Pofma actions on Covid-19 falsehoods: 27 uses, and where they landed

Twenty-seven is the figure that gets quoted. It is correct, and on its own it is misleading, because it counts instruments rather than falsehoods. This page separates the two.

What was counted

Every use of the Act on a claim relating to the virus outbreak, taken from the government’s own clarifications site, which publishes each direction alongside the statement it concerns.

The 27 split into two instruments that do different things. Twenty-two were correction directions, which require the person who made a statement to carry a notice saying it contains a falsehood, with a link to the clarification. Removal is not required. Five were designations of a page as a Declared Online Location, which cuts off its ability to earn revenue and can be paired with an order to disable it.

The second count, and why it is separate

At least eighteen further claims were answered by a published clarification and no Pofma instrument at all. Others never reached the site and were dealt with directly, through a post or a statement.

Those are not in the 27, and they must not be added to it. An action and a clarification are different responses: some falsehoods drew both, some drew only one, and the sets overlap in a way that the published record does not resolve. Two figures reported separately are honest; one figure made by adding them is not.

Where the actions landed

Twenty of the twenty-two correction directions went to Facebook or to individuals posting there. One went to a forum. Fifteen of the twenty-seven actions concerned a single operator and the pages he ran.

That concentration is the most useful thing in the set. It says the instrument was aimed at published, attributable content with an identifiable author — which is what it was built for — and that a single actor accounted for more than half of its use on this subject.

What it cannot tell you

Whether the statements were false. The count records that an instrument was used, not the merits. An order can be appealed to the courts, and the record of a use is not a record of a finding.

The size of the problem. The instrument reaches what is published and leaves what is forwarded. The eighteen clarifications are evidence that the second category is at least as large, and nobody publishes a count of what circulated in closed messaging.

Whether it worked. No measure of reach, correction or belief change is published alongside the orders, so the set supports nothing about effect.

One statement, two recipients

The twenty went to Facebook or to individuals posting there, and the distinction matters when counting. A direction about a single statement can be issued to the person who wrote it and, separately, to the platform carrying it. Read as a list of statements, that looks like two falsehoods. Read as a list of instruments, which is what the record publishes, it is two instruments about one claim.

The rule used here counts instruments, so both appear. Anyone recounting from the clarifications site and arriving at a lower number has almost certainly collapsed the pairs to one row each, which is a defensible rule — it is simply not this one, and the two totals should not be compared.

How to reuse this

Quote the instrument with the number. “Twenty-seven Pofma actions” is checkable. “Twenty-seven falsehoods” is not the same claim and the record does not support it.

If the comparison is with another period or another subject, confirm the other count uses the same rule: actions, virus-related, from the clarifications site. Most published figures on this topic do not say which of the three they used.

The file

The download carries the rows above, one measure per line. Both counts are in it — the actions and the clarifications — kept as separate rows precisely so that nobody sums them by accident.

What it does not carry is a per-order list. Each direction is published individually on the government’s clarifications site with the statement it concerns, and that is where a row-level working set belongs. This file holds the totals that were published here.

Checking a single figure

Take the split between instruments. The clarifications site lists each action with its type, so filtering to virus-related entries and tallying corrections against designations reproduces the 22-and-5 split directly.

If a later count differs, the likely cause is the window rather than the arithmetic: actions continued after the piece was written, and a figure quoted without its window will drift from one quoted with it.