Educational gains for students in dual language programs: Effects of language proficiency and program model

• In Event: Do dual language experiences benefit the academic and literacy outcomes of Spanish speaking ELs?

Abstract

Dual language education (DLE) programs reduce the achievement gap for English Learners (ELs) (August & Shanahan, 2006). Spanish-English DLE programs typically employ a 50/50 model (instruction time is half in Spanish and half in English), or a 90/10 model (90% of instruction in Spanish in kindergarten and decreasing over grades until reaching a 50/50 distribution). Prior comparisons of 90/10 and 50/50 models indicate that literacy, reading, and math achievement in English initially favors the 50/50 model, but these differences remit by the upper elementary grades. In contrast, the 90/10 model confers an advantage in Spanish oral language and literacy for ELs and non-ELs, with no negative impact on English achievement outcomes (Lindholm-Leary & Howard, 2008). The present study focused on Spanish-English DLE programs and examined whether: 1) students in 50/50 vs. 90/10 models made differential gains on English and Spanish academic outcomes from fall to spring; and 2) gains varied depending on student initial proficiency in English or Spanish.

Analyses were conducted using 131 kindergarten-3rd grade students enrolled in Spanish-English DLE programs in the Southeastern US. Students were 57% female; 32% Latino, 25% White, 33% Multiethnic, and 9% Black. Distribution by grade was: K (35%), G1 (32%), G2 (18%), or G3 (15%). Home language was mostly Spanish (29%), both Spanish and English (22%), and mostly English (49%). Students were assessed in English and Spanish in the fall and spring on language and literacy, math, and expressive vocabulary (Muñoz-Sandoval et al., 2005; Schrank et al., 2014); receptive vocabulary (Martin, 2013; Martin & Brownell, 2011); and language proficiency (CTB/McGraw-Hill, 2006).

Structural equation models were conducted separately for each spring outcome assessed in either Spanish or English, controlling for student gender, grade, home language, parent education, fall language proficiency in the language in which the outcome was assessed, and initial fall scores. Predictors of interest were DLE program model (90/10 vs. 50/50) and the interaction between fall language proficiency (proficient or not) and DLE program model.

As shown in Table 1, students with low fall proficiency in Spanish made greater gains on Spanish language/literacy and Spanish math when in 90/10 classrooms compared to 50/50 classrooms (Figure 1). The same pattern was true for students with low fall proficiency in English for English receptive vocabulary only. In contrast, students with high fall proficiency made greater gains on the same outcomes in 50/50 classrooms compared to 90/10 classrooms. Thus, students with lower initial proficiency in the language of assessment did better in 90/10 classrooms, and students with higher initial language proficiency did better in 50/50 classrooms.

Consistent with past research, we found that the 90/10 model confers an advantage in Spanish oral language and literacy, but only for non-ELs. We did not replicate past findings favoring the 50/50 model for language/literacy or math outcomes in English. Overall, our findings suggest that it is important to consider how DLE program model (50/50 vs 90/10) might affect academic outcomes differentially for students with and without initial language skills in English or the partner language (e.g., Spanish).
 

Authors

• Doré R. LaForett, Child Trends - North Carolina
  Presenting Author

• Adam Winsler, George Mason University

• Ximena Franco-Jenkins, University of North Carolina at Chapel Hill

• Sabrina Zadrozny

• Ulda Gibelli Maldonado-Arias, George Mason University

• Diego Ordoñez Rojas, George Mason University