From a intuitive point of view, I might use these two figures of merit:

(Tail arm)/MAC;

(Mass moment of Inertia)/ (tail area*tail arm).

Each of these requires that you use coherent units. In pitch, use the arm from CG to the 1/4c point of the horizontal tail, mass moment of inertia in pitch, and are of the horizontal. Once we have computed these for a few well know airplanes, we can check it out. Perhaps someone already has...

The first puts how far back the tail is in terms of the wing size. There are some amazingly short airplanes that do not have a reputation for feeling close coupled. The entire Rutan canard family and all of its derivatives, the Questair Venture and its derivatives, and others.

The second puts the amount of tail power in proportion to mass it has to move.

From a more intellectual perspective, the airplanes that feel twitchy might have another set of metrics that would correlate. One of the most common is tail volume coefficient. Pazmany uses:

Pv = (Aht*Rht)/(Aw*MAC)

Another is tail damping coefficient:

Dv = (Aht*Rht^2)/(Aw*MAC)

And yet a third is static margin:

Lsm = Lnp - Lcg

The tail volume coefficient just indicates if you have enough tail. Tail damping coefficient indicates how deadbeat the response to control inputs, gusts, etc will be, and static margin indicates how much tail moment will be required to move the airplane off the equilibrium point a given amount.

Pazmany listed a bunch of the statistics in his book, and then calculated the volume coefficients. You could take his terms and also calculate the damping coefficients, and then get flight impressions on each airplane, and develop some correlation. Or perhaps it has already been done.

Billski