Introduction

The assessment of bilinguals is complex due to many diverse factors. Firstly, the definition of bilingualism is not clear and is influenced by multiple factors such as the age of acquisition of the second language (L2) and continued exposure to the first language (L1). Popular definitions of bilingualism conceptualize language knowledge as being a binary category—whether one is classified as having acquired two languages or not (Brutt-Griffler & Varghese, 2004; Valdez & Figueora, 1994). However, bilingualism should be thought of as being on a continuum as this more accurately reflects reality, where one can have varying levels of proficiency in two languages, regardless of how and when they were acquired. For a more detailed discussion of the definitions of bilingualism, please see the following entry: DEFINING BILINGUALISM For the purpose of this paper, bilinguals will be defined as sequential bilinguals who are learning the majority language as L2 at school and in the community. These learners are first and second generation immigrants, and are the most common type of bilingual learners in Canada (Statistics Canada, 2001).

Language and literacy skills are comprised of multiple subskills. This paper will demonstrate that some linguistic skills are highly related across languages, while other skills are less likely to be related. Factors that are highly related across languages could potentially be measured in either L1 or L2. Factors that are not related or are weakly related across languages are good candidates for assessment in both L1 and L2, where adequate measures and personnel are available.

Key Research Questions

1. What are some of the practical barriers to assessing children in two languages?
2. Which skills are related to reading in bilinguals?
3. How do specific reading-related skills transfer across languages?
4. How does the research inform whether skills should be assessed in one or both languages?

Practical reasons for considering L1 and L2 assessment

For bilingual learners many arguments exist about whether the person should be assessed in their first language only, in their second language only, or in both languages. The following section will review evidence to support both sides of the argument with regard to practical reasons for L1 and/or L2 assessment.

Why is L1 testing important? Adequate knowledge of a child’s L1 skills and abilities is advocated as a safeguard against falsely identifying English as a Second Language (ESL) speakers as having learning disabilities when they do not. The majority of research on misidentification of students as having learning disabilities has been conducted in the United States. This research has shown that more than 17 percent of Hispanic students are identified as having learning disabilities in comparison to only 12-13 percent of the total American population (Office of Special Education Programs, 2002). These proportions reflect an overrepresentation of the Hispanic population being identified as learning disabled. It is likely then that in Canada students who are acquiring a second language in the school system (one that is not primarily spoken at home—e.g., English) have a higher risk of being misdiagnosed as having a learning disability due to delays and differences that emerge in cognitive, language and reading assessments. However, the opposite effect may also occur if minority language students with learning disabilities are not assessed for fear of misidentification. For example, in many school boards, bilingual children who are struggling academically are not assessed until they have been in school for five to seven years.

Canadian research on misidentification has examined the accuracy of teacher assessments of L2 learners who were classified as at-risk for reading disability (Limbos & Geva, 2001). The researchers found that low oral language proficiency led teachers to misclassify L2 speakers as learning disabled. As a result, Geva (1999) emphasizes that assessments should be separated at the level of oral and written language. Assessment of word reading skills can be conducted early, when the children have a few years of English experience. However, higher-level oral language skills can sometimes take several years to develop (Cummins, 1991). Geva (1999) argues that inappropriate interpretation of low scores on oral language assessments might result in a student being misidentified as at-risk for deficits in reading as opposed to deficits in oral language (e.g., vocabulary).

Providing additional assistance and high quality, intensive reading instruction to children who show delays in reading acquisition based solely on being ESL speakers can improve achievement (D’Angiulli, Siegel, & Maggi, 2004). However, additional services are often provided only to children who are labeled as having a disability. Falsely labeling a child can result in inappropriate placement in special education classes.

Problems with L1 assessment: It is often difficult to find trained personnel who can provide assessments in the child’s L1 (Westernoff, Nilssen-Lalla, & Bismilla, 2000). In addition, L1 skills do not develop at the same rate in bilinguals as they do in monolinguals due to decreased exposure to the L1 when exposure to the L2 begins to increase (Tabors, Paez, & Lopez, 2003). This slower growth in L1 skills means that the norms on standardized tests are not appropriate for individuals who maintain their L1 in a minority context. These norms are based on skills of monolingual speakers of a given language who are living in their native countries. Testing in either the L1 or L2 will result in lower scores on both languages in comparison to monolingual speakers of each language because often some concepts are only known or acquired in one language. For example, home based vocabulary items, including names of household objects, are more likely to be known in the L1, while academic and curriculum based words are more likely to be known in the L2 (Gottardo Faroga, Mueller, & Irvine, 2006).

In addition to the lack of trained personnel, it is often difficult to find appropriate tests in the student’s first language. It is expensive and time-consuming to create standardized tests. Simply translating English tests into the student’s other language is not appropriate because each language emphasizes different language skills and structures. For example, specific vocabulary items might be very common in one language and obscure in the other language. Additionally, there are differences between languages in how sound units are recognized, for example, the letter-sound relationship for consonants is relatively consistent in English while the letter-sound relationships for vowels are consistent in Spanish.

Theoretical reasons for considering L1 and L2 assessment: Transfer of reading-related skills across languages

The initial step in understanding assessment of reading is to determine the skills related to reading in bilinguals. Researchers also attempt to determine how languages are learned and organized in the brain of bilinguals. Specifically, do language skills in one language assist in the acquisition of similar skills in another language (positive transfer)?

First and foremost, oral language proficiency, that is, the ability to understand and produce oral language, is important to the acquisition of proficiency in a second language (Cummins, 1991; Geva & Yaghoub-Zadeh, 2006). However, bilingual children might begin acquiring L2 literacy and L2 oral language skills at approximately the same time. Once children have proficiency in skills related to word reading, such as phonological awareness (the awareness that words contain smaller parts: syllables, sounds) and naming speed, other oral language skills do not contribute much to differences in word reading skills (e.g., Castles & Coltheart, 2004; Durgunoglu, Nagy & Hancin-Bhatt, 1993; Geva & Siegel, 2000). In addition, understanding print conventions and concepts as well as the ability to decode (word level reading) is important to literacy acquisition in any alphabetic language. Other aspects of language related to reading include metalinguistic awareness, which is typically defined as thinking about and reflecting upon the functions of language (Pratt & Grieve, 1984). In addition to the awareness of sounds in words (mentioned above), awareness of grammatical markers in sentences can aid in word level reading (e.g. knowing that sign and signal are related) and in reading comprehension. And yet, the end goal of reading acquisition is skilled reading comprehension. Reading comprehension requires a wide array of word reading and oral language skills, in addition to experience with text and knowledge of the subject matter. For example, oral language proficiency, specifically, vocabulary development has an impact on reading comprehension skills and is typically delayed in second language learners (August, Carlo, Dressler & Snow, 1999; Proctor, Carlo, August & Snow, 2005).

Research-based arguments that encompass linguistic theory and experimental research suggest that to obtain the most accurate picture of a bilingual person’s ability, both L1 and L2 skills should be assessed. As early as 1978, a sociolinguist, Lavandera, suggested that assessment of bilinguals in just one language would not provide an accurate picture of their linguistic skills. For example, vocabulary items are assumed to be stored as a “unitary whole” and include the concept as well as the known pronunciations in each language (Valdez & Figueora, 1994). The idea is that bilinguals distribute vocabulary items across languages based on the environments in which the items are learned and used most frequently. As mentioned previously, L1 testing will underestimate academic vocabulary while L2 testing will underestimate basic or “home” vocabulary. Research by Oller and Eilers (2002) supported this claim even with middle class Spanish-speakers.

In contrast, other researchers suggest that in some cases it is only necessary to assess skills in one language. Some skills are thought to be universal and be highly related across languages while other skills are thought to be learned separately for each language (Durgunoglu, 2002). Literacy skills can be separated into pre-reading and early reading skills, specifically phonological skills, word level reading which includes how accurately and how quickly words are read, and reading comprehension. Reading comprehension can be further broken down into word reading and listening comprehension. Factors related to listening comprehension include, but are not limited to, knowledge of vocabulary and understanding of grammatical rules and grammatical markers (e.g. plural, past tense). Finally, factors unique to reading comprehension include understanding narrative structure and text genre. Research has shown that for each subskill, L1 and L2 abilities vary in how much they are related to each other. Therefore, research about each subskill will be examined and recommendations will be made separately regarding the necessity and utility of testing these skills in both languages.

Phonological awareness: Research has shown that phonological awareness in a child’s L1 is related to reading an alphabetic language, in this case English, regardless of the nature of the first language. For example, phonological awareness skills in languages with very different scripts (writing systems), such as Spanish and Chinese, are related to English reading (e.g., Durgunoglu, 1998; Lindsey, Manis, & Bailey, 2003; Gottardo, Yan, Siegel, & Wade-Woolley, 2001). Although L1 and L2 phonological awareness are related, for young children who learn to read both languages simultaneously, English phonological awareness is more strongly related to English reading (Gu & Gottardo, 2004). However, for older children who learn to read two languages sequentially and read their first language before learning to read their second language, L1 and L2 phonological awareness are equally related to English, L2, reading. Therefore, for younger bilingual children, assessment of phonological awareness skills should take place in English, the language of reading instruction. For older children, the assessment of phonological awareness can take place in English or the L1. For bilingual children with relatively good L2 proficiency, L2 phonological awareness might be more strongly related to L2 reading. Problems sometimes exist in finding standardized tests of phonological awareness in the L1. In cases where standardized tests of phonological awareness do not exist in the L1, English phonological awareness will give a reasonably accurate picture of phonological awareness skills that would have an impact on learning to read any alphabetic language.

Word reading: Word reading skills are also related across the L1 and L2 for alphabetic languages. However, L1-L2 literacy skills are not related for Chinese-speakers reading Chinese and English (Gottardo, Geva, Faroga, & Ramirez, 2006). The results of this research suggest that if a child is learning to read in two alphabetic languages, even languages with alphabets that are visually very different (English and Hebrew; Hebrew and Russian), assessment of word reading in either language could be used to understand reading potential in the other language. However, if the children are bilingual and biliterate in Chinese and English, assessing reading potential in one language is not likely to inform whether the child will experience difficulties or be successful in the other language. Therefore, word level reading skills are related across languages with the exception of word reading in alphabetic versus nonalphabetic scripts.

Oral language proficiency: When other linguistic skills are compared across languages, a different pattern emerges. Variables related to oral language proficiency in English and in Spanish were examined in detail in a group of Spanish-English bilinguals in first grade (Gottardo, 2002). The results suggested that grammatical knowledge is related but not completely overlapping across languages. Not surprisingly children performed better on L1 syntactic tasks and showed the highest level of performance on early developing L2 structures such as the use of regular plurals (e.g. cat/ cats) (Brown, 1973). Interestingly, only the use of regular plurals in English was related to the use of plurals in Spanish. However, the “rules” for forming these plurals are similar for both languages. This finding suggests that assessment of these skills in only one language will not provide an accurate picture of these skills across languages.

Vocabulary skills also show weak relations across L1 and L2. When vocabulary scores on standardized tests were compared across languages, Spanish and English, the skills were not related (Uccelli & Paez, 2007). Bilingual children showed greater growth in English vocabulary scores than in Spanish vocabulary scores from kindergarten to first grade. In addition, the total number of words and the total number of different words that bilingual children used in oral language story-telling tasks were not related across languages. Young children do not automatically use L1 vocabulary knowledge to deduce the meanings of similar sounding words in their L2 (Gottardo et al, 2006; Gottardo, van Daal, Charbonneau, Meberg & Skaatun, 2007). Therefore, in order to accurately measure vocabulary knowledge, both L1 and L2 skills should be tested.

Knowledge of story structure: In contrast to research on vocabulary and grammatical knowledge, research on story structure has shown that children’s skills in knowing what a “good story” needs are related across languages (Uccelli & Paez, 2007). A “good story” was defined as including the main events of the story in the right order, linking parts of the plot together to give the story flow and providing reasons why events take place. Children’s skills in producing a well-structured story were related across Spanish and English and showed growth in both languages over time (Uccelli & Paez, 2007). The results of research conducted on knowledge of higher-level narrative skills shows that these skills are related across L1 and L2, and can be assessed in only one language.

Conclusion

The above findings show that the assessment of bilinguals is more complex than the assessment of monolingual speakers. All assessments must consider multiple language and literacy related skills. In addition, the assessment of bilinguals must consider factors such as definitions of bilingualism and degree of proficiency in each language. Although existing research on reading acquisition in bilinguals does not provide us definitive answers as to whether students should be assessed in their L1 or L2, some guidelines can be gleaned based on research on relations between L1 and L2 skills. In general, if all conditions are favorable, such as the availability of good L1 and L2 measures and trained assessors fluent in each language, assessment of both L1 and L2 skills is recommended in order to obtain a complete picture of the student’s language and literacy skills. However, these conditions rarely exist. Therefore, guidelines discussed above regarding the relations among L1 and L2 skills can allow assessors to make the best decisions. In particular, some lower level skills such as phonological awareness and word reading are related across alphabetic languages. Many oral language skills are not strongly related across L1 and L2. Higher-level skills that tap text and story structure knowledge transfer across languages. Research also suggests that for skills that are highly related across L1 and L2, assessments can be conducted in the L2 when the student has a couple of years experience with the language, while other skills might require several years of experience with the L2 to develop. Therefore, practitioners assessing bilingual students should be aware of, and consider the many variables that can influence their choice of assessment measures and strategies.