We've had an analysis piece reach an advanced draft stage, built around a specific thesis, only to have further reporting during the fact-checking process reveal that the thesis didn't actually hold up as well as the earlier drafting had assumed - and we killed the piece rather than publishing a weaker version of the original argument.
Why this happens even with careful planning
An analytical piece often starts from a genuinely reasonable working thesis based on available evidence at the time, and further verification during the writing and editing process sometimes surfaces additional evidence that complicates or undermines that original thesis in ways not apparent when the piece was first conceived.
Why killing the piece was the right call, not a failure
Publishing a piece built around a thesis the reporting itself has since complicated, simply because significant work has already gone into it, prioritizes sunk cost over accuracy - exactly the trap we'd criticize in other contexts. The honest response to reporting that undermines your own thesis is to let the thesis go, not to publish it anyway with the inconvenient parts minimized.
What we did with the underlying research
Where the research itself remained valuable even though the original thesis didn't hold, we sometimes repurposed it into a different, more accurate piece reflecting what the reporting actually supported, rather than discarding genuinely useful work entirely.
Why we think this discipline matters
A willingness to kill a piece when the evidence stops supporting it, even late in the process, is part of what we think separates genuine analysis from advocacy dressed up as analysis. The thesis has to answer to the evidence, not the other way around.