***** To join INSNA, visit http://www.insna.org ***** Normalizing the measurement of citation performance: Principles for comparing sets of documents Authors: Loet Leydesdorff, Lutz Bornmann, Rüdiger Mutz, Tobias Opthof (Submitted on 20 Jan 2011) Abstract: Using citation analysis, sets of documents can be compared as independent samples; for example, in terms of average citation counts using potentially different reference sets. From this perspective, the size of samples matters only for the identification of significant differences and estimating margins of error. Using the percentile rank approach, differences among citation distributions can be studied non-parametrically and in a single scheme. Comparison among the sets clarifies that the different sizes of samples affect the weighing of the probabilities and therefore the rankings. We distinguish among (1) the normalization of papers against external reference sets, (2) normalization in terms of frequencies relative to the margin-totals of independent versus dependent samples, and (3) the potentially normative definition of percentile rank classes for the evaluation (e.g., top-1% most highly cited, median, etc.). When the sets to be evaluated are considered as subsamples of a single sample, the consequent citation indicator can be negatively correlated to citation indicators used hitherto. Preprint version available at http://arxiv.org/abs/1101.3863 ** apologies for cross-postings ________________________________________ Loet Leydesdorff Professor, University of Amsterdam Amsterdam School of Communications Research (ASCoR) Kloveniersburgwal 48, 1012 CX Amsterdam. Tel. +31-20-525 6598; fax: +31-842239111 [log in to unmask] ; http://www.leydesdorff.net/ Visiting Professor, ISTIC, Beijing; Honorary Fellow, SPRU, University of Sussex _____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.