To: House Committee on Education
From: Dave Porter
Subject: Question about the “State Longitudinal Data System”
Why have a “longitudinal data system” if longitudinal data studies are not going to be used to set state policies and programs? Why spend the money if it is not going to be used?
The 2012-15 “Study of Dual-Language Immersion in the Portland Public Schools” by RAND, the American Councils for International Education, and Portland Public Schools found that “Immersion programs as implemented in Portland appear to be a cost-effective strategy for raising English reading performance of both native English speakers and native speakers of other languages.”
The Rand study is a longitudinal data analysis. It has had zero impact on state policies and programs.
The study included 27,741 students who enrolled in kindergarten in Portland Public Schools in 2004-05 through 2010-11. These students’ academic performance on the Oregon Assessment of Knowledge and Skills (OAKS) was tracked in reading, mathematics, and science through 2013-14, meaning that the youngest cohort was tracked through grade 3, and the oldest two cohorts through grade 8.
Within the sample, the main focus was on 1,625 students who were randomized to immersion or a control group via Portland’s immersion pre-K and Kindergarten immersion lotteries in 2004-05 through 2010-11. Through collaboration with the Oregon Department of Education, the study was able to include 200 randomized students (about 10.3%) who never enrolled in Portland after entering a pre-K or kindergarten immersion lottery. This left a similarly modest levels of sample attrition for students who won the lottery (13%) versus those who did not (19.3%).
The study found that students randomly assigned to immersion outperformed their peers in English reading by about 7 months in grade 5, and about 9 months in grade 8.
Further, DLI program costs about the same as non-DLI programs. The study found no evidence that observable peer, teacher, and class size characteristics were driving the effects of immersion on achievement. Class sizes appeared similar for immersion lottery winners and those who did not win immersion slots. Based on interviews in 2013-14 with 14 of 19 immersion school principals, they found that school-level resources for DLI and non-DLI Programs were proportional to immersion enrollments. DLI operating costs were concentrated at the district level. In 2013-14, they represented about 0.1% of the district’s operating budget (excluding grant dollars), though this represented a considerable increase in DLI costs from prior years. These expenditures, which were applied toward teacher professional development and curriculum support, amounted to $137 per immersion student in 2013-14 (n=4,108).
More on the study can be found at http://www.pps.net/Domain/85.