Cognitive Resemblance and Citation Relations in Chemical Engineering Publications.

In this study we have measured word‐profile similarities between citing and cited publications, as well as between publications citing specific highly cited papers. This “cognitive resemblance” was operationalized by different similarity measures using various kinds of terms and classification types. This study focuses on publications of internationally recognized chemical engineering scientists for the year 1982 as “source” publications, and subsequent publications (of other scientists) citing to these pubications. This study empirically shows that publications with a citation relationship are significantly more content‐related than other publications. It also shows that highly cited documents are mainly cited within their own research area. Thus, at least in chemical engineering, publications sharing citations to the same highly cited article, represent work of the same subject‐matter research area. This is certainly not caused by the “narrowness” of the field, as we also show that there is a clear distribution of publications over many (sub)fields so that chemical engineering can be characterized as a broad, interdisciplinary research field. A weak relationship between word‐profiles and type of classification was found, and this relationship differs between various types of classification. Mapping based on correspondence analysis clearly visualizes content‐related groups of citing and cited publications. Our findings are contrary to the results of some earlier studies and to opinions in circles of sociologists of science that authors refer to publications in a rather arbitrary way mainly for adornment of their claims. These differences can be explained simply with statistical arguments. © 1995 John Wiley & Sons, Inc.