Assessing the Impact of Washington State's Transitional Bilingual Instructional Program

• In Event: 34.054 - Washington State: Investigating Outcomes and Experiences of Bilingual and English Learner Students and Programs

Abstract

Purpose
The objective of this study is to analyze the effectiveness of Washington State’s Transitional Bilingual Instructional Program (TBIP), the specialized service offered to English learner (EL) students to assist in developing their language proficiency and enable access to grade-level academic content. This paper seeks to explore the effectiveness of TBIP and the validity of cutoff scores used to determine eligibility for such services.
Perspective
Given that eligibility for TBIP services is annually determined by an English proficiency test offered by the state, the main analytical framework used here is the one introduced in Robinson (2011). That is, I will be focusing on whether current score thresholds used for determining program eligibility are validly set such that there is a smooth transition (in academic achievement) across the cut points.
Methods
I use the standard regression discontinuity design (RDD) to analyze Washington State TBIP data (Lee & Lemieux, 2010). The RDD essentially compares achievement of students distributed around the current score threshold. These students are arguably similar in all ways except for TBIP program assignment.
Data Sources
I utilize student-level information from Washington State. I compare students who score a “Level 3” (students eligible for TBIP) to those who score a “Level 4” (students no longer eligible for TBIP) on the English proficiency test. Table 1 provides descriptive statistics for these two groups. It should be noted that I do not intend to use all of these students in the analysis, only those who score within an optimally determined bandwidth of the threshold (Calonico, Cattaneo, & Titiunik, 2014).
Results
I present preliminary findings from the analysis in Table 1 and Figure 6. Descriptively, it appears that the cut point determining eligibility into TBIP is validly set. This holds for both students enrolled in bilingual and immersion programs. The null effects appear to indicate that EL students are smoothly transitioned out of their programs into the mainstream classroom as there is no clear “jump” in test scores following reclassification.
However, an interesting finding is when we look two years out from reclassification. While there appears to be no effects of entering the mainstream classroom for students initially enrolled in English immersion programs, those who were in bilingual education classrooms appear to show an improvement in math test scores two years after reclassification. This might suggest that, if long-run outcomes were to be taken into account, EL students in bilingual education would perform better had they been reclassified earlier in their program. However, this graph only provides descriptions of trends, but offers no test of statistical significance. Future analysis to be completed in the next several months will include a RDD approach to test hypotheses of differences between the two groups.
Significance
Preliminary findings from this study indicate that setting cut points for program eligibility might want to take into account both the initial instructional program and long-run outcomes of the EL.

Author

• Alec Kennedy, University of Washington