Prescott Valley, AZ Correspondent — This stopped being hypothetical four weeks ago and the timing matters.

Draft rules went to the Federal Register last month that would give political appointees authority over how research grants are awarded and the power to terminate active projects. That is not a replication standard. That is the political test the question warns about, arriving before the method arrived.

I still want the method. Preregistration where it fits, data access, and a portfolio score for whether anything in the last five years ever failed to replicate.

But I have to answer the question honestly given what’s on the table. Viewpoint neutrality written as a ban on political scoring is method. Viewpoint neutrality implemented as appointee control over individual awards is patronage, and the second is what’s being drafted.

Write the standard so a flipped administration has to lie to turn it into patronage. What’s in the Register does not require anybody to lie.

Novi, MI Correspondent — Quality systems either measure a part or they measure a vibe.

Replication is measuring the part: can a hostile lab follow it, does the claim survive.

Appointee authority over individual awards is not measuring anything. It’s a purchasing decision made by somebody with a political interest in the result, and in manufacturing we’d call that a conflict and remove the person from the panel.

I’ll support a replication score and a ban on political criteria. I’ll oppose a structure where the award depends on who is in the building, because that’s the disease under a different coat.

Clermont, FL Correspondent — Apply both standards, and write them so an ordinary administrator can tell method from politics.

The safeguard I named a year ago was a published rubric and inspector general review. I’d say now that the safeguard has to be in the rule rather than alongside it.

If the final version gives appointees discretion over individual awards without a published standard and a written justification, then the answer to this question is the second clause. It becomes a political test, and the record will show it did.

Wheeling, WV Correspondent — Yes, if the rule is short enough for people outside the agency to understand.

Can the data be checked. Can the claim survive another competent lab. Was political agreement part of the score. That’s enough to start.

I have no interest in replacing one loyalty test with another, and a teaching hospital in this state should not need the fashionable paragraph of the year or next year’s unfashionable one.

What’s drafted now looks like it could produce the second kind. I’d rather my own side got that right than won it fast.

Orange County, CA Correspondent — A neutrality rule becomes a political test if it’s implemented as a culture memorandum. It stays method if it’s a checklist a hostile reviewer can use.

The distinction is entirely about who holds discretion. A rubric applied by career reviewers under published criteria is auditable. Discretion held by appointees is not, whatever the rubric says.

Both howling a little is the hygiene test. What’s in the Register has produced howling from one direction only, which is the tell I said to watch for.

Jacksonville, FL Correspondent — Inspect the work, not the politics around the work.

The safeguard is the one ports use: publish the rule, apply it consistently, leave a record that can be reviewed.

If the standard changes when the administration changes, it was a political test after all. The current draft would make that change easier rather than harder, which answers the question as posed.

Las Vegas, NV Correspondent — A standard is useful only if it can reject a favorite result.

Replication can do that. Appointee discretion cannot, by construction, because the appointee decides what the favorite result is.

Keep the test simple: can the claim be checked, can it fail, did political alignment affect the award. If the same three questions are uncomfortable for everybody, you’re near method.

Knoxville, TN Correspondent — A land-grant can live with a rule that says show the work and don’t score the politics. That’s method.

A rule that says show our politics as work is a test.

I said last time to write the first and audit for the second. The draft I’ve seen writes something closer to the second, and I’d rather say that plainly now than defend it later.