Correlation of Faculty Score with Number of Titles and Cost

With the sets defined and the measures of ST value constructed and validated, an investigation was undertaken to map out the structure of the library market for ST serials as this market was revealed by the needs stated by the LSU faculty in the SRP survey. For this purpose, the desired universe of serials was used, which consisted of all those serials listed by the faculty on the SRP questionnaires regardless of whether these serials were on subscription at LSU Libraries. A fundamental feature of this market is the phenomenon of concentration, which arises from the skewed distributions caused by the probabilistic mechanisms underlying this market. This concentration manifests itself both in the relationships among the curriculum cores and within the curriculum cores themselves.

Concerning the former, five curriculum cores—Biology, Chemistry, Mathematics, Physics, and Zoology & Physiology—accounted for the bulk of the titles, cost, and faculty score of the serials listed by the LSU faculty in the SRP survey. Although these cores represented only 15.2% of all the cores, they contained 38.0% of the titles that accounted for 55.0% of the total costs and 40.1% of the aggregate faculty score. The five dominant cores were most intimately related to six LSU academic units—the Departments of Chemistry, Mathematics, Microbiology, Physics & Astronomy, Plant Biology, and Zoology & Physiology—that had furnished 93 (23.7%) of the 392 respondents to the SRP survey.

Under these conditions, the question naturally arose as to whether there were any imbalances between the resources requested and the faculty making these requests. For example, the Mathematics curriculum core had the greatest number of titles, with 220 (9.9%) of the 2,226 listed titles, but only 22 (5.6%) of the 392 respondents were from the Department of Mathematics; whereas the Physics curriculum core was the highest in total costs, accounting for $203,873 (15.1%) of the $1,349,350 cost of the desired universe, but there were only 25 (6.4%) respondents from the Department of Physics & Astronomy.

To test for the imbalance, the total faculty scores of the curriculum cores were correlated with their number of titles and total costs, and the answer was a resounding no—there were no major imbalances between the number and costs of the titles in the various curriculum cores on the one hand, and their value to the LSU faculty as a whole on the other. The correlation of total faculty score was 0.84 with number of titles and 0.79 with total cost. Only one outlier was found, and this concerned the core Food Science, which accounted for 1.5% of the titles but only 0.5% of the faculty score. When this outlier was excluded, the correlation of faculty score with number of titles rose to 0.86.

Two factors account for this lack of imbalance of the number and costs of the serials in the various curriculum cores with their value to the faculty as whole, despite the discrepancy with the number of faculty officially related to the cores. First, there is Garfield's law of concentration on the interrelationship among disciplines. Second, there is the heavy dependence of applied technology on basic science. All five of the dominant curriculum cores were for the most part constructed from the LC Q schedules. Moreover, the six LSU academic units to which they were most closely related were in the Colleges of Arts and Sciences and Basic Sciences—which are oriented more toward pure science—and not in the Colleges of Agriculture and Engineering —which are oriented more toward applied technology. These correlation tests stand as further proof that the management of ST serials should be based within the library on library parameters and not allocated out to the faculty on departmental bases.