Go on dpreview.com and see what they talk about. Then condense it a lot, obviously – their reviews are pretty technical but you could simplify.
They do stuff that would be fun to measure and quantify in a science class. Like they have these charts with fine converging lines printed next to a measurement scale. You take a picture and see at what point you can’t differentiate between the lines any more – this indicates the effective resolution ie what detail it can actually record, as opposed to the number of pixels. There are also black and white squares in the middle and the edge of the shot – towards the edge you can see some coloured fringing which is worse on some cameras than others. Then you could also just take a picture and zoom in a lot – make sure there’s lots of fine detail in the picture like a bookshelf on the other side of the classroom. If you zoom right in you’ll be able to read some of the text on some cameras and not others.
Because you are testing very different cameras the differences will be really apparent.
But also what his Mungness says – usability and intended purpose will bring design, technology and even marketing into it as well as scientific comparison methods.