A New Approach to Corrections
Rethinking criticism management and how to make it easier for readers to understand what's disputed in a given story.
Subscriber note: Today’s post is just a policy update with a bit of commentary on how my thinking on corrections has clarified over the course of this project. But as readers are split on how interesting they find this kind of tactical stuff, I’ve included a TLDR. While I’m still unsure how many stories I have left in me with my current health and competing priorities, hopefully this adjustment will help make them count.
TLDR: I’m no longer going to engage with criticisms of my work on Twitter or related platforms. I’m going to create a FAQ-style “dispute doc” for each story to house criticisms and replies that can then be linked across those platforms. The intent is solving for social media’s warped incentives, poor discovery, and fragmented threading—making it easier for the curious reader to quickly find all raised disputes alongside the most current counterpoints.
You can find my ongoing corrections log here.
Background
For a few years now, I’ve had a corrections policy. While I’ve gone into the logic of it at length before, the gist is pretty simple:
It’s impossible to write about something fast enough to hit a story’s main attention window while not getting at least a few things wrong. The world is too complex, our expertise too limited, and our sources and judgment too fallible. While good process can limit errors, it can’t possibly eliminate them.
If the point of journalism is giving readers accurate information about the world upon which to form useful judgments, it follows that readers deserve an easy means of knowing where an older story was wrong or disputed.
My original intuition was thus that journalists1 should give readers an explicit incentive to call out mistakes and blind spots, where that feedback then follows a transparent loop back into the original story in the form of marked corrections.
For my own work, this has meant cash payouts for corrections (and really for any meaningful feedback that helped make the story more accurate and helpful).
The Problems
The first flaw with my system has been that—independent of recent health setbacks—I have profound ADHD, and managing feedback across multiple channels has frequently overwhelmed me, leading to significant delays in that loop working as intended. While I have a designated path for reader inputs (a Typeform), few make use of it. Mostly I’m flagged to discussions happening on Twitter / Reddit / Hacker News, on top of emails, DMs, and Substack comments. I then engage in each of those places, which is both an administrative nightmare and not very helpful for readers who are only likely to encounter a small percentage of those exchanges.
Second, all those platforms have user dynamics that aren’t overly well-aligned with the process of making stories clearer or more accurate. If someone on Twitter is going to dunk on one of my stories, it’s good that I answer their criticism(s), but it’s also an intensely draining and infuriating experience to try to do so within an exchange not premised on good-faith engagement.
Lastly, there was no obvious solution for would-be corrections where I just wasn’t convinced. While I’ve always offered to pass those disputes to neutral mediators, people who like dunking on Twitter don’t seem to actually want that2. But this doesn’t itself make them wrong, and those disputes should be captured somewhere.
The Solution
In coming up with a revised system, I prioritized two things:
Readers should be exposed efficiently to all raised criticisms, not just those that happen to surface in their feeds
I should have a chance to respond to these criticisms in enough detail for readers to make informed judgment calls, without making posts unreadable3
So, the solution:
When I publish a story, I’ll also create an FAQ-style “dispute doc” that I’ll link alongside the Typeform4 in my default corrections preface
When made aware of criticisms, I’ll paste them into said doc
I’ll then give various time-stamped responses to those criticisms over time5 (using Google Docs’s edit-log to keep me honest)
For mostly-but-not-entirely-duplicate criticisms, I’ll factor in anything that’s new / additional using marked edits to make sure I’m steelmanning my critics
If I feel a particular thread on Twitter etc needs replying to, I’ll just link the dispute doc and point out which section(s) covers the criticism(s) in question
For at least one of the larger stories left in queue, I’m also planning to experiment with paying an admin to search for criticisms to include (and to help me manage any resulting payouts). Because really I think that’s the natural endpoint for maximum reader trust: not just rewarding critical feedback, but actively looking for it.
I've never known how to classify my own work. Am I a citizen journalist? My interest is mostly in re-examining stories, not telling them from scratch. I think "journalism critic" fits that better. But where re-examining a story includes looking for and surfacing new information, I’m to some degree doing journalistic work. So I definitely think what I write should be judged against the same standards, whether or not the title actually fits.
Unresolved “was this actually a correction” disputes have actually been pretty rare. I don’t have a huge ego about being wrong, and I’ve been quick to take and reward corrections over quite minor / subjective things, even from overly uncharitable people. But taking this example, once in a while someone will insist “c’mon how can you be this naive” without meaningfully engaging with my position. In 100% of these cases I’d be happy to turn things over to a neutral panel. It’s possible in some cases that I really am naive and that more informed people closer to the subject would readily agree. That’s fine! It’s not very hard to find people willing to serve on said panels. But the process requires the person on the other side to be more interested in that correction happening than they are in the performative aspects of the social internet. And I think roughly anyone who has ever engaged with a journalist addicted to Twitter’s status games can tell you how that goes.
While footnotes kinda-sorta work for this, they impose deep formatting limitations. And including too many of them makes pieces appear to be too long, which unfairly scares off readers. It's just much better to host this commentary in a centralized FAQ with the odd footnote pointing the way for those who want the extra detail (e.g., "see section x at this link for more discussion").
I’m keeping the Typeform because it offers easy anonymity. It doesn’t capture anything about the sender other than what they opt to share in the text fields.
This time delay can be crucial. Another shortcoming of the default Twitter model of engagement is that replying too slowly becomes tactically indistinguishable from not replying at all. But sometimes the right response is really just “yeah, interesting point, I’m going to look into this more and it may take me a day”. Twitter doesn’t really have tools or norms to support that. While linking out to a Google Doc obviously doesn’t itself solve this, it at least shifts any resulting dialogue to a slower and more thoughtful place.
Seems like a good system, hope this becomes standardized someday.