You need to get a grasp of how video and audio comperssion works to understand what you just said. So I'll give you a little quick walkthrough on this.
Let's take a source video, say a DVD video. Well, after you've ripped it you have a VOB file which is the pure DVD video, unaltered pure uncompressed form. When you encode something with divx you have an audio and video stream. The audio stream is usually encoded with mp3 or ogg vorbis. These are lossy audio formats. The video is encoded with DiVX. DiVX is a lossy video codec, meaning everytime you encode something with it, you lose something, data. DiVX does fairly well at mid-high bitrates at masking what it compresses and loses but there are still visible artifacts.
Now say you want to encode at 640x480 for a divx movie. Thats 307,200 pixels per frame that are rendered. Now say you want to do what you intended to by re-encoding at a larger resolution. First you have to uncompress the data from the video-stream back into mpeg or mpeg4 or some uncompressed video format. This will blow a movie up from 700 meg to around 4-5 gig usually, the original size of the unfiltered, uncompressed movie. Although the movie is the same size, data is still LOST because the compression is 'lossy'. Well, say you want to encode the data into a higher resolution, lets say 800x600, thats 480,000 pixels. Thats 172,800 more pixels you have to fit the EXACT same amount of data in as the original resolution. So how do you attempt this feat? Streching. The image is streched, warped and hung to dry. And on top of that, you have to recompress it using a lossy codec which makes it look worse, so you're looking at 2x lossy codec + streched image from resolution change. This is not going to be pretty.
Hope this helped a bit.