Dudley Henriques brought next idea :
I just want to make it absolutely clear that in my opinion, the ONLY
relationship between physical sensation and IFR is in understanding how
physical sensations can harm you and how to deal with them by instrument
referencing all the way through the scan down to primary panel.
I would NEVER attempt to verify an instrument reading by referencing a
physical sensation. In ANY situation where an instrument reading was suspect,
I would immediately extend my primary scan to include peripheral instruments
to verify the quality of the data that was suspect. Under NO circumstance,
would I EVER allow the time line necessary to include a physical sensation in
this equation. To do so in my opinion is dangerous not only in a possible
erroneous attitude input, but as well extends the time line to a recovery
input.

Being primarily creatures earthbound (land underfoot), where feelings
are our primary sources of instrument accuracy (speed in a car, wind in
our hair), its kewl to trust those sensory inputs. A lot of
day-in/day-out experiences too.