Academic development of Head Start children: Role of dual language learning status

Abstract

Using a large longitudinal dataset including children who attended Head Start over two years, this study examined academic growth trajectories during the period between Head Start entry and kindergarten (2.5 years), and whether those growth trajectories differ by children's dual language learning status. Analyses comparing three groups of children (i.e., Spanish-English bilinguals, Spanish-English emergent bilinguals [EBs], and English monolinguals) showed three noteworthy findings. First, bilinguals entering Head Start with English proficiency showed similar developmental trajectories in vocabulary and math to those of monolinguals. Second, EBs entering Head Start with limited English proficiency presented the lowest baseline skills in vocabulary and math. Whereas the initial vocabulary gaps generally persisted over time, gaps in math between EBs and monolinguals narrowed by kindergarten. Third, no difference was found between bilinguals and EBs in their Spanish vocabulary development. Results highlight needs for additional instructional support and resources for EBs especially in their vocabulary development.

Introduction

Rapid growth within the dual language learner (DLL) population, particularly among young children under the age of 6 (Child Trends, 2014; U.S. Department of Health and Human Services [U.S. DHHS] & U.S. Department of Education [U.S. DoE], 2017), has increased attention to research regarding the early educational needs of this diverse group of children. Dual language learners (DLLs) refers to children who are simultaneously developing two or more languages or those who are learning English as a second language while having some mastery of their home language already (U.S. DHHS & U.S. DoE, 2017). Across the research literature, DLL is used as an umbrella term encompassing a broad population of children ranging from those with limited English proficiency (LEP) who communicate most strongly in their home languages to bilinguals who communicate with some proficiency in both their home language(s) and English. Although each of these terms (e.g., LEP and bilinguals) can specifically represent a different population of DLLs with varying levels of English proficiency, these terminologies have often been interchangeably used and broadly referred to as DLLs. Other research has studied this diverse population without reference to level of English proficiency, using such terminology as language-minority children, children who speak a language other than English, and children with English as an additional language. As each terminology conveys different meaning and focus, the interchangeable use of varying terminologies has been noted as a major challenge in DLL literature (Castro, 2014; Hammer et al., 2014). Regardless, there is general consensus that DLLs in the U.S. often share the common characteristic of having less exposure to and opportunities for using English at home when they are compared to English-monolingual children.

Though recent developmental literature has uncovered important cognitive benefits for children who can communicate in more than one language (e.g., Barac, Bialystok, Castro, & Sanchez, 2014; Bialystok, 2001; Choi, Jeon, & Lippard, 2018), the predominant discourse within the literature is a view of DLLs as a risk group (August, Shanahan, & Escamilla, 2009; see Hakuta, 1986; Rathbun & West, 2004). For example, in a report using the Early Childhood Longitudinal Study, Kindergarten Class of 1998–1999 (ECLS-K) published by the National Center for Education Statistics (NCES; Rathbun & West, 2004), children's status of using non-English as their primary home language was considered as a family risk indicator along with other factors including household poverty status, low maternal education, and single-parent household. Similarly, a book entitled “Developing literacy in second language learners—Report of the National Literacy Panel on Language-Minority Children and Youth” (August & Shanahan, 2006), prepared by a panel of experts with support from federal government funds, has been criticized for its portrayal of DLLs using a “deficit orientation” because such a perspective conveys the idea that DLLs are “other than normal” (August et al., 2009).

This disadvantaged and deficit viewpoint of the DLL population could partially be a reflection of average DLLs' underperformance in academics, especially when they are assessed in English. Empirical studies, mostly conducted with exclusively Spanish-speaking DLL samples of young children before they enter formal schooling in US settings, find that average DLL students achieve lower skills than non-DLL students in various academic areas including language, math, and reading (Hammer, Lawrence, & Miccio, 2008; Mancilla-Martinez & Lesaux, 2011; Reardon & Galindo, 2009). Further, these early achievement gaps are often found to persist throughout the primary school years and beyond (Child Trends, 2014; Halle, Hair, Wandner, McNamara, & Chien, 2012; Han, 2012; Kieffer, 2008; Mancilla-Martinez & Lesaux, 2011; Reardon & Galindo, 2009). Historically, it was viewed that bilingualism could confuse young children and hinder their subsequent learning (see Hakuta, 1986). DLLs have been further considered as students experiencing double challenges for their learning (Greenfader & Miller, 2014) because DLLs disproportionally experience poverty, which is a well-known predictor of early disadvantages in learning and other risks (Child Trends, 2014).

While the ECE field has often focused on children's primary home language as the key difference between DLLs and English-monolinguals (Castro, 2014), more recent literature has increased attention to heterogeneous characteristics of DLLs. Such literature identifies variations among DLLs in demographic characteristics (e.g., race/ethnicity, country/culture of origin, immigration status and timing, parental English proficiency, and social-economic status (SES); De Feyter & Winsler, 2009; Han, Lee, & Waldfogel, 2012; Hernandez, Denton, & Blanchard, 2011; Winsler et al., 2014) and diversity in young DLLs' language use (e.g., primary home language(s)/dialect(s), types of language(s) a child use at home and school, and the timing of first exposure to the second language; Castro, 2014; Hammer et al., 2008; Kim, Lambert, & Burts, 2016; Mancilla-Martinez & Lesaux, 2011).

Perhaps most fruitful has been attention to home-language and English-language proficiency levels in light of the highly predictive nature of DLLs' English proficiency on school readiness and later school adjustment outcomes (Halle et al., 2012; Lambert, Kim, Durham, & Burts, 2017; Strand, Malmberg, & Hall, 2015; Whiteside, Gooch, & Norbury, 2017). Given the fact that the facilitation of English-language skills could be reasonably addressed within the structures of our current formal education settings (Ackerman & Tazi, 2015), this line of research is particularly relevant. Acknowledging that overlooking differences on English proficiency among young DLLs can obscure important developmental differences (Hammer, Jia, & Uchikoshi, 2011), recent studies, in fact, have made important distinctions among characteristics of language acquisition, such as the sequence of language learning in English and home language(s) (i.e., simultaneous or sequential learner; e.g.,Hammer et al., 2008) and the age/timing of achieving sufficient English skills (e.g.,Halle et al., 2012). Although the specific approaches and number of group distinctions vary, these studies collectively chart a new course for research to better understand heterogeneity among DLLs based on their English proficiency.

Recent research with school-aged children has documented differences in academic achievement across DLLs based on English proficiency. Results indicate that DLLs who acquire English proficiency early present more positive growth trajectories in academics and social-emotional skills, and those trajectories differ from their English monolingual peers (Halle et al., 2012; Han, 2012; Han et al., 2012; Kieffer, 2008; Mancilla-Martinez & Lesaux, 2011; Whiteside et al., 2017). One study using the ECLS-K (Halle et al., 2012), for example, found that DLLs who entered kindergarten with English proficiency demonstrated achievement in math and reading that was comparable or better than their English monolingual peers during the period between kindergarten and eighth grade. Those who entered kindergarten with limited English proficiency, however, presented poorer academic performance and growth, with the worst outcomes evidenced for children who did not achieve English proficiency by first grade.

Although this relatively new line of research is particularly scant within the ECE literature, preliminary evidence suggests achievement gaps also exist among preschool-aged DLLs with varying levels of English proficiency or exposure (Choi et al., 2018; Hammer et al., 2008; Kim et al., 2016; Lambert et al., 2017; Mancilla-Martinez & Lesaux, 2011). For example, Lambert and colleagues empirically classified 4-year-old DLLs based on their English and Spanish use at home and in the classroom. Using Latent Class Analyses, these researchers found three groups of DLLs: bilinguals, emergent bilinguals [EBs], and heritage language speakers. At the beginning of the academic year, these three DLL groups, relative to English monolingual peers, were generally found to present lower skills in various teacher-rated developmental areas (i.e., cognitive, language, math, and social-emotional skills). However, the three groups of DLLs presented diverse developmental patterns by the end of the school year: Bilinguals tended to catch up or even outperform English monolinguals in teacher-rated academic skills, while EBs and heritage language speakers were not able to catch up with their English-monolingual peers in academics, even though they showed faster rates of growth in their academic achievement than the English monolingual children.

Choi et al. (2018) expanded on these findings within the area of mathematics achievement by focusing on an exclusively low-income sample of children enrolled in Head Start at age 4 and followed into kindergarten (1.5 years). Findings indicated that bilinguals presented the highest math achievement by the middle of kindergarten followed by English-monolingual peers and then EB (or DLLs with limited English skills) identified at age 4. Notably, this study also found that these achievement gaps were explained by the variations in children's developmental trajectories in inhibitory control skills (i.e., baseline skills and the rate of growth), an area where bilinguals have evidenced particular benefits in development (Bialystok, 2001). These recent findings from preschool-aged children generally agree with those from school-aged children, suggesting that earlier, rather than later, achievement of age-appropriate English proficiency has potential benefits for DLLs. In line with studies from school-aged children, these findings from younger children further refute the deficit-focused, homogenous view of DLLs as a universally at-risk population.

While laudable improvements have been made, early childhood research continues to be limited for informing programmatic and policy changes that adequately address the needs of America's increasing population of DLLs (seeCastro, 2014). With few exceptions, most studies exclusively sample either English-monolingual children or sub-populations of DLLs who are proficient in English; DLLs as a whole group or sub-populations of DLLs who do not meet minimum English proficiency standards skills are often excluded (Hammer et al., 2011). Furthermore, while the need to consider variations in English proficiency among DLLs has been noted, heterogeneity in their home language skills or conceptual skills that incorporate both languages is not often investigated. Frequently, these studies rely on child assessments conducted only in English, which is a clear confound with DLL status. Scholars further acknowledge methodological concerns in DLL studies (see Castro, 2014; Hammer et al., 2014), such as the reliance on small homogenous samples or cross sectional or short-term longitudinal data with limited capacity for longitudinal examination. ECE research, specifically, is limited by a small quantity of studies that often focus on only one domain of academic skills and rely exclusively on parent/teacher ratings of child outcomes and/or language use. Thus, while recent research expanding beyond the simple dichotomous categorization of children's DLL status based on their use of English in home is promising, additional studies with preschool children are needed to better understand how early English proficiency predicts DLLs' developmental trajectories across multiple academic domains (Han, 2012).

The current study aims to extend this emerging ECE literature by conducting a longitudinal study of the academic growth trajectories of preschoolers with varying baseline English proficiency levels. Using the Head Start Family and Child Experiences Survey 2009 cohort (FACES 2009; U.S. DHHS, 2009–2013) dataset, this study examined academic developmental trajectories of 3-year-old children with varying DLL status across four points in time between Head Start entry and kindergarten (2.5 years). Such a repeated measures approach enhances statistical power compared to studies which include only two time points reflecting baseline and outcomes. Further, growth trajectory models such as this allow for an examination of non-linear patterns of development over time. As previous studies have shown that learning does not exclusively occur in a linear manner (Choi, Elicker, Christ, & Dobbs-Oates, 2016; Hindman, Cromley, Skibbe, & Miller, 2011; Peisner-Feinberg et al., 2001; Wen, Bulotsky-Shearer, Hahs-Vaughn, & Korfmacher, 2012), this approach has the capacity to reveal a more precise understanding of the academic developmental patterns among the three language groups of children attending Head Start.

To accomplish the goals of the study, this research deliberately utilized the FACES dataset which include a representative sample of low-income children enrolled in Head Start at age 3. Given national estimates indicating that >25% of children enrolled in Head Start are DLLs, with 92% of those speaking Spanish as the primary home language, this is an ideal target population to study language diversity (Moiduddin, Aikens, Tarullo, West, & Xue, 2012). FACES includes two years of data in Head Start when children were ages 3 and 4, which is not the case for many other preschool studies that focus exclusively on 4-years-olds. To address limitations in prior research, we investigate multiple academic domains using scores that were assessed or estimated in consideration of DLLs' home language (i.e., Spanish) to avoid sole reliance on English in child assessments. Four key early academic skills were examined: English receptive vocabulary, Spanish receptive vocabulary, conceptual expressive vocabulary (assessed in both English and Spanish if home language is Spanish), and math (assessed in English if children passed language screening or in Spanish if children failed to pass the screening).

Two primary aims guided the study. First, we examined children's growth patterns in academic skills during the period between Head Start entry and kindergarten (approximately 2.5 years), and explored whether those growth trajectories vary by three groups of children identified based on parent-reported primary home language use and direct child assessed English proficiency at the entry of Head Start around age 3. The three groups identified for this study were (1) Spanish-English bilinguals, (2) Spanish-English emerging bilinguals (EB), and (3) English monolinguals (see “Measures” for detailed descriptions). Second, we examined whether academic achievement in kindergarten differed by these three groups of children. Specific research questions are summarized below.

• Do children present different growth trajectories in academics (i.e., baseline skills and the rate of growth) during the period between Head Start entry and kindergarten (2.5 years) based on their DLL status identified at Head Start entry?

• Do children present different academic achievement in kindergarten based on their DLL status identified at Head Start entry?