Hierarchical Regression« Back to Glossary Index
Hierarchical regression examines the relation between independent or predictor variables (e.g., age, sex, disease severity) and a dependent (or outcome) variable (e.g., death, exercise capacity). Hierarchical regression differs from standard regression in that one predictor is a subcategory of another predictor. The lower level predictor is nested within the higher level predictor. For instance, in a regression predicting likelihood of withdrawal of life support in ICUs participating in an international study, city is nested within country, and ICU is nested within city.