Teachers' Hub

The magazine to cross ideas in early childhood education

Long-Term Outcomes of Bilingual Education Models:

What Does the Research Tell Us?


Parents, educators, local communities, researchers, and policy-makers have an interest in promoting the academic achievement of children whose first language is not considered the primary or majority language of the country. In the United States where English is the primary (though not official) language, language-minority children who speak a language other than English at home are typically referred to as English Language Learners (ELLs) or Dual Language Learners (DLLs) - the most rapidly growing segment of the student population of the country (Wolf et al., 2008). Although many DLLs excel in school, especially after they attain English proficiency (with former DLL students often outperforming native speakers of English Ardasheva, Tretter, & Kinny, 2012)), on average, sizeable disparities in achievement exist favoring native English-speaking students. Thus, a critical question is how schools can best support the academic achievement of, and ensure educational equity for, DLL students.


When DLLs begin their formal education in the U.S., they are typically enrolled in some kind of second language instruction program, most of which prioritize learning the majority language in terms of both oral and written English as fast as possible, resulting in English monolingualism. The instructional approach in many of these programs continues to be English-only. Though some programs will use the child’s home language for 1-3 years to provide accessible instruction in academic content areas, use of the home language is gradually phased out in favor of English-only instruction (Center for Applied Linguistics [CAL], 2016). Other programs that deliver instruction in English and a partner language have increased more than tenfold since the 1980s, and aim for students to attain and maintain advanced levels of proficiency in two or more languages (CAL, 2016).


Parents and educators should be aware of the different types of programs that exist and what the research says regarding the academic performance of DLLs who enroll in them. Programs that schools implement can vary greatly, but there are several main models that are used often in the U.S, in order of the extent to which learners' home language is used (for more details, see Tedick, Christian, and Fortune (2011) as well as online resources, CAL (2016) and “What are common program models for ESL education?” (2014)):


  • English immersion/English-only programs – English is the only language of instruction (sometimes called ‘submersion’), and the goal is for the student to attain oral and written English proficiency as quickly as possible in mainstream classrooms and without regard for the other language (and without other bilingual education services). The underlying assumption is that 100% exposure to English is the fastest, the best way to do this.


  • Pull-out – DLLs are pulled out of regular class time by a trained English as a second language (ESL) specialist at various times during the school day in order to receive ESL instruction (typically in English), causing learners to miss core content instruction. Children can be pulled out of class across different grade levels and classrooms in order to group students with similar or mixed levels of English proficiency together.


  • Inclusion/Pull-in – A trained specialist enters the classroom to work with DLLs in small groups and/or individually in the classroom setting. The specialist may modify the lesson that the rest of the class is learning and use comprehension aids such as pictures. English is the main language of instruction in this model as well though the specialist may use the students' L1 to the extent needed to facilitate comprehension.


  • Transitional bilingual education – Students are instructed in the home language for 1-3 years ('early exit') or 4-6 years ('late exit'), and it is gradually replaced with English-only instruction, with the ultimate goal of developing English proficiency as soon as possible.


  • One-way immersion – Often referred to as developmental bilingual education, students in these programs are mainly DLLs, or native speakers of the minority language, and receive 80-90% of instruction in their home language (see 90/10 and 80/20 models below).


  • Two-way immersion (TWI) – Also known as dual language programs, these programs deliver at least 50% of instruction in the partner/minority language in addition to English, and aim to enroll equal numbers of native English-speaking students and native speakers of the partner language. In 50/50 TWI models, students study language arts and academic core content (e.g., math, science, social studies) in both languages, ideally 50% of the time in English and 50% of the time in the other language. TWI models may also follow 90/10 or 80/20 distributions where native English and DLL students are initially instructed in the partner language the majority of the time and instructional time in English gradually increases every year until both languages are used 50% of the time. The ultimate goals of TWI programs, which are offered for a minimum of five years and often extended into middle school, are for all students to develop bilingualism and bi-literacy, grade-level academic performance, and cross-cultural awareness.


Understanding the empirical evidence for the relative effectiveness of each type of program listed above is essential, as debates on the role of bilingual education are typically embedded in the larger conflicting ideologies of the time. Bilingual education in the U.S. has a long history of oscillating between pro- and anti-bilingual education, on the one hand touting the cognitive, linguistic, affective, cultural, social and economic benefits associated with bilingualism, while on the other, framing the speaking and/or maintenance of home language other than English as a threat to national unity (Crawford, 1999; Flores & García, 2017; Leeman, 2004; Kim, Hutchison, & Winsler, 2015; Ovando, 2003). When discourse about bilingual education concerns the acquisition of an additional language by (white, economically advantaged) language-majority students, it tends to be viewed favorably as a desirable resource and economic asset, whereas when it concerns maintenance of a language other than English by ethnically diverse language-minority learners (likely a person of color, in poverty), it is more often treated as an obstacle or problem to be overcome (Aparicio, 1998). Language can be seen as a problem, a right, or a resource when it comes to language planning and policy-making in the U.S. depending on the underlying ideological orientation (Hult & Hornberger, 2016; Ruíz, 1984).


These debates have perhaps the strongest impact in education, as policy decisions at the federal and state level can positively or negative affect equal access to, and the quality of, education for different groups of students, particularly those who are underserved. For instance, the passage of the Bilingual Education Act (1968, Title VII, Elementary and Secondary Education Act) and the Supreme Court case of Lau vs. Nichols (1974) in the U.S. asserted the need for schools to provide additional language support to ensure educational equity for students whose home language was not English, under the logic that without it, their right to a quality education was being violated. However, clearly not all models of language education ensure educational equity, as they vary greatly in their objectives and eventual outcomes (August & Hakuta, 1998; Baker, 2011; Christian, 2016; García, 2009; Lindholm-Leary, 2001). Thus, a chief concern among educators and policy-makers (and the topic of this article) is to assess what the research says about which program(s) lead to the best academic and linguistic outcomes for DLLs. We summarize several key findings from this research that parents and educators should be aware of in order to understand how different types of bilingual education programs relate to later academic outcomes for students.


Effectiveness of Bilingual Education Models


To compare the effectiveness of different models, research tends to compare outcomes for DLLs enrolled in different programs, with a focus on their speed in acquiring English and/or their academic performance on standardized tests of reading and math. To date, the majority of research suggests that TWI models are the most efficient solution to promoting English proficiency and reducing the achievement gap, supporting some degree of home language instruction (Collier & Thomas, 2004; Greene, 1998; Lindholm-Leary, 2001; Lindholm-Leary & Howard, 2008; Marian, Shook, & Shroeder, 2013; Slavin & Cheung, 2005; Sugarman & Howard, 2001; Steele et al., 2017; Thomas & Collier, 2002; Willig, 1985; Winsler, Diaz, Espinosa, & Rodriguez, 1999). However, previous methodological limitations (discussed below) prevent unequivocal statements in support of TWI (Crawford, 2004; Krashen, 2005; Salazar, 1998) and some previous research has reported similar or superior outcomes in reading for structured English immersion compared to transitional bilingual education (Rossell, 1990; Rossell & Baker, 1996). However, the latter studies have been critiqued for problematic definitions and ideologically motivated discourse (Cummins, 1999; Krashen, 1996).


Dual language immersion has received overwhelmingly positive support based on work by Wayne Thomas and Virginia Collier who carried out several cross-sectional and longitudinal studies in grades K-12 in U.S. public schools since the early 80's, analyzing more than 700,000 DLL student records within 5 large, geographically diverse school districts (1997, 2002; Collier & Thomas, 2004) as well as over 7.5 million student records from a total of 36 school districts in 16 U.S. states (2012, 2014; Collier & Thomas, 2017). In these studies, the researchers consistently report that TWI bilingual education programs, in comparison with one-way immersion, content-based and traditional transitional bilingual education, and ESL inclusion/pull-in and pull-out models, are the most effective in attaining better long-term academic outcomes as measured on state-mandated, standardized tests of reading. Specifically, language minority students in TWI programs who begin U.S. schooling with no English proficiency in kindergarten, reach age-appropriate grade-level norms (50th percentile) in reading earlier (by 5th grade) than language minority students in other programs, and subsequently sustain that level of achievement later on. Importantly, significant positive effects for TWI only emerge after 3-5 years of being enrolled in well-implemented, high-quality, dual language programs (Collier & Thomas, 2017).


When interpreting these strong effects, it is important to consider certain aspects of the research methodology, such as whether student background variables were statistically controlled in order to better isolate the effect of program type. Thomas and Collier examined cohorts of similar students grouped by several key demographic student-level variables (first and second language proficiency, age, students' first language, level of formal schooling in the first language, and socioeconomic status). However, without random assignment, a selection bias may still have influenced results since this research examines students in existing programs/schools and other differences between students that did and did not choose to enroll in TWI programs may explain DLL students’ enhanced academic achievement. Several other studies report positive effects for TWI without either a longitudinal design (Marian, Shook, & Schroeder, 2013), including a control group (Sugarman & Howard, 2001), or performing controlled comparisons between program types (Lindholm-Leary, 2001).


Fortunately, some true experimental studies exist that randomly assign children to different types of bilingual education models and compare different outcomes. For example, a recent study conducted in Portland, Oregon, USA showed that students randomly assigned to dual immersion programs in kindergarten (both one-way and two-way) not only gained English proficiency faster, but they also significantly outperformed students enrolled in non-immersion, English-only programs on state-wide reading tests in 5th and 8th grade, with no effects (positive or negative) on mathematics and science performance (Steele et al., 2017). Previous research also supports positive effects of bilingual instruction on reading, though not all studies were truly randomized. For example, Slavin and Cheung (2005) reviewed the results of 17 empirical studies comparing English-only and bilingual reading programs and found that students in programs with instruction in both the home language and in English made greater advances in both languages.


As part of the Miami School Readiness Project (MSRP; Winsler et al., 2008), which collected data on five cohorts of 4-year-old prekindergarten (pre-K) children over five years, Winsler, Kim, and Richard (2014) examined child-level predictors (e.g., socioemotional skills, L1 competence) of English acquisition during the transition from pre-K to kindergarten for a subset of low-income, Latino immigrant children. They reported that greater facility in Spanish at age 4 predicted the attainment of English proficiency in kindergarten, as measured on a state-approved oral proficiency test. In an expanded longitudinal study spanning from kindergarten through fifth grade, Kim, Curby, and Winsler (2014) investigated similar child-level factors as well as family-level (e.g., parental education) and school-level (e.g., class size) factors that may predict low-income children's success in acquiring English proficiency. Crucially, L1 competence in Spanish also emerged as a key predictor of achieving district standards of English proficiency. However, the effect of types of language instruction that DLL children received in early elementary school remains unclear as it was not included among school-level factors.


Taken together, the evidence suggests that, rather than being a problem or liability, children’s first language appears to play a facilitative role in acquiring the second language, English, and in promoting academic achievement over the long-term for DLLs. Research indicates that additive bilingual models of education in which students learn core content through two languages appear to be superior to subtractive models like English-only immersion and transitional bilingual models and have the added benefit of promoting bilingualism, biculturalism, and biliteracy among both language-minority and language-majority students (García, 2009). However, more large-scale, statistically robust research is needed in order to unequivocally state that dual language immersion is the most effective avenue for promoting equal access to meaningful education for all students and to differentiate the effects of different immersion models (i.e., one-way versus two-way). Furthermore, we need more evidence pertaining to the effects of TWI not only on academic performance in reading but on other content areas as well, like math and science.


Reasons for the Effectiveness of Additive Models of Bilingual Education


A variety of social, linguistic, psychological, and practical reasons may underlie the apparent superiority of programs that provide content instruction in both the home language and the majority language. From a social standpoint, García and García (2012) noted that Vygotsky’s sociocultural learning theory provides a strong theoretical framework from which to understand the effectiveness of dual language instruction, given that language acquisition is largely a social process and native/nonnative speaker integration provides substantial opportunities for natural interactions to occur on a daily basis in ways that facilitate language learning. In other words, bilingual programs enable peer-learning by creating a setting in which children are encouraged to interact with each other in more than one language (García & García, 2012) and to serve as leaders or models for their peers (de Jong & Howard, 2009; Lindholm-Leary, 2001). This sort of continuous peer interaction likely offers benefits that go beyond language learning compared to the segregation that is often present when DLLs (who are more likely to be in poverty) are all placed together in different classrooms or schools.


Collier and Thomas (2004) noted that TWI programs may work so effectively because they go beyond the remedial support that English-only models offer for a few years. The pull-out model removes DLLs from the classroom during times when they are receiving important grade-level instruction in math, science, language arts, and other subjects, which can pose a challenge to providing consistent instruction (“What are common program models for ESL education?” 2014). When removed from the regular curriculum, DLLs are likely to fall behind standards for grade-level performance, and they must makeup in some cases over a year’s progress each year to eventually reach the achievement levels of their language-majority peers. In contrast, TWI programs do not isolate children from the regular classroom based on language proficiency, and provide grade-level academic curriculum in two languages (Collier & Thomas, 2004).


Another likely mechanism underlying enhanced L2 learning in TWI programs is that first language competence appears to help in the acquisition of L2, and academic skills in particular are known to transfer between languages. Thus, there is a hypothesized interdependence between cognitive and linguistic development in both languages (Cummins, 1991) such that, rather than detracting from developing L2 competence, strong skills in L1 can actually aid in the learning of a second language (Kim et al., 2014; Winsler et al., 2014), and one does not have to pick one language over the other – it is possible to have high-quality early childhood bilingual education where both English and Spanish, for example, can grow over time (Winsler et al., 1999).




The academic achievement of DLLs is a critical component for the success of educational systems throughout the world. Robust longitudinal research conducted in the United States, where DLLs represent 10% of the K-12 public school population in 2014-2015 and are projected to make up 40% of the population by 2030 (National Center for Education Statistics, 2017), indicates that enabling children to strengthen both their home language and English in a formal education setting not only helps them acquire the majority language, but also predicts stronger academic outcomes years later. Ready access to this cumulative empirical evidence is crucial in order for stakeholders, including parents, educators, and policy-makers, to be able to make informed decisions about the education of students who speak a home language other than English.

Ellen J. Serafini

Assistant Professor

Applied Linguistics

George Mason University

United States

Nadine Rozell

Foreign Languages and Psychology

George Mason University

United States

Adam Winsler

Associate Chair Professor,

Applied Developmental Psychology

George Mason University

United States



Aparicio, F. R. (1998). Whose Spanish, whose language, whose power? An ethnographic inquiry into differential bilingualism. Indiana Journal of Hispanic Literatures, 12, 5-26.

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

August, D., & Hakuta, K. (Eds.) (1998). Educating language minority children. Washington, DC: National Academy Press.

Baker, C. (2011). Foundations of bilingual education and bilingualism (5th ed.). Bristol, UK: Multilingual Matters.

Christian, D. (2016). Dual language education: Current research perspectives. International Multilingual Research Journal, 10, 1-5.

Center for Applied Linguistics. (2016). Glossary of terms related to dual language/TWI in the United States. Retrieved from http://www.cal.org/twi/glossary.htm.

Collier, V. P., & Thomas, W. P. (2004). The astounding effectiveness of dual language education for all. National Association for Bilingual Education Journal of Research and  Practice, 2(1), 1-20.

Collier, V. P. & Thomas, W. P. (2017). Validating the power of bilingual schooling: Thirty-two years of large-scale, longitudinal research. Annual Review of Applied Linguistics, 37, 1-15.

Crawford, J. (1999). Bilingual education: History, politics, theory, and practice. Los Angeles, CA: Bilingual Educational Services.

Crawford, J. (2004). Has two-way been oversold? Bilingual Family Newsletter, 21(1), 98-100.

Cummins, J. (1991). Interdependence of first- and second-language proficiency in bilingual children. In E. Bialystok (Ed.), Language processing in bilingual children (pp. 70–89). Cambridge, UK: Cambridge University Press.

Cummins, J. (1999). The ethics of doublethink: Language rights and the bilingual education debate. TESOL Journal, 8(3), 13-17.

de Jong, E., & Howard, E. (2009). Integration in two-way immersion education: Equalizing linguistic benefits for all students. International Journal of Bilingual Education and Bilingualism, 12(1), 81-99.

Flores, N. & García, O. (2017). A critical review of bilingual education in the United States: From basements and pride to boutiques and profit. Annual Review of Applied Linguistics, 37, 14-29.

García, E. E., & García, E. H. (2012). Understanding the language development and early  education of Hispanic children. New York: Teachers College Press.

García, O. (2009). Bilingual education in the 21st century: A global perspective. UK: Wiley-Blackwell.

Greene, J. P. (1998). A meta-analysis of the effectiveness of bilingual education programs. Claremont, CA: Tomas Rivera Policy Institute.

Krashen, S. (1996). 'The case against bilingual education'. Paper presented at the Georgetown University Roundtable on Languages and Linguistics, Washington, DC.

Krashen, S. (2005). The acquisition of English by children in two-way programs: What does the research say? In V. Gonzales & J. Tinajero (Eds.), Review of Research and Practice (pp. 3-19). Mahwah, NJ: Lawrence Erlbaum Associates.

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

Hult, F. M. & Hornberger, N. H. (2016). Revisiting orientations in language planning: Problem, right, and resource as an analytical heuristic. Bilingual Review/Revista Bilingüe, 33, 30-49.

Kim, Y. K., Curby, T. W., & Winsler, A. (2014). Child, family, and school characteristics related to English proficiency development among low-income, dual language learners. Developmental Psychology, 50(12), 2600-2613.

Kim, Y.K., Hutchison, L.A., & Winsler, A. (2015). Bilingual education in the United States: An historical overview and examination of two-way immersion. Educational Review, 67 (2), 236-252.  Doi:10.1080/00131911.2013.865593

Leeman, J. (2004). Racializing language: A history of linguistic ideologies in the US Census. The Journal of Language and Politics, 3(3), 507-534.

Lindholm-Leary, K.J. (2001). Dual language education. Clevedon, England: Multilingual Matters.

Lindholm-Leary, K. J., & Howard, E. R. (2008). Language development and academic achievement in two-way immersion programs. In T. W. Fortune & D. J. Tedick (Eds.), Pathways to multilingualism: Evolving perspectives on immersion education (pp. 177–200). Oxford, England: Blackwell.

Marian, V., Shook, A., & Shroeder, S. R. (2013). Bilingual two-way immersion programs benefit academic achievement. Bilingual Research Journal, 36, 167-186.

National Center for Education Statistics. (March 2017). English Language Learners in Public Schools. Washington, DC: US Department of Education. Retrieved from https://nces.ed.gov/programs/coe/indicator_cgf.asp

National Education Association, NEA Education Policy and Practice Department. (2008). An  NEA policy brief: English Language Learners Face Unique Challenges. Retrieved from  http://www.nea.org/assets/docs/HE/ELL_Policy_Brief_Fall_08_(2).pdf

Ovando, C.J. (2003). Bilingual education in the United States: Historical development and current issues. Bilingual Research Journal, 27(1), 1-24.

Rossell, C. H. (1990). The effectiveness of educational alternatives for limited-English-proficient children. In G. Imhoff (Ed.), Learning in two languages (pp. 71–121). New Brunswick, NJ: Transaction Publishers.

Rossell, C. H., & Baker, K. (1996). The educational effectiveness of bilingual education. Research in the Teaching of English, 30(1), 7–74.

Ruíz, R. (1984). Orientations in language planning. NABE Journal, 8, 15-34.

Salazar, J. J. (1998). A longitudinal model for interpreting thirty years of bilingual education research. Bilingual Research Journal, 22, 1-12.

Slavin, R. E., & Cheung, A. (2005). A synthesis of research on language of reading  instruction for English Language Learners. Review of Educational Research,75(2), 247- 284.

Steele, J., Slater, R., Zamarro, G., Miller, T., Li, J., Burkhauser. S., & Bacon, M. Dual-Language  Immersion Programs Raise Student Achievement in English. Santa Monica, CA: RAND  Corporation, 2017. https://www.rand.org/pubs/research_briefs/RB9903.html.

Sugarman, J. & Howard, L. (2001). Two-way immersion shows promising results: Findings from a new study. Washington, DC: Center for Applied Linguistics.

Tedick, D. J., Christian, D., & Fortune, T. W. (Eds.). (2011). Immersion education: Practices, policies, and possibilities. Clevedon, UK: Multilingual Matters.

Thomas, W.P., & Collier, V.P. (1997). School effectiveness for language minority students. Washington, DC: National Clearinghouse for English Language Acquisition (NCELA) Resource Collection Series, No. 9.

Thomas, W.P., & Collier, V.P. (2002). A national study of school effectiveness for language minority students' long-term academic achievement. Santa Cruz, CA: Center for Research on Education, Diversity and Excellence, University of California-Santa Cruz.

Thomas, W. P., & Collier, V. P. (2012). Dual language education for a transformed world. Albuquerque,NM, Dual Language Education of New Mexico-Fuente Press.

Thomas, W. P., & Collier, V.P. (2014).  English learners in North Carolina Dual Language Programs: Year 3 of this study: School Year 2009-2010.  Raleigh, NC: North Carolina Department of Public Instruction.

What are common program models for ESL education? (2014, June 12). Retrieved from  https://ed.psu.edu/pds/elementary/intern-resources/esl-handbook/common- models

Willig, A. C. (1985). A meta-analysis of selected studies on the effectiveness of bilingual education. Review of Educational Research, 55(3), 269–317.

Winsler, A., Diaz, R.M., Espinosa, L., & Rodriguez, J. L. (1999). When learning a second language does not mean losing the first: Bilingual language development in low-income, Spanish-speaking children attending bilingual preschool. Child Development, 70, 349-362.

Winsler, A., Tran, H., Hartman, S., Madigan, A., Manfra, L., & Bleiker, C. (2008). School readiness gains made by ethnically diverse children in poverty attending center-based childcare and public school prekindergarten programs. Early Childhood Research Quarterly, 23, 314– 329.

Winsler, A., Kim, Y. K., & Richard, E. (2014). Socio-emotional skills, behavior problems, and Spanish competence predict the acquisition of English for English language learners in poverty. Developmental Psychology, 50, 2242–2254.

Wolf, M.K., Kao J., Griffin N., Herman J.L., Bachman P.L., Chang S.M., & Farnsworth T. (2008). Issues in assessing English language learners: English language  proficiency measures and accommodation uses. CRESST Report 732. National Center  for Research on Evaluation, Standards, and Student Testing; Los Angeles, CA: