How the video 'looks' with bad data depends on how the decoder has been programmed.
A decoder might try to 'conceal' a bad block by repeating the block from the previous (or next) frame. If there are many bad blocks, it might repeat the whole frame. As a last resort it might return black frame(s) and hope the stream returns to normal soon. A different decoder may do more error concealment, or less.
Point is, there is no particular way a video with bad data will look, unless you control the decoding.
BTW the transmission medium (eg videotape, blu-ray, satellite) might scramble the blocks in space and time, so a big data dropout is broken into many small chunks which are more easily concealed. This again changes how the video 'looks'.
EDIT if you don't like the answers you have gotten so far, perhaps you should explain what you are trying to do exactly.
For my part, my answer was intended to give you some search terms to find more detailed information.
I supposed that if you were compiling a reference codec and injecting errors into the data stream, you were hardly a
newbie.