Ten Difference Score Myths

Abstract

Difference scores have been widely used in studies of fit, similarity, and agreement. Despite their widespread use, difference scores suffer from numerous methodological problems. These problems can be mitigated or avoided with polynomial regression analysis, and this method has become increasingly prevalent during the past decade. Unfortunately, a number of potentially damaging myths have begun to spread regarding the drawbacks of difference scores and the advantages of polynomial regression. If these myths go unchecked, difference scores and the problems they create are likely to persist in studies of fit, similarity, and agreement. This article reviews 10 difference score myths and attempts to dispel these myths, focusing on studies conducted since polynomial regression was formally introduced as an alternative to difference scores.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 685 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.