Abstract: We validate the research pyramid model of research evolution. Moreover, we propose and evaluate two algorithms to identify research pyramids. Finally, we improve publication scores in terms of accuracy and separability via publications’ research pyramids. Accurately ranking publications enables users to aggregate pertinent results quickly and easily. Studies show that citation-based publication-importance functions, e.g., PageRank and Citation Count, are extremely skewed and have accuracy problems. Based on the notion of research pyramids we propose a priori technique to assign more effective and accurate publication importance scores. We showed that the proposed technique provides more accurate and significantly less skewed publication scores than citation-based techniques. Our experiments showed 16-25% improvement in search outputs accuracy measured for the top-k search results.