Even on different movies of the same file sizes, the degree that each is compressed won't necessarily match up. Consider the overall length of the film, for example. For arugment's sake, lets say that you're comparing Pearl Harbor to First Blood. (This is ONLY for the sake of argument, guys!) Let's further assume that both movies have the same file sizes. Pearl's close to the three hour mark. Blood's only 90 minutes.
If First Blood was encoded using a high bit rate, you'll be able to whittle away at it quite a bit before really noticing very many defects in the copy. That's because there's more picture information present than there would have been with a lower bitrate. Therefore, you'll have a better result if this is the movie you're copying (again, keeping in mind that I'm being hypothetical).
Let's say that Pearl Harbor uses a fairly low bitrate. (To fit on the same size of file as First Blood, it'd have to!) Since you don't have much to chew on, ANY reduction is going to be obvious.
But it's entirely inconcievable to have two movies with very different content (more dark areas, action scenes, ect), different lengths, and different overall bitrates, seem like they're the same when all you're viewing is the file sizes. The file size only states the volume of information that they contain, as it applies to whatever will be reading them. It doesn't really provide much of an indicator as to properties of the content.