Effects of an elementary Two Way Bilingual Spanish-English Immersion school program on junior high and high school student achievement
Author(s)Vega, Luis Diego
Full recordShow full item record
This study explores the effects of a Two-Way Bilingual Immersion (TWBI) program on language majority and minority students. The fundamental hypothesis was that the process of receiving instruction in two languages (English and Spanish) throughout elementary school (i.e., attendance at a TWBI school) would help the native Spanish-speaking students and not have a negative effect on the native English-speaking students in the performance of core academic areas (reading, mathematics, writing), and that this beneficial effect would carry through Junior High and High School in which instruction was delivered through a "business as usual" English-only model. This is a longitudinal quasi-experimental study with an ex post facto, non-randomized, matched-pairs design. A multi-level matching procedure was used to match students from the TWBI elementary school (treatment group) with comparable students from throughout the school district (control group) beginning in third grade. Eleven annual cohorts of students from the treatment school were matched on a student-by-student basis on seven variables - cohort year, student's primary language, years of enrollment in the program, ethnicity, gender, socioeconomic status, and 3rd grade performance - with comparable students from within the school district. These eleven cohorts of 3rd graders were then tracked to the end of elementary school, middle and high school and measured on their reading, writing, and math achievement scores at each year. ACT scores were also collected in 11th grade. We found that students who graduated from the TWBI program had significantly higher CSAP reading, writing and math scores at the end of their elementary school when compared with their matched pairs. We also observed a consistent main effect on program type over time across all three outcome domains, indicating the strength and breadth of the intervention across Junior High and High School. Native Spanish-speaking students who graduated from the TWBI program achieved significantly better in reading and math, and somewhat better in writing across Junior High through 10th grade than the matched control group. Native English-speaking students who graduated from the treatment program achieved as well as their matched counter parts in writing and math across Junior High through 10th. Furthermore, in the reading area, native English-speaking students who graduated from the treatment program achieved significantly better than their matched controls. We found that the overall program main effect was small in all three CSAP areas (reading, writing, and math), with at least three interesting trends. First, effect sizes (ESs) tended to be higher for native Spanish-speaking than for native English-speaking students in all three domains, and especially in grades 8, 9 and 10. Second, ESs tended to get bigger for native Spanish-speaking students and smaller for native English-speaking students across Junior High and High School (time) in all three domains. Third, ESs for native Spanish-speaking students in math were the biggest ones at each grade level, with only the exception of 9th grade. Also, math ESs for Spanish-speaking students were bigger than reading and writing ESs for this language group. ESs for native Spanish-speaking students in math were bigger than all ESs for English-speaking students. The treatment program had its biggest effect in the math area for native Spanish-speaking students. Results also indicate that all students who attended the TWBI program performed better in ACT English, reading, and math scores when compared with their matched pairs. ACT Reading scores were significantly higher for native Spanish-speaking students than for their matched pairs (d = .72), but this was not the case for English, math and science. Native English-speaking students from the treatment group performed equal to or better than their matched counterparts. Furthermore, students from the treatment program obtained mean ACT scores significantly higher than the control group in English (d = .28), reading (d = .36), and math (d = .35) but not science (d = .22). Effect sizes were medium and large for native Spanish-speaking students in English and Reading while they were small to medium for native English-speaking students in these areas, a pattern that is similar to the one that was observed in grades 6 to 10. Findings suggest consistent support for the two-way immersion program over matched control students across all three achievement areas in Junior High and in three of the four areas evaluated in High School. It appears the greatest effect for native English speakers may be in reading, while native Spanish speakers may benefit more in writing and mathematics. Limitations to generalizability and causal inferences due to the small sample sizes and inherent weaknesses of the research design are noted. The analysis of attrition revealed that native Spanish speakers from the TWBI program were more likely to stay in the school district than native Spanish speakers from other programs. This was an unexpected but important finding. It could be possible that native Spanish speakers who attended the TWBI program received the benefits of a coherent and theory-based program that successfully helped them improve their academic achievement and allowed them to pursue and navigate their secondary level of instruction.