Shared resources and NSF funding strategy

Eos has an article about NSF funding strategies for ocean sciences: “A Transformational Path Forward for the Ocean Sciences Community”. Ocean research requires some expensive investments in infrastructure projects, including a fleet of research ships, offshore drilling projects, and research centers. The costs of maintaining this vast infrastructure have increased during the last 10 years, so that more NSF funding is going to support infrastructure than basic research grants to institutions and researchers. In broad terms, NSF is helping to maintain a fleet of research vessels but is thereby unable to fund research grants to put those ships to their best use. So a new funding strategy would reduce the infrastructure and reallocate funding into research by investigators.

The investment in ocean sciences must balance the interests of independent research with shared facilities. From the article:

The second major revelation is the recommendation to immediately reduce infrastructure expenses by 10% and implement a further 10%–20% reduction within 5 years, with the savings invested back in the core science programs. This rebalancing will eventually return the portfolio to a ratio nearing 60% science and 40% facilities.

I’m linking to this article because it made me think about how NSF has approached paleoanthropology over the last 20 years. We do not have the kind of shared infrastructure that ocean sciences require, but we do have a great need for large projects to generate shared datasets that can be used by other scientists. In my opinion, NSF should allocate its funds in ways that will create the greatest net benefit for the science. It is obvious that the science goes more slowly when funds are invested in peripheral subjects that do not generate data useful to the broader scientific community.

Independent research projects during the last 20 years have been required to outline strategies to share their data, but I have never seen any assessment of the effects of those requirements or rates of compliance by investigators. I have been browsing through funded NSF grants over the last 20 years, and have found a number referencing the creation of datasets or databases that do not presently exist in any web-accessible form. I found one large grant to build a web resource that I cannot presently find online.

Meanwhile, NSF has invested a large amount of money in a small number of graduate training programs devoted to paleoanthropology during the last 20 years, but I cannot find any assessment of the impact of those programs upon the field—such as success of underrepresented minorities in faculty positions, or representatation of the graduates of such programs in fieldwork projects. NSF funds these programs as a way of prompting institutions to change their administrative structure in a question-focused way (e.g., the interdisciplinary focus of the IGERT program), and so they bring money from outside the traditional Biological Anthropology and Archaeology funding programs. Seems like a good thing, at least on the surface. But how exactly do such programs compare to traditional graduate programs, when controlled for the funding they provide? Should we be redirecting the effort of graduate students away from mainstream questions—especially the I suspect that providing the same amount of money across paleoanthropology to support student fieldwork would have a massively greater effect on the science, and would spread funding to a much larger cross-section of early researchers.