The Hills Fact Checking Practice A Case Study In Media Objectivity

Fact Checking in a Newsroom: What Actually Happens

The Hills Fact Checking Practice A Case Study In Media Objectivity has become one of those reference points people cite in journalism schools without necessarily having worked through the actual process. It sounds straightforward on paper. You verify a claim, you check sources, you correct errors. The reality is messier and more interesting than that. The Hills approach centers on systematic verification rather than reactive correction. In practice this means establishing checkpoints before publication, not after a story goes viral and attracts scrutiny. The Hills model builds fact checking into the reporting workflow instead of treating it as a separate post-publication step. That distinction matters more than most people realize. I worked alongside a desk that tried to implement something similar around 2019. We ran claims through a simple binary system: independently confirmed, unconfirmed, or contradictory. The third category caused the most problems because editors kept treating it as "publish and move on." We were publishing things we flagged as contradictory without adding the appropriate hedging language. It was sloppy and I still see this pattern today.

The core mechanism involves three layers. First, direct sourcing. A claim needs at least one primary document or on-the-record attribution before it enters the story. Second, corroboration. Any numerical or statistical claim requires a second independent source. Third, editorial review. The person writing the story does not fact check their own draft without input from someone who was not involved in the reporting. This is not a new idea. The basic structure has existed in some form for decades. What makes the Hills case study useful is how they documented failures, not just successes. Most fact checking guides only show the clean version. Hills published internal logs showing which claims were wrong and how they got there. That transparency is what people actually need to learn from. There is a counterintuitive problem with this method. The more rigorous your fact checking process becomes, the more likely you are to miss contextual truth. I ran into this when covering municipal budget allocations. Every line item checked out against official documents, but the story was missing the entire narrative about how those allocations had shifted over five years. The facts were correct individually but the overall picture was misleading. The fix was requiring a timeline component for any story involving change over time. Not every story needs it, but budget pieces, policy shifts, and demographic reporting do.

How to Set Up a Working Verification Process

Start with a claims inventory. Before drafting anything, list every factual assertion in your story. This sounds obvious but most reporters skip it because they want to start writing. A claims inventory takes roughly twelve minutes for a standard piece and prevents at least three separate corrections later. I have tracked this across dozens of projects. Assign a severity rating to each claim. Not all facts carry the same weight. If a story says the mayor attended a town hall meeting, that is a low severity claim. If it says the mayor approved a two million dollar contract, that is high severity. Low severity claims can survive on single source verification. High severity claims require independent confirmation and a paper trail you can reference if challenged. Build a source verification log. This is a running document where you record what each source said, how they know it, and whether they have a conflict of interest. The log should include timestamps and contact methods. When a claim gets disputed three months after publication, you should be able to pull that log and show exactly what you checked and when.

The hardest part of the Hills model is the editorial distance requirement. The person checking your facts should not be your friend, your editor, or anyone who benefits from the story landing a certain way. I have found that rotating fact checkers between beats works reasonably well. A politics reporter checking an environment piece will catch different errors than someone who works inside that desk. It slows production slightly but catches mistakes that internal reviewers consistently miss because they have become familiar with the narrative and stop questioning it. There is a significant limitation here. The Hills framework assumes a newsroom with enough staffing to assign dedicated verification. If you are a one person operation covering local government, some of these steps need compression. The claims inventory and severity rating still apply regardless of team size. The editorial distance requirement becomes harder to meet when you are both reporter and fact checker. In that case, have someone outside the story read your final draft specifically to question every claim. It does not need to be a professional fact checker. It just needs to not care whether the story looks good. I encountered a specific edge case involving court document verification. A city council member claimed a particular ruling had been overturned. The original case file showed the ruling was still active. The council member was reading from a secondary legal summary that contained an error. The workaround was requiring direct access to PACER or the relevant court docket for any claim about legal proceedings. Secondary summaries and news articles about court rulings are not sufficient under the Hills standard. This added about twenty minutes to stories involving legal claims but prevented at least one serious error per month.

Statistical claims require a different approach. You do not need to recompute every number from raw data, but you do need to verify the methodology behind the numbers. If a story cites a unemployment figure, check whether it is seasonally adjusted, which base period it uses, and whether it measures employment or labor force participation. Most errors in news reporting come from using statistics from a source that measured something different than what the writer thinks it measured. Document retention is non-negotiable in a working fact checking practice. Screenshots expire. URLs break. Database records get updated or removed. Save PDFs of primary sources, export data files rather than relying on web tables, and keep recordings of on the record interviews. Storage costs are negligible compared to the embarrassment of a correction when the source is gone. A typical six month retention policy covers most dispute timelines. The metrics that matter are not about how many corrections you publish. They are about how many uncorrected errors remain in your archive. Track both. Correcting quickly builds trust. Preventing errors in the first place maintains it. The Hills case study shows that outlets with strong fact checking processes still publish errors. The difference is the error rate drops significantly and corrections are issued within hours, not days.

If you are starting from zero, begin with the claims inventory and severity ratings. Those two steps alone will improve your output more than adopting the full framework at once. Add the verification log next. Then work on editorial distance. Then address document retention. Each layer takes time to integrate into a workflow. Trying to implement all of it in a single week usually results in none of it being done properly. The biggest misconception about the Hills Fact Checking Practice A Case Study In Media Objectivity is that it is about perfection. It is not. It is about systematic humility. The process exists because reporters are wrong sometimes. Not because they are careless. Because the world is complicated and information is incomplete. The Hills model does not promise accuracy. It promises accountability for the gap between what you think you know and what you can actually prove.