When it first became common for college students to write on computers, researchers noticed some changes in the quality of their writing. (I think I read about this first in Michael Heim’s 1989 book Electric Language, though I could be misremembering.) On the sentence level, they tended to improve, largely because it is so easy on screen to make changes to improve word choice, sentence structure, and basic grammar and syntax. By contrast, before electronic word-processing the making of such changes was so laborious — involving the delicate application of White-Out at best, retyping of whole pages at worst — that most of us just didn't bother. We had significant incentive not even to notice such shortcomings.
But word-processed papers, at least in those early days, tended to be more poorly organized than their typed or handwritten predecessors. The use of multiple sheets of paper, coupled with most writers’ need to produce drafts that had to be retyped at least once, allowed for and even encouraged visualizing of the paper as a whole. You could lay out pages before you and just see how the words on one page related to the words on others. You could non-metaphorically cut and paste, and then rearrange paragraphs or whole sections. But when the writer’s view was cut down to a few lines on the screen, and the rest of the paper was invisible without scrolling — which of course didn't make more of the paper visible, just some different part of it — it became harder to discern the overall shape of the paper, and so while papers got better in stylistic and mechanical ways they got organizationally worse.