OK - I missed this link. Now there's something to talk about.
Since you're just measuring airflow in an artificial situation, I don't really understand how you deducted this:
"The Noctua NF-F12, in whatever iteration you may find it, was not designed as a pressure-optimized fan."
Moreover this is incorrect:
"They have their NF-P and new NF-A series (for 120 mm fans) as pressure-optimized cooling solutions "
NF-F and NF-P are marketed as preasure-optimized. NF-A is an all-rounder.
Testing Noctua's claims would certainly be interesting, but it can't be done with your setup.
As for the methodology itself: you're measuring airflow in the middle of the exit, right?
This will affect your results. The tunnel is quite short and you can't assume the airflow is uniform at the exit.
The difference won't be huge in absolute terms, but even few % could seriously change how the ranking looks.
Measuring in multiple points would give a more accurate result and it can be done with an Arduino board and parts from an old printer. ;-)
Also there's just no way this:
"[setup] consists of a 1 m tall cylinder made out of acrylic (6" OD, 5.75" ID) with two flow straighteners inside out of a metal wire screen placed 50 mm and 100 mm from the exhaust"
describes this:
the tunnel's length is ~4x the fan diameter.
