Crowd and Poll Claims Checklist for Political Journalism

Pull fire code postings, seating charts, public safety permits, and prior event manifests to define hard caps and typical layouts. Store them in a labeled folder so you can publish capacity context within minutes.

Coordinate with the photo desk on drone, helicopter pool, or fixed rooftop positions, and verify FAA Part 107 and local no-fly restrictions. If airspace is restricted, lock in a high vantage point with time-lapse capability.

Create copy blocks that explain your method, margin of error, and source links so editors can drop in numbers fast. Include standard language for estimate ranges and methodology notes.

Maintain a roster of firms with AAPOR Transparency Initiative membership, clear methodology PDFs, and recent accuracy, plus a red flag list for advocacy and push operations. Require methodology links before using any topline.

Get contacts at network research desks and vendors like Nielsen and Comscore, and note embargo windows. Prepare a one-pager on metric definitions so the newsroom uses consistent terms.

Use a Slack form or shared drive to collect permits, venue emails, poll PDFs, and ratings reports with consistent file naming. Make a simple dashboard so editors can see what evidence is in hand.

Shoot from fixed points with overlapping frames and keep EXIF data intact, noting the exact timestamp against the program schedule. This anchors density estimates in time and space.

In QGIS or Google Earth, draw polygons for the stage zone, plaza, streets, and overflow, then record square meters per polygon. Save the project file and export a screenshot for the receipts folder.

Use a Jacobs-style grid with density assumptions like 0.5, 1, 2, and up to 4 persons per square meter for very packed zones. Justify any higher assumption with photographic evidence and safety reports.

Request turnstile or magnetometer totals from security or the venue, and save emails or statements as screenshots. Note whether counts include re-entries or staff to avoid inflation.

Station a camera at a main entrance and log arrivals in intervals so you can reconcile peak occupancy against flow. This prevents mistaking high churn for high attendance.

Mark risers, camera pens, tree cover, and off-limits areas on your footprint map. This combats cherry-picked angles and explains gaps in the visual record.

Pull the methodology PDF and verify when interviews were conducted, who was sampled, and how (live-caller, IVR, online, or hybrid). Record weighting variables like age, education, race, region, and party ID.

Publish both the stated margin of error and design effect if available, and note the effective sample size. Avoid spurious precision by rounding toplines sensibly.

Compare the poll to a median of recent nonpartisan surveys and note consistent lean. Flag outliers and avoid rewriting the narrative on a single data point.

If a poll is funded by an advocacy group or campaign, label it as such and check for leading language or priming. Treat push instruments as messaging, not measurement.

Do not mix national approval with battleground results, and clarify whether data are from traditional exit polling or AP VoteCast. Use rolling windows to present trend rather than cherry-picking a spike.

When n is under 200, mark results as directional and avoid headline treatment. Note higher uncertainty for subgroup estimates and any weighting instability.

Report whether numbers are live plus same day or live plus 3, and whether the universe is P2+, A25-54, or another demo. Include the vendor and any embargo timing.

When a program airs across cable, broadcast, and streaming, sum only unique audiences and disclose the method. Avoid adding re-airs to live figures unless clearly labeled as cumulative.

Publish both peak concurrent and total unique viewers, with methodology on autoplay suppression and suspected bot traffic. Keep platform breakouts separate to avoid double counting.

Grab platform dashboards or vendor reports with UTC timestamps and archive them for public reference. Screenshots should include the URL bar and retrieval time.

Do not attribute highlight reel views to live program audience. Label short-form engagement metrics clearly so readers understand reach versus viewing.

If content ran on FAST channels, embedded players, or partner feeds not captured by standard panels, say so explicitly. Note that such channels are not directly comparable to panel-based ratings.

Identify the speaker, event, and timestamp, and link to primary video with a transcript marker. Save a web archive copy in case the original is altered or deleted.

Publish low-mid-high estimates with the density rationale and area measurements, not a single absolute. A methods note anticipates challenges and builds trust.

Link methodology PDFs, venue emails, permits, and analytics pages in a bulleted list. Keep originals in a newsroom drive with a naming convention like YYYYMMDD_event_source_type.ext.

Stick to verifiable facts, explain the basis for your determination, and avoid imputing motive. Phrases like unsupported by venue counts and poll methodology help minimize risk.

In a sidebar or callout, define margin of error, design effect, likely voter screens, and ratings universes. This reduces confusion and preempts accusations of bias.

If a venue issues revised counts or a pollster updates weights, update the story and social posts with a revision log. Note the exact change and time of update.