An examination of the types of publishers involved in these trade-offs corroborated in general the findings of Bensman (1996). For purposes of this examination, publishers were first divided into U.S. and foreign. Then these two sets were each further divided into the following subsets: commercial, association, university press, and miscellaneous nonprofit (which contained academic departments, institutes, museums, etc., acting as publishers). It should be emphasized that serials published under association auspices but produced and marketed by commercial publishers were defined as commercial.

Table 6 presents the findings on the types of publishers involved in the trade-off in costs versus ST value on a global basis by aggregating the data from all the curriculum cores. In this table, the dominance of the U.S. association, U.S. commercial, and foreign commercial publishers at the upper ends of the cost and ST value distributions is visible. These three types of publishers produced 92.0% of the serials that accounted for 75.0% of total costs per curriculum core and 87.4% of the serials that accounted for 75.0% of total faculty score per core. However, their shares in these two sets were remarkably different.

On the one hand, of the titles that accounted for 75.0% of total costs in the different curriculum cores, the U.S. associations published only 11.1%, whereas the U.S. commercial publishers and foreign commercial publishers respectively produced 14.5% and 66.4%, which together totaled 80.9% of these titles. On the other hand, of the serials that accounted for 75.0% of total faculty score in the various curriculum cores, the U.S. association serials represented 32.1% of the titles, whereas the U.S. commercial publishers and foreign commercial publishers were respectively responsible for 16.3% and 39.0%, which added up to 55.3% of these titles.

Thus, in a comparison of total cost to total faculty score, the share of the U.S. associations rose from 11.1% to 32.1%, and the proportion of the commercial publishers dropped from 80.9% to 55.3%, still a respectable amount that demonstrates the need to take the output of the commercial publishers seriously into consideration. It should be noted that in a small way, the performance of the U.S. associations is mirrored by the foreign associations, whose share rose from 3.5% of the titles accounting for 75% of total costs to 6.9% of the titles accounting for total faculty score.

The basic dichotomy in the library market for ST serials is emphasized by the Evaluator recommendations for cancellation or nonpurchase. Here the foreign commercial publishers alone produced 70.6% of the titles recommended for such treatment, dwarfing the shares of all other types of publishers. Together, the U.S. and foreign commercial publishers were responsible for 84.5% of the titles recommended by the Evaluator for cancellation or nonpurchase.

The same patterns emerge in table 7, which summarizes the data by curriculum core. Once again the dominant role of the U.S. associations, U.S. commercial publishers, and foreign commercial publishers at the upper ends of the cost and ST value distributions is visible. The serials of these publishers not only were represented in a consistently higher percentage of curriculum cores but also accounted on the average for a much higher percentage of the serials responsible for 75% of the total cost and faculty score in the curriculum cores in which they were represented than those of the other publishers.

However, what is particularly striking in table 7 is the divergence of the serials of the U.S. associations and foreign commercial publishers in terms of costs and ST value. Thus, in terms of the serials accounting for 75% of the costs, U.S. association serials were represented in 69.7% of the curriculum cores, being on the average 17.5% of these titles in their respective cores, whereas foreign commercial publishers were present in 100.0% of the curriculum cores and produced on the average 64.9% of these titles in the cores. When the serials responsible for 75.0% of total faculty score were considered, U.S. associations were represented in 100.0% of the curriculum cores and accounted for on the average 36.9% of these serials in the cores, whereas foreign commercial publishers had serials in 97.0% of the curriculum cores and produced on the average 37.0% of these titles in the cores. As is evident, the average share of the U.S. associations rose dramatically in the transition from cost to faculty score, whereas precisely the opposite happened with respect to the foreign commercial publishers.

This dichotomy between cost and ST value was emphasized in the cancellation or nonpurchase recommendations of the Evaluator for each curriculum core. Whereas the titles of the U.S. associations averaged 18.0% of the cancellation or nonpurchase recommendations in 42.4% of the curriculum cores, those of the foreign commercial publishers averaged a stunning 71.3% of these recommendations in 100.0% of the curriculum cores. However, it once again must be emphasized that the foreign commercial publishers produce enough serials of sufficient ST quality to merit serious consideration.

The basic divergence between cost and value in the library market for ST serials became even more pronounced when the anomaly presented by the Computer Science and Electrical Engineering curriculum cores was taken into account. In both of these cores, U.S. association titles predominated in the Evaluator recommendations for cancellation or nonpurchase, and this type of title accounted for 50.0% of such recommendations in the former and 46.2% in the latter. Altogether these two curriculum cores contained 16 (48.5%) of the 33 U.S. association publications recommended by the Evaluator for cancellation or nonpurchase.

The two departments at the basis of these cores, Computer Science and Electrical and Computer Engineering, were closely linked in the SRP survey, and as a result of the high overlap in the serials selections on the part of the faculty of these two departments, it was decided to make the serials classed in the computer engineering part of the LC schedules (TK7885–7895) part of the Computer Science curriculum core to compensate for the low response rate of the Department of Computer Science to the SRP survey. A closer analysis of the Evaluator recommendations for cancellation or nonpurchase revealed that the anomaly was caused by the serials published by the Institute of Electrical and Electronics Engineers (IEEE).

The large and complicated bibliographic structure of the IEEE publications had a large role in the failure of perception that had led to the zero class in faculty score, and this structure had probably also led to the systematic underscoring of these serials. However, the ST value of the IEEE publications was demonstrated by the concentration of NOTIS use upon them. The importance of the IEEE serials manifested itself in the Evaluator cancellation or nonpurchase recommendations, because the divergence of cost and ST value was more pronounced for the titles of the commercial publishers than for these IEEE titles. Thus, in Computer Science the trade-offs in terms of budgetary reduction for ST value loss in terms of faculty score were 5.8 to 1 for U.S. commercial serials and 4.5 to 1 for foreign commercial ones, whereas for IEEE titles it was only 2.8 to 1. This phenomenon was repeated in Electrical Engineering. Here the trade-offs of budgetary reduction for ST value loss in terms of faculty score were 4.6 to 1 for U.S. commercial publications and 9.7 to 1 for foreign commercial titles, whereas for IEEE serials it was merely 1.9 to 1.

With the exclusion of the Computer Science and Electrical Engineering cores, the share of U.S. association serials in the total number of Evaluator cancellation or nonpurchase recommendations drops from 33 (8.0%) of 411 to 17 (4.3%) of 395. As for the average number of titles per curriculum core in which such recommendations were made, it fell from 2.4 to 1.4 in contrast to the that of the foreign commercial publishers, which actually rose from 8.8 per core to 9.1. With respect to the latter, cancellation or nonpurchase recommendations of U.S. association titles were once again made in less than half of the curriculums cores (12) than such recommendations were made for foreign commercial ones (31).