Going back to the OP, every model given by FE'rs predicts that the apparent size of the Sun and Moon will be less than half around dawn or dusk compared with its apparent size at noon. This simple fact comes from the proposed path of the light, that in every FE model is two to five times longer at dawn than at noon.
On the other hand, modern science states that the apparent size of the Sun and Moon changes only minimally due to refraction of the air, which will only make a barely measurable difference (less than 2%).
The experiment is simple: look at the Moon when it is close to the horizon; extend your arm and hold your thumb up; see if your thumb can just barely cover the image of the Moon; if it covers too much or too little, adjust the distance from your eyes to your finger accordingly. Then repeat the experiment when the Moon is close to the zenith.
To do the experiment with the Sun without burning your retinas, make a pinhole camera and look only at the image in the camera. Measure the size of the image at noon and dawn or dusk. Also check whether the image is well defined or just a lot of glare.
If there is any ground under any of the FE models, the Sun and the Moon will seem much smaller when close to the horizon. If modern science is right, you will not be able to find any difference in apparent size with the crude method explained.
Need a simpler experiment?