This is a diffuser for Gemmy-style three-LED spotlights. These are often sold by Home Depot, Lowe's, Kroger, etc. It will probably not work for any of the fire-and-ice or similar.
The diffuser is the white circle in the photos. It makes the light a bit less narrowly focused. It also cuts the brightness/glare a bit, which is a good thing if the light is pointed directly at your porch. (The white part around the edge is a test print for a different part that is not yet completed.)
Choosing a Model
The original model has a face thickness of 0.4mm. After testing, this thickness (when printed in White PLA) blocks enough light that it is indistinguishable beyond about 3 - 4 feet, at least in the ambient lighting conditions in my yard. If your setup is darker, it might work better. This thickness works pretty well when printed in a natural PLA, but it might be a bit low -- the light still casts shadows from props 6+ feet away onto walls twice that distance.
As such, I have created a couple more versions of the diffuser, designated by the number at the end of the filename, and by the number of small holes on the side. The 0.3 mm face thickness version is intended for White PLA. The 0.5 and 0.6 mm versions are intended for natural PLA (experiments to find a size that is somewhere between 0.4 mm natural and 0.4 mm white in terms of throw and brightness).
The layer height that any of these diffusers should be printed at is "initial layer height plus height of subsequent layers" in any combination that can add up to exactly the listed number. For example, the 0.3 mm model could be printed at 0.1 mm layer height with a 0.2 mm initial layer. Or it could be printed with all layers at either 0.1 mm or 0.15 mm. Or a 0.3 mm initial layer (with any value for the other layers). All of these combinations can add up to 0.3 mm.
The reason is because when your layer height(s) don't evenly add up to the thickness of the face, the slicer will be forced to either print a thinner part or a thicker one, depending on how its slicing algorithm prioritizes dimensional accuracy in the Z axis. If you specify 0.2 mm height for all layers, you are telling the slicer it is only allowed to print either a 0.2 mm face or a 0.4 mm face ... not any value in-between (such as 0.3 mm).
Note: In reality, material choice and the thickness of the plastic in front of the light aren't the only things that can affect diffusion. FDM printing doesn't produce a monolithic piece of plastic. There are inevitable voids caused by the curvature of the filament traces and the uneven texture of the top of each layer (and bottom, if you aren't printing on glass). This is the same reason transparent filaments don't print completely transparent parts.
The consequence is that the exact number of layers, the fill patterns used by the slicer, under/over extrusion, and even the nozzle size can all affect your results. However material and part thickness are among the more obvious and easily changed factors. (Fill patterns are also easily adjusted, but depend on what your slicer supports.)
I am currently experimenting with concentric infill, to provide a more even diffusion, somewhat like a Fresnel lens. To do this in Slic3r Prusa edition, I discovered that you must set Top and Bottom horizontal shell to no more than 1. If you go above this number, the extra horizontal shells (those not literally on the top or bottom of the part) will not respect either pattern setting on the Infill options page. They will invariably use recitlinear pattern instead.
This part was designed to be printed in white or natural PLA (I use Inland PLA, if you are curious). You'll have to experiment if you want to use another color/type of plastic. One thing you might try is printing the part in something like transparent PETG in order to use it as a gel with a white spotlight (or to tweak a colored light).
The diffuser sits in the recess on the front of the spotlight. On my lights it will stay in place when the spot is staked in the yard or similarly mounted, but can easily be pulled out. It may fall out if you try to mount your lights upside-down in a tree or in a roof or whatever (I suggest scaling it slightly upwards in the X and Y axes to increase the diameter and see if that works better.)
Also, if you scale the part in the Z axis it will affect the thickness of the face of the diffuser. You could, in theory, use this to fine tune the amount of diffusion. There are limits, of course. If you scale the Z axis too high, the diffuser will be too thick to allow much light through. If you scale it too low, the face of the diffuser may be too thin to be of much use, too fragile to come off the bed intact, and the edges of the part will be below the edge of the recess (which may make it harder to get out of the spotlight if the fit is too tight)