English skills contribute more to English spelling skill than Dutch skills in pupils learning English as a second language, and language skills more than cognitive skills within each language

(1)

English skills contribute more to English spelling skill

than Dutch skills in pupils learning English as a second

language, and language skills more than cognitive skills

within each language

C. M. Gerbrandy

11711930

Assessor: dr. P. J. F. Snellings Supervisor: N. L. Leona, Msc Developmental Psychology,

part of the faculty of Social and Behavioural Sciences University of Amsterdam

Submission date: 19th_{of June 2020}

Wordcount: 5981

Contact information author

(2)

English skills contribute more to English spelling skill

than Dutch skills in pupils learning English as a second

language, and language skills more than cognitive skills

within each language

Gerbrandy, C. M. June 19th_{, 2020}

Abstract

Due to the globalization of the economy, the English language has become a universal language. The education of the English language as a second language is therefore important yet challenging, mainly because the orthography of the English language is a complex system: mapping phonemes to associated graphemes, which defines spelling, is complicated. Theories about the impact of native language competence on the process of learning a second language state that a transfer of native language skills will arise that either enhances this process or causes problems related to the native language. Previous research found different language and cognitive skills in both native and second language that contribute to the prediction of English spelling skill. The contribution of these language and cognitive skills has been researched separately. Therefore, the present study investigated the effect of English skills on English spelling skill while controlling for Dutch skills in Dutch primary school pupils, and the contribution of cognitive skills while controlling for language skills. The pupils performed the PPVT, PPVT-4, SVT, WRAT, CB & WL, CTOPP – RAN, FAT-R, CTOPP - Blending Words, NWR and CTOPP – NWR to respectively measure Dutch and English vocabulary, spelling skill, RAN with numbers and colours, phonological awareness and phonological memory. The regression revealed a high contribution of English vocabulary, English RAN with numbers and English phonological memory to English spelling skill in addition to the high contribution of Dutch vocabulary, Dutch spelling and Dutch phonological awareness. Specifically, Dutch spelling and English vocabulary were shown to contribute most. Therefore, it is concluded that English language and cognitive skills highly contribute to English spelling skill besides native language and cognitive skills in Dutch primary school pupils, and that cognitive measures besides language measures contribute to a small extent. These findings emphasize the importance of a pupil’s native language skills and a pupil’s second language skills for second language acquisition. It is crucial that we adapt our educational system accordingly because a well-studied method of native and second language teaching is essential in our increasingly multilingual society.

Keywords

English language; Dutch primary school pupils; phonological awareness; phonological memory; predictors of spelling; rapid automatized naming; second language learning; vocabulary

Introduction

Due to the globalization of the economy and our increasingly multilingual society, the English language has become a universal language. Of the world’s population, one

third uses English as their first or secondary language (Crystal, 2008). In addition to the economy, science contributed to this global spread as well, where English is the modern “lingua franca”. In the Netherlands, primary

(3)

schools are required to start teaching English language from the fifth grade, in Dutch ‘groep 7’. However, due to the fact that no strict guidelines exist about this education, schools are free to formalize it independently. Hence, the skill level of reading and writing in English varies between children from different schools after completing primary school (Ministry of Education, Culture and Science, 2019).

The English language is thought to have a complex system of orthography (Miller, 2019): mapping phonemes to associated graphemes, which defines spelling (Adams, 1990), is complicated. Hence, spelling skill in English is a good measure for general English skills. Due to the complexity, the education of English spelling could be challenging for Dutch pupils learning English as a second language. Children with dyslexia for example, experience problems with spelling in general (Vellutino, 1979) and will experience problems during this English

language learning process. In the

Netherlands, the prevalence of dyslexia among children ranging from seven to twelve years amounted eight percent in 2016, while this was six percent in 2009 (Central Bureau for Statistics, 2016). Apparently, the number of children having dyslexia in this age group is increasing. However, this could also be the consequence of improved diagnostics. Next to children with dyslexia, other children also face difficulties during the education of English spelling skill because of this complexity. At the same time, there will always be a group having no trouble learning the English language. Hence, individual differences among Dutch pupils appear in the process of learning to spell in English as a second language. It is important to define these individual differences to predict which factors contribute to the spelling skill of a pupil and to adapt our educational system accordingly.

Research has focused on the use of first language knowledge to the learning

process of a second language, since transferring knowledge is the goal of the educational system (Badford & Schwartz, 1999). Different theories about the impact of first language competence on second language learning have been put forward: the Linguistic Coding Difference Hypothesis (LCDH) proposed by Ganschow, Sparks and Javorsky (1998), states that native language skills influence learning the skills of a foreign language, meaning problems that will arise when learning a foreign language, are related to problems with native language acquisition. Furthermore, the Common Underlying Proficiency (CUP) theory states that there is a common underlying cognitive proficiency across languages that provides the transfer of knowledge and skills across languages (Cummins, 1980). According to these theories, the knowledge of someone’s first language either enhances learning a second language or interferes with it (Yan, 2010). The CUP would provide the transfer of developed knowledge and skills of someone’s native language to someone’s second language, causing enhancement in learning a second language. According to the LCDH, however, problems with learning a second language are related to the native language, since native language problems would influence learning a foreign language. A pupil’s command of native language skills would therefore have a predictive function towards the learning process of a second language.

Regarding the transfer of one’s native language to a second language, research found a cross-language influence of Dutch vocabulary on reading comprehension in English in Dutch pupils with an average age of 12 years (Schoonen, Hulstijn & Bossers, 1998). This is supported by research showing a correlation between Chinese vocabulary knowledge and English spelling skill in eight to ten years old Chinese pupils (Li, McBride-Chang, Wong & Shu, 2012). Additionally, vocabulary knowledge of a pupil’s second

(4)

comprehension in that language, as shown by Van Gelderen and colleagues (2007) who found that English vocabulary contributed

substantially to English reading

comprehension in Dutch pupils. Besides vocabulary knowledge, writing proficiency in a pupil’s native language also shows a strong contribution to the prediction of English writing proficiency as a second language, as shown in Chinese pupils (Li, McBride-Chang, Wong & Shu, 2012) and Dutch pupils (Schoonen et al., 2003).

Besides these important language factors, cognitive factors are also important to take into account since learning a language is a cognitive process. Especially in children cognitive factors are important to analyse because of the developing brain. Children in primary school are early adolescents, which means their cognitive functions are still evolving. In the early development of a child, phonological awareness begins to develop, which is a cognitive skill that allows an individual to recognize, distinguish and manipulate sounds in language (McBride-Chang, 1995). Deficits with phonological awareness are associated with dyslexia (Fawcett & Nicolson, 1995). This cognitive skill shows a strong contribution to the prediction of spelling skill in English for eight-year-old native Canadian English speakers (Harrison, Goegan, Jalbert, McManus, Sinclair & Spurling, 2016). Additionally, for pupils learning English as a second language, phonological awareness of one’s native language contributes to the prediction of spelling in English as a second language, as was shown in eight-year-old Spanish pupils (Sun-Alperin & Wang, 2011). Furthermore, in

non-native English pupils, English

phonological awareness contributes to the prediction of English spelling as well, as Wolff (2014) found in nine-year-old Swedish pupils learning English as a second language. These findings suggest that phonological awareness in both a pupil’s native language and English

as a second language contribute to the prediction of English spelling skill.

Phonological awareness does not only help to discriminate sounds, it is also an important component of processing spoken and written language, a process known as phonological processing (Wagner & Torgesen, 1987). A task often used to measure phonological processing is the rapid automatized naming task (RAN), during which participants have to name words as quickly as possible. In native British English-speaking pupils, RAN strongly correlates with English spelling skill (Moll et al., 2014). For pupils learning English as a second language, RAN in one’s native language also correlated strongly with English spelling skill, as was shown in Chinese pupils ranging from 8 till 10 years old (McBride-Chang, Wong & Shu, 2012). Besides the contribution of a pupil’s native RAN score to English spelling skill, a pupil’s English RAN score likewise contributes to English spelling skill, in the case of English as a second language (Jonge, Verhoeven & Siegel, 2007). This was also reported by Adams, Everatt, Ocampo & Smythe (2000), who showed that English second language learning (ESL) pupils with an average English spelling skill, outperformed ESL-pupils with a poor English spelling skill on RAN with English words.

Phonological processing,

subsequently, relies on phonological working memory. The phonological loop is in turn part of working memory, stated by Baddely (1992) in his working memory model. In the German language, 9-year-old pupils with an age-adequate spelling skill outperformed poor spellers on phonological loop skills (Brandenburg et al., 2015). Besides

this influence of native language

phonological processing to a pupil’s native spelling skill, also phonological processing skills in English correlated strongly with English spelling skill in eleven-year-old native Mandarin children who learned English as a second language (Yeong, Fletcher & Bayliss,

(5)

2014). Moreover, research found that English phonological processing explained 34.2 percent of the variance in English spelling skill in 5-year-old pupils learning English as a second language (Keilty & Harrison, 2015). Hence, also a pupil’s second language phonological processing influences spelling skill in that language. Because phonological processing is a function of phonological working memory, these findings provide evidence for phonological working memory being a reliable predictor for English spelling skill as a second language.

Summarizing the previous research above, language and cognitive skills in both native and second language seem to contribute to the prediction of spelling skill in English as a second language for primary school pupils. While the contribution of these language and cognitive skills to English spelling skill has been shown separately, it remains unknown what the contribution of language and cognitive skills would be together. Analysing the additional value of cognitive skills while controlling for language skills within a language, would provide more information about a pupil’s learning process of a second language. The current method of teaching a second language to pupils could therefore be improved in our educational

system. Because of the unknown

contribution of language and cognitive skills together to the learning process of English spelling skill and the influence of native and foreign language skills on each other, the present study will further investigate the effect of English skills while controlling for Dutch skills on the prediction of English spelling skill in Dutch primary school pupils, and the additional contribution of cognitive skills while controlling for language skills.

In order to explore these predictors, Dutch and English language skills and cognitive skills were measured in Dutch

primary school pupils with different Dutch and English language and cognitive tests, to determine the importance of each predictor for English spelling skill. Based on previous research it is expected that respectively Dutch and English vocabulary, Dutch spelling skill, Dutch and English phonological awareness, Dutch and English performance on RAN and Dutch and English phonological memory, will be important predictors for English spelling skill. It is hypothesised that these English language and cognitive skills will together show a high additional contribution to the prediction of English spelling skill while controlling for the Dutch language and cognitive skills. Regarding the additional contribution of cognitive skills, it is also hypothesised that within both languages the cognitive skills will explain additional variance in English spelling skill besides the variance explained by the language skills. Dutch spelling skill will furthermore show a high contribution to the prediction of English spelling skill.

To specify, based on earlier research, English vocabulary, RAN, phonological awareness and phonological memory are expected to show a high contribution to the prediction of English spelling while controlling for these Dutch language and cognitive skills. Moreover, English

phonological memory, phonological

awareness and RAN are expected to explain an additional variance in English spelling skill besides the explained variance of English vocabulary. Dutch phonological memory, phonological awareness and RAN are likewise expected to explain an additional variance to English spelling skill besides the explained variance of Dutch vocabulary and spelling skill. Lastly, Dutch spelling is expected to show a high contribution to the prediction of English spelling skill.

(6)

Materials and methods

The present study included 276 Dutch primary school pupils (average age 11.75 years) in the sixth grade, in Dutch ‘groep 8’, with parental consent, of whom 52% were boys and 48% were girls. Pupils were included from six elementary schools in the province of North Holland and from one in the province of South Holland. All skills were measured in the sixth grade, with the exception of Dutch phonological awareness being measured one year before, in the fifth grade.

All pupils had to complete one test for each language or cognitive skill, resulting in ten different tests in total. The tests were either measured in class or individually. When a test was measured in class, a tablet was used; individually measured online tests were performed with laptops. For the individually measured tasks, three different test setups were used: setup 1 was used for online tasks during which the pupil and test administrator sat next to each other with the pupil behind the laptop. Setup 2 was used for offline tasks with the test administrator and pupil sitting across from each other. Setup 3 was used for online tasks during which a pupil was not allowed to see the screen. The test administrator and the pupil also sat across from each other in this setup. During all individually measured tests, a voice-recorder was used to record the answers given by the pupils.

Vocabulary

Dutch vocabulary was measured in class with the Peabody Picture Vocabulary Test (PPVT) (Schlichting, 2005) during which the pupils had to match a Dutch spoken word with a picture, which they chose out of four pictures. This test normally consists of 17 sets, each containing 12 words, but for the present study only set 9 up to and including set 13 were performed, resulting in a total of 60 test items. The number of correctly chosen pictures was used as a measure for

the Dutch vocabulary of a pupil. English vocabulary was also tested in class with the Peabody Picture Vocabulary Test Fourth Edition (PPVT-4) (Dunn & Dunn, 2007). This test followed the same procedure as the test measuring Dutch vocabulary, with English words spoken by a native British English speaker instead of Dutch words. Also, this test normally consists of 17 sets, each containing 12 words. Since this test is being developed for measuring the vocabulary of native British English speakers, Dutch pupils only had to perform set three up to and including set nine. This resulted in a total of 84 test items. The same method for obtaining a score was applied as in the Dutch version. Spelling

Dutch spelling skill was measured in class with the Dutch ‘Schoolvaardigheidstoets Spelling’ (SVT) (Braams & de Vos, 2015), which was divided into two categories: nouns and verbs. During these tests, a sentence was spoken with a repeated word of the sentence at the end, which the pupils had to write down. For the noun category, pupils had to finish two blocks of 15 sentences, resulting into a total of 30 sentences. For the verb category, pupils had to finish two blocks of 16 sentences, resulting into a total of 32 sentences. These blocks were of ascending difficulty. The number of correctly spelled words and verbs was used as a measure for the Dutch spelling skill of a pupil. Afterwards, English spelling skill was measured in class with the Wide Range Achievement Test (WRAT) (Wilkinson & Robertson, 2006) during which the pupils were instructed to complete 20 trials. Originally, this test is developed for native British English speakers and contains 42 trials with an ascending difficulty. Since the test is being performed by non-native British English pupils, it was chosen to only measure the first 20 trials, of which one was a practice trial. In each trial, a pupil heard a recording of a native British English speaker, during which they first heard

(7)

the target word, followed by a sentence containing this target word and afterwards a repetition of the target word. The pupil then had to type this target word. The same method for obtaining a score was applied as in the Dutch version.

RAN

RAN with Dutch words was tested individually with the Dutch ‘Continu Benoemen en Woorden lezen’ (CB&WL) (van den Bos & Lutje Spelberg, 2007), using test setup 2. This test was divided into two categories: colours and numbers, each containing five different items. For each category, a map was made with five columns of ten items, resulting in 50 items in total. The five different items were placed in a random order on the colour and number map. The pupils had to name the items on both maps as quickly as possible. The naming time of each map and the mistakes per map were

recorded separately by the test

administrator. A mistake was made when a pupil missed an item, named an item wrong or when an item was dictated to the pupil after five seconds. If a pupil named an item wrong, but instantly corrected himself, it counted as a right answer. RAN with English words was tested individually with the Comprehensive Test of Phonological Processing Rapid Automatized Naming (CTOPP) (Wagner, Torgesen, Rashotte & Pearson, 1999), using test setup 2. This test was also divided into a colour and number category with six different items each. First, the pupils heard the six different English colours and numbers spoken by a native British English speaker. Then the pupils practiced with these items. Thereafter, the procedure followed the same steps as in the test measuring RAN with Dutch words, with four columns containing nine items instead of five containing ten.

Phonological awareness

Phonological awareness in Dutch was tested individually using test setup 3 with the Dutch ‘Fonemische Analyse Test’ (FAT-R) (de Groot, van den Bos & van der Meulen, 2014), during which pupils heard two words played out by a laptop. They were instructed to swap the first letters of both words and to pronounce the new formed words. A good answer was given when the new formed words were correctly pronounced by a pupil. The test administrator recorded the reaction times per trial and the amount of good answers by using the laptop. The pupils had to finish three practice trials, followed by 12 test trials. A score for each pupil was formed based on age, the reaction times per word and the amount of good answers. Phonological awareness in English was measured individually with the Comprehensive Test of Phonological Processing (CTOPP) Blending Words (Wagner, Torgesen, Rashotte & Pearson, 1999), also in test setup 3. Pupils were instructed to combine different sounds to create an existing English word. An audio fragment could be repeated once, only when the pupil asked. They had to complete six practice trials and twenty test trials. If a pupil gave three wrong answers in a row, the test ended. For each good answer, a pupil obtained one point. The total amount of points was used as a measure for the phonological awareness in English of a pupil. Phonological memory

Phonological memory in Dutch was tested individually with the Dutch ‘Non-woord Repetitietaak’ (NWR) (French & O’Brien, 2008), in test setup 3. During this task, pupils were instructed to repeat nonsense words ranging from three till five syllables, after a computer played them out. The pupils had to repeat three practice words followed by 22 test words and obtained points if they pronounced the nonsense words correct. A spoken word could be repeated once, only when a pupil asked. The number of correctly

(8)

pronounced words was used as a measure for the Dutch phonological memory of a pupil. Phonological memory in English was

measured individually with the

Comprehensive Test of Phonological Processing Nonword Repetition (CTOPP – NWR) (Wagner, Torgesen, Rashotte & Pearson, 1999). This test followed the same procedure as the test for phonological memory in Dutch, with English nonsense words ranging from one till seven syllables, spoken by a native British English speaker instead of Dutch words. The pupils had to complete 18 trials after completing three practice trials. If a pupil repeated all three practice items wrong, the test trials could not be performed by the pupil. The test also ended when a pupil incorrectly repeated three nonsense words in a row. The same scoring method was used as in the Dutch version.

Data analysis

The data used in the present study was previously collected in the ORWELL-project (ORal and Written English Language Learning), which is a project of the Rudolf Berlin Center at the University of Amsterdam. To test whether each measured predictor contributed to the variance in the dependent variable, English spelling skill, a multiple hierarchical regression was performed in R studio (version 1.1.463). Prior to running the regression, a relevant descriptive analysis was performed, all intercorrelations were calculated and all relevant assumptions were tested. The assumption of multivariate normality was tested with the Shapiro-Wilk test. The assumptions of linearity, no multicollinearity and homoscedasticity were visually analysed by plotting the data. The regression contained all predictors, with the exception of Dutch RAN with colours, due to the non-significant correlation with English spelling skill. To control for Dutch measures

and to test the additional variance explained by English measures on English spelling skill, all Dutch measures were entered first. Furthermore, to control for language measures and to test the additional variance explained by cognitive measures within a language of English spelling skill, all languages measures were entered first within a language. This resulted in the following sequence of predictors in the regression: Dutch vocabulary, Dutch spelling, Dutch RAN

with numbers, Dutch phonological

awareness, Dutch phonological memory, English vocabulary, English RAN with numbers, English RAN with colours, English

phonological awareness and English

phonological memory.

Results

The goal of this study is to investigate the effect of Dutch and English skills on the prediction of English spelling skill in Dutch primary school pupils and the additional contribution of cognitive skills while controlling for language skills. Different Dutch and English language and cognitive tests were measured in Dutch primary school pupils to determine the importance of each predictor for English spelling skill. A total of 12 variables were measured, of which 11 predictors and the dependent variable: English spelling skill. The relevant scores of all pupils for each variable is shown in table 1. These relevant scores are respectively maximum possible score, range of scores, the average score, the standard deviation and

the percentage of average score.

Importantly, as shown in table 1, the maximum scores of both vocabulary and spelling differ between Dutch and English, causing the high difference in means between these Dutch and English variables. Therefore, the percentage of the average score was also calculated to give insight into the performance of each category.

(9)

Variable Maximum possible score

Range of scores

(min-max) Mean (SD) Percentage of average score Dutch vocabulary 60 12-52 33.225 (7.451) 55.38 English vocabulary 84 23-84 62.975 (10.577) 74.97 Dutch spelling skill 66 12-64 46.717 (9.922) 70.78 English spelling skill 19 1-19 9.598 (3.909) 50.52 Dutch RAN (numbers) NA 13.25-45.91 22.285 (4.410) NA Dutch RAN (colours) NA 22.81-79.18 38.411 (8.445) NA English RAN (numbers) NA 10.7-38.5 20.201 (5.248) NA English RAN (colours) NA 16.4-70.41 31.362 (7.801) NA Dutch phonological awareness NA 20-63 49.486 (8.188) NA English phonological awareness 20 4-19 13.659 (2.932) 68.30 Dutch phonological memory 22 5-22 15.351 (3.783) 69.78 English phonological memory 18 6-18 12.949 (2.492) 71.94

Table 1: The relevant scores of all participants per variable. This table displays all measured predictors for English spelling skill with their corresponding maximum possible score, range of scores, mean, standard deviation and percentage of average score.

Afterwards, the intercorrelations between these 12 variables were calculated and are displayed in table 2. As shown in table 2, English spelling skill of Dutch primary school pupils was significantly correlated with Dutch vocabulary, Dutch spelling, Dutch RAN with numbers, Dutch phonological awareness, Dutch phonological memory, English vocabulary, English RAN with numbers, English RAN with colours, English

phonological awareness and English

phonological memory. Only the performance

on Dutch RAN with colours was not significantly correlated with English spelling skill. Additionally, within the Dutch language high and significant correlations were found between Dutch vocabulary and spelling (.42), Dutch phonological awareness and spelling (.47) and Dutch RAN with colours and numbers (.54). Within the English language,

the correlations between English

phonological memory and phonological awareness (.43) and between English RAN with colours and numbers (.43) was high and

(10)

significant. Moreover, between both languages, high and significant correlations were found between Dutch and English vocabulary (.38), Dutch and English RAN with numbers (.48), Dutch and English RAN with colours (.63), Dutch and English phonological awareness (.40) and Dutch and English phonological memory (.42). A notable pattern of associations was found for phonological memory, RAN and vocabulary measures. The strength of the correlation for Dutch phonological memory, Dutch RAN with numbers, Dutch RAN with colours and Dutch vocabulary with English spelling skill was considerably weaker (.13, -.15, -.08 and .24) than for these English measures (.34, .38, -.37, and .56).

To address the research question, a ten-step multiple hierarchical regression analysis was run to determine the influence of the Dutch and English predictors on English spelling skill in Dutch primary school pupils. Prior to running the multiple hierarchical regression, all relevant assumptions were tested. Plotting the correlation of each factor with the outcome variable, English spelling skill, revealed linearity between each factor and English spelling skill. Secondly, the assumption of multivariate normality was met since the residuals were normally distributed (W = .99, p = .13). The calculated correlations furthermore revealed no multicollinearity since the correlations between all independent variables was under .7 (table 2). Lastly, plotting the standardized residuals versus the predicted values showed a homogeneity of variance, hence the assumption of homoscedasticity was met.

Since all relevant assumptions were met, the ten-step multiple hierarchical regression was performed. The variable of Dutch RAN score with colours was left out due to the non-significant correlation with English spelling skill. In the first five steps, only Dutch predictors were entered and consecutively English predictors. Moreover, language skills (vocabulary and spelling) were

entered first followed by cognitive skills. The ten-step multiple hierarchical regression analysis is displayed in table 3. As shown at step one, Dutch vocabulary contributed significantly to English spelling skill and explained 5.3% of the variance (F (1, 274) = 16.26, p < .001, R2_{= .053). Adding Dutch}

spelling in step 2 also revealed a significant contribution to English spelling skill and accounted for another 16.1% in English spelling skill (F (2,273) = 28.43, p < .001, ΔR2

= .161, after the first step, delta R2_{is used to}

illustrate the additional explained variance for each step). This addition interestingly caused Dutch vocabulary to lose its significant contribution to English spelling skill (β = .081, t (273) = .831, p = .407). The

addition of Dutch RAN with numbers in step 3 did not reveal any significant contribution to the model (F (3,272) = 26.77, p = .089, ΔR2

= .005). In step 4, Dutch phonological awareness was entered in the model and contributed significantly and explained another 2.7% of the variance in English spelling skill (F (4,271) = 23.45, p = .001, ΔR2 ₌

.027). The addition of Dutch phonological memory in step 5 did not lead to a significant contribution and resulted in a decrease in the total R2_{(F (5,270) = 18.71, p = .809, ΔR}2₌

-.003). This was however not significant. At step 6, the addition of English vocabulary showed a significant contribution and explained an additional 24.2% of the variance in English spelling skill (F (6,269) = 44.2, p < .001, ΔR2_{= .242). This step}

interestingly resulted in Dutch RAN with numbers (β = -.005, t (269) = -2.308, p = 0.022) and Dutch vocabulary (β = -0.228, t (269) = -2.629, p = 0.009) to become significant contributors for English spelling skill. Entering English vocabulary furthermore caused Dutch phonological awareness to increase in significant contribution to the model (β = .004, t (269) = 3.362, p < .001).

Adding English RAN with numbers in step 7, explained an additional 1.6% of the variance in English spelling skill (F (7,268) = 40.5, p =

(11)

.002, ΔR2_{= .016). Because of the addition of}

English RAN with numbers, Dutch RAN with numbers lost his significance (β = -.001, t (268) = - .496, p = .620). This addition also caused Dutch vocabulary (β = -.201, t (268) = -2.350, p = .020) and Dutch phonological awareness (β = .004, t (268) = 3.252, p = .001) to contribute less significantly to the model. At step 8, English RAN with colours was added and did not significantly contribute to English spelling skill (F (8, 267) = 35.67, p = .235, ΔR2_{= .001).}

The addition of English phonological awareness in step 9 showed a significant contribution to English spelling skill (F (9,266) = 32.78, p = .024, ΔR2_{= .008) and revealed an}

additional explained variance of .8%. This step caused Dutch vocabulary to become a more significant contributor (β = -.224, t (266) = -2.619, p = .009) and Dutch phonological awareness to become a less significant contributor (β = .003, t (266) = 2.247, p = .025). At the last step, step 10, English phonological memory was added and revealed a significant addition of .6% in explained variance (F (10,265) = 30.32, p = .037, ΔR2_{= .006), causing English}

phonological awareness to lose its significant contribution (β = .120, t (265) = 2.100, p =

.037). The addition of English phonological memory furthermore resulted in Dutch vocabulary (β = -.218, t (265) = -2.564, p = .011) and English RAN with numbers (β = -.005, t (265) = -2.509, p = .013) to contribute less significantly to the model.

At the last step, all independent variables together accounted for 51.6% of the variance in English spelling skill. This step revealed that Dutch vocabulary (-.22), Dutch spelling skill (.42), Dutch phonological awareness (.003), English vocabulary (.71), English RAN with numbers (-.005) and English phonological memory (.16) were significant predictors for English spelling skill. The regression furthermore revealed that these significant English predictors explained 26.4% of the variance in English spelling skill after the significant Dutch predictors were added, which explained 24.1% of the variance. Before the English variables were added, Dutch phonological awareness accounted for an additional 2.9% besides the variance explained by vocabulary (5.3%) and spelling skill (16.1%). Within the English language, the additional variance explained by RAN with numbers and phonological memory was 3.1% besides the 24.2% variance explained by the language predictor English vocabulary.

(12)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 1. Dutch phonological awareness -

2. English phonological awareness .40*** -

3. Dutch RAN (numbers) -.19** -.03 -

4. Dutch RAN (colours) -.22*** -.01 .54*** -

5. English RAN (numbers) -.21*** -.17** .48*** .34*** -

6. English RAN (colours) -.37*** -.13* .39*** .63*** .43*** -

7. Dutch phonological memory .30*** .39*** <.001 .01 -.16** -.05 -

8. English phonological memory .25*** .43*** -.01 <.01 -.21*** -.14* .42*** -

9. Dutch vocabulary .30*** .28*** .14* .01 <.01 -.13* .18** .19** -

10. English vocabulary .15* .22*** .11 -.03 -.25*** -.25*** .10 .31*** .38*** -

11. Dutch spelling .47*** .26*** -.16** -.09 -.16** -.27*** .15* .17** .42*** .19** -

12. English spelling .39*** .32** -.15* -.08 -.38*** -.37*** .13* .34*** .24*** .56*** .47*** -

Table 2: The calculated intercorrelations between all measured variables. This table displays the correlation coefficient r of all calculated intercorrelations between the 12 measured variables. All variables are significantly correlated with English spelling skill, except for Dutch RAN with colours which has a correlation of -.08 with English spelling skill. One asterisk shows a significant correlation with α < .05, two with α < .01 and three with α < .001.

(13)

English spelling skill Variables β t _R2 ∆R2 Step 1 .053 .053*** Dutch vocabulary .392 4.033*** Step 2 .214 .161*** Dutch vocabulary .081 .831 Dutch spelling .607 7.567*** Step 3 .219 .005 Dutch vocabulary .120 1.203 Dutch spelling .573 6.948***

Dutch RAN (numbers) -.004 -1.706

Step 4 .246 .027**

Dutch vocabulary .068 .687

Dutch spelling .470 5.413***

Dutch phonological awareness .005 3.265**

Step 5

Dutch vocabulary .066 0.662 .244 -.003

Dutch phonological memory .016 0.242

Step 6 .485 .242***

Dutch vocabulary -.228 -2.629**

Dutch RAN (numbers) -.005 -2.308*

Dutch phonological awareness .004 3.362***

Dutch phonological memory .002 0.037

English vocabulary .867 11.302***

Step 7 .501 .016**

Dutch vocabulary -.201 -2.350*

Dutch RAN (numbers) -.001 -.496

Dutch phonological memory -.023 -.428

English RAN (numbers) -.006 -3.120**

Step 8 0.502 .001

Dutch RAN (numbers) < -.001 -.194

Dutch phonological memory -.017 -.308

(14)

Table 3: The multiple hierarchical regression analysis predicting English spelling skill. This table shows the multiple

hierarchical regression analysis predicting English spelling skill In Dutch primary school pupils. The last step, step 10, consists of all variables (with the exception of Dutch RAN with colours), and explains 51.6% of the variance in English spelling skill. One asterisk shows a significant correlation with α < .05, two with α < .01 and three with α < .001.

English RAN (colours) -.002 -1.189

Step 9 .510 .008*

Dutch vocabulary -.223 -2.619**

Dutch RAN (numbers) -.001 -.220

Dutch phonological awareness .003 2.247*

Dutch phonological memory -.053 -0.931

English phonological awareness .159 2.276*

Step 10 0.516 .006*

Dutch RAN (numbers) < -.001 -.284

Dutch phonological awareness .003 2.242*

Dutch phonological memory -.090 -1.523

English RAN (numbers) -.005 -2.509*

English phonological awareness .120 1.683

(15)

Discussion

The present study examined English spelling skill of Dutch primary school pupils to gain insight into the degree of prediction of English and Dutch language and cognitive skills towards English spelling skill. Research showed both language and cognitive skills to be important predictors for spelling separately. However, the predictive value of cognitive measures while controlling for language measures has been understudied. Therefore, this study specifically addressed the additional contribution of cognitive measures besides language measures for the prediction of English spelling skill and the additional predictive value of English skills besides Dutch skills.

The regression revealed that English vocabulary, English RAN with numbers and English phonological memory significantly contributed to the model predicting English spelling skill, besides the significant contribution of Dutch vocabulary, Dutch spelling skill and Dutch phonological awareness to this prediction. Additionally, within the Dutch language, phonological awareness contributed to the prediction of English spelling skill besides the contribution of vocabulary and spelling. Within the English language, RAN with numbers and phonological memory likewise contributed to this prediction, besides the contribution of vocabulary. These results are consistent with the hypotheses stated in the introduction. It can therefore be concluded that English language and cognitive skills contribute highly to the prediction of English spelling skill in Dutch primary school pupils besides Dutch language and cognitive skills. It is secondly concluded that cognitive skills additionally contribute to the prediction of English spelling skill, besides the language skills within each language. Lastly, performance on Dutch spelling contributes highly to the prediction of English spelling skill in Dutch primary school pupils.

The additional significant contribution of English vocabulary, RAN with numbers and phonological memory accounted for 26.4 percent of the variance in addition to the variance explained by Dutch vocabulary, spelling

skill and phonological awareness, which accounted for 24.1 percent. This additional explained variance of these English skills seems substantial compared to the explained variance of the native language. This emphasizes the importance of second language knowledge and skills on second language acquisition. However, the high beta-value of Dutch spelling skill (.42) in the last step of the regression also emphasizes the importance of native language skills on second language acquisition. This is consistent with the LCDH and the CUP theory stating that during second language learning, a transfer of native language knowledge and skills to the second language learning process arises. Furthermore, the additional explained variance of Dutch phonological awareness besides vocabulary and spelling accounted for only 2.7 percent. Within the English language, phonological memory and RAN with numbers explained an additional 2.2 percent besides vocabulary. Thus, cognitive measures do show an additional contribution in addition to language measures, but it is much smaller.

Analysing the regression step by step reveals that entering Dutch spelling skill in step two interestingly caused Dutch vocabulary to lose its significance. Dutch vocabulary and Dutch spelling are significantly correlated (.42) and therefore could share explained variance of English spelling skill. Dutch spelling is likely to explain the majority due to the higher correlation with English spelling skill (.47 compared to .24), hence Dutch vocabulary loses its significance. Contrarily, entering English vocabulary in step six caused Dutch vocabulary to become significant again. Since both vocabulary measures correlate with .38, it is likely that they share explained variance of English spelling skill. Moreover, adding more variables to a model causes the model to improve in predictiveness. When adding English vocabulary, the model already consists of five variables and only explains 24.4 percent of the variance in English spelling. Entering English vocabulary causes the model to improve because an additional 24.2 percent of variance in English spelling is explained. Because this

(16)

additional explained variance is substantial and Dutch and English vocabulary are likely to share explained variance, Dutch vocabulary becomes and remains a significant predictor. This result is consistent with the CUP theory because it shows that a native language skill contributes to a second language skill, so a transfer of knowledge and skills arises.

Entering English vocabulary also causes Dutch RAN with numbers to become significant. However, when entering English RAN with numbers in step seven, Dutch RAN with numbers is no longer significant. English and Dutch RAN with numbers are significantly correlated (.48), meaning that a pupil with a high performance on the Dutch version is likely to perform well on the English version. This high correlation may cause both variables to share explained variance, with English RAN explaining the majority because of the higher correlation with English spelling skill (-.38 compared to -.15). In addition to this shared explained variance, English RAN could also explain unique variance of English spelling skill because the English version provides specific English information. Hence, the Dutch RAN variable loses its significance because it no longer provides new information and becomes redundant.

At the last step of the regression, English phonological memory is entered and causes English phonological awareness to lose its significance. These variables are significantly correlated (.43). This could implicate a shared variance of English spelling skill. Considering, however, both have a strong correlation with English spelling skill, .32 with English phonological awareness and .34 with English phonological memory, it does not appear that phonological memory provides the majority of the shared explained variance. The reason phonological awareness loses its significance could lie in the function of both brain components: phonological awareness relies on phonological memory, so adding phonological memory causes phonological awareness to not provide new information anymore, it therefore becomes redundant. This last step furthermore reveals that within the Dutch language, vocabulary, spelling skill and phonological

awareness are significant predictors for English spelling skill and within the English language, phonological memory, RAN with numbers and vocabulary. This implicates that besides language specific skills, also general language skills, such as phonological awareness and phonological memory, from either the primary or secondary language are important during the learning process of a second language.

There are multiple limitations of the present study important to note. Dutch phonological awareness was measured one year before all other predictors. Cognitive functions are not evolved in pupils yet, so measuring phonological awareness in the same year with all other predictors could make a difference in the contribution of phonological awareness to English spelling skill because it would accordingly be more developed. Furthermore, all schools that participated were located in the province of North Holland and South Holland. Spreading this schools more evenly in the country, would give a more valid and reliable view of Dutch primary school pupils in general. Also, more valid and reliable results could be obtained by measuring pupils originally from the Netherlands, since not all participating pupils were born in the Netherlands and speak Dutch at home. Moreover, both the Dutch and English language are West-Germanic languages (Kufner, 1972). As these languages share the same root, the Dutch language might show resemblance to the English language, making it easier to learn than pupils with for instance a native language with Latin roots. Replicating the procedure of the present study with pupils with a non-Germanic native language would therefore be interesting, to see whether languages have to share a common root in order to be of use when learning a second language.

The findings of the current study emphasise the importance of knowledge and skills of a pupil’s native and second language for second language acquisition. It is therefore crucial to focus on a well-studied method for teaching both and to adapt our educational system accordingly. As the current study shows that both Dutch vocabulary and spelling skill are important predictors, an improved method

(17)

should prioritise both and should investigate other language components like grammar. For the English language as second language, an improved method should prioritise vocabulary

and should examine other language

components, also like grammar. A more detailed research into the importance of these Dutch and English predictors would give more insight into the mechanism of the contribution and whether different factors could increase or decrease this contribution. The effect of starting age of teaching English vocabulary and the effect of English vocabulary knowledge per age to English spelling skill, for instance, could be important factors to analyse. Besides these language specific skills, also general language skills such as phonological awareness should be included and examined in an improved method, because the current study also showed that these skills are important for second language acquisition.

To summarise, the results of the present study contributed to learning theories stating that a transfer of native knowledge and skills arises

during second language acquisition.

Additionally, second language skills and knowledge were shown to be of high importance besides native language skills. Within both, cognitive measures revealed a small contribution to second language acquisition besides the language measures. General language skills were furthermore shown to be important besides language specific skills. This study therefore highlights the importance of native and second language teaching and the contribution of cognitive measures and other general language skills. In a country like the Netherlands a well-studied method for teaching both languages has to be formalized, to ensure everyone is able to navigate in our increasingly multilingual society.

(18)

Literature

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

Baddeley, A. (1992). Working memory. Science, 255(5044), 556-559.

Van den Bos, K. P., & Lutje Spelberg, H. C. (2007). CB&WL. Continu Benoemen & Woorden Lezen. Amsterdam: Boom

Braams, T., & De Vos, T. (2015). Schoolvaardigheidstoets Spelling. Amsterdam: Boom Test Uitgevers.

Bradford, J. D., & Schwartz, D. L. (1999). Rethinking transfer: A simple proposal with multiple implications. Review of Research in Education, 24, 61–100.

https://doi.org/10.3102/0091732X024001061

Brandenburg, J., Klesczewski, J., Fischbach, A., Schuchardt, K., Büttner, G., & Hasselhorn, M. (2015). Working memory in children with learning disabilities in reading versus spelling: Searching for overlapping and specific cognitive factors. Journal of learning disabilities, 48(6), 622-634. https://doi.org/10.1177/0022219414521665

Central Bureau for Statistics. (2016, October 6). Slight increase children with dyslexia[Children with dyslexia of 7-11 years old to period, gender, age and origin]. Retrieved from

https://www.cbs.nl/nl-nl/nieuws/2016/40/lichte-toename-kinderen-met-dyslexie Crystal, D. (2008). Two thousand million? English today, 24(1), 3-6.

Cummins, J. (1980). The construct of language proficiency in bilingual education. In Alatis, J. E. (ed.), Current issues in bilingual education: Georgetown University Round Table on Languages and Linguistics (GURT) 1980, pp. 81–103. Washington, DC: Georgetown University Press

DePalma, M. J., & Ringer, J. M. (2011). Toward a theory of adaptive transfer: Expanding

disciplinary discussions of “transfer” in second-language writing and composition studies. Journal of Second Language Writing, 20(2), 134-147.

Dunn, L. M., & Dunn, D. M. (2007). Peabody Picture Vocabulary Test, Fourth Edition. Bloomington: Pearson.

Everatt, J., Smythe, I., Adams, E., & Ocampo, D. (2000). Dyslexia screening measures and bilingual- ism. Dyslexia, 6(1), 42–56. DOI:

10.1002/(SICI)1099-0909(200001/03)6:1<42::AID-DYS157>3.0.CO;2-0

Fawcett, A. J., & Nicolson, R. I. (1995). Persistence of phonological awareness deficits in older children with dyslexia. Reading and Writing, 7(4), 361-376. https://doi.org/10.1007/BF01027724 French, L. M., & O’Brien, I. (2008). Phonological Memory and Children’s Second Language Grammar Learning. Applied Psycholinguistics, 29(3),

(19)

Van Gelderen, A., Schoonen, R., Stoel, R. D., De Glopper, K., & Hulstijn, J. (2007). Development of adolescent reading comprehension in language 1 and language 2: A longitudinal analysis of constituent components. Journal of Educational Psychology, 99(3),

477-491. https://doi.org/10.1037/0022-0663.99.3.477

Georgiou, G. K., Torppa, M., Manolitsis, G., Lyytinen, H., & Parrila, R. (2012). Longitudinal

predictors of reading and spelling across languages varying in orthographic consistency. Reading and Writing, 25(2), 321-346. https://doi.org/10.1007/s11145-010-9271-x

de Groot, B. J. A., van den Bos, K. P., & van der Meulen, B. F. (2014). Fonemische Analyse Test Revised.

Harrison, G. L., Goegan, L. D., Jalbert, R., McManus, K., Sinclair, K., & Spurling, J. (2016). Predictors of spelling and writing skills in first-and second-language learners. Reading and Writing, 29(1), 69-89. https://doi.org/10.1007/s11145-015-9580-1

Jongejan, W., Verhoeven, L., & Siegel, L. S. (2007). Predictors of reading and spelling abilities in first-and second-language learners. Journal of educational psychology, 99(4), 835-851.

https://doi.org/10.1037/0022-0663.99.4.835

Keilty, M., & Harrison, G. L. (2015). Linguistic and literacy predictors of early spelling in first and second language learners. Canadian Journal of Applied Linguistics, 18(1), 87-106.

Kufner, H.L. (1972). The grouping and seperation of the Germanic languages. In F. van Coetsem & H.L. Kufner (Eds.), Toward a Grammar of Proto-Germanic languages (p. 72). Tübingen: Max Niemeyer Verlag.

Li, T., McBride-Chang, C., Wong, A., & Shu, H. (2012). Longitudinal predictors of spelling and reading comprehension in Chinese as an L1 and English as an L2 in Hong Kong Chinese

children. Journal of Educational Psychology, 104(2), 286-301. https://doi.org/10.1037/a0026445 McBride-Chang, C. (1995). What is phonological awareness? Journal of Educational

Psychology, 87(2), 179-192. https://doi.org/10.1037/0022-0663.87.2.179

Miller, R. T. (2019). English Orthography and Reading. The TESOL Encyclopedia of English Language Teaching. 1-7. https://doi.org/10.1002/9781118784235.eelt0461

Ministry of Education, Culture and Science. (2019). Einde basisonderwijs 2017-2018. Retrieved from https://www.onderwijsinspectie.nl/documenten/rapporten/2019/11/08/peil.engels-einde-basisonderwijs-2017-2018

Moll, K., Ramus, F., Bartling, J., Bruder, J., Kunze, S., Neuhoff, N., ... & Tóth, D. (2014). Cognitive mechanisms underlying reading and spelling development in five European

orthographies. Learning and Instruction, 29, 65-77. https://doi.org/10.1016/j.learninstruc.2013.09.003

Savage, R., Pillay, V., & Melidona, S. (2008). Rapid serial naming is a unique predictor of spelling in children. Journal of Learning Disabilities, 41(3), 235-250.

(20)

Schlichting, L. (2005). Peabody Picture Vocabulary Test-III-NL. Amsterdam: Harcourt Test Publisher.

Sparks, R. L., & Ganschow, L. (1991). Foreign language learning differences: Affective or native language aptitude differences? The modern language journal, 75(1), 3-16.

https://doi.org/10.1111/j.1540-4781.1991.tb01076.x

Schoonen, R., Hulstijn, J., & Bossers, B. (1998). Metacognitive and language-specific knowledge in native and foreign language reading comprehension: An empirical study among Dutch students in grades 6, 8 and 10. Language learning, 48(1), 71-106. https://doi.org/10.1111/1467-9922.00033 Schoonen, R., Gelderen, A. V., Glopper, K. D., Hulstijn, J., Simis, A., Snellings, P., & Stevenson, M. (2003). First language and second language writing: The role of linguistic knowledge, speed of processing, and metacognitive knowledge. Language learning, 53(1), 165-202.

https://doi.org/10.1111/1467-9922.00213

Sun-Alperin, M. K., & Wang, M. (2011). Cross-language transfer of phonological and orthographic processing skills from Spanish L1 to English L2. Reading and Writing, 24(5), 591-614.

https://doi.org/10.1007/s11145-009-9221-7

Vellutino, F. R. (1979). Dyslexia: Theory and research. Cambridge, MA: MIT Press. https://doi.org/10.1017/S0142716400001752

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.

https://doi.org/10.1037/0033-2909.101.2.192

Wagner, R., Torgesen, J., Rashotte, C., & Pearson, N. A. (1999). CTOPP: Comprehensive Test of Phonological Processing–Second Edition.

Wilkinson, G. S., & Robertson, G. J. (2006). Wide range achievement test. Psychological Assessment Resources.

Wolff, U. (2014). RAN as a predictor of reading skills, and vice versa: Results from a randomised reading intervention. Annals of Dyslexia, 64(2), 151-165. https://doi.org/10.1007/s11881-014-0091-6

Yan, H. (2010). The role of L1 transfer on L2 and pedagogical implications. Canadian Social Science, 6(3), 97-103. http://dx.doi.org/10.3968/j.css.1923669720100603.012

Yeong, S. H., Fletcher, J., & Bayliss, D. M. (2014). Importance of phonological and orthographic skills for English reading and spelling: A comparison of English monolingual and Mandarin-English bilingual children. Journal of Educational Psychology, 106(4), 1107-1121.