I can offer a hypothesis. When I was still in the military, I was part of a lot of these surveys. I put my hand up for everything that came down. Some of them are done by paper and some of them are done by town halls. When you're in the town hall variant, you can see this happen in real time: The facilitators will reject, out of hand, certain comments that don't fit the scope of the study. A lot of times we're told, oh, that's a budget thing and it's not going to happen, or, oh, that's just a gripe and the hardship of the service. The comments get rejected out of turn.
It's the same with the paper copies. When we do those paper surveys, we get together afterwards and ask each other about what we put. When the eventual report comes out months later and we all get together, we see that the comments we brought up are nowhere to be found in the highlighted section. You have to go all the way back to the raw data.