The way I've always thought about this is that with a real image, you can imagine taking a screen, putting where the image is, and see the image projected onto the screen. That's really equivalent to what you said, but you can do some easy (and potentially spectacular) demos using the idea.

With a lens, demonstrate the difference between magnifying text written on a piece of flash paper (a virtual image), and creating a small (and very brief!) image of the Sun on that same piece of flash paper (a decidedly *real* image).

For mirrors, try the virtual image of the student's face in a convex mirror, vs the real image of the sun at the focus of a parabolic solar cooker.

(At the heart of every (ex-)physicist is a pyro waiting to come out. :-D)

Real images are formed when light rays converge; virtual images result when light rays *appear to* converge. Convex lenses and concave mirrors produce an optical image which is "real" in the sense that it can be mapped onto a specific focal plane which is independent of the observer's location. A screen placed at that plane will reveal the same image to all observers.

Virtual images are produced by the divergence of light rays, as with concave lenses and convex mirrors. None of the light rays making up the image actually pass through the perceived location of the image, and the observer becomes an element of the optical system, meaning that no two observers see the virtual image in the same apparent location.

A simple experiment might be to project a real image and have students describe its location relative to the rest of the room, then set up a convex mirror and have the students map the virtual images they see in the same way.