First is understanding the horizontal stabilizer's function relative to its partner, the wing.
Jockey and horse really, perhaps director and symphony. For passenger rated aircraft (and big cargo transports who really don't want to spill the goods), a general approach to design parameters may begin with the following criteria:
To control wing AOA and prevent the wing from reaching a stall AOA uncommanded.
This means the Hstab must be adequately sized enough to act as a "weathervane",
keeping wing AOA (relative wind) constant.
To keep net center of lift from drifting too far forward as angle of attack increases.
As the wing pitches up, it's Clift moves forward, but as Hstab begins to contribute tail
up/nose down torque, properly balanced, the effects cancel.
In the event of a full blown stall/sink, to pitch nose down faster than vertical descent
increases AOA. Failure to do so puts plane into "deep stall".
Designers have a choice of the airfoil approach, the larger flat plate approach, or a combination of the two. A low aspect flat plate might be a safe choice for 3, as it acts in a manner similar to a parachute to push the nose down even if it is fully stalled. Airfoils will work for 1 and 2, but must stall at a higher AOA than the wing.
Once the right area is established, determination of maximum deflections may be as follows: no one deflection should be so great as to not be overcome by two others. This is why 3 pitch controllers may be better: Hstab pitch trim (very slow), elevator (weaker than Hstab in normal flight, higher rates available for emergencies), and a fine trim tab (for minor pitch adjustments such as a few knots of airspeed). The pilot(s) control the elevator.
If Hstab is designed properly and thrust line is correct, this should make for a very safe and stable aircraft. Although obviously influenced by recent events, a review of basic designs can help engineers stay anchored near the highest of safety standards required for passenger carriers.