Beginning to read across languages varying in orthographic consistency: Comparing the effects of...

Learning and Instruction 19 (2009) 466e480www.elsevier.com/locate/learninstruc

Beginning to read across languages varying in orthographic consistency:Comparing the effects of non-cognitive and cognitive predictors

George Manolitsis a,*, George Georgiou b, Kathy Stephenson b, Rauno Parrila b

a University of Crete, Department of Preschool Education, University Campus of Gallos, 74100 Rethymnon, Crete, Greeceb University of Alberta, Department of Educational Psychology, Education North 6-102, Edmonton-Alberta T6G 2G5, Canada

Received 16 October 2007; revised 14 July 2008; accepted 29 July 2008

Abstract

We examined whether the effect that different non-cognitive and cognitive factors have on reading acquisition varies as a function oforthographic consistency. Canadian (n¼ 77) and Greek (n¼ 95) children attending kindergarten were examined on general cognitive ability,phonological sensitivity, and letter knowledge. The parents of the children responded to a questionnaire on home literacy activities and theteachers reported on children’s task-focused behaviour. In Grades 1 and 2 the children’s word decoding and reading fluency were assessed.Results indicated that direct teaching of letter names and sounds at home was associated with better letter knowledge in both languages. Task-focused behaviour and letter knowledge in kindergarten predicted significantly nonword decoding in Grade 1, but their effect was stronger inEnglish than in Greek. This pattern was not replicated for reading fluency in Grade 2.� 2008 Published by Elsevier Ltd.

Keywords: Home literacy environment; Task-focused behaviour; Emergent literacy; Reading development; Cross-linguistic comparison

1. Introduction

The importance of phonological sensitivity and letterknowledge on reading development has been well established(Holopainen, Ahonen, & Lyytinen, 2001; de Jong & van derLeij, 1999; Kirby, Parrila, & Pfeiffer, 2003; Lepola, Pos-kiparta, Laakkonen, & Niemi, 2005; Parrila, Kirby, &McQuarrie, 2004; Wagner et al., 1997). Despite theacknowledged importance of these emergent literacy skills,several issues remain unresolved, including what their originsare, how their predictive strength is influenced by many othernon-cognitive and/or environmental factors known to predictreading development, and whether they predict reading simi-larly across languages that vary in orthographic consistency.Thus, the purpose of the current study was to examine: (a) theeffect of non-cognitive (task-focused behaviour) and envi-ronmental (various aspects of home literacy environment)

* Corresponding author. Tel.: þ30 2810 327161.

E-mail address: [email protected] (G. Manolitsis).

0959-4752/$ - see front matter � 2008 Published by Elsevier Ltd.

doi:10.1016/j.learninstruc.2008.07.003

factors on phonological sensitivity and letter knowledge, (b)the unique and joint contribution of task-focused behaviour,environmental factors, and emergent literacy skills on worddecoding and reading fluency, and (c) the effect of ortho-graphic consistency on these relationships.

1.1. Home literacy environment and beginning to read

Home literacy environment (HLE) is an umbrella conceptthat is normally used to describe a variety of childeparentactivities related to literacy (Burgess, 2002; Leseman & deJong, 1998; Senechal, 2006). Shared reading is one of the bestknown HLE variables in the literature. Shared reading refers tothe frequency of storybook reading by the parents to theirchildren. Previous studies have provided conflicting findingswith respect to the effects of shared reading on emergentliteracy skills (Bus, van Ijzendoorn, & Pellegrini, 1995; Evans,Shaw, & Bell, 2000; Foy & Mann, 2003; Scarborough &Dobrich, 1994). For example, Bus et al.’s (1995) meta-analysisreported a positive association between shared reading and

mailto:[email protected]

http://www.elsevier.com/locate/learninstruc

467G. Manolitsis et al. / Learning and Instruction 19 (2009) 466e480

children’s literacy development whereas Scarborough andDobrich (1994), based on review of 30 years of literature,concluded that shared reading was not associated with literacyskills as consistently or strongly as could be expected. Severalrecent studies have provided support for Scarborough andDobrich’s (1994) argument by demonstrating that sharedreading is not associated with phonological sensitivity, letterknowledge or word decoding in either inconsistent (Evanset al., 2000; Foy & Mann, 2003; Frijters, Barron, & Brunello,2000; Senechal & LeFevre, 2002; Stephenson, Parrila, Geor-giou, & Kirby, 2008) or consistent orthographies (de Jong &Leseman, 2001; van Steensel, 2006; Stoep, Bakker, & Ver-hoeven, 2002). However, it is possible that shared bookreading is related to the development of vocabulary knowledge(Frijters et al., 2000; Senechal & LeFevre, 2002; Torppa et al.,2007), and vocabulary knowledge is then associated with laterreading comprehension (de Jong & Leseman, 2001; Senechal,2006; Storch & Whitehurst, 2002).

Other aspects of HLE seem to be associated more directlywith reading skills than shared reading. Informal teachingactivities (teaching reading, letters, or writing) taking place athome have been shown to predict phonological sensitivity(Foy & Mann, 2003), letter knowledge (Evans et al., 2000),and early reading skills (Senechal & LeFevre, 2002) inEnglish. In contrast, studies conducted in orthographicallyconsistent languages have found no direct effect of parents’teaching activities on phonological sensitivity (Senechal,2006; van Steensel, 2006) and a small (Leseman & de Jong,1998) or nonsignificant effect on early reading achievement(Senechal, 2006; van Steensel, 2006). However, Torppa et al.(2006) showed with Finnish children that less frequentteaching of letter names at home at 4.5 years of age wasassociated with delayed letter knowledge development from 4to 6 years of age. Finally, studies that have examined child’slevel of engagement during his/her literacy experiences(Burgess, Hecht, & Lonigan, 2002; Levy, Gong, Hessels,Evans, & Jared, 2006) have indicated that English-speakingchildren who are engaged actively by their parents in literacyactivities have better emergent literacy skills than childrenwho are engaged in more passive activities (e.g., listeningstorybooks, watching TV).

Similar to direct teaching and frequency of storybookreading, parents’ beliefs and expectations of their children’sliteracy development can be considered as part of HLE. It isnot only the physical interaction with the children that mayaffect their performance, but also what their parents believeand expect from them. Several researchers have shown thatparents’ positive beliefs about their children’s cognitive oracademic ability play an important role on children’s schoolperformance (Aunola, Nurmi, Niemi, Lerkkanen, & Rasku-Puttonen, 2002). To date, no study has examined howparents’ beliefs and expectations interact with orthographyand other HLE measures to influence later reading achieve-ment. It is possible, for example, parents’ beliefs andexpectations about their children’s future reading ability toelicit more intensive direct teaching or more frequent sharedreading.

To summarize, the literature review on the effects ofdifferent aspects of HLE on phonological sensitivity, letterknowledge, and reading suggests that in inconsistent orthog-raphies, shared reading may be associated with more generallanguage skills, such as vocabulary knowledge, while moreactive involvement between parents and their children may berequired to directly affect phonological sensitivity and letterknowledge or early word-reading development. However, inconsistent orthographies, the relationship appears to be morelimited and in most cases insignificant.

1.2. Task-focused behaviour and learning to read

Motivational processes have been shown to affect child-ren’s cognitive performance (Wilson & Trainin, 2007).Particularly, achievement strategies that a child deploys inschool tasks, such as mastery orientation or task-focusedbehaviour, are associated with better scholastic outcomes andhelplessness beliefs or task-avoidant behaviour are associatedwith poorer scholastic outcomes (Aunola et al., 2002; Midg-ley, Arunkumar, & Urdan, 1996). For example, Aunola,Nurmi, Lerkkanen, and Rasku-Puttonen (2003) demonstratedthat children’s task-avoidance behaviours at school predictedlower achievement in mathematics later on.

In terms of learning to read, Fyrsten, Nurmi, and Lyytinen(2006) found that task-focused behaviour at the age of 6.5years predicted spontaneous reading acquisition before anyformal reading instruction. In addition, several studies havefound that elementary school children’s reading achievementwas predicted by their task-avoidant (Hagtvet, 2000; Onatsu-Arvilommi, & Nurmi, 2000) or task-focused behaviour (Dally,2006; Lepola, Salonen, & Vauras, 2000; Poskiparta, Niemi,Lepola, Ahtola, & Laine, 2003) in kindergarten.

Importantly, some studies have indicated that the effectof emergent literacy skills on reading acquisition may bemediated by task-focused behaviours. For example, Fyrstenet al. (2006) showed that children who were verballyskilful at age 5 showed a higher level of task-focusedbehaviour at age 6.5 years that then predicted spontaneousreading acquisition. Similarly, Lepola et al. (2005) showedthat task orientation in preschool predicted phonologicalawareness in kindergarten, which then predicted taskorientation at the beginning of Grade 1. Task orientation atthe beginning of Grade 1 was a significant predictor ofword recognition at the end of Grade 1. Furthermore, somestudies have found that task-focused behaviour in kinder-garten was a significant predictor of word-reading skill inGrade 1 (Dally, 2006; Salonen, Lepola, & Niemi, 1998),even after controlling for the effects of kindergartenphonological awareness.

In sum, the existing evidence generally supports a signifi-cant role of task-focused behaviours on reading achievementin both consistent and inconsistent orthographies. However, nostudy has directly compared whether the effects of task-focused behaviours on phonological sensitivity, letter knowl-edge, and reading ability are similar across languages that varyin orthographic consistency.

468 G. Manolitsis et al. / Learning and Instruction 19 (2009) 466e480

1.3. The present study

The present study is a follow-up of our previous work onthe effects of preschool HLE and task-focused behaviours onphonological sensitivity, letter knowledge, nonword decoding,and word-reading fluency in Grade 1 in English (Stephensonet al., 2008). Two important additions have been made in thisstudy. First, we have reassessed the same English-speakingchildren on their reading performance in Grade 2. Second, andmost importantly, we compare directly the effects of differentaspects of HLE and task-focused behaviour on phonologicalsensitivity, letter knowledge, nonword decoding, and readingfluency (word- and text-reading fluency) across two languagesthat vary in orthographic consistency.

Based on the findings of previous studies (Foy & Mann,2003; Leseman & de Jong, 1998; Senechal, 2006; Senechal& LeFevre, 2002), we hypothesized that more direct teachingof letters and sounds at home, but not more shared reading,will be associated with better performance on phonologicalsensitivity and letter knowledge tasks (Hypothesis 1).Moreover, this relationship will be stronger in English thanin Greek (Hypothesis 2). We reasoned that children learningto read English need more direct teaching at home in orderto learn early the inconsistent graphemeephoneme relation-ships of the English writing system compared to childrenlearning the same consistent relationships of the Greekorthography.

For the same reason, we expected that motivationalprocesses (i.e., task focus) play a more important role (a) forchildren’s phonological sensitivity and letter knowledge inEnglish than in Greek (Hypothesis 3.1) and (b) for childrenlearning to decode English words than for children learning todecode Greek words (Hypothesis 3.2). Task-focused behav-iours are likely needed more by a child who struggles toovercome the peculiarities of a writing system than by a childthat can decode every single word relatively quickly aftera few months of reading instruction. However, task-focusedbehaviours are expected to contribute to reading fluencydevelopment irrespective of the orthography because to ach-ieve a higher level of fluency a child needs repeated practice,and, consequently, more motivation (Hypothesis 4).

Finally, in line with existing evidence, we hypothesized thatboth phonological sensitivity and letter knowledge will predictnonword decoding and reading fluency in Grades 1 and 2(Hypothesis 5). However, based on the findings of previouscross-linguistic studies (Georgiou, Parrila, & Papadopoulos,2008; Mann & Wimmer, 2002), we expected that phonologicalsensitivity and letter knowledge will be more strongly relatedto nonword decoding and reading fluency in English than inGreek (Hypothesis 6).

2. Method

2.1. The context of the study e reading instruction

Learning to read Greek seems to differ from learning toread English. Greek orthography is much more regular than

English orthography. In Greek, pronunciation of most words ispredictable from print although the phonemeegraphemecorrespondence is not always perfect, especially for vowels.However, even the few existing inconsistencies are rule-learned and apply in almost every case in which the particularspelling pattern occurs (Porpodas, 2001).

Reading instruction in Greece begins in Grade 1 in whichchildren register at the age of 6 years (maximum at 6.5 years).Generally, most children begin school by entering kinder-garten at the age of 5. The Language section of the currentGreek National Kindergarten Curriculum uses an emergentliteracy perspective in which children are trained mainly inconcepts about print, reading whole words and emergentwriting activities. Children are taught reading skills in Grade 1‘‘as one group, all engaged in the same activity at a giventime’’ (Papoulia-Tzelepi, 2001: 201). All teachers use thetextbook and must follow the National Curriculum guidelinesthat mandate a phonics-based method.

Reading instruction in Alberta (Canada), where the data forthe English-speaking children were collected, begins at the ageof 5 when the children attend kindergarten and becomes moresystematic at the age of 6 when children enter Grade 1. Themethod of reading instruction places emphasis on both gra-phemeephoneme correspondences and on whole-wordrecognition strategies. This method of reading instruction isknown as Balanced Literacy Program and it has been thor-oughly described in previous studies (Senechal & LeFevre,2002; Senechal, LeFevre, Thomas, & Daley, 1998).

2.2. Participants

2.2.1. ChildrenLetters of information describing the study were sent to

parents of 223 kindergarten students in six suburban schools inSt. Albert, Alberta, Canada, and to parents of 232 kindergartenstudents in four schools in Rethymno, Crete, Greece. Fromthis initial pool, 161 Canadian students and 177 Greekstudents were given parental permission to participate in thestudy. Of them, 77 (39 males and 38 females; meanage¼ 66.84 months, SD¼ 3.88) Canadian students and 95Greek (50 males and 45 females; mean age¼ 67.01 months,SD¼ 2.93) students were randomly selected to be part of thepresent study. No students were excluded on the basis oflinguistic or other grounds. All the participating children werenative speakers of English and Greek, respectively.

Of the 77 Canadian students, 10 students had missing dataon the parent questionnaire and one student had missing dataon letter-sound knowledge. Of the 95 Greek students, 12students had missing data on the parent questionnaire, and twohad missing data on phoneme elision, blending, letter-soundknowledge, and letter-name knowledge. To examine if theperformance of the 10 Canadian and 12 Greek children withmissing data on parents’ questionnaire differed significantlyfrom the performance of the children with no missing data,two multivariate analyses were performed on the kindergartenmeasures. The MANOVAs showed that the two groups did notdiffer from each other on any measure in kindergarten for


either Canadian, Pillai’s trace¼ .09, F(8, 67)¼ .88, p> .05, orGreek children, Pillai’s trace¼ .14, F(8, 80)¼ 1.66, p> .05.Of the Canadian students, 97% were Caucasian and 50% hadattended preschool. All Greek students were Caucasian and65% had attended preschool.

By Grade 2 the sample consisted of 55 Canadian and 70Greek children. Of them 22 Canadian (28.5% of the sample)and 25 Greek children (26.3% of the initial sample) withdrewfrom the study. These withdrawal numbers are very similar toones reported in previous longitudinal studies (e.g., Senechal,2006). To examine if the performance of the children whowithdrew from the study differed significantly from the rest ofthe children, we performed MANOVAs on their measures inkindergarten. None of the MANOVAs reached significance foreither Canadian, Pillai’s trace¼ .10, F(8, 67)¼ .97, p> .05, orGreek children, Pillai’s trace¼ .11, F(8, 80)¼ 1.25, p> .05.All Canadian and Greek children attended single-gradeclasses.

2.2.2. ParentsParents of all the kindergarten children received two

questionnaires along with the letters of information andconsent forms. Ten Canadian and 12 Greek parents agreed fortheir children to participate but did not return the question-naire. The scale for parent’s educational level ranged from0 (completed elementary school) to 5 (completed postgraduatedegree). The modal educational level for Canadian mothersand fathers was ‘‘completed community college,’’ theminimum was ‘‘some high school,’’ and the maximum was‘‘completed graduate school’’. The modal educational level forGreek mothers and fathers was ‘‘completed high school’’, theminimum was ‘‘completed elementary school’’ and themaximum for mothers was ‘‘completed graduate school’’ andfor fathers ‘‘completed university.’’ Canadian parents hadhigher educational level than the Greek parents, t(148)¼ 4.14,p< .001, Cohen’s d¼ .68.

2.2.3. TeachersAll nine Canadian and 12 Greek kindergarten teachers from

the participating schools gave written consent to participate inthe study. The teachers in each country filled out a question-naire regarding task-focused behaviour (see below for details)for all the children participating in the study. All of theCanadian and Greek schools were part of the same schooldivision and all of the teachers were female. Because childrenwere nested within classes, we calculated point-biserialcorrelations between class membership and the dependentmeasures. No significant correlations emerged in eitherlanguage.

2.3. Measures

2.3.1. General cognitive abilityVocabulary was assessed in both languages using the

Peabody Picture Vocabulary Test-Third Edition (PVTT-IIIA;Dunn & Dunn, 1997). Participants were shown four picturesand the examiner said a word to describe one of the four

pictures. The participant was required to point to the correctpicture for the word given by the examiner. Items wereadministered in sets of 12. Testing was discontinued after eightor more errors within the highest set of items administered.Participants’ score was the number of correct items. Split-halfreliability coefficient in our English- and Greek-speakingsample was .95 and .92, respectively.

Nonverbal cognitive ability was assessed in both languagesusing the Coloured Progressive Matrices (Sets A, Ab, and B;Raven, 1956). Each set contained 12 items and a participant’sscore was the total items correct on all three sets. Cronbach’salpha reliability coefficient was .74 and .82 for our English-speaking and Greek-speaking sample, respectively.

2.3.2. Phonological sensitivityElision in English was modified from the Comprehensive

Test of Phonological Processing (CTOPP) (Wagner, Torgesen,& Rashotte, 1999) by adding six test items, four two-syllablewords and two words that required the participant to say theword without saying a designated sound in the word, to makea total of 29 items. There were three practice items and 29 testitems: three items required the participant to say a compoundword without saying one of the words (e.g., cowboy withoutthe /cow/ is boy), six test items required the participant to saythe word without saying one of the syllables, and theremaining 20 items required the participant to say a wordwithout saying a designated sound in the word. The position ofthe phoneme to be removed varied across those 20 items; ineight cases it involved the initial phoneme (e.g., farm withoutthe /f/ is arm); in six cases, the medial phoneme (e.g., winterwithout the /t/ is winner), and in six cases, the final phoneme(e.g., sheep without the /p/ is she). Testing was discontinuedafter three consecutive errors. Participant’s score was thenumber of correct items. Split-half reliability coefficient in ourEnglish-speaking sample was .92.

The Greek version of Elision had the same number of itemsas the English task and was administered in the same way.However, given that the items after the deletion of a phonememust produce a real word and that in contrast to the majority ofthe English words, the typical Greek words are two or moresyllables, complete matching of the words in the twolanguages was impossible. Instead, we devised a list of itemsthat required the children to delete the same sound (initial,medial, final) from the word as it was required from theEnglish-speaking children. Specifically, four test itemsrequired the participant to say the word without saying one ofthe syllables (e.g., l3moni /lemoni/ [lemon] without the /le/ ismonh /moni/ [alone]), eight test items required the participantto delete the initial phoneme (e.g., polh /poli/ [town] withoutthe /p/ is oloi /oli/ [all]); six test items required the participantto delete the medial phoneme (e.g., dınu /Cino/ [give] withoutthe /n/ is dyo /Cio/ [two]), and six test items required theparticipant to delete the final phoneme (e.g., zua /zoa/[animals] without the /a/ is zu /zo/ [live]). Split-half reliabilitycoefficient in our Greek-speaking sample was .94.

Blending in English was measured with the CTOPPBlending Words task (Wagner et al., 1999), which required the


examinee to listen to a series of separate sounds and then putthe sounds together to make a whole word. There were 20 testitems. Stimuli were modified to Canadian pronunciations.Testing was discontinued after three consecutive errors weremade and a participant’s score was the number of correctitems. In Greek, Blending contained the same number of itemsthat were matched to the English ones in terms of the numberof provided sounds in each word. Split-half reliability coeffi-cient in English and Greek was .89 and .92, respectively.

2.3.3. Letter knowledgeLetter-Name Knowledge in English was assessed by

administering the Letter Identification test (Clay, 1993).Participants were asked to identify each of the upper andlowercase letters. Two lowercase letters, a and g, were pre-sented in two different fonts, so the total possible score was54. In Greek, the participants were asked to name each of theupper and lowercase Greek letters. The maximum score inGreek was 48. Cronbach’s alpha reliability coefficient in ourEnglish- and Greek-speaking sample was .96 and .95,respectively.

Letter-Sound Knowledge in both languages was assessed byhaving participants give the sound of each uppercase letterpresented in random order on a laptop screen. Testing wasdiscontinued after six consecutive items were incorrect.Participants’ score was the total number correct. Themaximum score is 26 for English and 24 for Greek. Cronba-ch’s alpha reliability coefficient in our English- and Greek-speaking sample was .94 and .95, respectively.

2.3.4. Home literacyHome literacy in both languages was assessed with seven

Likert-type scale questions. Parents were asked (a) how oftentheir child was taught to identify letters; (b) how often theirchild was taught letter sounds; and (c) how often their childwas taught to read words when the child was 2e4 years of age.In addition, they were asked (d) how often their child is read toat home, (e) how many children’s books are in the home, (f)how many books are in the home, and (g) how well do youthink your child will do in reading in the future. For the firstfour questions, the six-point Likert-type scale ranged from 1(never) to 6 (more than once a day); for the fifth and sixthquestion, from 1 (less than 10) to 6 (more than 200) and 1 (lessthan 100) to 6 (more than 1000), respectively; for the lastquestion, from 1 (not at all well) to 6 (very well).

In the present study, the two questions about books at homewere correlated (English, r¼ .63; Greek, r¼ .69) and so theirstandardized scores were summed to make a single Books atHome score. Cronbach’s alpha coefficient for Books at Homewas .75 for English and .80 for Greek.

Similarly, parents’ reports of their children being taught (a)letter names, (b) letter sounds, or (c) to read words werecorrelated (for letter names and letter sounds, r¼ .76 andr¼ .81 for English and Greek, respectively; for letter soundsand to read words, r¼ .60 and r¼ .67 for English and Greek,respectively; and for letter names and to read words, r¼ .35and r¼ .69, for English and Greek, respectively). The question

about frequency of reading, however, correlated only weaklywith teaching letter names, letter sounds, and word reading(English, r¼ .23, .30, and .28; Greek, r¼ .27, .25, and .14,respectively). Therefore, rather than making a single readingactivity score, the standard scores of the three teaching ques-tions were summed together to make a single variable, DirectTeaching, whereas the reading frequency question was keptseparate. Cronbach’s alpha coefficient for Direct Teaching was.78 for English and .76 for Greek. Thus, we report fourdifferent home literacy indicators below: Books at Home,Direct Teaching, Reading Frequency, and Parent’sExpectations.

The Greek version of the questionnaire was translated intoGreek by the second author who is fluent in both English andGreek and who is familiar with home literacy research. Inaddition, it was back-translated by three professional trans-lators. The independent translations resulted in generallinguistic agreement, with only minor differences in wordingof the items. Disagreements over wording were resolvedthrough discussion with the first author. Before its formal use,it was administered to a group of 16 Greek parents to ensureclarity of instructions. No difficulties responding to the ques-tionnaire were reported.

2.3.5. Task-focused behavioursKindergarten teachers were asked to evaluate the behaviour

of each child using the Behavioural Strategy Rating Scale-II(BSR-II; Aunola, Nurmi, Parrila, & Onatsu-Arvilommi, 2000).Teachers were asked to think of a specific classroom situationand then rate the child’s behaviour using seven statementsassessed with a five-point Likert-type scale that ranged from 1(very much/easily) to 5 (not at all). Five questions assessedchildren’s use of task-focused versus task-avoidant behaviours(e.g., ‘‘Does the student give up easily?’’ For a list of the itemssee Stephenson et al., 2008).

Two additional questions assessed children’s defensiveness(e.g., ‘‘Does the student blame him/herself readily when he/she fails?’’). More than half of the participants (55%) in bothlanguages received a score of three (middle of the scale) onboth questions, indicating problems with the scale. Thesequestions were dropped from further analyses. Cronbach’salpha reliability coefficient for the five items that remained inthe study was .96 and .93 for our English- and Greek-speakingsample, respectively.

The Greek version of the questionnaire was translated intoGreek by the second author and back-translated by threeprofessional translators. Before its formal use, it was admin-istered to a group of eight Greek kindergarten teachers whoreported that there were no ambiguous words and that theinstructions for the completion of the questionnaire were clear.

2.3.6. Word decodingThe Woodcock Reading Mastery Tests-Revised (WRMT-R;

Woodcock, 1998) was used to assess decoding in English inGrade 1. Form H of Word Attack was utilized. Participants wereasked to read 45 nonwords presented on a laptop screen as ifthey are real English words. The number of characters in the


English Word Attack task was 215. Testing was discontinuedafter six consecutive errors. Participant’s score was the numberof items correct. Split-half reliability coefficient in our English-speaking sample was .94.

Similar to its original version in English, the Greek WordAttack task consisted of 45 pronounceable nonwords that werederived from real words after changing two or three letters(either by substituting them or using them backwards). It wasoriginally piloted in a study by Papadopoulos and Georgiou(2000) and was later used in other studies, showing satisfac-tory psychometric properties (Papadopoulos, 2001). Thenumber of characters in Greek Word Attack task was 218.Testing was discontinued after six consecutive errors. Partic-ipant’s score was the number of items correct. Split-half reli-ability coefficient in our Greek sample was .86.

2.3.7. Word- and text-reading fluencyTest of Word Reading Efficiency (TOWRE; Torgesen,

Wagner, & Rashotte, 1999) was used a measure of word-reading fluency in English. The child was given a list of 104words, divided into four columns of 26 words each, and askedto read them as fast as possible. A short, eight-word practicelist was presented first. The number of words read correctlyand the number of errors made within a 45-s time limit wasrecorded. The score was the number of words read correctly.Torgesen et al. (1999) reported testeretest reliability of .95 forthe ages of 6e9 years. The Greek version of this task had thesame format as the English one. It consisted of 104 wordsbeginning with one syllable words and ending with threesyllable words. However, the Greek version of the task hadlonger words compared to the English task. More specifically,although the Greek word-reading fluency task contained 644characters, the English word-reading fluency task contained607 characters.

Gray Oral Reading Test (GORT; Wiederholt & Bryant,2001) was used to assess children’s text-reading fluency inEnglish. The participants were asked to read as fast and asaccurately as possible two short stories. The readingcomprehension questions that follow each story were notadministered because it was not the intention of this study toexamine reading comprehension. Individual’s score was thetime taken to read the two short stories. Wiederholt and Bryant(2001) reported test-retest reliability for GORT to be .93. TheGreek version of this task was a translation/back translation ofthe English GORT. Although the number of characters wasexactly the same as in the English GORT texts, the number ofwords was smaller than in English. The Greek text-readingfluency task contained 53 words whereas the English versioncontained 61 words.

2.4. Procedure

All participants in both languages were tested individuallyin their respective schools during school hours by trainedexperimenters in both countries (two Canadian graduateresearch assistants and four Greek graduate research assis-tants). The children in both countries were tested in April/May

of kindergarten, in April/May of Grade 1, and in November/December of Grade 2. In kindergarten the tests were admin-istered in two sessions lasting approximately 35 min each. Inthe first session, Raven’s Matrices, Blending, and Letter-NameKnowledge were administered whereas in the second sessionPPVT-IIIA, Elision, and Letter-Sound Knowledge wereadministered. In Grade 1, Word Attack, word-reading fluency(WRF), and text-reading fluency (TRF) were all administeredin one session. In Grade 2 the children were administered onlyWRF and TRF. We decided not to administer Word Attack inGrade 2 on the basis of the findings of previous studies con-ducted in Greek showing that word decoding is close to ceilingby Grade 2 with minimal variability in the distribution ofscores (see Georgiou et al., 2008; Porpodas, 1999).

2.5. Statistical analyses

The statistical analyses were conducted in three steps. First,regression analyses were used to examine the unique contri-butions of predictor variables to the dependent variables.Because age of the participants and preschool attendance werenot significantly related to any dependent variable in the studythey were left out of the analyses. Model 1 contained thecontrol variables, PPVT-IIIA and Raven’s Matrices, togetherwith the Home Literacy variables (entered as a block at Step2). Model 2 contained the two control variables along withparental education. Model 3 contained the two control vari-ables together with Task-Focus. Model 4 was estimated onlywhen reading outcomes in Grades 1 and 2 were the dependentvariable and contained the two control variables and phono-logical sensitivity and letter knowledge (entered as a block atStep 2).

Next, the independent variables that survived the statisticalcontrol of PPVT-IIIA and Raven’s Matrices in each one of thethree/four initial models were entered together in path anal-yses. Path analysis is a type of structural equation modelingwhere observed variables are used instead of latent factors(Loehlin, 2004). In these path analyses, the two control vari-ables were allowed to correlate with each other but not withthe rest of the independent variables, and the independentvariables were allowed to correlate with each other.

The final step in the statistical analyses involved theexamination of the cross-linguistic differences in the predic-tors of emergent literacy skills, Word Attack, WRF, and TRFby performing multi-group analyses in two steps. First, wetested the fit of a model in which no cross-group constraintswere imposed. The unconstrained multi-group model includedthe paths from the predictor variables to the outcome measuresthat were either significant in one language but not in the other(based on the results of the path analysis in each languageseparately) or paths that were significant in both languages.Regression paths that were not statistically significant in eitherlanguage were left out. Second, we tested the invariance of theregression paths across the two language groups by imposingequality constraints on the direct and indirect effects of thepredictor variables on the criterion variables. In testing for theinvariance of the regression paths, we compared the c

2 value


of the constrained model with that of the initial multi-groupmodel in which no cross-language constraints were imposed. Ifthe difference in c2 values, given the difference in the degreesof freedom between the two models (df constrained� dfunconstrained), was significant then this indicated that thespecific predictor was contributing in a different way to theoutcome variable across the two languages.

Because we were interested in examining the predictors ofphonological sensitivity, letter knowledge, nonword decoding,and reading fluency, the results of the path analyses are pre-sented below according to the dependent variables at hand andnot according to the effect of each one of the predictors. Wewill elaborate on the contribution of each of the predictorvariables in Section 4.

3. Results

3.1. Preliminary data analyses

Table 1 presents the descriptive statistics for all the measuresused in the study in the two languages. An initial examination ofthe distributional properties of the variables revealed someproblems. Text-reading fluency (TRF) was positively skewed inboth grades and across languages. Log transformations wereused to achieve normality (Tabachnick & Fidell, 2001).

Table 1

Descriptive statistics for all the measures used in the study.

Greek

M SD Min Ma

Control variables (K)

PPVT-IIIA 96.35 25.63 40 157

Raven’s matrices 15.55 3.12 10 26

Phonological awareness (K)

Elision 3.46 5.40 0 29

Blending 4.53 4.45 0 20

Letter knowledge (K)

Letter-name knowledge 7.53 11.63 0 48

Letter-sound knowledge 12.85 7.93 0 24

Home literacy environment (K)

Number of children books 3.15 .99 1 5

Number of books in home 2.54 1.39 1 5

Frequency of reading to child 3.60 .86 2 5

Teach to identify letters 2.40 1.24 0 5

Teach letter sounds 2.46 1.24 0 5

Teach to read words 1.68 1.47 0 4

How well child read in future 4.41 .79 2 5

Parental Education (K) 2.43 1.15 0 5

Achievement strategies (K)

Task-focus 3.69 1.18 1 5

Grade 1 reading

Word-reading fluency 26.17 11.07 0 49

Text-reading fluency 92.43 49.78 32 303

Word attack 32.26 9.43 0 43

Grade 2 reading

Word-reading fluency 33.06 11.07 2 57

Text-reading fluency 52.42 37.12 18 233

K¼Kindergarten.

However, even after the log transformations, the performanceof some children in TRF was more than 2 SDs from the groupmean (English: two children in Grade 1 and four in Grade 2;Greek: three children in Grade 1 and two in Grade 2). Theresponses of these outliers were replaced by a value equal to thenext highest non-outlier-score plus one unit of measurement(Tabachnick & Fidell, 2001). In addition, Word Attack in Greekwas negatively skewed. Reflection and log transformation wereused to improve the distribution. The transformed scores wereused in further analyses. As Word Attack scores were reflected,results were corrected for direction.

To limit the effect of task-specific variability, we combineddifferent indicators of a processing skill into construct scores.Elision and Blending were highly correlated (English: r¼ .74;Greek: r¼ .48) and their z-scores were summed to makea single variable, Phonological Sensitivity, which was used inall the remaining analyses. Similarly, Letter Name Knowledgeand Letter Sound Knowledge were highly correlated (English:r¼ .64; Greek: r¼ .48) and their z-scores were combined toform a single variable, Letter Knowledge, which was used inall subsequent analyses. The English-speaking children out-performed the Greek-speaking children on Letter-NameKnowledge, t(168)¼ 17.52, p< .001, Cohen’s d¼ 2.68, butthe differences were minimal for Letter-Sound Knowledge,t(167)¼ 2.42, p< .05, Cohen’s d¼ .37. This pattern of results

English

x N M SD Min Max N

95 93.40 15.14 56 130 77

95 19.22 4.28 12 31 77

93 6.92 4.81 0 18 77

93 6.17 2.74 0 13 77

93 39.74 12.29 6 54 77

93 15.78 7.66 0 26 76

83 4.18 .76 3 5 67

83 3.21 1.01 1 5 67

83 4.01 .77 2 5 67

83 3.21 1.20 0 5 67

83 2.68 1.34 0 5 67

83 1.98 1.30 0 5 66

83 4.41 .72 2 5 66

83 3.12 .85 1 5 67

95 3.36 1.34 1 5 77

74 35.72 14.77 0 69 69

73 65.79 49.85 20 272 63

74 16.77 9.22 0 37 69

70 48.38 14.66 15 78 55

70 40.00 24.34 16 158 55


confirmed previous findings (Tafa & Manolitsis, 2008) and itwas expected because children in Greece are not taught theletter names before the end of Grade 1. In contrast, they aretaught the letter sounds because these are the primary units ofanalysis in reading.

The Home Literacy Environment (HLE) questionnaireindicated that on average the Canadian parents reported havingmore books at home (Mdn¼ 300e499) and more children’sbooks at home (Mdn¼ 100e199) than the Greek parents(Mdn¼ 100e299, for books at home, and Mdn¼ 25e99, forchildren’s books). Canadian parents reported further thatstorybook reading occurred in the home about once a day. Incontrast, Greek parents reported reading to their children a fewtimes a week. Parents in both countries reported teaching theirchildren to read words a few times a month. Finally, there wereno differences across the two languages in the parent’sexpectations for their children’s future reading. The majorityof the parents (59% of the Greek parents and 53% of theCanadian parents) reported expecting that their child will dovery well in reading in the future.

Finally, an examination of the distributional properties ofthe Behavioural Strategy Rating Scale-II indicated that theTask-Focus score across language groups was negativelyskewed and showed a ceiling effect (about 16% of samplescored 5 in both languages). To normalize the distribution, thescores were first reflected and then log transformed (Tabach-nick & Fidell, 2001). The log transformed reflected scoreswere used in further analyses. As Task-Focus scores used in allfurther analyses were reflected, results were corrected fordirection to simplify their interpretation.

3.2. Correlations between the variables

Table 2 displays the correlations obtained for all themeasures used in the study in each language group. We foundboth similarities but also differences in the relationshipsbetween the different measures used in this study acrosslanguages. In terms of similarities, children’s age and preschoolattendance did not correlate significantly with any of thedependent variables and frequency of shared book reading;number of books was not related to any reading outcome ineither language. In addition, letter knowledge was stronglyrelated to reading in both languages. However, also noticeablewere the differences across languages with respect to thecontribution of parent’s beliefs and expectations and phono-logical sensitivity to the reading outcomes. Parental educationand parent’s beliefs and expectations appear to be moreimportant for reading in Greek than in English and phonolog-ical sensitivity appears to be more important for reading inEnglish than in Greek. Finally, Task-Focus correlated morestrongly with the dependent variables in English than in Greek.

3.3. Path analyses

3.3.1. Phonological sensitivityThe results of the initial regression analyses indicated that

Task-Focus was the only variable associated significantly with

phonological sensitivity in English. Along with the controlvariables (namely PPVT-IIIA and Raven Matrices scores) itaccounted for 18% of the variance. In Greek, Raven’sMatrices, Direct Teaching, parental education, parent’s beliefs,and number of books were significantly associated withphonological sensitivity in the initial regression models. Whenput together in the same path analysis model, only the firstthree remained significant predictors accounting for 48% ofthe variance in phonological sensitivity. Notably, the rela-tionship between Direct Teaching and phonological sensitivitywas negative suggesting that poorer phonological sensitivitywas associated with more direct teaching of letter sounds,letter names, and words. This finding is particularly interestingand we will discuss it further in Section 4. Fig. 1 shows theresults of path analyses with phonological sensitivity as thedependent variable.

Next, we examined if there were any significant differencesin the predictors of phonological sensitivity across the twolanguages. The analyses showed that significant changes in c

2

were observed only when the effects of Direct Teaching wereconstrained to be equal across the two languages, Dc2(1)¼6.23, p< .05. In sum, these findings suggest that phonologicalsensitivity is associated with different non-cognitive variablesacross languages. However, only the association of DirectTeaching with phonological sensitivity was significantlydifferent across languages.

3.3.2. Letter knowledgeFig. 2 shows the results of path analyses with Letter Knowledge

as the dependent variable. In English, PPVT-IIIA, DirectTeaching, and Task-Focus were significantly related with LetterKnowledge. In total, the model accounted for 35% of the variance.In Greek, Raven’s Matrices, Direct Teaching, parent’s beliefs, andparental education related significantly with Letter Knowledge. Intotal, the model accounted for 32% of the variance. Next, weexamined if PPVT-IIIA, Raven’s Matrices, Direct Teaching,parent’s beliefs, Task-Focus, and parental education were associ-ated with Letter Knowledge differently in the two languages. Theanalyses showed that significant changes in c2 were observedwhen the effects of Raven’s Matrices, parent’s beliefs, and DirectTeaching were constrained to be equal across the two languages,Dc2(1)¼11.56, p< .001, Dc2(1)¼ 4.09, p< .05, Dc2(1)¼ 4.22,p<.05, respectively. In sum, the relationship of Raven’s Matricesand parent’s beliefs with Letter Knowledge was significantlygreater in Greek than in English whereas the reverse was true forDirect Teaching.

3.3.3. Word decodingOnly Task-Focus and Letter Knowledge were significant

predictors of Grade 1 Word Attack in their initial regressionmodels in the two languages. When considered together in thesame path analysis model with PPVT-IIIA and Raven’sMatrices, they accounted for 39% and 20% of the variance inWord Attack in English and Greek, respectively. Fig. 3 showsthe results of the path analyses in each language separately.

The multi-group analyses showed that significant changesin c2 were observed when the effects of Letter Knowledge or

Table 2

Correlations between predictor and outcome variables in Greek (above the diagonal) and in English (below the diagonal).

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

1. Age .07 .23* .03 .19 .03 �.06 .24* .06 .16 .10 .13 .15 �.18 .01 .06 �.22

2. PAT �.06 �.07 .08 �.01 �.09 �.05 .02 �.04 �.04 .04 �.08 �.15 .00 .14 �.01 .01

3. PAREDU �.01 .25 .47** .34** �.07 .27* .55** .33** .19 .52** .41** .26* �.15 .14 .09 �.15

4. PPVT-IIIA .38** .07 .21 .31** �.00 .35** .35** .22 .13 .31* .24* .03 .04 �.02 �.01 .03

5. Raven’s .16 �.03 .14 .36** �.05 .06 .17 .38** .24* .53** .49** .31** �.25* .18 .34** �.28*

6. Direct Teachinga �.14 �.07 .05 .17 .18 .25* .05 �.13 �.16 �.32** .17 .03 �.01 �.13 .03 �.10

7. Read to Childa �.08 .06 .48** .17 �.06 .32** .47** .27* .15 .02 .20 .02 .01 .09 .03 .07

8. Books at Homea �.19 �.05 .13 �.00 .07 .36** .14 .43** .11 .33** .19 .20 �.14 .09 .14 �.19

9. PBFRa .06 �.37** .06 .32* .21 .19 .19 .14 .16 .35** .38** .29* �.30* .19 .34** �.26*

10. Task-Focusa .27* �.07 .25 .46** .30** .28* .23 .00 .36** .18 .19 .05 �.19 .27* .10 �.01

11. PSa .09 �.13 .21 .33** .33** .26* .17 .14 .21 .38** .43** .28* �.22 .22 .17 �.13

12. LKa .17 �.05 .26 .39** .29* .45** .35** .19 .12 .47** .73** .47** �.39** .38** .42** �.35**

13. WRF-G1 .10 �.07 �.03 .22 .35** .20 �.01 .09 .08 .46** .42** .51** �.62** .35** .65** �.66**

14. TRF-G1 .01 �.17 �.06 �.33** �.30* �.22 �.13 �.18 �.02 �.52** �.43** �.59** �.81** �.47** �.77** .77**

15. WAT-G1 .11 .03 .16 .41** .35** .23 .01 .09 .06 .53** .52** .59** .60** �.59** .32** �.31**

16. WRF-G2 .02 .08 .06 .37** .30* .31* .21 .07 .14 .51** .38** .53** .69** �.84** .62** �.89**

17. TRF-G2 .04 �.10 .03 �.40** �.36** �.14 �.09 �.06 �.10 �.47** �.37** �.45** �.80** .90** �.58** �.88**

PAT¼ Preschool Attendance; PAREDU¼ Parental Education; PBFR¼ Parent’s Beliefs about Future Reading; PS¼ Phonological Sensitivity; LK¼ Letter

Knowledge; WRF¼Word-Reading Fluency; TRF¼ Text-Reading Fluency; WAT¼Word Attack; G1¼Grade 1; G2¼Grade 2.

*p< .05; **p< .01.a Measured in kindergarten.


Task-Focus were constrained to be equal across the twolanguages, Dc2(1)¼ 12.89, p< .001, and Dc2(1)¼ 4.97,p< .05, respectively. Both measures were stronger predictorsof nonword decoding in English than in Greek.

3.3.4. Word-reading fluencyIn English, Raven’s Matrices, Task-Focus, and Letter

Knowledge were significant predictors of Grade 1 word-reading fluency (WRF) in initial regression models. Inaddition, Letter Knowledge was a significant predictor ofGrade 2 WRF beyond the effect of Grade 1 WRF. As wasexpected on the basis of previous longitudinal studies (de

PPVT

Raven

Books

PhonologicalSensitivity - K

.15

.21

.26*

English

DirectTeaching

Task-Focus

Parent’sBeliefs

ParentalEducation

a b

Fig. 1. The baseline path model of predictors of Phonological Sensitivity in kinderg

c2(17, N¼ 77)¼ 27.17, p< .05, CFI¼ .80, RMSEA¼ .08; for Greek, c2(21, N¼p< .001.

Jong & van der Leij, 1999; Wagner et al., 1997), Grade 1WRF had a strong autoregressive effect on Grade 2 WRF.When Task-Focus and Letter Knowledge were put togetherin the same path analysis model with PPVT-IIIA andRaven’s Matrices, the model accounted for 36% of thevariance in Grade 1 WRF and 55% of the variance in Grade2 WRF. In this final model, Raven’s Matrices, Task-Focus,and Letter Knowledge predicted significantly Grade 1 WRFand Letter Knowledge and Grade 1 WRF predicted Grade 2WRF (see Fig. 4).

In Greek, only Letter Knowledge was a significantpredictor of Grade 1 WRF in the initial regression models.

PPVT

Raven

Books

DirectTeaching

.02

.42**

.12

-.32**

Greek

Task-Focus

Parent’sBeliefs

.00

ParentalEducation

PhonologicalSensitivity - K

.35**

arten in English (a) and Greek (b). K¼Kindergarten. Fit indices: for English,

95)¼ 48.82, p< .001, CFI¼ .78, RMSEA¼ .11. * p< .05; ** p< .01; ***

PPVT

Raven

Task-Focus

PPVT

Raven

Task-Focus

DirectTeaching

DirectTeaching

LetterKnowledge - K

LetterKnowledge - K

.22*

.09

.29**

.36***

-.02

.36**

.26**

English Greek

Parent’sbeliefs

Parent’sbeliefs

.20*

ParentalEducation

ParentalEducation

.25*

a b

Fig. 2. The baseline path model of predictors of Letter Knowledge in kindergarten in English (a) and Greek (b). K¼Kindergarten. Fit indices: for English, c2(13,

N¼ 76)¼ 29.66, p< .001, CFI¼ .73, RMSEA¼ .13; for Greek, c2(12, N¼ 93)¼ 27.40, p< .001, CFI¼ .75, RMSEA¼ .15. * p< .05; ** p< .01; *** p< .001.


Similar to English, there was a strong autoregressive effect onGrade 2 WRF. When combined with the general cognitiveability measures the model accounted for 21% of the variancein Grade 1 WRF and 44% of the variance in Grade 2 WRF.Fig. 4 presents the model in which WRF in Grades 1 and 2 wasthe dependent variable in each language.

Next, multi-group analyses showed that significant changesin c2 were observed when the effect of Letter Knowledge orTask-Focus was constrained to be equal across the twolanguages, Dc2(1)¼ 7.31, p< .01, and Dc2(1)¼ 5.45,p< .05, respectively. These findings suggest that LetterKnowledge was a stronger predictor of Grade 1 WRF in Greekthan in English, whereas Task-Focus was more important inpredicting Grade 1 WRF in English than in Greek.

3.3.5. Text-reading fluencyTask-Focus and Letter Knowledge were significant

predictors of Grade 1 text-reading fluency (TRF) in English intheir respective initial regression models. When they werecombined, they accounted jointly with the control variables for

Word Attack - G1

.09

.14

.27*

.41***

English

PPVT

Raven

Task-Focus

LetterKnowledge

a b

Fig. 3. The baseline path model of predictors of Word Attack in Grade 1 in

N¼ 69)¼ 25.33, p< .001, CFI¼ .73, RMSEA¼ .26; for Greek, c2(4, N¼ 74)¼ 2

42% of the variance in Grade 1 TRF and 80% of the variancein Grade 2 TRF. Both of them remained significant predictorsof Grade 1 TRF. Only Grade 1 TRF predicted significantlyGrade 2 TRF. In Greek, Raven’s Matrices and Letter Knowl-edge were significant predictors of Grade 1 TRF, and Task-Focus and Grade 1 TRF were significant predictors of Grade 2TRF. When combined together in the final model, Raven’sMatrices was no longer a significant predictor of Grade 1 TRF,but both Letter Knowledge and Task-Focus were significantpredictors of Grade 1 and Grade 2 TRF, respectively. In total,the model accounted for 15% of variance in Grade 1 TRF and62% of variance in Grade 2 TRF. Fig. 5 presents the model inwhich TRF in Grades 1 and 2 were the dependent variables ineach language.

When equality constraints were imposed across languages,the analyses showed that significant changes in c2 were observedonly when the effects of Letter Knowledge were constrained tobe equal across the two languages, Dc2(1)¼ 11.32, p< .001.Letter Knowledge was a stronger predictor of Grade 1 TRF inEnglish than in Greek. This finding appears to be in conflict with

Word Attack - G1

.33***

-.14

-.02

.22*

Greek

PPVT

Raven

Task-Focus

LetterKnowledge

English (a) and Greek (b). G1¼Grade 1. Fit indices: for English, c2(4,

8.74, p< .001, CFI¼ .79, RMSEA¼ .25. * p< .05; ** p< .01; *** p< .001.

-.14

PPVTRavenTask-Focus LetterKnowledge

WRF-G1 WRF-G2 WRF-G2

PPVTRavenLetterKnowledge

WRF-G1

.30**

.22*.06 .15 .43***.36***

.17.16

-.09

-.10

.59*** .62***

Task-Focus

English Greek

.22*

a b

Fig. 4. The baseline path model of predictors of word-reading fluency in Grade 2 in English (a) and Greek (b). WRF¼Word-Reading Fluency; G1¼Grade 1;

G2¼Grade 2. Fit indices: for English, c2(5, N¼ 55)¼ 27.10, p< .001, CFI¼ .80, RMSEA¼ .24; for Greek: c2(7, N¼ 70)¼ 30.67, p< .001, CFI¼ .72,

RMSEA¼ .19. * p< .05; ** p< .01; *** p< .001.


what was reported above for WRF, namely that Letter Knowl-edge was more important in Greek than in English.

A closer look at the correlation coefficients between WRF/TRF and letter knowledge in the two languages revealed thatthere was a difference of r¼ .04, in favor of English in WRFand a difference of r¼ .20 in favor of English in TRF.Although these differences could explain the stronger effectsof letter knowledge on TRF in English compared to Greek,they could not explain the stronger effects of letter knowledgeon WRF in Greek compared to English. Thus, we shifted ourattention to the contribution of the rest of the predictors inthe path analyses models. Notably, the effects of Raven’sMatrices were dissimilar across reading fluency outcomes inEnglish. On one hand, an increase in the predictive strength ofRaven Matrices’ in Grade 1 WRF was accompanied bya similar decrease in the predictive strength of letter knowl-edge. On the other hand, a decrease in the predictive strengthof Raven’s Matrices in Grade 1 TRF was accompanied bya similar increase in the predictive strength of letter knowl-edge. Thus, the differential predictive strength of the rela-tionship between Letter Knowledge and reading fluencymeasures across languages may be explained by the differ-ential contribution of Raven’s Matrices on Grade 1 WRF andTRF in English.

4. Discussion

The present study examined the concurrent and longitu-dinal effects of different aspects of home literacy environment(HLE) and task-focused behaviour on phonological sensitivity,

TRF-G1 TRF-G2

-.29** -.44***

-.14-.07 -.08

.01

.87***

.01

PPVTRavenTask-FocusLetter

Knowledge

Englisha

Fig. 5. The baseline path model of predictors of text-reading fluency in Grade 2

G2¼Grade 2. Fit indices: for English: c2(6, N¼ 55)¼ 28.04, p< .001, CFI¼RMSEA¼ .20. * p< .05; ** p< .01; *** p< .001.

letter knowledge, nonword decoding, and reading fluency inchildren learning to read in an orthographically inconsistentlanguage (English) and in children learning to read in anorthographically consistent language (Greek). The resultsindicated that both direct teaching and parent’s beliefs weredifferentially related to phonological sensitivity and letterknowledge across languages. In addition, the effect of task-focused behaviour on nonword decoding and reading fluencywas stronger in English than in Greek. Letter knowledgeaccounted for unique variance in all Grade 1 reading outcomeswhile phonological sensitivity did not make a unique contri-bution in either language.

4.1. The impact of HLE aspects

The first aim of our study was to examine the contributionof different aspects of HLE on phonological sensitivity andletter knowledge. Our findings verified Hypothesis 1 and are inline with previous studies showing that shared book readingwas not associated significantly with either phonologicalsensitivity or letter knowledge (Evans et al., 2000; Foy &Mann, 2003; Frijters et al., 2000; Scarborough & Dobrich,1994; Senechal & LeFevre, 2002; Torppa et al., 2007). On theother hand, direct teaching of letter sounds, letter names, andwords was significantly related to phonological sensitivity inboth English and Greek. However, in Greek, the direction ofthe relationship was opposite to what would have beenexpected on the basis of previous studies conducted primarilyin English (Burgess, 2002; Evans et al., 2000; Foy & Mann,2003; Senechal & LeFevre, 2002). More frequent teaching of

TRF-G1 TRF-G2

-.35***

-.13

-.15 .03

.15

.75***

-.15*

PPVTRavenLetterKnowledgeTask-Focus

Greekb

in English (a) and Greek (b). TRF¼ Text-Reading Fluency; G1¼Grade 1;

.87, RMSEA¼ .22; for Greek, c2(6, N¼ 70)¼ 30.02, p< .001, CFI¼ .76,


letter sounds, letter names, and words in Greek was associatedwith lower scores on phonological sensitivity. One possibleexplanation may be that while English-speaking parents teachletter names and words to children irrespective of theirperformance on pre-literacy tasks, Greek parents start teachingtheir children letter names or words more frequently onlywhen they actually notice a potential learning difficulty onbehalf of their children. Thus, more frequent direct teachingon behalf of the parents in kindergarten may, in fact, be anearly indication of expectation of reading difficulties.

In addition, letter knowledge was significantly associatedwith direct teaching and not with shared book reading ornumber of books at home. Our findings confirmed Hypothesis2 and extended previous work in English (Evans et al., 2000;Frijters et al., 2000; Senechal & LeFevre, 2002) showing thatparents’ direct teaching had a stronger relationship with letterknowledge in English than in Greek. Thus, it seems safe toconclude that children, especially those who are learning toread an opaque orthography, acquire letter knowledge whenthey are taught letters directly by either parents or school andnot by just enjoying general literacy experiences, such as bookreading.

Beyond the contribution of direct teaching on phonologicalawareness and letter knowledge, parental education wassignificantly related to both skills in Greek, but not in English,although the difference was not statistically significant.Previous studies in Finnish also reported significant linksbetween mother’s education and phonological awareness(Leppanen, Aunola, Niemi, & Nurmi, 2008; Torppa et al.,2007). It is possible that parental education matters only if it isbelow a certain threshold level. The majority of the parents inAlberta (on average 71%) had completed at least somecollege, whereas the corresponding percentage in Rethymnowas much lower (on average 45%). However, the differencesmay also reflect the fact that parental education in Rethymnohad more variability than parental education in Alberta, andthus, better chances to correlate significantly with the criterionvariables.

4.2. The impact of task-focused behaviour

Task-focused behaviour was associated with letter knowl-edge in English, but not in Greek. However, the difference inthe strength of the path weights was not significant acrosslanguages. Thus, Hypothesis 3.1 that task-focused behaviourwould be associated more with phonological sensitivity andletter knowledge in English than in Greek was not confirmed.

4.3. The joint contribution of cognitiveand non-cognitive factors

The second aim of this study was to examine the jointcontribution of various HLE dimensions, task-focusedbehaviour, phonological sensitivity and letter knowledge onnonword decoding and reading fluency in Grades 1 and 2.Notably, none of the HLE components had an effect on any ofthe reading skills and this finding was consistent across

languages replicating the findings of previous studies (de Jong& Leseman, 2001; Senechal & LeFevre, 2002; Senechal et al.,1998). Task-focused behaviour predicted significantly allGrade 1 reading outcomes in English and nonword decodingin Greek. Letter knowledge, but not phonological sensitivity,predicted significantly all reading outcomes in both languagesand verified Hypothesis 5 only for the letter knowledge part.Finally, it is worth mentioning that letter knowledge predicteddirectly Grade 2 WRF in English verifying Hypothesis 6 andtask-focused behaviour predicted directly Grade 2 TRF inGreek, beyond the effects of autoregressors (reading ability inGrade 1).

The significantly higher contribution of task-focusedbehaviour in English than in Greek Grade 1 reading decoding,which confirmed Hypothesis 3.2, and reading fluencymeasures, which rejected Hypothesis 4, may reflect the factthat reading in English is a much more demanding task thanreading in Greek (Aro & Wimmer, 2003; Ellis et al., 2004). Ifchildren learning to read English take four times longer toachieve the same level of word-reading accuracy than childrenlearning to read Greek (Seymour et al., 2003), it seemsreasonable to expect that they will also need more motivationto overcome the difficulties imposed by the irregular ortho-graphic system.

The present study revealed also some important findingsconcerning the contribution of phonological sensitivity andletter knowledge on later reading achievement. Letter knowl-edge was the best predictor of reading outcomes in Grade 1 forboth language groups. In turn, phonological sensitivity did notpredict significantly any of the reading skills in eitherlanguage. This finding was in contrast to the vast majority ofstudies who have emphasized the importance of phonologicalsensitivity on reading development in English (Bryant,MacLean, & Bradley, 1990; Kirby et al., 2003; Parrila et al.,2004) as well as in some more consistent orthographies(Caravolas, Volin, & Hulme, 2005; Papadopoulos, 2001). Thisdiscrepancy may be attributed to several reasons. First, thenonsignificant effect of phonological sensitivity may have todo with the type of phonological sensitivity tasks used in thepresent study which are closely related with the onset ofreading. However, reading at the age of 5.5 years of age isunusual in both Canada and Greece. Particularly, the Greekchildren achieved a very low score on the phonologicalsensitivity task resulting in some restriction of range. Easierphonological sensitivity tasks, such as onset and rime tasks,would probably generate greater variability in scores andmaybe better chances for a significant contribution. Second,phonological sensitivity and letter knowledge were signifi-cantly correlated in both languages and inclusion of letterknowledge in the path analysis models necessarily limits therole of phonological sensitivity. It is noteworthy that most ofthe previous studies (Bryant et al., 1990; Caravolas et al.,2005) reporting significant relationships between phonologicalsensitivity and early reading skills did not control for child-ren’s letter knowledge. Similar to our findings, Leppanen et al.(2008) showed that when both letter knowledge and phono-logical awareness are included in the same model, letter


knowledge predicts reading skills in Grade 1 better thanphonological awareness.

In summary, the findings of the present study suggest thathow strongly different non-cognitive and cognitive factors arerelated to reading acquisition depends on the orthographychildren are learning to read. Number of books and sharedreading appear to be only weakly related to phonologicalsensitivity and letter knowledge across languages. More directteaching of letters appears to have a positive effect on letterknowledge across languages and on phonological sensitivity inEnglish. Nevertheless, only task-focused behaviour, among thenon-cognitive factors, survived as a predictor of reading oncethe effect of letter knowledge was taken into account. Letterknowledge predicted nonword decoding and reading fluencybetter than any other measure including phonologicalsensitivity.

4.4. Limitations

Some limitations of the current study are worth mentioning.First, our findings can be generalized only to the languagesunder investigation and for the ages of the participants we hadin our sample. Many researchers have used Greek as anexample of a consistent orthography (Ellis et al., 2004; Sey-mour et al., 2003) but there are still objections to thisconceptualization (Miles, 2000). To the extent that the purposeof this study was to identify differences in processes involvedin reading in a consistent orthography and in an inconsistentorthography, then using Greek might not have been an optimalsolution compared to languages such as Finnish or Turkish,which are considered close to 100% consistent.

Second, the reading accuracy and fluency tasks were notstrictly matched in the two languages. Greek and Englishbelong to different families of languages and have differentorthographic and phonological characteristics. For example,there is a large body of short single-syllable words in English,whereas there are only few of these words in Greek. Given thenumber of single-syllable words used in existing reading testsin English (see, e.g., Woodcock Reading Mastery Tests-Revised; Woodcock, 1998), it is not possible to constructword-reading tasks in Greek that would be strictly parallel interms of length and word frequency to the English ones. Onthe other hand, using more multi-syllabic words in Englishlikely will not create an equal task because of significantdifferences in the syllable structures between Greek andEnglish.

Third, information on HLE was collected by sending outa questionnaire to the parents. This may result in two kinds ofproblems: (a) as Bus et al. (1995) pointed out, the question‘‘How frequently do you read to your child?’’ may beambiguous and interpreted inconsistently by the parents. Forexample, some parents may classify two books read in one dayas a single reading event, whereas other parents may considertwo books read in a one day as two reading events; (b) therequirements to indicate frequency of reading at home,number of books, and frequency of teaching letter sounds and

words is subject to a social-desirability bias if parents attacha high value to these aspects of HLE.

Fourth, some children in our sample may have been readersalready by kindergarten. This may be particularly true forsome of our Greek participants as indicated by these children’sceiling or close to ceiling performance on the emergentliteracy skill measures. Consequently, pre-reading abilitycould have been the best predictor of their reading ability inlater grades. However, we did not measure reading inkindergarten and therefore we could not control for its effects.

In addition, two measures in our study, HLE and Task-Focus, may be culturally biased. Although some cultural biasis inevitable when conducting cross-linguistic studies, we tookall the necessary measures to ensure that the tasks and also theadministration procedures would be as similar as possibleacross languages following international standards (van deVijver & Hambleton, 1996). Beyond the culture bias, the Task-Focus questionnaire may also be biased (and therefore corre-lated with reading) by children’s behaviour/results in class.Perhaps an external instrument, such as observation of child-ren’s classroom behaviour would be preferable.

Sixth, because the two groups of children differed signifi-cantly on their reading performance, the predictive models fornonword decoding may be biased in Greek and the predictivemodels for fluency may be biased in English. Ideally, the twogroups should be matched on their reading ability before werun any cross-linguistic comparisons. However, given thedifferent reading development rates across languages (Sey-mour et al., 2003), matching the two groups on reading wouldintroduce other problems, such as differences in the age of theparticipants and unequal print exposure.

Finally, a larger sample size in both language groups wouldbe desirable. For the type of analyses we ran in our study thesample size was barely adequate. Thus, future studies shouldstrive to replicate these findings with a larger sample size.

4.5. Psycho-educational implications

The findings of the present study suggest that more frequentdirect teaching of letter names, letter sounds, and words athome is beneficial for the development of letter knowledge. Inturn, letter knowledge has been shown to be the best predictorof word decoding and reading fluency. Thus, parents as well askindergarten teachers should foster children’s letter knowledgeby teaching them directly the letter names and letter sounds.However, we should acknowledge that training letter namesalone may be ineffective on preventing reading failure (fora review see Adams, 1990) unless it is accompanied bytraining in phonological sensitivity (Schneider, Roth, &Ennemoser, 2000). Given the strong correlations betweenletter knowledge with phonological sensitivity in bothsamples, direct teaching should foster the links between printand sound.

Moreover, it is suggested that an early intervention to enhancetask-focused behaviour would be useful for children learning toread in both consistent and inconsistent orthographies. Recently,Andreassen, Knivsberg, and Niemi (2006) showed that intensive


individual counselling improved the reading performance ofa clinical sample of dyslexic students and that the observedprogress was not related to age, cognitive level, place of resi-dence, or previous special education received, but instead toimproved motivation. Given that remediation programsdesigned to increase phonological awareness have not alwaysbeen successful (Torgesen & Davis, 1996), it would be inter-esting to examine whether their success could be improved bysupplementing these programs with additional componentsfocusing on task-focused behaviour.

References

Adams, M. J. (1990). Beginning to read: Thinking and learning about print.

Cambridge, MA: MIT Press.

Andreassen, A. B., Knivsberg, A. M., & Niemi, P. (2006). Resistant readers 8

months later: energizing the student’s learning milieu by targeted coun-

seling. Dyslexia, 12, 115e133.

Aro, M., & Wimmer, H. (2003). Learning to read: English in comparison to six

more regular orthographies. Applied Psycholinguistics, 24, 621e635.

Aunola, K., Nurmi, J.-E., Lerkkanen, M.-K., & Rasku-Puttonen, H. (2003).

The roles of achievement-related behaviours and parental beliefs in

children’s mathematical performance. Educational Psychology, 23,

403e421.

Aunola, K., Nurmi, J.-E., Niemi, P., Lerkkanen, M.-K., & Rasku-Puttonen, H.

(2002). Developmental dynamics of achievement strategies, reading

performance, and parental beliefs. Reading Research Quarterly, 37,

310e327.

Aunola, K., Nurmi, J.-E., Parrila, R., & Onatsu-Arvilommi, T. (2000).

Behavioural strategy rating scale II - BSR-II. Unpublished test.

Bryant, P. E., MacLean, M., & Bradley, L. (1990). Rhyme, language, and

children’s reading. Applied Psycholinguistics, 11, 237e252.

Burgess, S. R. (2002). The influence of speech perception, oral language

ability, the home literacy environment, and pre-reading knowledge on the

growth of phonological sensitivity: a one-year longitudinal investigation.

Reading and Writing: An Interdisciplinary Journal, 15, 709e737.

Burgess, S. R., Hecht, S. A., & Lonigan, C. J. (2002). Relations of home

literacy environment (HLE) to the development of reading-related

abilities: a one-year longitudinal study. Reading Research Quarterly, 37,

408e426.

Bus, A. G., van Ijzendoorn, M. H., & Pellegrini, A. D. (1995). Joint book

reading makes for success in learning to read: a meta-analysis on inter-

generational transmission of literacy. Review of Educational Research, 65,

1e21.

Caravolas, M., Volin, J., & Hulme, C. (2005). Phoneme awareness is a key

component of alphabetic literacy skills in consistent and inconsistent

orthographies: evidence from Czech and English children. Journal of

Experimental Child Psychology, 92, 107e139.

Clay, M. M. (1993). An observation survey: Of early literacy achievement.

Portsmouth, NH: Heinemann.

Dally, K. (2006). The influence of phonological processing and inattentive

behaviour on reading acquisition. Journal of Educational Psychology, 98,

420e437.

Dunn, L. M., & Dunn, L. M. (1997). Peabody picture vocabulary test - Third

edition. Circle Pines, MN: American Guidance Service.

Ellis, N. C., Natsume, M., Stavropoulou, K., Hoxhallari, L., Van Daal, V. H. P.,

Polyzoe, N., et al. (2004). The effects of orthographic depth on learning to

read alphabetic, syllabic, and logographic scripts. Reading Research

Quarterly, 39, 438e468.

Evans, M. A., Shaw, D., & Bell, M. (2000). Home literacy activities and their

influence on early literacy skills. Canadian Journal of Experimental

Psychology, 54, 65e75.

Foy, J. G., & Mann, V. (2003). Home literacy environment and phonological

awareness in preschool children: differential effects for rhyme and

phoneme awareness. Applied Psycholinguistics, 24, 59e88.

Frijters, J. C., Barron, R. W., & Brunello, M. (2000). Direct and mediated

influences of home literacy and literacy interest on prereaders’ oral

vocabulary and early written language skill. Journal of Educational


Fyrsten, S., Nurmi, J.-E., & Lyytinen, H. (2006). The role of achievement

beliefs and behaviours in spontaneous reading acquisition. Learning and

Instruction, 16, 569e582.

Georgiou, G., Parrila, R., & Papadopoulos, T. (2008). Predictors of word

reading accuracy and fluency in English and Greek: a cross-linguistic

comparison. Journal of Educational Psychology, 100, 566e580.

Hagtvet, B. E. (2000). Prevention and prediction of reading problems. In

A. Badian (Ed.), Prediction and prevention of reading failure (pp. 105e132).

Baltimore, MD: York Press.

Holopainen, L., Ahonen, T., & Lyytinen, H. (2001). Predicting delay in

reading achievement in a highly transparent language. Journal of Learning

Disabilities, 34, 401e413.

de Jong, P. F., & Leseman, P. P. M. (2001). Lasting effects of home literacy on

reading achievement in school. Journal of School Psychology, 39, 389e414.

de Jong, P. F., & van der Leij, A. (1999). Specific contributions of phonological

abilities to early reading acquisition: results from a Dutch latent variable

longitudinal study. Journal of Educational Psychology, 91, 450e476.

Kirby, J. R., Parrila, R. K., & Pfeiffer, S. L. (2003). Naming speed and

phonological awareness as predictors of reading development. Journal ofEducational Psychology, 95, 453e464.

Lepola, J., Poskiparta, E., Laakkonen, E., & Niemi, P. (2005). Development of

and relationship between phonological and motivational processes and

naming speed in predicting word recognition in Grade 1. Scientific Studiesof Reading, 9, 367e399.

Lepola, J., Salonen, P., & Vauras, M. (2000). The development of motivational

orientations as a function of divergent reading careers from pre-school to

the second grade. Learning and Instruction, 10, 153e177.

Leppanen, U., Aunola, K., Niemi, P., & Nurmi, J.-E. (2008). Letter knowledge

predicts Grade 4 reading fluency and reading comprehension. Learning

and Instruction, 18(6), 548e564.

Leseman, P. M., & de Jong, P. F. (1998). Home literacy: opportunity,

instruction, cooperation and social-emotional quality predicting early

reading achievement. Reading Research Quarterly, 33, 294e318.

Levy, B. A., Gong, Z., Hessels, S., Evans, M. A., & Jared, D. (2006).

Understanding print: early reading development and the contributions of

home literacy experiences. Journal of Child Experimental Psychology, 93,

63e93.

Loehlin, J. C. (2004). Latent variable models: An introduction to factor, path,and structural equation analysis (4th ed.). Mahwah, NJ: Erlbaum.

Mann, V., & Wimmer, H. (2002). Phoneme awareness and pathways into

literacy: a comparison of German and American children. Reading andWriting: An Interdisciplinary Journal, 15, 653e682.

Miles, E. (2000). Dyslexia may show a different face in different languages.

Dyslexia, 6, 193e201.

Midgley, C., Arunkumar, R., & Urdan, T. C. (1996). If I don’t do well

tomorrow, there’s a reason’’: predictors of adolescents’ use of academic

self-handicapping strategies. Journal of Educational Psychology, 88,

423e434.

Onatsu-Arvilommi, T. P., & Nurmi, J.-E. (2000). The role of task-focused and

task-avoidant behaviours in the development of reading and mathematical

skills during the first school year: a cross-lagged longitudinal study.

Journal of Educational Psychology, 92, 478e491.

Papadopoulos, T. C. (2001). Phonological and cognitive correlates of word-

reading acquisition under two different instructional approaches. European

Journal of Psychology of Education, 16, 549e567.

Papadopoulos, T. C., & Georgiou, G. K. (2000). Parameters of reading

development in Greek language [In Greek]. In S. Georgiou, L. Kyriakides,

& K. Christou (Eds.), Contemporary research in educational studies (pp.

241e248). Nicosia, Cyprus: University of Cyprus Press.

Papoulia-Tzelepi, P. (2001). Continuity and change in Language Arts text-

books in Greek primary schools. L-1-Educational Studies in Language and

Literature, 1, 195e207.

Parrila, R., Kirby, J. R., & McQuarrie, L. (2004). Articulation rate, naming

speed, verbal short-term memory, and phonological awareness:


longitudinal predictors of early reading development. Scientific Studies ofReading, 8, 3e26.

Porpodas, C. (1999). Patterns of phonological and memory processing in

beginning readers and spellers of Greek. Journal of Learning Disabilities,

32, 404e416.

Porpodas, C. (2001). Cognitive processes in first grade reading and spelling of

Greek. Psychology: The Journal of the Hellenic Psychological Society, 8,

384e400.

Poskiparta, E., Niemi, P., Lepola, J., Ahtola, A., & Laine, P. (2003). Motiva-

tional-emotional vulnerability and difficulties in learning to read and spell.

British Journal of Educational Psychology, 73, 187e206.

Raven, J. C. (1956). Coloured progressive matrices: Sets A, AB, B. London:

Lewis.

Salonen, P., Lepola, J., & Niemi, P. (1998). The development of first graders’

reading skill as a function of pre-school motivational orientation and phonemic

awareness. European Journal of Educational Psychology, 2, 155e173.

Scarborough, H. S., & Dobrich, W. (1994). On the efficacy of reading to

preschoolers. Developmental Review, 14, 245e302.

Schneider, W., Roth, E., & Ennemoser, M. (2000). Training phonological skills

and letter knowledge in children at risk for dyslexia: a comparison of three

kindergarten intervention programs. Journal of Educational Psychology,

92, 284e295.

Senechal, M. (2006). Testing the home literacy model: parent involvement in

Kindergarten is differentially related to Grade 4 reading comprehension,

fluency, spelling and reading for pleasure. Scientific Studies of Reading, 10,

59e87.

Senechal, M., & LeFevre, J. (2002). Parental involvement in the development

of children’s reading skill: a five-year longitudinal study. Child Develop-

ment, 73, 445e460.

Senechal, M., LeFevre, J., Thomas, E. M., & Daley, K. E. (1998). Differential

effects of home literacy experiences on the development of oral and

written language. Reading Research Quarterly, 33, 96e116.

Seymour, P. H., Aro, M., & Erskine, J. M. (2003). Foundation literacy

acquisition in European orthographies. British Journal of Psychology, 94,

143e174.

van Steensel, R. (2006). Relations between socio-cultural factors, the home

literacy environment and children’s literacy development in the first years

of primary education. Journal of Research in Reading, 29, 367e382.

Stephenson, K., Parrila, R., Georgiou, G., & Kirby, J. (2008). Effects of home

literacy, parents’ beliefs and children’s task-focused behavior on emergent

literacy and word reading skills. Scientific Studies of Reading, 12, 24e50.

Stoep, J., Bakker, J., & Verhoeven, L. (2002). Parental and teacher commit-

ment to emergent literacy development. In L. Verhoeven, C. Elbro, &

P. Reitsma (Eds.), Precursors of functional literacy (pp. 249e264).

Amsterdam: Benjamins.

Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related

precursors to reading: evidence from a longitudinal structural model.

Developmental Psychology, 38, 934e947.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (3rd

ed.). New York: Harper Collins College Publishers.

Tafa, E., & Manolitsis, G. (2008). A longitudinal literacy profile of Greek

precocious readers. Reading Research Quarterly, 42, 165e185.

Torgesen, J. K., & Davis, C. (1996). Individual difference variables that predict

response to training in phonological awareness. Journal of Experimental

Child Psychology, 63, 1e21.

Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1999). Test of word reading

efficiency. Austin, TX: PRO-ED.

Torppa, M., Poikkeus, A.-M., Laakso, M.-L., Tolvanen, A., Leskinen, E.,

Leppanen, P. H. T., et al. (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, 73e103.

Torppa, M., Poikkeus, A.-M., Laakso, M.-L., Eklund, K., & Lyytinen, H.

(2006). Predicting delayed letter knowledge development and its relation

to Grade 1 reading achievement among children with and without familial

risk for dyslexia. Developmental Psychology, 42, 1128e1142.

van de Vijver, F. J. R., & Hambleton, R. K. (1996). Translating tests: some

practical guidelines. European Psychologist, 1, 89e99.

Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1999). Comprehensive testof phonological processing. Austin, TX: PRO-ED.

Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, A. S., Barker, T. A.,

Burgess, S., et al. (1997). Changing relations between phonological

processing abilities and word-level reading as children develop from

beginning to skilled readers: a 5-year longitudinal study. Developmental


Wiederholt, J. L., & Bryant, B. R. (2001). GORT 4: Gray oral reading test.Austin, TX: PRO-ED.

Wilson, K., & Trainin, G. (2007). First-grade students’ motivation and

achievement for reading, writing, and spelling. Reading Psychology, 28,

257e282.

Woodcock, R. W. (1998). Woodcock reading mastery test-revised-normative

update: Examiner’s manual. Circle Pines, MN: American Guidance

Service.

Beginning to read across languages varying in orthographic consistency: Comparing the effects of...

Documents

Transcript of Beginning to read across languages varying in orthographic consistency: Comparing the effects of...