Skip Navigation

Using classification and regression tree (CART) analysis to identify students at risk of adverse outcomes: A guide for state and local data leaders

Region:
Appalachia
Abstract:

Description: An important part of education research is the identification of students who may need additional supports or interventions. Traditional statistical methods can be difficult to implement or understand, but classification and regression tree (CART) analysis is a type of decision tree that can be used to visually represent a set of rules to identify students of interest. This report will show state and local educators how CART analysis can provide insights into relevant policy questions by describing factors that educators need to consider when developing research questions that are answerable using CART and how to communicate these questions to data analysts. The report will also provide data analysts with resources to conduct CART analysis by including two relevant educational examples to explain the steps of a CART analysis along with code and synthetic data used to conduct the analyses.

Study Design: This project will use data from a publicly available synthetic data set. This student-level data includes demographic and achievement information that will be used to identify students at risk of adverse educational outcomes at the end of high school. The resulting report will have two parts: the first will introduce CART analysis and the types of questions that practitioners and policymakers could answer using it, and the second will provide detailed examples of implementing CART analysis along with supporting data and analytic code.

Projected Release Date: Summer 2021

Related Products: Applied Research Methods

Principal Investigators & Affiliation:

Dr. Neil Seftor, SRI International