Skip Navigation
archived information
REL Appalachia

[Return to Blogs]

A Life Raft in a Sea of Data

April 22, 2019

SRI International
   Stephanie Nunn, REL Appalachia
   Todd Grindal, REL Appalachia
Goochland County Public Schools
   Sean Campbell

School leaders can feel as if they are drowning in data. On the first day a child registers for school, we begin recording information about their backgrounds to support participation in programs and services, such as those that support learning for students with disabilities, help children receive free or reduced priced lunch, or gifted education services. Over time, schools collect data about student attendance, which classes they take, and how they perform on dozens of assessments of academic achievement, all with the intention of using these data to make informed decisions throughout their education system. The promise of data-driven decisionmaking has motivated state departments of education and large school districts to direct substantial resources to support the collection, analysis, and interpretation of relevant data. Making effective use of these data is often more challenging in smaller school districts, which have fewer staff to dedicate to examining student data and too few students to support some “big data” analyses.

For districts large and small, classification and regression tree (CART) analyses may provide a life raft in this sea of data. Long used in public health, CART analyses require fewer assumptions than many commonly used statistical approaches and can be used with smaller samples. The results of CART analyses are presented as a series of straightforward “if-then” statements that can help teachers and school leaders best support their students. In the sample in figure 1, test scores included in the model are represented in the blue ovals. Based on those scores, students are categorized as at risk or not at risk for an outcome, for example, failing an end-of-grade assessment. In this example, students who score above a 66 on test 1 are not at risk. If they score at or below 66, AND score higher than 250 on test 2, then they are also not at risk. However, students who score at or below 66 on test 1 and at or below 250 on test 2 are identified as at risk.

Diagram depicting sample classification and regression tree (CART) analysis used to help teachers & school leaders best support students.

Figure 1. Sample CART analysis to help teachers and school leaders best support their students.

REL Appalachia (REL AP) staff members recently collaborated with leaders in Goochland County Public Schools (GCPS) to leverage the CART method to answer critical questions about student learning. Located about 40 miles west of Richmond City, Virginia, GCPS serves a diverse group of approximately 2,500 students. The school leaders had been looking for ways to improve supports for young students who struggle with reading. Specifically, they wanted to identify early the students with a high probability of struggling with reading and provide them with targeted supports. GCPS also had a wealth of data available to learn more about patterns of student difficulties.

In partnership with REL AP researchers, Goochland staff used CART analysis to identify the early literacy skills that predict whether a GCPS student would earn a score of proficient or higher on the state test of reading skills at grade 3. The results provided new insights into linkages between early reading skills and students' later achievement, particularly the skills that may be early indicators that students are at risk of struggling with reading before grade 3.

The next task is to turn these insights into action. GCPS instructional leaders and REL AP staff are developing professional development materials on critical early literacy skills that are linked with grade 3 reading outcomes as identified through the CART analysis. The professional development will provide teachers with knowledge about key strategies for struggling readers and ongoing coaching to support their implementation of those strategies in the classroom.

Resources for ongoing learning

Interested in learning more about effective early literacy practices and CART analysis? Check out these resources.