Repairing the System: Discoveries from Research-Practice Partnership with a Diverse, Urban District

• In Event: Crossing the Decoding Threshold: Research, Remedy, and Repair for Older Students With Word Recognition Needs

Abstract

We report results from an 18-month research-practice partnership aimed at improving adolescent reading outcomes through foundational reading skills intervention. The project evolved in three phases, during which the partnership examined the systematic need for intervention, implemented and evaluated a curricular approach, and examined how that curriculum impacted a particular subgroup of students.

Phase 1: District Wide Analysis of Foundational Skills Needs
First, a district-wide foundational skills needs assessment determined that enough high school students were likely to need instruction in foundational reading skills, based on the decoding threshold identified by Wang et al (2019). The partners used the ReadBasix assessment (Sabatini et al., 2019b) to systematically identify the proportion of adolescents who might need specific support in word reading. Across grades 5-11, 1,215 students across the district took the ReadBasix assessment of Word Reading and decoding. We then calculated the percentage of students in each grade under the word reading threshold. Results are in Table 2. Because existing reading intervention efforts and infrastructure were in place for the district’s middle school, but not the high school, the partnership decided to focus on 9th grade foundational skills.

Phase 2: Within-Students Experimental Evaluation of a One-Semester 9th Grade Implementation of the Strategic Adolescent Reading Intervention
In phase 2, based on the district-wide ReadBasix screening, 55 9th graders were selected for foundational skills intervention. Based on the district’s desire for an evidence-based, low-cost, replicable curriculum, we decided to adopt the Strategic Adolescent Reading Intervention (STARI; Kim et al. 2017; Troyer, 2019). The researcher and district leaders designed a within-students experiment. Table 3 explains the within-students two-semester design, exploiting the mid-semester schedule change and ensuring group equivalence on observed and unobserved characteristics since each student was in both the treatment and control group. Students were pre- and post-tested on ReadBasix foundational skills assessments (word recognition and decoding, vocabulary, morphology, and sentence processing) as well as the NWEA MAP reading comprehension assessment. Results showed no significant differences between treatment (STARI) and control (electives) students for all 55 students assigned to STARI. However, for students who completed at least 60% of the STARI curriculum (n=36), STARI students significantly outperformed control students on reading efficiency (p=0.04, ES=0.45), but not the other measures.

Phase 3: Correlating ReadBasix and AAVE
The Reading Reimagined research initiative also focuses on improving reading achievement for students of color. During the STARI implementation, the researcher noticed that the students’ African American English (AAE) appeared to shape the instruction. The research team, guided by a codebook from Washington and Craig (2012) then coded the density of students’ written AAE from a short informal writing sample (See Figure 4 for an example). Density of AAE ranged between 0%-18% of written words.
Then, density of written AAE was correlated with the 6 assessments (five ReadBasix plus MAP). Results are in Table 4, showing that all six assessments are negatively correlated with written AAE density, four strongly so (p<0.01). Students may be paying a particular tax for their use of AAE.

Author

• Dan Reynolds, SRI International