The math is easy enough. A 12mp 3:2 sensor is about 4200 pixels on the long side; an 8mp crop of the same sensor would be about 3500 pixels. 4200/3500 means the crop factor is about 1.2, so it has a wider angle of view than a sensor with a crop factor of 1.6. The hypothesis can be rejected. J
To get a crop factor of 1.6 (and the same angle of view as the 30D), you’d have to crop the 5D to less than 5mp.
I’m pretty sure of my math. It is possible that we are answering different questions…
If you put a 5D on a tripod, lock the head, take a photograph, take the lens off the 5D, put it on a 30D, put the 30D on the same tripod, and take another photograph, and then view both photographs at 100% in Photoshop (which I think is equivalent to what you mean by ‘actual size’: each pixel on the screen corresponds to a pixel in the photograph), they will not be the same.
Objects in the 5D image will be smaller. If you have a 30D image that you like, you know exactly where and how it was taken, and you want to use a 5D to produce a pixel-by-pixel copy, you can’t do it by cropping. You have to change the focal length of the lens. Specifically, you have to zoom in: I think increasing the focal length by a third would be about right, but I won’t swear to the exact value. (Increasing it by a factor of 1.6 would be too much.)
Another way to phrase it (basically what Steve said in the first place) is that an image covers fewer pixels when it is projected on to the sensor of a 5D than when it is projected on to the sensor of a 30D with identical lenses. The pixels on the 5D sensor are physically larger and their centers are spaced farther apart.