The Research Base

Our extensive exploratory work with TRANSLATE has allowed us to try out a wide range of theory and research-based instructional techniques and then use these findings to propose a viable instructional model. This model capitalizes on research described above, and provides MLs with rigorous, guided, and challenging instruction presented in meaningful contexts that engage them in strategic work and collaborative peer learning (Jiménez et al., 2015). Our research team developed these ideas with teachers and students in grades 2 through 8 over a period of more than ten years. This included 2 months with Spanish, Somali and Kurdish speaking middle school students, 6 months with Spanish-speaking middle school students; and 9 months with second-, third-, fourth-, and fifth-grade classrooms where teachers worked with Spanish, Arabic and Kurdish speaking students.

​

In the last two years, we conducted a design-based research study with middle grade teachers (4-6 grades) focused on examining how educators make use of our single model unit that attempts to integrate the four elements of the TRANSLATE curriculum described above within a 25-week instructional cycle. Currently, we are conducting a two-year evaluation study involving over 30 educators and over 700 students. 

We previously conducted a quasi-experimental study examining the impact of integrating translation into teachers’ existing curriculum. In working with 78 students (N=31 TRANSLATE, N=47 control) within six classrooms, TRANSLATE was found to significantly predict performance on a district literacy measure controlling for initial performance. The Fountas & Pinnell Benchmark Assessment System was used by the district in which we worked to assess students three times a year (Fall- prior to TRANSLATE, Winter- during TRANSLATE, & Spring-post TRANSLATE) to determine their instructional reading level. At each of these timepoints, the student’s level was conveyed by a letter, with A representing the lowest level. Because this was a pilot study focused on TRANSLATE development, we did not have power to perform multilevel modeling. With that said, ANCOVA analyses suggest that on average, TRANSLATE students performed significantly better than control students at both the Winter (p=.045) and Spring (p=.017) timepoints controlling for their fall performance. On average, TRANSLATE students were 0.65 levels in the Winter and 1.22 levels in the Spring ahead of control students controlling for their prior performance. Using the small sample size adjustment for Hedges g, these results indicate that TRANSLATE had a small to moderate effect on achievement (g=0.18) between Fall and Winter assessments and a moderate effect (g=0.34) between Fall and Spring. As Hill, Bloom, Black, and Lipsey (2008) have showed that a year’s typical growth on standardized reading tests for 4th graders is approximately 0.4, these initial pilot results indicate TRANSLATE students were able to achieve almost an additional year of growth as compared to control students. This effect size is based on teachers implementing a rudimentary version of TRANSLATE.