I, too, found the flutter discussion unnerving. Having a plane shed mechanical parts in-flight is one thing, (I'm used to that) but airframe components is entirely different.
As Joe suggests, accurately reproducing a known design seems the safest way to go, assuming we've heard of that design's concerns.
Clearly, geometry and rigging must be accurate. But how accurate? Is it possible that the tolerances within which a design is flutter resistant are very small?
Sid's observation about elevator stop settings on the Luscombe suggest that a change in the stops which produces 1/4 inch greater deflection at the trailing edge can cause more interesting stalls. This should provide a clue.
So, flutter experts. What would you look for in flight testing these planes?
