OSTP Can’t Finish Its Homework
Like bright but scattered schoolkids, OSTP produces interesting reading, but doesn't do the assignment
Within weeks of its introduction in 2022, the OSTP public access policy guidance was challenged by the House Science, Space and Technology Committee. The pressure remained percolating behind the scenes, resurfacing about a year later, and then again in March, when OSTP was given 100 days via a stopgap spending bill to provide relevant committees in Congress with a report on its public access policy’s financial costs and related impacts.
This week, well after 100 days had elapsed (making the homework late, as well), the OSTP released its report to Congress on financing mechanisms for OA publishing of federally funded research.
It’s a letdown in a number of ways, and surprising in other ways.
In short, it’s the kind of homework product you might have turned in by a bright student who didn’t stick to the assignment, who didn’t want to answer the questions, and who spent a lot of time exploring tangents and hoping diversionary tactics work to generate a passing grade.
Buried about halfway through the 30-page report, we come across the admission:
OSTP’s November 2023 Report to the Committees outlined key limitations to calculating OA fees in Sections 4 and 5 and the additional data needs for more accurate and detailed analyses. In the months between November 2023 and the development of this current report, these data needs have not been met.
In fact, there is a huge piece of information missing from the ensuing pages — the number of research publications generated by the US government research and development budgets writ large, and how a complete adoption of OA via APCs could impact the overall budget.
Instead, we get Table 2 in the report, which tallies up the projected amount spent on APCs from 2016-2022, but without sufficient detail about how many articles were included, what percentage of government research outputs this represents, what APC rates they used specifically, and how the calculations were made.
Despite praising data sharing in another section, they do not show their work.