Testing neighborhood, information seeking, and attitudes as explanations of environmental knowledge using random forest and conditional inference models

Abstract

This article tests the explanatory power and interactions among five alternative explanations of environmental knowledge: (1) local information availability, (2) neighborhood characteristics, (3) environmental attitudes, (4) personal empowerment, and (5) information seeking. Using random forest and conditional inference trees, the article analyzes survey responses and finds that attitudes about personal empowerment and frequent information seeking are the strongest predictors of knowledge.The study offers random forest and conditional inference trees as statistical tools for complex data sets and studies that test hypotheses generated from multiple theories. We discuss the influence of knowledge differences over inclusive sustainability discussions. © Copyright 2013 by Association of American Geographers.