American Science Needs a Scorecard That Survives an Election

American science needs a scorecard that survives an election. Pew Research Center’s January 15 report^[1] found that in 2023, during Joe Biden’s presidency, 44% of Republicans and Republican leaners said the United States was losing ground in scientific achievements relative to other countries, against 37% of Democrats and Democratic leaners. By October 2025, under Donald Trump, Democrats were the more pessimistic group, 65% to 32%. Yet majorities in both parties still called U.S. science leadership very important.

Views of U.S. science's global standing, by party, 2023 and 2025

A useful scorecard distinguishes output from influence. The National Science Foundation’s July 2025 review^[2] credits China with more than twice the U.S. output of science and engineering papers in 2023, using the Scopus database and dividing credit for international collaborations among countries. On influence, America retained an advantage: 1.7% of its 2022 journal articles reached the world’s top 1% by field- and year-adjusted citations, against 1.3% of China’s. That lead had narrowed since 2008. I see a strong research system losing its margin. School preparation presents a separate warning: the 2024 National Assessment of Educational Progress put a record 45% of 12th graders below Basic in mathematics, while performance at the 90th percentile held steady from 2019 to 2024.^[3] The weakness concerns the breadth of preparation. A weak result in school mathematics deserves a response aimed at school mathematics, not automatic enlistment in whichever argument Washington happens to be having about laboratories.

The Trump administration’s May 2025 budget request^[4] sought $3.9 billion in discretionary funding for NSF, against its $8.8 billion fiscal 2025 estimate.^[5] NSF’s stated aim was to “complement private-sector R&D”^[6] with its investments. Congress set a different number: the fiscal 2026 spending bill that passed the House on January 8^[7] and the Senate on January 15^[8] provides $8.75 billion for NSF.^[9] The argument will return with the next budget, when a defender of federal grants can anticipate lost discoveries and a supporter of cuts can anticipate a productive reallocation. Pew found that 76% of Republicans considered government research investment worthwhile,^[1] while 54% expected private investment to ensure enough progress without it. I take the reallocation case seriously. If grant procedures consume resources better spent on research, shifting work to faster-moving private funders could buy more science. Investors also have a reason to stop paying for work that leads nowhere.

Businesses already support discovery: NSF’s February 2025 tables put their share of U.S. basic-research funding in 2022 at 35%, against the federal government’s 41%.^[10] Basic research is not a government monopoly. But an existing funding share is different from a pledge to replace a particular grant. Consider an investor choosing between a marketable medicine and an engineering method whose benefits would spread among competitors. The second project can be useful and still offer too little return to its funder. I favor cutting administrative waste without pretending that every worthwhile question has a paying customer.

I accept the cost of backing research that might fail. That does not entitle every recipient to keep its grant. Before the next funding decision, I want each side to put a number and a date on its forecast. For basic research, start with the share of U.S.-authored papers reaching that top 1%, specifying the publication year and allowing time for citations to accumulate. A funding defender should also name a failure that would justify ending a program. A cutter should name who will finance the displaced work, on what terms and when. “The market” is not a funding commitment.