On the objectives of citation analysis: Problems of theory and method

Citation analysis can go beyond its present preoccupations and explore new areas if it follows several rules of research design—some of them known and generally accepted, others specific to this discipline: (a) a rigorous definition of the research objectives is essential in deciding on its design, including the selection and definition of variables and the measurement of their effects; (b) the ascertainment of content‐related variables enhances the theoretical interest and practical usefulness of citation analysis, although it may entail the use of smaller samples; (c) the set of papers to be compared with respect to citation frequency should be stratified in order to make them as similar as possible to each other; (d) the dependent variable, citation frequency, may comprise more than one kind of citation; (e) the elementary methods of analysis based on stratification (or matching) and adjustment should be complemented by model‐based methods which could accomodate larger numbers of variables and would take into account the skewness of citation count distributions. © 1992 John Wiley & Sons, Inc.