Home Articles Reader Opinion Editorial Book Reviews Discussion Writers Guide About TCRecord
transparent 13
Topics
Discussion
Announcements
 

The Long-Term English Language and Literacy Outcomes of First-Generation Former Child Immigrants in the United States


by Becky H. Huang & Alison L. Bailey — 2016

Background/Context: Children from Asian ethnic backgrounds currently constitute the second largest group of child immigrants in the United States. Although stereotyped as model minority students due to their academic and economic success, studies have revealed that many Asian immigrant students struggle in school. Research has also shown that, compared to child immigrants from an Indo-European language background such as Spanish and French, Asian child immigrants experience more challenges in learning English as a second language (L2) due to greater cross-linguistic differences. However, little is known about the long-term English language outcomes of first-generation Asian child immigrants.

Purpose/Objective/Research Question/Focus of Study: The present study examines the effects of learner-level and input-level factors on first-generation Asian child immigrants’ long-term English outcomes.

Research Design: Data for the current study are selected from a larger correlational and cross-sectional study that examined the effect of the age of arrival variable on Chinese immigrants’ English L2 outcomes. We used two criteria to select participants from the larger study: (1) those who had arrived in the United States between the ages of 5 to 18 (to qualify as a “child immigrant”), and (2) those who had resided in the United States for at least 10 years (to examine long-term outcomes). These criteria resulted in the current sample of 69 participants. The English language proficiency data include participants’ phonological production ratings, performances on a grammaticality judgment task, and their self-ratings of English proficiency in listening, speaking, reading, and writing.

Conclusions/Recommendations: The current study showed a complex interplay of factors affecting former child immigrants’ English L2 acquisition. Although age of arrival played a critical role in their L2 learning outcomes, it exerted varying degrees of influence by L2 domain. Age of arrival was a strong predictor of L2 phonological production, grammar knowledge, and oral language proficiency, but not literacy skills. L2 input, language learning aptitude, and child immigrants’ affective status also contributed to their L2 outcomes, and carried more weight than age of arrival. We interpreted the results to be in line with the multiple sensitive period hypothesis in developmental psycholinguistics research. The results also suggested that literacy is not susceptible to age-related effects in the same way in which oral language and more specifically the phonological and syntactic systems are. Literacy as a cultural construct rather than a biologically unique human system is intensively taught throughout the school years and curriculum. Malleable factors, such as instruction and reading strategies, are thus perhaps more important in determining child immigrants’ long-term literacy outcomes.



The current study focuses on the long-term English language outcomes of a sample of first-generation child immigrants from Asian, specifically Chinese, ethnic backgrounds. Children from Asian ethnic backgrounds currently constitute the second largest group of child immigrant/English Language Learner (ELL) students in the United States (U.S. Census Bureau, 2013), and have consistently done so since the 1970s, covering the period when the former child immigrant participants in this study were growing up in the United States (National Center for Education Statistics, 1992).


First-generation, foreign-born children arrive in the United States prior to college at varying ages (birth to 18 years)1  and have varying length of residence in the United States. There is also great diversity in these children’s first language (L1) proficiency, English language proficiency, socioeconomic status, and prior schooling experiences. In contrast with their second generation U.S.-born peers, foreign-born child immigrants struggle more to acquire the English language and adjust to the new culture (Cortes, 2006; Kim & Suarez-Orozco, 2014; Suárez-Orozco, Suárez-Orozco, & Todorova, 2008). They also tend to underperform their native-English speaker (NES) or second-generation U.S.-born peers (Conger, 2009; Kieffer, 2008; Thomas & Collier, 2002).


Although stereotyped as model minority students due to their academic and economic success (Cheryan & Bodenhausen, 2000; Lee, 1998), studies have revealed that many Asian immigrant students struggle in school (Hammer, Jia, & Uchikoshi, 2011). Research has also shown that, compared to child immigrants from an Indo-European language background such as Spanish and French, Asian child immigrants experience more challenges in learning English as a second language due to greater cross-linguistic differences (Bialystok, McBride-Chang, & Luk, 2005; Bialystok, Luk, & Kwan, 2005). However, little is known about the long-term English language outcomes of first-generation Asian child immigrants.


The present study examines the effects of learner-level (e.g., gender, socioeconomic status, age of arrival) and input-level factors (e.g., exposure to English, length of ESL instruction) on first-generation Asian child immigrants’ long-term English outcomes. Given the limited research on the long-term English language outcomes of first-generation Asian child immigrants in particular and of first-generation child immigrants in general, the current study fills an important gap in the literature. Although the results may not generalize to all first-generation child immigrant populations in the U.S., they help shed light on the processes and outcomes of English language learning in a second-language (L2) speaking context. The results inform both L2 learning theories and educational practices for child immigrants. In the following sections, we first review the existing literature on former child immigrants’ long-term English language and literacy outcomes. We also describe our data source, the descriptive and bivariate correlation results as well as regression models. We then discuss the results and conclude with implications for educational practices and future research.


LITERATURE REVIEW


SECOND LANGUAGE RESEARCH ON CHILD IMMIGRANTS’ LONG-TERM ENGLISH LANGUAGE OUTCOMES


Research on child immigrants’ long-term L2 outcomes are generally drawn from two fields: L2 acquisition research and educational linguistics research. The former has produced abundant research that examine the predictors of immigrants’ long-term L2 outcomes, yet the majority focuses on addressing the critical period hypothesis debate via testing the effect of the age of arrival (AoA) variable (e.g., Birdsong & Molis, 2001; DeKeyser, 2000; Flege, Yeni-Komshian, & Liu, 1999; Hakuta, Bialystok, & Wiley, 2003; Johnson & Newport, 1989). The critical period hypothesis originated from biological research and suggests biological, maturational constraints on the learning of an L2. This body of research is typically grounded in the structural linguistics perspective that views language as an innate symbolic system consisting of discrete and hierarchical domains such as phonetics, phonology, semantics, and grammar (Chomsky, 1986; Pinker, 1994). The most commonly studied domains are phonology (Asher & Garcia, 1969; Flege, Munro, & Mackay, 1995; Flege, Yeni-Komshian, & Liu, 1999; Huang & Jun, 2011) and grammar (Bialystok & Miller, 1999; Birdsong & Molis, 2001; DeKeyser, 2000; Johnson & Newport, 1989). These two domains have been hypothesized to be subject to effects of brain maturation and loss of plasticity whereas the learning of vocabulary utilizes higher order association mechanism that is not susceptible to an aging effect (Pulvermüller, & Schumann, 1994). Furthermore, this research also precludes literacy, which is considered a cultural construct rather than a biologically endowed system unique to human beings (Chomsky, 1986; MacSwan & Pray, 2005; Pinker, 1994).


In general, this body of research shows that AoA exerts a robust effect on immigrants’ long-term L2 outcomes. Those with younger AoAs achieve higher L2 proficiency than those with older AoAs. The negative correlation between AoA and L2 outcomes is globally linear among child immigrants (DeKeyser, 2000; Johnson & Newport, 1989; Flege et al., 1999; Hakuta et al., 2003; Huang, 2013). Although the negative relationship between AoA and L2 outcomes is consistent across studies, L2 acquisition researchers debate the cause of age-related effects. Proponents of the critical period hypothesis attribute the AoA effect to the loss of brain plasticity due to the closing of the critical window of opportunity (Abrahamsson & Hyltenstam, 2009; DeKeyser, 2000; Johnson & Newport, 1989; Patkowski, 1990), while opponents of the hypothesis argue that L2 input or socio-cultural variables, not AoA per se, are responsible for the negative correlations between AoA and long-term L2 outcomes (Bialystok & Hakuta, 1994; Jia, Aaronson, & Wu, 2002). L2 acquisition researchers have also found that child immigrants do not necessarily catch up with their NES peers even after prolonged immersive exposure to the L2 in an L2-speaking context (Flege et al., 1999; Granena & Long, 2013; Huang, 2013), and that the AoA variable affects speech production outcomes more strongly than grammar outcomes (Flege et al., 1999; Huang, 2013).


In addition to the AoA variable, some studies have examined the effects of other predictors such as length of residence in the L2-speaking country, self-estimated L2 input, L1 proficiency, self-reported motivation, and language learning aptitude, and showed that these variables correlated with long-term L2 outcomes to varying degrees (Birdsong & Molis, 2001; Flege et al., 1999; Huang, 2013; Jia et al., 2002; Johnson & Newport, 1989; Moyer, 2004, 2007). For example, Flege et al. (1999) assessed the speech production and grammar outcomes of 240 Korean immigrants in the United States whose ages of arrivals (AoAs) ranged from 1 to 23. All immigrants had lived in the United States for at least 8 years. Grammar knowledge was measured with a grammaticality judgment task, and phonological production was determined by a panel of NES raters’ perceptions of the global foreign accents in participants’ speech recording. Various information such as participant’s length of residence, L2 input, and language learning aptitude, was also collected via a questionnaire. Results revealed a robust effect of AoA for both grammar and speech production outcomes. However, the effect of AoA on grammar outcome disappeared after controlling for participants’ years of education in the U.S., whereas AoA’s effect on speech production remained significant, suggesting potential maturational constraints on L2 speech production but not on grammar learning. Additionally, participants’ self-estimated language learning aptitude, English media input, and motivation to learn English were also found to contribute to their English grammar and speech production outcomes.


In a study that focused on predictors of long-term English grammar outcomes, Jia and colleagues tested 112 immigrants in the United States with a wide range of AoA (1–38) and from a variety of native language backgrounds. Grammar knowledge was assessed via a reading and a listening grammaticality judgment task, and predictors were collected via a language background questionnaire. The results revealed AoA to be a significant predictor for immigrants’ grammar outcomes. However, mother’s English proficiency and number of English speakers at home also significantly predicted English grammar outcomes, suggesting the importance of L2 input.  


To summarize, this body of research was grounded in a structural linguistics theoretical framework and focuses on L2 phonological and grammar outcomes. The results revealed a robust effect of AoA on child immigrants’ long-term L2 phonological outcomes, but the effect on L2 grammar outcomes was less clear. Language learning aptitude, L2 input, L1 proficiency, and affective variables were also found to correlate with long-term L2 outcomes in some studies.


EDUCATIONAL LINGUISTICS RESEARCH ON CHILD IMMIGRANTS’ LONG-TERM ENGLISH LANGUAGE AND LITERACY OUTCOMES


Educational linguistics and policy researchers have contributed to a second body of research examining child immigrants’ long-term L2 outcomes (Conger, 2009; Collier, 1987; Cummins, 1981; Hakuta, Butler, & Witt, 2000; Slama, 2011). In contrast to the dominance of structural linguistics perspective in L2 acquisition research, educational linguistics research conceptualizes language from several different viewpoints, such as the systemic functional linguistics perspective that considers language as meaning- and context-based rather than structure-based, and users’ choices of specific lexical and grammatical registers are driven by the functional purposes of language tasks (Halliday, 1994; Schleppegrell, 2002). Another widely adopted view of language is the distinction between social and academic language (Bailey, 2007; Cummins, 1980, 2000; Scarcella, 2003; Snow, 2010). Cummins (1980, 2000) argued that the day-to-day languages that learners use to socially interact with others, i.e., Basic Interpersonal Communication Skills (BICS), is typically context embedded and cognitively undemanding. In contrast, Cognitive Academic Language Proficiency (CALP) refers to formal, academic language uses that are both context-reduced and cognitively demanding.


In contrast to the wide uses of researcher-developed grammaticality judgment tasks and phonetic/phonological analyses in L2 acquisition research, educational linguistics researchers generally use results from state-mandated standardized language tests as measures of L2 proficiency (for exceptions see MacSwan & Pray, 2005). Additionally, L2 acquisition researchers select immigrant participants based on their AoA and LoR. However, educational linguistics researchers are mainly concerned with improving child immigrants’ academic achievements, and thus tends to focus on a subset of struggling child immigrants, known as or categorized as English language learners (ELL) or Limited English Proficiency (LEP) students depending on the designations of their school districts, who do not have grade-level appropriate English proficiency in oral language and literacy, and/or are at risk for academic failure (Bailey & Carroll, 2015).


To inform education policy, linguists in educational settings have gone beyond determining if child immigrants’ English language proficiency is comparable to the NES norms to further investigate how long it takes for children to catch up with their NES peers. As reviewed in the previous section, results from the L2 acquisition research suggest discrepancies across L2 domains; child immigrants are observed to perform comparably with NES peers in grammar (Birdsong & Molis, 2001; Flege et al., 1999; Huang, 2013, but see Granena & Long, 2013) but not in phonological production (Flege et al., 1999; Granena & Long, 2013; Huang, 2013). However, results from the educational linguistics research are more mixed. While some showed that child immigrants/ELL students caught up with NESs by 7 years of immersive exposure in an L2-speaking context (Conger, 2009; Cortes, 2006; Hakuta et al., 2000; MacSwan & Pray, 2005), others have found that they still fell behind their NES peers (Kieffer, 2008; Klesmer, 1993; Slama, 2011; Wright & Ramsey, 1970). These discrepancies may be partly attributed to the differences in the language measures used. Studies that found comparable performances between child immigrants/ELL students and NESs tend to assess grammatical knowledge, such as Bilingual Syntax Measure (BSM) used in MacSwan and Pray (2005), or social language proficiency, such as Language Assessment Battery (LAB) and Idea Proficiency Test (IPT) (Butler & Stevens, 2001) used in Conger (2009) and Hakuta et al. (2000), respectively. In contrast, those that observed child immigrants’/ELL students’ persistent, lower performances generally compared their academic language proficiency (Slama, 2011) or reading achievement (Kieffer, 2008) with NESs. The mixed findings could also result from differences in the sample characteristics (e.g., first-generation foreign-born child immigrants vs. second-generation U.S.-born ELL students) or in the input/instruction child immigrants received (e.g., sheltered English immersion, bilingual education, or two-way immersion programs). Unfortunately, specific information about child immigrants’ generation status and instruction/program type was missing in many of the existing educational linguistics studies.


To illustrate, one of the most relevant and widely cited studies was conducted by Hakuta, et al. (2000), who analyzed test data from two school districts (n =1800) to examine the amount of time required for ELL students to attain proficiency in oral (social) and academic language. The ELL students in the study were in grades 1 to 6 at the time of testing and were classified as ELL since they were in kindergarten. They were from Spanish and Vietnamese backgrounds. Hakuta and colleagues adopted the BICS/CALP distinction while acknowledging that the distinction may be oversimplifying the construct of language proficiency. The language proficiency measures used in District A were the Idea Proficiency Test (IPT) as an oral (social) language test, and the MacMillan Informal Reading Inventory and a district-developed writing test as measures of academic English proficiency. District B used the Woodcock-Muñoz Language Survey-Revised (WMLS-R), a standardized, norm-referenced test that evaluated both oral (social) and academic language. Hakuta et al. found that it took ELL students an average of 3 to 5 years to develop oral proficiency and 4 to 7 years to develop academic English proficiency. However, it was unclear whether the ELL students in the study were born in the United States or were foreign-born child immigrants.  


Although the two bodies of literature differ in their findings about whether immigrant participants perform comparably to their NES peers in English language proficiency tasks, the two fields converge to show an older learner advantage (older biological age) in the initial rate of acquisition (MacSwan & Pray, 2005) and a younger learner advantage (younger ages of initial L2 exposure in an L2-speaking context) in long-term L2 outcomes in phonology (Flege et al., 1999; Huang, 2013), grammar (DeKeyser, 2000; Johnson & Newport, 1989), and basic oral language proficiency (Conger, 2009).


In sum, L2 acquisition research and educational linguistics research converge to reveal an AoA effect but diverge in whether or not immigrants catch up to NES norms in various L2 domains. In addition to AoA, a variety of other learner-level variables, such as their socioeconomic status, motivation, and language learning aptitude, were also found to play a role in immigrants’ L2 outcomes. However, although first-generation child immigrants were shown to be at higher risks for low English proficiency and academic failure (Cortes, 2006; Kieffer, 2008; Slama, 2009), none of these empirical studies focused on first-generation child immigrants’ long-term L2 linguistic outcomes after they have lived in the L2-speaking country for over a decade. L2 acquisition research generally included adult immigrants (immigrants who arrived in the L2-speaking country as adults) in the sample or even focused solely on adult immigrants (Bongaerts, 1999; Lardiere, 2007) rather than on child immigrants. These studies also aimed at addressing the theoretical debate on the critical period hypothesis rather than understanding learner-level predictors of L2 outcomes. On the other hand, educational linguistics research that examined immigrants’ long-term L2 outcomes are relatively limited, and none of the existing studies tracked L2 outcomes beyond 7 years of exposure in an L2-speaking context. Several of the existing studies were conducted in the 1970s and 1980s (Collier, 1987; Ramsey & Wright, 1974; Wright & Ramsey, 1970), and the results need to be updated. Furthermore, virtually none of these educational linguistics studies included an NES comparison group, but inferred an ELL-NES comparison based on child immigrants/ELL students’ performances on standardized L2 proficiency tests with NES norming samples.


THE PRESENT STUDY


As discussed earlier, the number of language minority students in the United States has grown exponentially in the past few decades, yet relatively little is known about their long-term English L2 outcomes. In particular, Asian child immigrants constitute the second largest group of child immigrant/ELL students in the U.S., but research on L2 outcomes of this subgroup is very limited (Hammer et al., 2011). The present study thus aims to fill this gap in the literature and provide a descriptive account of the predictors of long-term English language and literacy outcomes for first-generation former child immigrants from Asian backgrounds. The study also addresses an important education question that is still under debate in the literature: whether or not first-generation child immigrants catch up with their monolingual NES peers after decades of potential exposure to English. Although the current results may not generalize to other subgroups of child immigrants in the United States from different socioeconomic or native language backgrounds, they nonetheless provide a much-needed exploratory understanding of a topic that is critical for both L2 acquisition and educational linguistics researchers.


In addition to the main contributions described above, the study also presented three improvements from previous studies. First, while existing studies utilized either researcher-developed psycholinguistic instruments or self-report census data (Hakuta et al., 2003) or secondary data obtained from school districts or states (e.g., Hakuta et al., 2000; Selma, 2011), the current study included both subjective (self-report) and objective (researcher-developed) L2 measures. As other researchers have suggested (Luk & Bialystok, 2013), combining both objective and subjective measures yields more informative results than using either measure alone. Also, very few studies have compared child immigrants’ L2 proficiency concurrently across multiple domains and skills or used multiple language measures (for exceptions see Flege et al., 1999; Hakuta et al., 2000; Huang, 2013). In addition to examining two discrete L2 domains (i.e., phonological production and grammar knowledge), the present study added self-ratings of L2 skills in reading, writing, speaking, and listening. Finally, many of these empirical studies only included a narrow range of learner-level variables rather than surveying a comprehensive list of predictors that may contribute to L2 outcomes (Hakuta et al., 2000; Slama, 2011). Important information such as child immigrants’/ELL students’ L1 proficiency, L2 exposure and input, and program of language instruction was missing in these educational linguistics research studies. The current study addressed this issue and surveyed a wide range of learner variables, including cognitive, social, and affective variables, as predictors of L2 outcomes. These improvements allowed for a more comprehensive and in-depth empirical investigation of the language proficiency outcomes of first-generation child immigrants from Asian backgrounds, as well as the predictors of their language outcomes.


The current study is guided by two research questions: (1) What are the predictors of first-generation former child immigrants’ (age of arrival: 5 years to 18 years) long-term English language proficiency outcomes in phonological production, grammatical knowledge (grammar hereafter), oral language proficiency, and literacy? (2) Do first-generation former child immigrants catch up with their monolingual English-speaking peers after more than two decades of prolonged exposure to English in the United States?


Based on the literature review, we expected to see a robust effect of Age of Arrival (AoA) on phonological production, grammatical knowledge, and oral language proficiency. In contrast, literacy domain will not be susceptible to the AoA effect because it is a cultural construct. We also predicted that AoA would have a stronger effect on L2 phonological production than on grammar outcomes and that child immigrants would perform comparably to NESs in grammar, oral language proficiency, and literacy, yet maintain a nonnative accent in phonological production even after prolonged residence in the United States.


In terms of the other predictors of the four outcomes, we expected to see L2 input having a strong influence on all four domains, and that language learning aptitudes would predict all but literacy outcomes because empirical evidence linking language learning aptitude and literacy is absent in the literature. We also anticipate an inverse relationship between Mandarin language proficiency and L2 outcomes. Finally, we expect to observe significant effects of psychological factors, such as positive/negative attitudes toward L2 learning, on child immigrants’ long-term L2 outcomes.


METHOD


PARTICIPANTS


Data for the current study are selected from a larger correlational and cross-sectional study that examined the effect of the Age of Arrival (AoA) variable on Chinese immigrants’ English L2 outcomes. To ensure that participants had an opportunity to reach “ultimate attainment” based on the predictions of the critical period hypothesis for L2 acquisition, all participants had resided in the United States for at least 5 years. Because the larger study aimed to examine successive L2 acquisition rather than the simultaneous acquisition of two first languages (i.e., bilingual first language acquisition, Genesee & Nicoladis, 2006), only participants who spoke their L1 (i.e., Mandarin) before age five and had no prior immersion exposure to English before they arrived in the United States were included. To compare the results with previous studies such as Johnson and Newport (1989) and Birdsong and Molis (2006), the sample in the larger study was also selected purposively based on participants’ education level. All participants were at least college-educated or were current college students. Participants were recruited from major coastal cities in the United States through flyers posted on university campuses, Mandarin heritage language classes, personal contacts, and referrals from participants, and were screened via email or over the phone to ensure that they met the selection criteria before they were invited to participate. The original sample included 118 participants who arrived in the United States between age 5 to 27.


The shared goal of examining Asian immigrants’ English L2 outcomes between the larger and the current study justified using a subsample from the larger study for the current investigation. However, in contrast to the larger study that covered a wide range of ages of arrival from early childhood to adulthood (i.e., 5–27) to specifically test age of arrival (AoA) effects on such factors as English phonology, the current study focused on understanding the predictors of long-term English L2 outcomes of child immigrants’ (i.e., only those participants who arrived in the U.S. prior to age 18).  Based on the specific purpose of the current study to document the long-term outcomes of child immigrants from Asia, we used two criteria to select participants from the larger study: (1) those who had arrived in the United States between the ages of 5 to 18 (to qualify as a “child immigrant”), and (2) those who had resided in the United States for at least 10 years (to examine long-term outcomes). These criteria resulted in the current sample of 69 participants. All participants were first-generation, foreign-born former child immigrants, currently aged between 20–45 years, who were originally from China or Taiwan, and reported Mandarin Chinese as their L1. To answer our research question about whether child immigrants catch up with their NES peers after prolonged exposure to English, we also used the language data of 20 NESs from the larger study for comparison. The NESs were recruited from the same universities and regions, and were comparable to the child immigrants on gender, X2 (1) = .267, p = .606, current age, t(87) = -1.236, p = .220, and educational level, i.e., total years of education received in the United States (t[87] = -.292, p = .771)2. We also examined the ranges and distributions of the continuous variables (current age and educational level) for the two groups to ensure comparability.3


Background Survey Data (Learner-Level Predictors)


The survey data include participants’ responses to questions in nine main constructs: Demographic, L2 Input, English Language Proficiency, Mandarin Language (L1) Proficiency, Language Aptitude, Motivation, Use of Language Learning Strategy, Cultural Affiliation, and Psychological-Affective Attitude. For questions in the Demographic construct, participants self-reported their demographic information, such as their age of arrival, current age, length of residence in the United States, years of education, etc.4 For questions in the L2 Input construct, participants rated their parents’ English proficiency on a scale of 1 (No proficiency) to 9 (Native-like), and estimated their own English language input in different modalities (literacy, oral language, and media) in percentages. For the English Language and Mandarin Language Proficiency constructs, participants self-rated, on a scale of 1 (No proficiency) to 9 (Native-like), their English and Mandarin language proficiency in reading, writing, speaking, and listening when they first arrived in the United States and at the time of testing. Their self-ratings of the four skills were averaged to create a composite score. For the Language Aptitude construct, participants evaluated their own sound processing ability, musical ability, and language learning ability on a scale of 1 (Poor) to 9 (Excellent).


Data for the remaining four constructs (Motivation, Use of Language Learning Strategy, Cultural Affiliation, and Psychological-Affective Attitude) involved participants’ responses to each question on a scale of 1 (Strongly Disagree) to 9 (Strongly Agree), for both the initial few years and the most recent few years of their residence in the United States. To illustrate, one of the questions in the Cultural Affiliation construct was “I have a sense of belonging to American culture.” Participants answered the question on a scale of 1 (Strongly Disagree) to 9 (Strongly Agree) for their initial time period and for their most recent time period in the United States. The distinction between initial and recent time periods was motivated by empirical findings that suggested discrepant patterns between predictors in the two time periods and language outcomes (see Birdsong & Molis, 2001). See Appendix A for additional sample questions from the questionnaire.


Due to the constraints of the sample size, we tried to reduce the large number of variables derived from the survey for correlational analysis. See Table 1 for the complete list of survey variables after data reduction.


ENGLISH LANGUAGE AND LITERACY DATA (OUTCOME VARIABLES)


The English language proficiency data include participants’ phonological production ratings, performances on a grammaticality judgment task, and their self-ratings of current English proficiency in listening, speaking, reading, and writing on a 1–9 scale (1= poor, 9 = native-like). Phonological production ratings were derived from NES raters’ ratings of each participant’s read-aloud speech. All participants were recorded reading a paragraph from the Speech Archive website (Weinberger, 2013) (See Appendix B), and their recordings were then all evaluated by the same five NES raters on a 1–9 scale (1 = strong foreign accents, 9 = like a native English speaker). All NES raters were also recruited from coastal cities in the United States and the average age was 25. They were all born and raised in the United States. The five raters’ ratings were averaged as the outcome variable.


The grammaticality judgment data include participants’ percentage correct scores on a grammaticality judgment task. All participants judged the grammaticality (Correct vs. Incorrect) of 112 English sentences on a computer, and their percent correct scores on the task were used as the outcome variable. The structures evaluated include standard English structures such as determiners, past tense, plurals, particle movement, and pronominalization. See Appendix C for a list of sample sentences.  


Participants’ self-ratings of listening and speaking proficiency were averaged to create an oral language proficiency composite, and so were their ratings of reading and writing as a literacy composite (both ranges = 0–9). The two composite scores are highly correlated with the original ratings (r = .91–.94).


RESULTS


SAMPLE CHARACTERISTICS


Table 1 presents the descriptive information for the learner-level variables. The current sample of former child immigrants arrived in the United States at the mean age of 10 and have lived here for an average of 20 years. Their English proficiency was fairly low when they first arrived, and their initial native language proficiency was high, but it had declined over the years. Based on self-evaluations, they had above-average language learning aptitude. They were also highly motivated English learners and frequently used language learning strategies. They identified with the American culture more strongly now than they first arrived, and the strength of identification with their own cultural heritage had also increased since they immigrated to the United States. Compared to their first post-immigration years, they were less likely to feel self-conscious about their English language proficiency or avoid using English.


Overall, they reported an English-dominant environment with at least 70% of English exposure in daily life in multiple domains (oral language, literacy and media). Their parents’ English proficiency was rated as slightly below average. However, compared to the language environment when they first arrived, their exposure to oral language in the L2 increased from 50% to 70%, possibly because of the increase in their own L2 proficiency. Their exposure to literacy and media in the L2 were similarly high both in the initial years upon arrival and in their current daily life.


Table 1. Means, Standard Deviations and Ranges of the Background Variables/Predictors from Survey


 

Child Immigrants (n = 69)

Demographic

Age of Arrival

10.12 (3.39) [5-18]

Age

30.15 (5.30) [20-44]

Length of Residence

19.56 (5.21) [9.3 - 30]

Gender

30 female, 39 male

Years of U.S. Education

13.67 (3.64) [5-23]

Years of ESL in the U.S.

2.02 (1.44) [0-6]

Total Years of Education

18.43 (2.45) [13-24.5]

English Language Proficiency

 

Initial English Proficiency Composite

1.61 (1.01) [1-5]

Mandarin Language Proficiency

 

Initial Mandarin Proficiency Composite

8.32 (0.93) [5-9]

Current Mandarin Proficiency Composite

6.01 (1.83) [2-9]

Language Aptitude

 

Language Aptitude Composite

6.40 (1.44) [2.20-9.00]

Motivation

 

Initial Motivation Composite

6.81 (2.45) [1-9]

Current Motivation Composite

6.66 (2.16) [1-9]

Use of Language Learning Strategy

 

Initial Language Learning Strategy Composite

5.76 (2.49) [1-9]

Current Language Learning Strategy Composite

7.14 (1.73) [2-9]

Cultural Affiliation

 

Initial American Culture Identification Variable

3.20 (2.28) [1-9]

Current American Culture Identification Variable

6.87 (2.03) [1-9]

Initial Heritage Culture Appreciation Variable

7.03 (2.18) [1-9]

Current Heritage Culture Appreciation Variable

7.99 (1.31) [4-9]

Psychological-Affective

 

Initial Embarrassment Variable

6.55 (2.70) [1-9]

Current Embarrassment Variable

4.25 (3.20) [1-9]

Initial Avoidance Variable

6.03 (2.83) [1-9]

Current Avoidance Variable

2.09 (2.08) [1-9]

L2 Input

 

Parents’ English Proficiency Composite

4.04 (2.03) [1-9]

Initial Oral English Input (Avg. %)

51.94 (16.76) [23.90 - 90.30]

Current Oral English Input (Avg. %)

72.28 (20.79) [4-100]

Initial English Literacy Input (Avg. %)

87.07 (15.39) [36.67-100]

Current English Literacy Input (Avg. %)

89.39 (14.14) [20-100]

Initial English Media Input (Avg. %)

78.14 (27.60) [8.33-100]

Current English Media Input (Avg. %)

79.26 (20.42) [0 -100]


LONG-TERM ENGLISH LANGUAGE AND LITERACY OUTCOMES

 

As shown in Table 2, after over a decade of residence in the United States, child immigrants in the current sample were still perceived as speaking with a mild nonnative accent (M = 6.40 out of 9) and received significantly lower foreign accent ratings than the NES controls, t(87) = 7.490; p < .001. They scored high on the grammaticality judgment task, although still lower than NES controls, t(87) = 6.867; p < .001. Child immigrants’ self-ratings of oral language and literacy skills were at ceiling, though still fell short of native-like (i.e., a rating of 9)5. We also examined the proportions of participants who rated themselves as native-like by L2 domain; approximately 40% of the participants reported native-like proficiency in receptive skills (Listening and Reading), and 30% reported native-like proficiency in productive skills (Speaking and Writing). In other words, only about one third of the child immigrants in the study believed they had attained native-like L2 proficiency.


The four language outcomes were all significantly correlated with each other, though the strength of association varied among the language dimensions (r = .268–.695, See Table 3). The positive correlations between the objective measures (foreign accents and grammar) and subjective measures (self-ratings of oral language and literacy) corroborate previous findings by Jia et al. (2002) and Hakuta and D’Andrea (1992). However, the current patterns appeared to suggest a distinction between objective and subjective measures: The association between the two objective measures was higher than their respective associations with subjective measures, and the same patterns held for objective measures. Comparing across the four language outcomes descriptively, former child immigrants reached slightly higher level of oral language than literacy proficiency. Their phonological production outcomes, as indexed by the foreign accent ratings, were also worse than both oral language and literacy outcomes. There was also greater variation in the foreign accent ratings than in their self-ratings of oral language and literacy proficiency. Although the foreign accents, oral language and literacy outcomes were all measured on a 1–9 Likert scale, they assessed different theoretical constructs, so we did not conduct any inferential statistics for comparison between them.


Table 2. Descriptive Results (Means and Standard Deviations) of the Language Outcomes

 

 

Child Immigrants

(n = 69)

Native Speakers

(n = 20)

t test

Foreign Accent Ratings

6.40 (1.46)

8.86 (0.20)

t(87) = 7.490;

p <.001

Grammar Scores

(% Correct)

85.44 (6.17)

95 (3)

t(87) = 6.867;

p <.001

Oral Language (Listening/Speaking)

     7.93 (.94)

    NA

     NA

Literacy (Reading/Writing)

     7.75 (1.10)

    NA

     NA


Table 3. Correlation Matrix for Language Outcomes


 

1

2

3

4

1. Foreign Accent Ratings

--

   

2. Grammar

.692**

--

  

3. Oral Language (L&S)

.434**

.460**

--

 

4. Literacy (R&W)

    .268*

.332**

.695**

--


PREDICTING LONG-TERM ENGLISH LANGUAGE AND LITERACY OUTCOMES


We first conducted bivariate correlation analyses to examine the relationships between the learner-level characteristics and their English language and literacy outcomes. As shown in Table 4, there were both similarities and differences in the correlation patterns across domains. In general, Age of Arrival (AoA), years of education in the United States, input-related variables (English media input and oral language input, etc.), and tendency to avoid using English were significantly correlated with all L2 outcomes.


The cross-domain differences included: (1) Language aptitude is positively correlated with grammar and oral language proficiency, but not with foreign accent ratings or literacy outcomes, (2) current Mandarin language proficiency is negatively correlated with foreign accent ratings and grammar scores, but not with oral language proficiency and literacy ratings, (3) degrees of identification with American culture is positively correlated with only oral language self-ratings and not with other L2 domains, (4) initial heritage culture appreciation is negatively correlated grammar, but not with foreign accent ratings or self-ratings of oral language proficiency and literacy, (5) current heritage culture appreciation is positively related to foreign accent ratings and oral language proficiency but not with grammar and literacy, (6) motivation is negatively correlated with all domain but self-ratings of oral language proficiency, (7) feelings of embarrassment is negatively correlated with all domain but foreign accent ratings.


Table 4. Correlations between the Language/Literacy Outcomes and Background Variables


 

Foreign  accent ratings

Grammar

Oral Language

Proficiency

Literacy

Demographic

    

Age of Arrival (AoA)

-.560**

-.489**

  -.355**

-.301*

Age

-.339**

-.365**

-.231

   -.080

Length of Residence

   -.009

   -.022

.067

.165

Years of U.S. Education

.441**

.404**

 .251*

.278*

Years of ESL in the U.S.

   -.096

    .002

.039

   -.063

Total Years of Education

   -.092

   -.028

-.138

   -.031

Initial English Proficiency

    

Initial English Proficiency Composite (oral)

   -.039

   -.092

-.007

   -.154

Mandarin Language Proficiency

    

Initial Mandarin Proficiency Composite

.031

.002

.117

.206

Current Mandarin Proficiency Composite

 -.439**

 -.499**

-.165

-.127

Language Aptitude

    

Language Aptitude Composite

.206

 .308*

 .475**

.213

Motivation

    

Initial Motivation Composite

-.132

    -.088

-.023

    -.062

Current Motivation Composite

 -.360**

-.423**

-.217

-.299*

Use of Language Learning Strategy

    

Initial Language Learning Strategy Composite

-.141

    -.057

-.013

-.041

Current Language Learning Strategy Composite

-.087

    .017

.123

.037

Cultural Affiliation

    

Initial American Culture Identification Variable

-.011

    .010

.000

-.039

Current American Culture Identification Variable

.102

    .186

  .253*

.048

Initial Heritage Culture Appreciation Variable

-.205

-.354**

.012

-.006

Current Heritage Culture Appreciation Variable

   .250*

     .214

   .327**

.141

Psychological-Affective

    

Initial Embarrassment Variable

     -.139

     -.105

 -.080

 -.117

Current Embarrassment Variable

-.134

-.289*

 -.305*

 -.297*

Initial Avoidance Variable

-.127

-.091

-.057

-.069

Current Avoidance Variable

 -.432**

  -.451**

  -.463**

  -.358**

L2 Input

    

Parents’ English Proficiency

.305*

 .407**

.188

.285*

Initial Oral English Input

  .318**

.348**

.179

-.112

Current Oral English Input

  .553**

.531**

.306*

.214

Initial English Literacy Input

.179

.233

.194

.066

Current English Literacy Input

 .238*

.210

.203

.009

Initial English Media Input

.230

  .351**

  .373**

  .343**

Current English Media Input

.270*

 .272*

.296*

.140

Note. *p < .05    **p < .01

Based on the results from bivariate correlations, we conducted multiple regression analyses to further understand the best learner-level predictors to long-term outcomes. We selected variables that yielded significant correlations (p < .05) with the L2 outcome variables to enter in the regression model, and used stepwise regression technique to select the best subset of predictors that would explain the maximum amount of variance.6 We opted to use stepwise regression because of the descriptive and exploratory nature of the current study and the constraint of the sample size in relation to the relatively large number of potential predictors. The default “enter” method to force all potential predictors in the linear regression model would have over fitted the model. Stepwise regression utilized mathematical algorithms to select predictor variables: variables that increased F value by at least 0.05 were included and variables that increased F value by less than 0.1 were excluded. Using the guidelines from Cohen, Cohen, West, and Aiken (2013), we also verified that the regression models met the assumptions of linearity, homoscedasticity, normality, and no excessive multicollinearity (tolerance value > .01 and VIF < 10).


We conducted separate stepwise regression analyses for the four language/literacy outcomes and present the results in Tables 5–8. The regression analyses with the four outcomes as the dependent variables revealed both similarities and differences in the set of predictors for different domains. Although AoA was a significant predictor for foreign accents, grammar, and oral language, it was not a significant predictor for literacy outcomes. The strength of AoA also varied among the three significant English L2 domains.


The analysis using foreign accent outcomes revealed a three-predictor model (see Table 5, Step 3): AoA, amount of current oral English language input, and motivation. Based on the magnitude of the standardized coefficients, current oral language input was more strongly related to foreign accents (β = .460; p < .001) than AoA (β = -.426; p < .001) and current motivation (β = -.184; p < .05). The three predictors combined accounted for about 50% of the variances in the foreign accent ratings. In contrast to the three-predictor model for foreign accent outcomes, the model for explaining the variance in grammar outcomes (55%) yielded five predictors. In addition to AoA (β = -.271; p = .003), current oral English language input (β = .378; p < .001) and current motivation (β= -.257; p = .004), language aptitude (β = .274; p = .002) and initial level of heritage culture appreciation (β = -.250; p =.004) were also found to predict grammar outcomes to varying degrees. Similar to the foreign accent outcomes, current oral language input was the strongest predictor to grammar outcomes. The other four predictors were of similar strength.


Table 5. Stepwise Regression Coefficients for Analysis Predicting Accent Ratings (n = 69)

 

Variable

B

SE

β

t

p

r

rpartial

Adj. R2

R2

ΔR2

Step 1

(Constant)

8.78

0.47

 

18.58

.000

 

 

.28

.29

 

Age of Arrival (AoA)

-0.23

0.05

  -0.54***

-5.22

.000

-0.54

-0.54

   

Step 2

(Constant)

5.96

0.67

 

 8.95

.000

 

 

.49

.51

.21

Age of Arrival (AoA)

-0.21

0.04

  -0.48***

-5.46

.000

-0.54

-0.56

   

Oral English language input (current)

0.04

0.01

   0.47***

 5.28

.000

 0.53

 0.55

   

Step 3

(Constant)

6.57

0.71

 

 9.21

.000

  

.51

.54

.03

Age of Arrival (AoA)

-0.18

0.04

  -0.43***

     -4.74

.000

-0.54

-0.51

   

Oral English language input (current)

0.04

0.01

   0.46***

5.35

.000

 0.53

 0.56

   
 

Motivation (current)

-0.12

0.06

    -0.18*

     -2.06

.044

-0.34

-0.25

   

B = Unstandardized beta coefficient, SE = Standard error of the unstandardized beta coefficient, β = Standardized beta coefficient,

* p < 0.05, ** p < 0.01, *** p < 0.001


Note. Excluded predictors (Beta/p value): Years of U.S. Education (β = -.040; p = .769); Current Mandarin Proficiency (β = -.016; p = .887); Avoidance (current) (β = -.138; p = .138); Parents’ English Proficiency (β = -.019; p =.845); Oral English Language Input (initial) (β = .132; p = .136); English Media Input (initial) (β = .039; p = .670); Heritage Culture Appreciation (current) (β = .040; p =.654); English Literacy Input (current) (β = .086; p = .328)


Table 6. Stepwise Regression Coefficients for Analysis Predicting Grammar (n = 69)

 

Variable

B

SE

β

t

p

r

rpartial

Adj. R2

R2

ΔR2

Step 1

(Constant)

73.78

2.67

 

27.61

.000

  

.23

.24

 

Oral English language input (current)

  0.16

0.04

  0.49***

  4.57

.000

0.49

0.49

   

Step 2

(Constant)

82.46

3.23

 

25.54

.000

  

.38

.40

.15

Oral English language input (current)

 0.14

0.03

  0.43***

 4.39

.000

0.49

0.48

   

Age of Arrival (AoA)

-0.72

0.18

 -0.40***

-4.03

.000

-0.46

-0.45

   

Step 3

(Constant)

86.19

3.28

 

26.26

.000

  

.45

.47

.08

Oral English language input (current)

 0.14

0.03

   0.43***

     4.66

.000

0.49

0.51

   

Age of Arrival (AoA)

-0.55

0.18

-0.31**

    -3.14

.003

-0.46

-0.37

   

Motivation (current)

-0.81

0.27

    -0.29**

    -3.02

.004

-0.41

-0.36

   

Step 4

(Constant)

80.56

3.69

 

   21.86

.000

  

.50

.53

.06

 

Oral English language input (current)

 0.13

0.03

     0.38***

     4.30

.000

 0.49

 0.48

   
 

Age of Arrival (AoA)

-0.58

0.17

    -0.32**

   -3.43

.001

-0.46

   -0.40

   
 

Motivation (current)

-0.78

0.26

    -0.28**

   -3.07

.003

-0.41

   -0.36

   
 

Language Aptitude

 1.05

0.37

     0.25**

   2.85

.006

 0.33

0.34

   

Step 5

(Constant)

83.73

3.63

 

 23.04

.000

  

.56

.59

.06

 

Oral English language input (current)

 0.12

0.03

     0.38***

  4.49

.000

 0.49

    0.50

   
 

Age of Arrival (AoA)

-0.49

0.16

   -0.27**

 -3.07

.003

-0.46

   -0.37

   
 

Motivation (current)

-0.72

0.24

   -0.26**

 -2.97

.004

-0.41

   -0.36

   
 

Language Aptitude

 1.15

0.35

    0.27**

  3.28

.002

 0.33

0.39

   
 

Heritage Culture Appreciation (initial)

-0.70

0.23

   -0.25**

 -2.97

.004

-0.34

   -0.36

   

B = Unstandardized beta coefficient, SE = Standard error of the unstandardized beta coefficient, β = Standardized beta coefficient,

* p < 0.05, ** p < 0.01, *** p < 0.001


Note. Excluded predictors (Beta/p value): Years of U.S. Education (β = -.036; p = .783); Current Mandarin Proficiency (β = -.103; p = .353); Embarrassment (current) (β = -.144; p = .110); Avoidance (current) (β = -.090; p = .345); Parents’ English Proficiency (β = .094; p =.331); Oral English Language Input (initial) (β = .063; p = .501); English Media Input (initial) (β = .022; p = .822); English Media Input (recent) (β = -.049; p =.608).

For oral language outcomes, the regression analyses revealed a three-predictor model. In the order of strength of association, oral language outcomes were predicted by language aptitude (β = .354; p = .001), AoA (β = -.308; p = .002), and tendency to avoid using English (β = -.305; p = .005). The three predictors combined accounted for approximately 40% of the variances in the sample.


In contrast to the other three domains, AoA was not a significant predictor of literacy outcomes. Instead, only avoidance (β = -.390; p = .001) and current motivation (β = -.251; p = .025) were found to predict literacy outcomes, and the two predictors combined explained approximately only 22% of the variance.


Table 7. Stepwise Regression Coefficients for Analysis Predicting Oral Language (n = 69)

 

Variable

B

SE

β

t

p

r

rpartial

Adj. R2

R2

ΔR2

Step 1

(Constant)

 8.43

0.14

 

58.56

.000

  

.25

.26

 

Avoidance (current)

-0.25

0.05

-0.51***

-4.84

.000

-0.51

-0.51

   

Step 2

(Constant)

 6.97

0.52

 

13.38

.000

  

.33

.35

.09

Avoidance (current)

-0.19

0.05

   -0.38**

    -3.47

.001

-0.51

-0.40

  

 

Language Aptitude

 0.21

0.07

    0.32**

     2.91

.005

 0.48

 0.34

  

 

Step 3

(Constant)

 7.63

0.53

 

   14.42

.000

  

.41

.44

.09

Avoidance (current)

-0.15

0.05

   -0.31**

   -2.89

.005

-0.51

-0.34

   

Language Aptitude

 0.23

0.07

    0.35**

    3.43

.001

 0.48

 0.39

   

Age of Arrival (AoA)

-0.09

0.03

   -0.31**

   -3.20

.002

-0.37

-0.37

   

B = Unstandardized beta coefficient, SE = Standard error of the unstandardized beta coefficient, β = Standardized beta coefficient,

* p < 0.05, ** p < 0.01, *** p < 0.001


Note. Excluded predictors (Beta/p value): Years of U.S. Education (β = -.116; p = .440); Embarrassment (current) (β = -.125; p = .219); Oral English Language Input (current) (β = .136; p = .177); English Media Input (initial) (β = .065; p = .557); Heritage Culture Appreciation (current) (β = .187; p =.054); Current American Culture (current) (β = .070; p =.481)

Table 8. Stepwise Regression Coefficients for Analysis Predicting Literacy (n = 69)

 

Variable

B

SE

β

t

p

r

rpartial

Adj. R2

R2

ΔR2

Step 1

(Constant)

8.24

0.18

 

46.87

.000

  

.17

.19

 

Avoidance (current)

   -0.25

0.06

    -0.43***

-3.88

.000

-0.43

-0.43

   

Step 2

(Constant)

     9.04

0.39

 

23.30

.000

  

.22

.25

.06

Avoidance (current)

    -0.23

0.06

      -0.39**

    -3.58

.001

-0.43

-0.41

  

 

Motivation (current)

-0.13

0.06

      -0.25*

    -2.30

.025

-0.31

-0.27

  

 

B = Unstandardized beta coefficient, SE = Standard error of the unstandardized beta coefficient, β = Standardized beta coefficient,

* p < 0.05, ** p < 0.01, *** p < 0.001


Note. Excluded predictors (Beta/p value): Age of Arrival (β = -.195; p = .088); Embarrassment (current) (β = -.112; p = .364); Years of U.S. Education (β = .107; p = .372); English Media Input (initial) (β = .147; p = .225); Parents’ English Proficiency (β = .144; p = .211)


DISCUSSION


Given the large and ever-increasing number of child immigrants in the United States, and the close relationships between their English language proficiency, academic achievement (Ardasheva & Tretter, 2013; Halle, Hair, Wandner, McNamara, & Chien, 2012; Kim & Suarez-Orozco, 2014; Suarez-Orozco et al., 2010), and psychological adjustment (Liu, Benner, Lau, & Kim, 2009; Noels, Pon, & Clément, 1996), it is surprising that specific research on child immigrants’ English language and literacy development, particularly long-term outcomes, is very limited (Saunders & O’Brien, 2006). The current study examined the long-term English language and literacy outcomes of 69 first-generation, foreign-born former child immigrants from Chinese ethnic backgrounds. All participants had lived in the United States for at least 10 years and on average 20 years. The study included both subjective and objective language proficiency data obtained from assessments/surveys and learner-level predictors obtained from a survey. The four English L2 domains under study included foreign accents (phonological production) and grammar knowledge measured through psycholinguistic, experimental tasks, and participants’ self-ratings of oral language and literacy proficiency. Twenty NESs selected to have comparable values for gender, current age and education level also provided objective language performance data as the baseline for comparison. The learner variables covered a comprehensive list of constructs, including Age of Arrival (AoA), years of U.S. education, motivation, cultural identity, language aptitude, language input, etc. We asked if the child immigrants had caught up with NESs in multiple language and literacy domains after an average of 20 years of residence in the United States and exposure to English. We also investigated the predictors of former child immigrants’ language outcomes.


Below, we summarize and interpret the results of our two research questions. We first discuss the divergent AoA effect we found across L2 domains, followed by the finding of multiple predictors for participants’ long-term outcomes and the comparison between child immigrants and their NES peers. Throughout our discussion, we also explain whether our expected outcomes are supported by the results, and how our results compare to previous research.


DIVERGENT EFFECTS OF AGE OF ARRIVAL (AOA) ACROSS L2 DOMAINS


Our first research question pertained to the predictors of first-generation former child immigrants’ long-term English language outcomes. We expected to find a robust Age of Arrival (AoA) effect on phonological production, grammar and oral language, but not necessarily on literacy. The current results supported our expectations and were in line with previous L2 acquisition research (e.g., Flege et al., 1999). Although AoA significantly correlated with all four English language domains in bivariate correlation analyses, once the effects of other variables were controlled for in a regression model, AoA was no longer a significant predictor of literacy outcomes. The current study is one of the first to provide empirical evidence for the prediction that critical period hypothesis applies to oral language outcomes only. As a cultural construct, literacy is not bound by the age effect, but is likely dependent on more malleable factors such as instruction and practice. However, we are cautious to interpret this nonsignificant finding in a conclusive way because the size of the effect of AoA on literacy may be small and our sample size may be too modest with which to detect a significant impact.


Comparing the predictive strength of AoA across domains, we found that AoA exerted a stronger effect on phonological production (β = -.426) and oral language (β = -.308) than on grammar outcomes (β = -.271). This particular finding also confirmed our expectation, and corroborated previous research comparing AoA effects on phonological production and grammar concurrently (Flege, MacKay, & Meador, 1999; Flege, Yeni-Komshian, & Liu, 1999; Huang, 2013). The divergent effects of AoA across L2 domains appeared to corroborate developmental psychology and psycholinguistics theory of “multiple critical/sensititve” period hypothesis (Newport, Bavelier, & Neville, 2001; Pulvermüller & Schumann, 1994; Singleton & Ryan, 2004). The multiple critical/sensitive period hypothesis argues for multiplicities of the critical period mechanism. The critical windows differ by areas of languages (phonetics/phonology, lexicon, syntax), and the closure for phonetics/phonology is generally believed to end the earliest, followed by the closure for syntax (Long, 2005; Newport et al., 2001).


MULTIPLE PREDICTORS TO SUCCESSFUL L2 OUTCOMES


In addition to Age of Arrival (AoA), we expected L2 input to be a strong predictor for all language/literacy outcomes. The current results supported this as well as corroborated previous results that showed a strong impact of L2 input on L2 linguistic outcomes (Flege et al., 2009; Huang, 2013). It is worth noting that, child immigrants’ self-reported English input was, in fact, a stronger predictor of both their phonological production and grammar outcomes than AoA was. This particular finding spelled good news for parents and educators of child immigrants, as the result highlighted the importance of a supportive, input-rich environment for successful long-term L2 outcomes.


The positive effects of language aptitude on child immigrants’ language outcomes also confirmed our expectations and were in line with results of numerous prior studies (Abrahamsson & Hyltenstam, 2009; DeKeyser, 2000; Flege et al., 1999; Purcell & Suter, 1980; Thompson, 1991). Although aptitude has traditionally been considered a static trait across life span (Ioup, Boustagui, El Tigi, & Moselle, 1994; Skehan, 1989), some researchers argued for its flexibility and learnability (Grigorenko, Sternberg, & Ehrman, 2000; McLaughlin, 1990). For example, Grigorenko et al. (2000) proposed a new conceptualization of language aptitude as information processing skills that can be trained and learned. The new perspectives thus open the possibility of facilitating and developing learners’ language aptitudes, making the language aptitude construct relevant for L2 learners and their educators.


Furthermore, based on previous research on Mandarin-speaking immigrants (Jia et al., 2002), we expected that psychological/affective factors would also predict the long-term L2 outcomes of child immigrants from Mandarin Chinese L1 background, and the expectation was supported. Those who reported avoiding using English more frequently also evaluated their oral language and literacy proficiency to be lower than those who reported less avoidance. The results also corroborated research with Spanish-speaking immigrants by Birdsong and Molis (2001), which showed that avoidance and self-consciousness correlated with immigrants’ English language proficiency. On the other hand, child immigrants’ self-reported level of motivation to learn English was also a negative predictor of their English grammar and literacy outcomes. We interpreted this finding to mean that their motivation served as a proxy of their English proficiency; those with lower language proficiency felt more concerned with and motivated to improve their proficiency.


Our prediction of an inverse relationship between child immigrants’ Mandarin (L1) language proficiency and L2 outcomes was partially supported. Current Mandarin language proficiency was significantly correlated with phonological production and grammar outcomes, but not with their self-evaluations of oral language proficiency and literacy. The results corroborated the study by Yeni-Komshian, Flege, and Liu (2000), in which they found an inverse correlation between Korean former child immigrants’ native-likeness in their pronunciation of their L1 (Korean) and their pronunciation of their L2 (English). However, in the current study, after controlling for the effects of other predictors in the regression model, the effect of former child immigrants’ Mandarin language proficiency was no longer significant, suggesting that L1 proficiency contributed to L2 outcomes in an indirect way.


CATCHING UP AND FALLING BEHIND


Our second research question asked if first-generation former child immigrants eventually catch up with their NES peers after prolonged residence in an English-speaking context. We expected child immigrants to perform comparably to NESs in all but phonological production domain. Results from the study, however, only partially supported our expectation. After an average of two decades of residence in the United States, child immigrants reached very high levels of proficiency across all domains. However, their proficiency levels were still not comparable to those of NES controls in phonological production and grammar knowledge. The majority of them (approximately two hirds) also self-reported less-than-native-like proficiency in oral language and literacy skills. In other words, after a prolonged period of exposure, child immigrants in the current study did not catch up with their NES peers in their English L2 proficiency.


The current results corroborated previous L2 acquisition research that showed L2 learners’ less-than-native-like proficiency in phonological productions and grammar knowledge (e.g., Flege et al., 1999; Granena & Long, 2013; Huang, 2013). However, the results contradicted some educational research studies that found child immigrants catching up with their NES peers within seven years of full immersive exposure in an L2-speaking context (Conger, 2009; Hakuta et al., 2000; MacSwan & Pray, 2005). The discrepancies between the current results and previous studies may be attributed to differences in methodology, specifically in participants’ backgrounds and language measures. While the current study employed both psycholinguistic measures that evaluated discrete L2 structures and participants’ self-reports of oral language and literacy proficiency, previous studies used standardized language assessments, such as IPT and LAB, to measure general L2 proficiency. It is possible that the psycholinguistic measures were more difficult than standardized language assessments. Additionally, whereas the participants in the current study were all first-generation, foreign-born child immigrants, previous studies may have included second-generation U.S.-born child immigrants. The inconsistencies could be due to the different sample characteristics since first-generation child immigrants are generally at higher risk for lower English proficiency than their second-generation counterparts (Kieffer, 2008; Slama, 2009). Furthermore, at the time of testing, these participants resided in major coastal cities in the United States where there were many other Chinese immigrants. The average self-estimated L2 oral input was approximately 70% for the current sample, though there was a wide range (4%–100%) and a great amount of variation (SD = 20%). The current sample may thus have less L2 exposure compared to child immigrants living in regions with far less Chinese immigrant population. As our results demonstrated, the amount of L2 input significantly predicted L2 outcomes, and the results may have differed had we included a sample who predominantly used English in their daily life.


Another possible explanation for child immigrants’ less-than-native-like L2 proficiency is the concept of fossilization (Han, 2004; Selinker, 1972), a construct in L2 acquisition research relating to the stagnation of progress in L2 development and applying to both child and adult L2 learners. Researchers believe that fossilization is a cognitive mechanism that affects L2 learning and is persistent and resistant to external interventions such as L2 learner’s motivation and efforts. Despite their prolonged full immersion experience in the L2-speaking country, child immigrants in the current study may have experienced fossilization in their L2 learning at one point and ceased to make progress toward the target language forms. Alternatively, they may have reached target forms at one point but regressed to non-target forms due to lack of stabilization of the linguistic forms.


It is important to note that, although the first-generation, foreign-born child immigrants in the study still fell behind their NES peers in L2 proficiency, they had nonetheless achieved very high levels of English language proficiency, and all of them had obtained or were in the process of obtaining a college degree at the time of testing. Their academic performances thus did not seem to have suffered from their less-than-native-like proficiency. In fact, several recent studies found that reclassified, English proficient child immigrants in elementary and middle grades either performed comparably (Kim & Herman, 2009) or outperformed their monolingual NES peers on content-area assessments (Ardasheva, Tretter, & Kinny, 2012). A growing body of psychological research has also claimed a bilingual advantage in cognitive functioning and metalinguistic awareness (Adesope, Lavin, Thompson, & Ungerleider, 2010; Bialystok, 1999; Bialystok & Martin, 2004, although see Morton, 2014, for a recent critique of this literature).


CONCLUSION AND FUTURE DIRECTIONS


To conclude, the current study showed that former child immigrants’ L2 acquisition is a complicated process involving multiple factors. Although age of arrival (AoA) played a critical role in their L2 learning outcomes, it exerted varying degrees of influence by L2 domain. AoA was a strong predictor of L2 phonological production, as manifested in child immigrants’ degrees of foreign accents in their speech. It also predicted grammar knowledge and self-reported oral language proficiency, but not literacy skills. Although AoA remained a significant predictor of phonological production, grammar knowledge and oral language proficiency controlling for the effects of other variables, it was not the strongest predictor. L2 input, language learning aptitude, and child immigrants’ affective status also contributed to their L2 outcomes, and weighed more than AoA. We interpreted the results to be in line with the multiple sensitive period hypothesis in developmental psycholinguistics research. While tentative due perhaps to a modest sample size, the results also suggested that literacy is not susceptible to age-related effects in the same way in which oral language and more specifically the phonological and syntactic systems are (MacSwan & Pray, 2005; Pinker, 1994). Literacy is a cultural construct rather than a biologically unique human system and as such is intensively taught throughout the school years and curriculum. Malleable factors, such as instruction and reading strategies, are thus perhaps more important in determining child immigrants’ long-term literacy outcomes.


The results entail practical implications for educators and parents of child immigrant ELL students. Specifically, ample support and encouragement should be provided for child immigrant ELLs to develop their L2 literacy skills. The positive effects of L2 input on former child immigrants’ L2 outcomes also demonstrated the importance of an input-rich environment for facilitating L2 development, and suggested collaborative efforts between schools and families to create such an environment for child immigrant ELL students (Aikens & Barbarin, 2008; Duursma et al., 2007; Molfese, Modglin, & Molfese, 2003).


Furthermore, we believed that the positive results about the influence of malleable factors, i.e., L2 input, language aptitude, and affective factors, spelled good news for educators and urged for research into these variables. To illustrate, former child immigrants’ self-reported frequency of avoiding using L2 (English) was negatively associated with their oral language and literacy outcomes. To improve current child immigrant ELL students’ English learning outcomes, it would be worth incorporating specific tasks and strategies throughout the school curriculum to increase their exposure to English and opportunities for meaningful use of English. The opportunity to use L2 meaningfully in various contexts is fundamental to successful L2 learning (Ortega, 2009). The responsibility should be shared by both the families and educators of child immigrant ELL students to provide a wide variety of routines and opportunities for formal and informal (i.e., out of school) exposure to and use of both L1 and L2 (Bailey & Osipova, 2016).


The study also found that first-generation, foreign-born former child immigrants did not catch up with NESs in their English L2 proficiency after a prolonged period of exposure. Nonetheless, they had achieved very high levels of proficiency across all L2 domains and had also successfully completed or were working toward a college degree at the time of testing. We concluded that the former child immigrants may have experienced fossilization in L2 development, and/or that their L2 learning environment was not optimal. It is important to note that, despite still falling slightly behind their NES peers, they had acquired impressive, advanced L2 proficiency.


In particular, the results on phonological production, combined with previous research, suggest potential age-related constraints for achieving native-like pronunciation. Although native-like pronunciation is not part of the K–12 academic curriculum and requirements, some assessments of reading fluency involve reading aloud, and nonnative-like pronunciation may thus play a role in the evaluations of child immigrants’ reading proficiency. Educators should take this finding into consideration and try to provide accommodations for such variations without compromising the validity of read-aloud assessments.


Alternatively, although falling short of native-like L2 proficiency in pronunciation as well as in other oral language domains in the L2, these child immigrant ELL students’ academic performances did not seem to be hindered, possibly because their L2 proficiency was sufficiently advanced. We thus urge researchers and educators to also reconsider the “native-like” expectations/standards for ELL students’ L2 outcomes. Because not all ELL students can achieve native-like proficiency in their L2, and their academic achievements did not seem to be compromised accordingly, it may neither be realistic nor necessary to expect native-like L2 outcomes for all ELL students. The “native speaker” criterion, which is used in some English language proficiency assessments, such as the Student Oral Language Observation Matrix (SOLOM), need to be modified or more clearly defined to avoid confusion and frustration for both ELL students and their families and educators.


As Grosjean (1989) argued, bilingual speakers are not two monolinguals in one. Comparing bilinguals’ proficiency in one language against that of monolinguals may thus be unfair to these bilingual/ELL child immigrant students. However, the limited research on the developmental trajectories of bilinguals/second language learners poses challenges to creating appropriate normative language proficiency assessments for this population (Bailey, 2007; Lesaux, 2006). In a recent study, Sanchez et al. (2013) proposed an alternative, i.e., the “multidimensional bilingual assessment approach,” to accurately measure bilingual/ELL students’ language and cognitive abilities. The researchers administrated multiple cognitive and academic language proficiency assessments in the two languages of the bilingual participants’ (i.e., Spanish and English). The results demonstrated that bilingual participants’ unique language development trajectories impacted the reliability and validity of other assessments. The multidimensional bilingual assessment approach appears to be a promising alternative to comparing bilinguals against monolingual norms. To ensure the reliability, validity, and fairness of the assessments for bilingual/ELL students, future research is needed to evaluate this and other alternative assessment approaches.


The research reported here is one of the few studies in the field to include first-hand data for multiple L2 domains, as well as a more comprehensive survey of learner-level predictors of L2 outcomes. The current results afford us a better understanding of child immigrants’ long-term English L2 outcomes, and the average length of immersive exposure to English (i.e., two decades) exceeded that of virtually all existing studies. However, several limitations need to be acknowledged and addressed by future research. First, the study utilized cross-sectional data that covered a wide range of Age of Arrival for child immigrants. To investigate child immigrants’ L2 developmental trajectory, future studies should use a longitudinal design combined with statistical techniques that model L2 growth over time. Further research using qualitative interviews would also help provide a more in-depth understanding of child immigrants’ L2 learning history and the influence of the various predictors on their L2 outcomes. Second, the current study did not include an objective measure of literacy but relied on participants’ self-reporting. Future research incorporating an objective literacy measure would offer additional insights into child immigrants’ L2 literacy outcomes. The potential reliability and validity issues with self-report measures also call for better measures of the three learner-level predictors, L2 input, language aptitude and psychological factors, which were found to be significant predictors of L2 outcomes. In particular, L2 input appeared to hold great potential for explaining child immigrants’ L2 outcomes.


Although the current study distinguished between oral, literacy and media domains, we did not examine the impact of input by context, such as informal social contexts versus formal classroom interactions, or the role of the L2 instruction/program types (e.g., sheltered English instruction, bilingual education) that child immigrants received. Carhill, Suárez-Orozco, and Páez (2008) found the amount of time immigrant youth spent speaking English in informal social contexts to be predictive of their English language proficiency. Several studies comparing different types of instructional programs had also shown the effect of L2 instruction on child immigrants’ L2 outcomes (Lindholm-Leary, 2014; Oller & Eilers, 2002). Further research is clearly needed to fully understand the roles of L2 input and L2 instructional program in former child immigrants’ L2 development.


As an example measure of L2 input, Flege (2009) has proposed using the Experience Sampling Method that involves asking participants at several randomly selected times during a day to report their language exposure in the immediate past hour. There are also standardized, objective measures of language aptitude that can be readily incorporated in future studies to verify and compare with the current findings. Those include the Modern Language Aptitude Test (MLAT) (Carroll & Sapon, 1959), the Pimsleur Language Aptitude Battery (PLAB) (Pimsleur, 1966), and the Defense Language Aptitude Test (DLAB) (Petersen & Al-Haik, 1976).


Finally, because the study focused on a subgroup of child immigrants, i.e., first-generation foreign-born child immigrants who spoke Mandarin as their first language and were college-educated, the results may not generalize to other subgroups of child immigrants with lower education levels or from other native language backgrounds, such as Vietnamese and Spanish. More research on other subgroups of child immigrant students from different education, socio-economic and native language backgrounds is clearly needed. Future studies, for example, might attempt to mine data available in the larger scale and more nationally representative sample provided by the Education Longitudinal Studies of 2002 (ELS: 2002).7 These data, sponsored by the National Center for Education Statistics of the Institute of Education Sciences, U.S. Department of Education, examined the content-area knowledge and educational experiences/opportunities of secondary school students in the United States, and involved a nationwide sample from diverse school types (public vs. private) and socioeconomic, racial/ethnic, and geographical backgrounds. Moreover, similar research efforts collecting new data specific to investigating child immigrants’ long-term English L2 outcomes would greatly help shed light on this important topic. The field would also benefit from a meta-analysis study that compares and contrasts results from different studies to derive a comprehensive picture of child immigrants’ English language development.


Notes


1.

We adopt the United States Department of Health and Human Services’ (2009) definition of “children” to refer to immigrants who arrive in the United States before age 18.


2.

We are aware of the controversies surrounding the “native speaker” construct (Davies, 2004; Ortega, 2013). However, because the purpose of the study is to examine whether child immigrants catch up with their NES peers in their L2 (English) proficiency, we opted to sample an educated NES group to ensure homogeneity in the baseline data for comparison. The NES comparison group indeed performed at ceiling on both of our language measures, and the variations within the NES group were also small (see Table 2), suggesting homogeneity in their English language proficiency.   


3.

The means, standard deviations and ranges of the current age variable are 28.45, 5.75, and 20–41 for the NES group, and 30.14, 5.30, and 20–44 for the Child Immigrant group. The values for the education level variable are 18.25, 2.17, and 15–24 for the NES group, and 18.42, 2.45, and 13–24.5 for the Child Immigrant group.


4.

NES participants also filled out a brief survey reporting their demographic information where applicable.


5.

As mentioned in Footnote 2, we use “native-like” while acknowledging the contested nature of the term in the applied linguistics field (see Davies, 2004, for definitions and discussion).


6.

Although “age” was significantly correlated with foreign accent ratings and grammar outcomes, because of the linear function between “age,” “age of arrival,” and “length of residence,” and the lack of this variable’s predictive power shown in previous studies, we did not select “age” to enter in the stepwise regression models.


7.

We thank an anonymous reviewer for bringing our attention to this database.


Acknowledgment


Data in the current study were collected as part of the first author's dissertation research at UCLA, which was funded by the UCLA Dissertation Year Fellowship, the Chiang Ching-Kuo Foundation for International Scholarly Exchange Dissertation Fellowship, and the Dissertation Grant Program of the journal Language Learning. We would like to thank Lyn Corno and three anonymous reviewers for their helpful comments. All remaining errors are our own.


References


Abrahamsson, N., & Hyltenstam, K. (2009). Age of onset and nativelikeness in a second language: Listener perception versus linguistic scrutiny. Language Learning, 59(2), 249–306.


Adesope, O. O., Lavin, T., Thompson, T., & Ungerleider, C. (2010). A systematic review and meta-analysis of the cognitive correlates of bilingualism. Review of Educational Research, 80(2), 207–245.


Aikens, N. L., & Barbarin, O. (2008). Socioeconomic differences in reading trajectories: The contribution of family, neighborhood, and school contexts. Journal of Educational Psychology, 100(2), 235.


Ardasheva, Y., & Tretter, T. R. (2013). Contributions of individual differences and contextual variables to reading achievement of English language learners: An empirical investigation using hierarchical linear modeling. TESOL Quarterly, 47(2), 323–351.


Ardasheva, Y., Tretter, T. R., & Kinny, M. (2012). English language learners and academic achievement: Revisiting the threshold hypothesis. Language Learning, 62(3), 769–812. doi:10.1111/j.1467-9922.2011.00652.x


Asher, J. J., & García, R. (1969). The optimal age to learn a foreign language. The Modern Language Journal, 53(5), 334–341.


Bailey, A. L. (Ed.). (2007). The language demands of school: Putting academic English to the test. Yale University Press.


Bailey, A. L., & Carroll, P. E. (2015). Assessment of English language learners in the era of new academic content standards. Review of Research in Education, 39(1), 253–294.


Bailey, A. L. & Osipova, A. (2016). Children’s multilingual development and education: Fostering linguistic resources in home and school contexts. Cambridge, UK: Cambridge University Press.


Bialystok, E. (1999). Cognitive complexity and attentional control in the bilingual mind. Child Development, 70, 636–644.


Bialystok, E., & Hakuta, K. (1994). In other word: The psychology and science of second language acquisition: New York: Basic Books.


Bialystok, E., Luk, G., & Kwan, E. (2005). Bilingualism, biliteracy, and learning to read: Interactions among languages and writing systems. Scientific Studies of Reading, 9, 43–61.


Bialystok, E., & Martin, M. M. (2004). Attention and inhibition in bilingual children: Evidence from the dimensional change card sort task. Developmental Science, 7, 325–339.


Bialystok, E., McBride-Chang, C., & Luk, G. (2005). Bilingualism, language proficiency, and learning to read in two writing systems. Journal of Educational Psychology97(4), 580–590.


Bialystok, E., & Miller, B. (1999). The problem of age in second-language acquisition: Influences from language, structure, and task. Bilingualism: Language and Cognition, 2(02), 127–145.


Birdsong, D., & Molis, M. (2001). On the evidence for maturational constraints in second-language acquisition. Journal of Memory and Language, 44(2), 235–249.


Bongaerts, T. (1999). Ultimate attainment in L2 pronunciation: The case of very advanced late L2 learners. Second language acquisition and the critical period hypothesis (pp. 133–159). Mahwah, NJ: Lawrence Erlbaum.


Butler, F. A., & Stevens, R. (2001). Standardized assessment of the content knowledge of English language learners K-12: Current trends and old dilemmas. Language Testing, 18(4), 409–427.


Carhill, A., Suárez-Orozco, C., & Páez, M. (2008). Explaining English language proficiency among adolescent immigrant students. American Educational Research Journal, 45(4), 1155–1179.


Carroll, J. B., & Sapon, S. (1959). The modern language aptitude test. San Antonio, TX: Psychological Corporation.


Cheryan, S., & Bodenhausen, G. V. (2000). When positive stereotypes threaten intellectual performance: The psychological hazards of “model minority” status. Psychological Science, 11(5), 399-402.


Chomsky, N. (1986). Knowledge of language: Its nature, origin, and use. Greenwood Publishing Group.


Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied multiple regression/correlation analysis for the behavioral sciences. Mahwah, NJ: Erlbaum.


Collier, V. P. (1987). Age and rate of acquisition of second language for academic purposes. TESOL Quarterly, 21(4), 617–641.


Conger, D. (2009). Testing, time limits, and English learners: does age of school entry affect how quickly students can learn English? Social Science Research, 38(2), 383–396.


Cortes, K. E. (2006). The effects of age at arrival and enclave schools on the academic performance of immigrant children. Economics of Education Review, 25(2), 121–132.


Cummins, J. (1980). The cross-lingual dimensions of language proficiency: Implications for bilingual education and the optimal age issue. TESOL Quarterly, 175–187.


Cummins, J. (1981). Four misconceptions about language proficiency in bilingual education. NABE Journal, 5(3), 31–45.


Cummins, J. (2000). Academic language learning, transformative pedagogy, and information technology: Towards a critical balance. TESOL Quarterly, 34(3), 537–548.


Davies, A. (2004). The native speaker in applied linguistics. In A. Davies & C. Elder (Eds.), Handbook of applied linguistics (pp. 431–450). Malden, MA: Blackwell.


DeKeyser, R. M. (2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition, 22, 499–533.


Duursma, E., Romero-Contreras, S., Szuber, A., Proctor, P., Snow, C., August, D., & Calderon, M. (2007). The role of home literacy and language environment on bilinguals' English and Spanish vocabulary development. Applied Psycholinguistics, 28(01), 171–190.


Flege, J. E. (2009). Give input a chance. In T. Piske & M. Young-Scholten (Eds.), Input matters in SLA (Vol. 35, pp. 175–190). Multilingual Matters.


Flege, J. E., MacKay, I. R., & Meador, D. (1999). Native Italian speakers’ perception and production of English vowels. The Journal of the Acoustical Society of America, 106(5), 2973–2987.


Flege, J. E., Munro, M. J., & MacKay, I. R. (1995). Factors affecting strength of perceived foreign accent in a second language. The Journal of the Acoustical Society of America97(5), 3125-3134.


Flege, J. E., Yeni-Komshian, G. H., & Liu, S. (1999). Age constraints on second-language acquisition. Journal of Memory and Language, 41, 78–104.


Genesee, F., & Nicoladis, E. (2006). Bilingual acquisition. In E. Hoff & M. Shatz (Eds.), Handbook of language development. Oxford, England: Blackwell.


Granena, G., & Long, M. H. (2013). Age of onset, length of residence, language aptitude, and ultimate L2 attainment in three linguistic domains. Second Language Research, 29(3), 311–343.


Grigorenko, E. L., Sternberg, R. J. & Ehrman, M. (2000). A theory-based approach to the measurement of foreign language aptitude: The CANAL-F theory and test. Modern Language Journal, 84, 390–405.


Grosjean, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain & Language, 36, 3–15.


Hakuta, K., Bialystok, E., & Wiley, E. (2003). Critical evidence a test of the critical-period hypothesis for second-language acquisition. Psychological Science, 14(1), 31–38.


Hakuta, K., Butler, Y. G., & Witt, D. (2000). How long does it take English learners to attain proficiency? (Policy report). Santa Barbara: University of California Language Minority Research Institute.


Hakuta, K., & d'Andrea, D. (1992). Some properties of bilingual maintenance and loss in Mexican background high-school students. Applied Linguistics, 13(1), 72–99.


Halle, T., Hair, E., Wandner, L., McNamara, M., & Chien, N. (2012). Predictors and outcomes of early versus later English language proficiency among English language learners. Early Childhood Research Quarterly, 27(1), 1–20.


Halliday, F. (1994). Rethinking international relations. New York: Palgrave Macmillan.


Hammer, C. S., Jia, G., & Uchikoshi, Y. (2011). Language and literacy development of dual language learners growing up in the United States: A call for research. Child Development Perspectives5(1), 4–9.


Han, Z.-H. (2004). Fossilization in adult second language acquisition. Clevedon: Multilingual Matters.


Huang, B. H. (2013). The effects of age on second language grammar and speech production. Journal of Psycholinguistic Research, 43(4), 397–420.


Huang, B. H., & Jun, S.-A. (2011). Specifying the effect of age on the acquisition of second language prosody. Language and Speech, 54(3), 387–414.


Ioup, G., Boustagui, E., El Tigi, M., & Moselle, M. (1994). Reexamining the critical period hypothesis. Studies in Second Language Acquisition, 16(1), 73–98.


Jia, G., Aaronson, D., & Wu, Y. (2002). Long-term language attainment of bilingual immigrants: Predictive variables and language group differences. Applied Psycholinguistics, 23(4), 599–621.  


Johnson, J. S., & Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21(1), 60–99.


Kieffer, M. J. (2008). Catching up or falling behind? Initial English proficiency, concentrated poverty, and the reading growth of language minority learners in the United States. Journal of Educational Psychology, 100(4), 851.


Kim, H. Y., & SuárezOrozco, C. (2014). The language of learning: The academic engagement of newcomer immigrant youth. Journal of Research on Adolescence, 25(2), 229–245.


Kim, J., & Herman, J. L. (2009). A three-state study of English learner progress. Educational Assessment, 14(3-4), 212–231.


Klesmer, H. (1993). ESL achievement project: Development of English as a second language achievement criteria as a function of age and length of residence in Canada. North York, Ontario: North York Board of Education.


Lardiere, D. (2007). Ultimate attainment in second language acquisition: A case study. Mahwah, NJ: Lawrence Erlbaum.


Lesaux, N. K. (2006). Building consensus: Future directions for research on English language learner at risk for learning disabilities. Teacher College Record, 108, 2406–2438.


Lee, J. (1998). Review of unraveling the "model minority" stereotypes: Listening to Asian American Youth. Teachers College Record99(4), 785–787.


Lindholm-Leary, K. (2014). Bilingual and biliteracy skills in young Spanish-speaking low-SES children: Impact of instructional language and primary language proficiency. International Journal of Bilingual Education and Bilingualism, 17(2), 144–159.


Liu, L. L., Benner, A. D., Lau, A. S., & Kim, S. Y. (2009). Mother-adolescent language proficiency and adolescent academic and emotional adjustment among Chinese American families. Journal of Youth and Adolescence, 38(4), 572–586.


Long, M. H. (2005). Problems with supposed counter-evidence to the Critical Period Hypothesis. International Review of Applied Linguistics in Language Teaching, 43(4), 287–317.


Luk, G., & Bialystok, E. (2013). Bilingualism is not a categorical variable: Interaction between language proficiency and usage. Journal of Cognitive Psychology, 25(5), 605–621.


MacSwan, J., & Pray, L. (2005). Learning English bilingually: Age of onset of exposure and rate of acquisition among English language learners in a bilingual education program. Bilingual Research Journal, 29(3), 653–678.


McLaughlin, B. (1990). The relationship between first and second languages: Language proficiency and language aptitude. In B. Harley, P. Allen, J. Cummins, & M. Swain (Eds.), The development of second language proficiency (pp. 158–178). Cambridge, U.K.: Cambridge University Press.

Molfese, V. J., Modglin, A., & Molfese, D. L. (2003). The role of environment in the development of reading skills: A longitudinal study of preschool and school-age measures. Journal of Learning Disabilities36(1), 59–67.


Morton, J. B. (2014). Sunny review casts a foreboding shadow over status quo bilingual advantage research. Applied Psycholinguistics35(05), 929–931.


Moyer, A. (2004). Age, accent and experience in second language acquisition. Clevedon, UK: Multilingual Matters.


Moyer, A. (2007). Do language attitudes determine accent? A study of bilinguals in the USA. Journal of Multilingual and Multicultural Development, 28(6), 502–518.


National Center for Education Statistics. (1992). The condition of education. Washington, DC: U.S. Department of Education.


Newport, E. L., Bavelier, D., & Neville, H. J. (2001). Critical thinking about critical periods: Perspectives on a critical period for language acquisition. In E. Dupoux (Ed.), Language, brain and cognitive development: Essays in honor of Jacques Mehler (pp. 481–502). Cambridge, MA MIT Press.


Noels, K. A., Pon, G., & Clément, R. (1996). Language, identity, and adjustment: The role of linguistic self-confidence in the acculturation process. Journal of Language and Social Psychology, 15(3), 246–264.


Oller, D. K., & Eilers, R. E. (2002). Language and literacy in bilingual children. Clevedon: Multilingual Matters.


Ortega, L. (2009). Understanding second language acquisition. London: Hodder Education.


Ortega, L. (2013). SLA for the 21st century: Disciplinary progress, transdisciplinary relevance, and the bi/multilingual turn. Language Learning, 63, 1–24.


Patkowski, M. S. (1990). Age and accent in a second language: A reply to James Emil Flege. Applied Linguistics, 11(1), 73–89.


Petersen, C. R., & Al-Haik, A. R. (1976). The development of the defense language aptitude battery. Educational and Psychological Measurement, 36(2), 369–380.


Pimsleur, P. (1966). The Pimsleur language aptitude battery. New York: Harcourt, Brace, Jovanovic.


Pinker, S. (1994). The language instinct: The new science of language and mind (Vol. 7529). London: Penguin UK.


Pulvermüller, F., & Schumann, J. H. (1994). Neurobiological mechanisms of language acquisition. Language Learning, 44(4), 681–734.


Purcell, E. T., & Suter, R. W. (1980). Predictors of pronunciation accuracy: A reexamination. Language Learning, 30, 271–287. doi:10.1111/j.1467-1770.1980.tb00319.x


Ramsey, C. A., & Wright, E. N. (1974). Age and second language learning. The Journal of Social Psychology, 94(1), 115–121.


Sanchez, S. V., Rodriguez, B. J., Soto-Huerta, M. E., Villarreal, F. C., Guerra, N. S., & Flores, B. B. (2013). A case for multidimensional bilingual assessment. Language Assessment Quarterly, 10(2), 160–177.


Saunders, W. M., & O'Brien, G. (2006). Oral language. In F. Genesee, K. J. Lindholm-Leary, W. M. Saunders & D. Christian (Eds.), Educating English language learners: A synthesis of research evidence (pp. 14–63). Cambridge: Cambridge University Press.


Scarcella, R. (2003). Academic English: A conceptual framework. UC Berkeley: University of California Linguistic Minority Research Institute.


Schleppegrell, M. J. (2002). Linguistic features of the language of schooling. Linguistics and Education, 12(4), 431–459.


Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics in Language Teaching, 10(1-4), 209–232.


Singleton, D. D. M., & Ryan, L. (2004). Language acquisition: The age factor (Vol. 47): Bristol, UK: Multilingual Matters.


Skehan, P. (1989). Individual differences in second-language learning. London: Edward Arnold.


Slama, R. B. (2011). A longitudinal analysis of academic English proficiency outcomes for adolescent English language learners in the United States. Journal of Educational Psychology, 104(2), 265–285.  


Snow, C. E. (2010). Academic language and the challenge of reading for learning. Science, 328, 450–452.


Suárez-Orozco, C., Gaytán, F. X., Bang, H. J., Pakes, J., O'Connor, E., & Rhodes, J. (2010). Academic trajectories of newcomer immigrant youth. Developmental Psychology, 46(3), 602.


Suárez-Orozco, C., Suárez-Orozco, M., & Todorova, I. (2008). Learning a new land: Immigrant students in American society. Cambridge, MA: The Belknap Press of Harvard University Press


Thomas, W. P., & Collier, V. P. (2002). A national study of school effectiveness for language minority students' long-term academic achievement. Retrieved from http://files.eric.ed.gov/fulltext/ED475048.pdf


Thompson, I. (1991). Foreign accents revisited: The English pronunciation of Russian immigrants. Language Learning, 41(2), 177–204.


United States Census Bureau. (2013). ACS 3-year estimate: 2013 [Data]. Age by language spoken at home by ability to speak English for the population 5 years and over. Available at http://factfinder2.census.gov


United States Department of Health and Human Services. (2009). Code of Federal Regulations: Protection of Human Subjects. Retrieved from http://www.hhs.gov/ohrp/humansubjects/guidance/45cfr46.html#46.402


Weinberger, S. H. (2013). Speech accent archive. Fairfax, VA: George Mason University. Retrieved from http://accent.gmu.edu.


Wright, E. N. & Ramsey, C. (1970). Students of non-Canadian origin: Age on arrival, academic achievement and ability. Research Report #88, Toronto Board of Education.  


Yeni-Komshian, G. H., Flege, J. E., & Liu, S. (2000). Pronunciation proficiency in the first and second languages of Korean-English bilinguals. Bilingualism Language and Cognition, 3(2), 131–149.


APPENDIX A


Sample Questions from the Background Survey

       Full Name: _______    Interviewer Initial:____    Date:_____ (mm/dd/yy)     Main lang: ____Chi  ____Eng       ID:___________

Language Learning History

 

1.

What language(s) do you and your parents speak? That is, languages that you can carry a conversation. (Check all that apply)

 

You

__ Mandarin   __Taiwanese   __ Cantonese    ___English  ___Other (specify languages/dialects&age of learning)

 

Mother

__ Mandarin   __Taiwanese   __ Cantonese    ___English  ___Other (specify languages/dialects&age of learning)

 

Father

__ Mandarin   __Taiwanese   __ Cantonese    ___English  ___Other (specify languages/dialects&age of learning)

 

2. Please rate your parents’ English proficiency. (Circle)

 

MOTHER

        

 

Listening & Speaking

Pronunciation/Accent

Reading & Writing

FATHER

Listening & Speaking

Pronunciation/Accent

Reading & Writing

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

No proficiency            1       2        3        4        5        6        7        8        9        Native-like

 

3. What language or languages do your parents usually speak to each other at home? (If not applicable, write NA in the “Other”)

 

____ Mandarin     ____Taiwanese     ____ Cantonese     ____English     _____________Other languages/dialects (specify)

 

4. At what age did you start learning English? (i.e., first study of one semester or more)   Age ______  

 

5. Where did you start to learn English?   _____ In my native country     _____ In the US       ______ Other (specify)

6. How did you start to learn English? (please check all the appropriate answers)

____   In school class taught by a native speaker                             _______ In school class taught by a non native speaker

_____ Taught by a native speaker (______________)                    _______ Taught by a non native speaker (___________)

_______ Self learning (please specify how________________ )            _______  Picked it up naturally

7. How frequent is the English instruction, and for how long? (Write NA if not applicable)

8. Have you received any intensive training in English pronunciation/accent correction?  ______Yes   _______  No

If yes, please describe the training you received (when/where/details of the training): ______________________________

9. Please list all places (city, country) in which you have lived for more than 3 months EXCEPT for the United States and your native country.  If you have not lived in other places for more than 3 months, please just leave the question blank.

(a) _____________________ from ________ (month, year) to _________ (month, year)

(b) _____________________ from ________ (month, year) to _________ (month, year)


APPENDIX B

Stimuli for the Phonological Production Task

Please call Stella. Ask her to bring these things with her from the store: six spoons of fresh snow peas, five thick slabs of blue cheese, and maybe a snack for her brother Bob. We also need a small plastic snake and a big toy frog for the kids. She can scoop these things into three red bags, and we will go meet her Wednesday at the train station.

APPENDIX C

Target Structure and Sample Sentences for Grammaticality Judgment Task

 

Structure Type

Examples

JN Items (Johnson & Newport, 1989; DeKeyser, 2000)

Determiner  

 Tom is reading a book in the bathtub.

*Tom is reading book in the bathtub.

Particle Movement   

 Kevin called Nancy up for a date.

*Kevin called Nancy for a date up.

Past Tense  


 Last night the old lady died in her sleep.

*Last night the old lady die in her sleep.

Plurals  


 A shoe salesman sees many feet throughout  the day.

*A shoe salesman sees many foots throughout the day.


Present Progressive  


 Janet is wearing the dress I gave her.

*Janet is wear the dress I gave her.


Pronominalization


 John took a sweater along but didn’t put it on.  

*John took a sweater along but didn’t put on.  


Subcategorization  


 The little boys laughed at the clown.

*The little boys laughed the clown.   


Third Person Singular  


 John’s dog always waits for him at the corner.

*John’s dog always wait for him at the corner.


Wh-Question  


 What is Martha bringing to the party?

*What Martha is bringing to the party?


Yes-No Question  


 Is the baby being held by his mother?

*Is being the baby held by his mother?


Complex NP - Noun Complement  


*What did Tom believe the claim that Ann saw?


Complex NP - Relative Clause  


*What did Sam see the man who stole?






Cite This Article as: Teachers College Record Volume 118 Number 11, 2016, p. 1-42
http://www.tcrecord.org ID Number: 21635, Date Accessed: 12/11/2018 3:36:03 PM

Purchase Reprint Rights for this article or review
 
Article Tools

Related Media


Related Articles

Related Discussion
 
Post a Comment | Read All

About the Author
  • Becky Huang
    University of Texas at San Antonio
    E-mail Author
    BECKY H. HUANG is Assistant Professor in the Department of Bicultural-Bilingual Studies at the University of Texas at San Antonio. Her areas of research include second language acquisition in both immersion and formal instruction contexts, and the assessment of second/foreign language proficiency. Her recent publications have appeared in Studies in Second Language Acquisition, International Journal of Multilingualism, and System: An International Journal of Educational Technology and Applied Linguistics.
  • Alison Bailey
    University of California, Los Angeles
    E-mail Author
    ALISON L. BAILEY is Professor of Human Development and Psychology in the Department of Education at the University of California, Los Angeles and a Faculty Research Partner at the National Center for Research on Evaluation, Standards, and Student Testing (CRESST). Her areas of research interest include first and second language development, and academic language pedagogy and assessment practices and policies with school-age English learners. Her most recent book from Cambridge University Press is Children's Multilingual Development and Education: Fostering Linguistic Resources in Home and School Contexts with Anna Osipova. She is currently Principal Investigator of the Dynamic Language Learning Progressions project funded by WIDA at the Wisconsin Center for Education Research.
 
Member Center
In Print
This Month's Issue

Submit
EMAIL

Twitter

RSS