In the Pipeline mentions backtracking in experimentation - going back and repeating what was done and checking all the details. While I have backtracked in cases such as this, I more often do so to fill in gaps that I didn't see in the original eperimental layout. In most cases, these gaps don't become apparent until the final report (or a progress report) is being written.
What is it about writing that causes such clarity? Certainly being able to express yourself clearly can only happen if you understand the concepts and results at hand, think about this: This whole concept is backwards, isn't it? How much writing is done in order to teach (textbooks, research articles), or if not teach, convince/persuade/sway (editorials and blogs)? In these cases, the "knowledge" is in the author and is passed on to the reader (yeah, this is the ideal case). But when learning occurs during the writing process, the conveyed knowledge loops around to the author again before eventually ending up with the final readers.
I recall an old bicycling racing teammate back at UIUC that was getting his master's degree in education that was writing his thesis on this area, so I've been kicking this idea around for a couple of decades. I've always wondered what would happen if I were to pre-write the final report before starting the experiment. It would require using fictictional number with reasonable values, but it might be interesting to try sometime.