A decision model for technical journal deselection with an experiment in biomedical communications

This report proposes a paradigm for technical journal deselection based on the Bradford Law of distribution, with a demonstration in biomedical communications illustrating various statistical derivations. The model employs the following operational prescriptives: (1) identification of fractional productivities in journal collections; (2) probability predictions for making successful retrievals of relevant items in any given productive fraction; (3) utilization by journal source language and origin; and (4) estimates of cost‐efficiencies of searches made in print and machine‐readable versions of serial publications. The results suggest that in approximately two out of three trials successful searches (“hits”) are likely to be made in a “critical” core of any technical subject collection. The highest probability for retention resides in this portion, about 20% of the publications. We infer that substantial amounts of the remaining publications are candidates for deselection since their information content, as needed, may be captured more cost‐effectively through electronic retrieval and document delivery access. © 1993 John Wiley & Sons, Inc.