"How realistic are published takeoff and landing distances for light GA airplanes?" — Fred C.
“This varies dramatically from one airplane type to another, among similar airplanes, and even in the same airplane under different circumstances.
What I suggest is that you compute the takeoff and landing distances in the ways you normally operate the airplane. Apply at least a 50-percent margin for less-than-perfect pilot technique or runway conditions. Now you know what’s normal for you in that airplane.
If you plan to do anything different from your normal—add a couple of passengers, take off at a higher density altitude, use a grass runway—compute the performance under those conditions, to the extent POH performance data exists, and apply at least a 100-percent margin (double the distances).
I use the 50-foot obstacle clearing distances, not the ground roll distances, as my deciding factor.
One last note: obtaining book performance requires you use the book technique. Most POH takeoff and landing data actually reflect short-field performance: Stand on the brakes until full throttle for takeoff; power off for a steep 1.3 times Vso approach on short final for landing. If the runway lengths are short enough that you need to double-check the airplane’s performance, you need to use the book takeoff and landing techniques.”
What are your actual takeoff and landing distances compared to book (POH) values?