What Random Forests Might Do for Educational Research: Counterculture Statistical Modeling

• In Event: Poster Session 7
  In Poster Session: Division D Section 2 Poster Session

Abstract

The main objective of the session is to explain and explore the potential of a statistical modeling procedure, random forest regression (RFR) (Breiman, 2001a), that is commonly used in several fields, including health sciences, medicine, machine learning, and more recently in psychology, but not commonly used in educational research yet (Strobl, Malley, & Tutz, 2009). We will also demonstrate RFR benefits and challenges to the larger educational research community by sharing an educational application in reading research using a machine-learning context. To date, RFR is largely positioned in a counter-culture to commonly employed educational research statistical modeling, but it is a statistical technique that could innovate the repertoire of educational statistical modeling.

Authors

• Jeffrey Elmore, MetaMetrics

• Jill Fitzgerald, MetaMetrics and UNC Emerita

• Heather Hughes Koons, MetaMetrics

• Kimberly C. Bowen, MetaMetrics

• Eleanor E. Sanford-Moore, MetaMetrics

• Elfrieda H. Hiebert, TextProject

• A. Jackson Stenner, MetaMetrics