It's not based on titles. Here [1] is the proposal. It included "Furthermore, by employing and mentoring students from underrepresented backgrounds in STEM, this project will aim at bridging the gap in institutions across the US. It will train the next generation of scholars from minority serving universities and marginalized communities in the fields of cybersecurity, utilization of renewable resources, and machine learning to address the pressing problems of this age."
Previously DEI adherence weighed into which proposals were awarded funding. They no longer do. It's unclear exactly how this is working but suspect they're flagging grants where either the entire point was DEI, or where the project was unrelated to DEI but the DEI stuff pushed it into the acceptable range (and/or drove the grant amount higher than necessary for the underlying science), and cancelling them.
Interesting, thank you for the addendum! Do you believe it is likely that inclusions of things such as "by employing and mentoring students from underrepresented backgrounds in STEM, this project will aim at bridging the gap in institutions across the US" result in a higher likeliness of funding? I also wonder if Hacker News would generally consider it to be ethical to use this to increase the likeliness of funding. In this case it does seem unrelated to DEI otherwise.
Definitely. The previous administration used executive power to direct various government organizations to factor DEI into all government funding, and this is also reflected in the last line of the proposal's abstract (extremely odd place for an administrative comment but that's where it is): "This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria."
If there is a point scheme being used behind the scenes, as seems reasonably probable, then selecting all grants to be eliminated was probably not much more than a single SQL query.
> The previous administration used executive power to direct various government organizations to factor DEI into all government funding,
Not the NSF. Provisions in NSF's organic statute to create programs that "expand STEM opportunities" were introduced by the American Innovation and Competitiveness Act, and were retained through the CHIPS Act.
> If there is a point scheme being used behind the scenes, as seems reasonably probable, then selecting all grants to be eliminated was probably not much more than a single SQL query.
No. The NSF review process does not use numeric ratings. Panels of peer reviewers get a tranche of proposals, provide comments individually, and then collectively sort them into competitiveness categories. There is no "DEI score" or "DEI component".
The NSF falls under Federal executive authority which is exactly why Trump is able to do what he's doing. Biden did the equal but opposite thing with his very first executive order. It effectively required all branches of the Federal government to institute DEI policies and policies aimed at furthering DEI ends. [1] In fact his executive order specifically worked to undo a previous Trump executive order [2] which forbade Federal agencies from discriminating against/for individuals/groups based on their race or sex.
To put a number to this, by the metric this report (from this topic) was using to measure DEI funding, 0.29% of NSF grants were for DEI stuff in 2021. By 2024, it was up to 27%! [3] Apologies for the excess citations here, but I think it's important on such a charged topic.
Then Trump's meddling with the NSF is executive authority without oversight – he cannot undo Biden-era legislative acts, which affect the NSF regardless of what Biden ordered elsewhere. Trump does not have complete authority over what Federal agencies do, despite legislation, as a matter of constitutional law – or, rather, JD Vance would like to argue that's the case within the doctrine of unitary executive theory, as well as the Heritage Foundation... we shall see, and the barrage of executive orders here is likely to give us a test case in the Supreme Court
And, of course, we've argued elsewhere that the report's methdology and results are... not a good assumption to begin this conversation on.
DEI stuff will never hold up under legal scrutiny - it runs face first into the Equal Protection Clause in the Constitution, and orders will also contradict the Civil Rights Act. Again, this is how race based admissions in universities were deemed unlawful.
The legislation you previously linked is solely mentioning a series of grants specifically aimed at increasing diversity, with specially allocated funding totaling $23 million. That's not only going to be separate from the NSF budget, but would be ~0.25% of their budget if not. It's unclear of such things would hold up under the Equal Protection Clause, but it's largely irrelevant one way or the other.
A practical but immeasurable issue is the scale of impact. Encouraging diversity is good, impactful and systematic discrimination is not. There's a not entirely well defined line between the two, but I think a program at this scale would have few claim it's the latter.
I downloaded the linked excel file and this is the description of the first project link I clicked on:
> "COLLABORATIVE RESEARCH: CISE-MSI: DP: CPS: CYBER RESILIENT 5G ENABLED VIRTUAL POWER SYSTEM FOR GROWING POWER DEMAND"
I feel like they just go by keywords like covid, 5g, gender, women, climate change etc....