At some point, you figure out that everything they told you was a lie...
Look in 'Theory of Wing Sections', at some of the airfoil data in the back. Stick to symmetrical airfoils to keep things simple.
Look at the curve of CM_c/4 vs. alpha. Theory says it should be zero and constant. What happens at stall?
For a cambered airfoil, CM_c/4 should not equal zero -- so the curve should be offset. If the AC is not at c/4, then we would expect to add a term that is proportional to cl - which should be a straight line with alpha, so we would expect CM_c/4 to perhaps be a straight line.
Look through the data -- I suspect you'll find some that are curved - and not just near stall.
I prefer to think about static margin as dCM/dCL -- and this is how VSPAERO calculates it (look at the bottom of the stab file). Run a target alpha - perturb one degree, run again, use those two points to compute dCM/dCL.
There is more fun to be had when you're working with a configuration with a large deltaZ between the lifting surfaces and the CG. Think about a high-wing aircraft with a heavy cargo pod rigidly strapped to the belly of the aircraft. What about a low-wing aircraft with a heavy pod high above the CG?
Rob