Measured · No. 3 · 2026-10-24

The Silence List, With Denominators

Of 2,057 stories three channels covered, MSNOW sat out 55.8% and BBC 56.9%, against Fox 20.0% and CNN 12.5% — and MSNOW has the most captured airtime, so this is format, not capture.

Take every story that at least three of the four main channels covered, then ask which one was missing. There are 2,057 such stories in the collection. The absences are not evenly spread.

CNN is the widest net in the record: it is missing from one story in eight that three others ran. Fox News is next. MSNOW and BBC each sit out more than half.

The obvious objection is that this is about capture, not choice — a channel that was recorded less has more absences. So here is the denominator, and it runs the other way:

MSNOW has the most recorded airtime of any channel in the collection and the highest absence rate. That is not a capture artefact. A long-format channel with fewer, longer segments covers fewer distinct stories in the same number of hours; the format does the excluding. BBC's figure has a different cause again — a rolling international feed simply does not carry every US domestic story, and it would be wrong to read 56.9% as a decision about American politics.

So the finding is narrower and more useful than "who ignored what": the shape of a channel's format sets the width of its story set, and someone comparing two channels' silence without normalising for airtime and segment length is measuring the format, not the editorial line.

What this does not say. It does not say any channel suppressed anything. Absence from a cluster is not absence from the channel's day — the story may have run before the cluster, or in a form the segmenter read as chatter. The desk's own coverage record has gaps, and the number above is what survives them, not a full account.

Method. Source: stories joined on cluster_id, kind='story'; a cluster counts when at least three of {foxnews, msnow, cnn, bbc} have a story in it — 2,057 clusters. Absence = channel has no story in that cluster. Airtime = sum of ts_end-ts_start. Window 2026-07-15 → 2026-10-23, recording stopped 2026-10-23. Figures pinned in data/measured/facts.json; rerun with python -m parrot.measured.

How to check every number. Figures are generated by python -m parrot.measured into data/measured/facts.json, each one carrying the query that produced it. The recording window closed on 2026-10-23; the desk does not add to it. Corrections are published against the facts file, not edited into the prose.

← All Measured pieces