“We recently published the results of a ... study of acknowledgment and coauthorship trends in 20th century psychology and philosophy [1]. The present paper describes the findings of a follow-on survey of chemistry. The methods used here are almost identical to those used in our earlier work.”

Whereas the authors’ earlier study examined an entire century’s worth of papers, this study examined a random sample, generated using a technique described in Appendix A, that included roughly 2.6 percent of the more than 110,000 research papers published in the Journal of the American Chemical Society (JACS) over the course of the last century. The size of this sample was somewhat larger than that of the complete data sets studied in the case of the psychology and philosophy journals in the authors’ earlier paper [1].

The findings of this study are that 88 percent of all papers in the sample were coauthored, ranging from 44 percent in the first decade of the twentieth century to 99 percent in the last. The difference between the ratio of single-authored to coauthored papers in the second half of the century and the same ratio for the first half of the century was statistically significant.

Acknowledgments were present in 75 percent of the papers in the sample, ranging from five percent in the first decade to 91 percent in the last. Each acknowledgment was classified by type, with many being of compound type. The types considered were financial (46 percent), instrumental or technical (34 percent), conceptual (18 percent), and editorial, moral, or unknown (all together two percent). As with the authors’ earlier paper [1], because the classification of acknowledgments into types is more of an art than a science, there was discussion on cases that were scored differently by the three authors, and a 12 percent random sample of the final classification demonstrated 90 percent intercoder reliability.

As I noted in my review of the earlier paper [1], “the authors make many fascinating asides.” The empirical research task here was harder because there were more papers, many more acknowledgments, and it was necessary to devise and apply a technique for generating a random sample of over 110,000 research papers. Moreover, its performance appears diligent and technically flawless. Nevertheless, I cannot conclude, as I did with that paper, that “this is a small, but good contribution to intellectual history.”

Here’s why: the empirical survey in the earlier paper [1] produced data that was of some interest, because it was neither obvious nor well known; it was, in short, a contribution to knowledge. The findings of this study, if not the asides, would be well known to anyone who is a regular reader of, say, the Chronicle of Higher Education or the Times Higher Education Supplement. More particularly, the fact that hard science has become “big science”--produced by laboratories that have a multitude of instruments and technical methods, and financed by industry and government agencies that alone have the funds to run such large enterprises, rather than by individuals working with pen and paper and a few beakers of “stuff”--is known even to people who know next to nothing about modern chemistry. All this study does (and does well) is to make explicit and precise just those facts. The precision might interest the editors of JACS, but for the rest of us, there is nothing genuinely new here. Although what the authors have published here does not stand on its own as a contribution to knowledge, it would have been very fine as part of a longer previous paper [1], and, indeed, the comparison of chemistry, psychology, and philosophy, especially Figure 4, is certainly interesting.

It is easy to criticize, and hard to undertake, a labor-intensive study of this magnitude, so let me conclude with a constructive suggestion. A still more detailed picture, using data the authors already have, would make “a small, but good, contribution to intellectual history,” and would stand on its own. Rather than a total of how many papers are coauthored or single-authored, the average and median number of authors per paper in each decade would be of considerable interest, as would the number of coauthored papers in the early decades with more than two and more than three authors, and the number of coauthored papers with only two coauthors in the final two decades. Likewise, rather than totaling up the types of acknowledgments across papers, the number of separate funding, instrumental/technical, or conceptual thanks per paper (of those that have any, excluding the zeros from the averages and medians) could be tracked over the decades. This information would reveal something about the structure and complexity of chemistry over the twentieth century, beyond what is generally known. The authors have already collected this body of data. Let them analyze it further and publish it.