grade

8.

van Bergen, E., de Jong, P.F., maassen, B., & van der Leij, A. (submitted). The

effect of parents’ literacy skills and children’s preliteracy skills on the

risk of dyslexia. Chapter 4 of the present thesis


184

Curriculum Vitae

Elsje van Bergen was born in Landsmeer, the

Netherlands, on 18th December 1981. After

completing secondary education at Pascal

College in Zaandam in 2000, she lived for a year in

Zaķuimuiža, Latvia, and attended the Rīgas Doma

Kora Skola. Back in the Netherlands, she studied

Sciences and Innovation management for a

year at the University of Utrecht (Propaedeutic

diploma, 2002), after which she proceeded to

study human movement Sciences at the VU University Amsterdam (Bachelors,

2005, cum laude; masters, 2006, cum laude) and a minor in Romanian at the

University of Amsterdam (2008, cum laude). In 2006 Elsje carried out her six-

month master’s project at the School of Psychology, University of Aberdeen,

Scotland, working with Professor mark mon-Williams. She investigated

reach-to-grasp movements in children (with and without developmental

coordination disorder), healthy adults, and stroke patients. From 2004 to 2008

she was employed by the Faculty of human movement Sciences, VU University

Amsterdam: first as a teaching- and research-assistant and the final year as a

PhD-student, studying sources of interlimb interactions. Subsequently she

switched to the field of reading research. From 2008 to 2012 she carried out her

PhD project at the Research Institute of Child Development and Education at the

University of Amsterdam. her research focused on the development of reading

and its cognitive underpinnings in children born into families with and without

a history of dyslexia. During her PhD Elsje obtained the BKO, the certificate for

teaching at universities. In 2012 she moved to England to join Professor Dorothy

Bishop’s group at the Department of Experimental Psychology, University

of Oxford. her postdoctoral research project is called ‘Reading fluency: Like

parent like child?’. In this project she studies the cognitive, environmental, and

genetic factors that influence reading ability and disability. moreover, she aims

to identify parental skills that can predict their offspring’s reading skills. Elsje

is funded by a Rubicon grant from the Netherlands Organisation for Scientific

Research. She has also been awarded a Junior Research Fellowship at Oriel

College, Oxford.


omslag_Elsje-van-Bergen_v2.indd 1 20-12-12 19:37

who will develop dyslexia? - uva · who will develop dyslexia? cognitive precursors in parents and...

Documents