Clinical studies to evaluate the wear of posterior composite restorations are complicated by the large number of factors which influence the findings. A multi-factorial equation has been developed which successfully normalizes the effects of these factors within studies. This equation is not capable of normalizing these effects for other investigations reported in the literature because study characteristics essential to the analysis are rarely reported. Quantitative estimates of wear rate often differ dramatically between studies and between groups of investigators. The objective of this study was to investigate a potential solution to this problem by ranking materials within studies, using common materials to relate rankings across studies, and achieving an overall pooled ranking for products. An intensive literature search disclosed 78 articles and 46 abstracts on clinical posterior composite wear. All studies were analyzed which involved: (1) more than five restorations per material, (2) Class I or II restorations in posterior adult teeth, (3) characterized commercial products, (4) 2 or 3-year wear data, and (5) information on more than one composite material per study. There were 10 2-year studies involving 25 materials, and 10 3-year studies with 26 materials. Within each study the materials were ranked by wear, and the rankings were converted to centered modified ridits. A meta-analysis combining the data across studies was conducted using ANOVA. Although caution is needed for interpreting significance levels because of the small numbers of products evaluated per study, a high level of agreement in rank correlation of 28 products across qualifying studies was observed.(ABSTRACT TRUNCATED AT 250 WORDS)