It depends on how much sample you loaded and it also depends on how you standardized it for each sample. Or at the final stage do you normalize everything using any calculations.
Well, it's a bit of an unknown outcome, it would help to label your lanes so we can determine if there is a pattern for which sample types give which gel outcomes. I agree with talking to another lab member, and you certainly have some kind of degradation and/or RNA carryover. What are you hoping to do with your samples?