Yesterday the UK was hit by another wave of public sector strikes over pensions, with workers across the country taking part.
Seven Unions were purported to be involved: the Public and Commercial Services Union (PCS), the Rail, Maritmie and Transport Union (RMT), Unite, the Immigration Service Union (ISU), University and College Union (UCU) Prison Officers' Association (POA) and the Northern Ireland Public Service Alliance (NIPSA).
But news reports featured wildly varying estimates of the number of workers involved. So which were right?
Part of the confusion caused by the estimates yesterday was that the Cabinet Office, who published the Government's estimates of how many workers were involved, put out two different figures, one for the estimated number of public sector workers overall and another for civil servants in the Cabinet Office alone.
The BBC sums up everyone's position succinctly:
"Union leaders say up to 400,000 workers are on strike, while Cabinet Office Minister Francis Maude puts the total number of public sector workers who joined strikes at "close to 150,000" which includes 102,244 civil servants."
The Cabinet Office confirmed that 102,244 civil servants were on strike yesterday, based on - as they informed Full Fact - their own staff database. Furthermore, Cabinet Office Secretary Francis Maude esimtated that the number overall was around 150,000 - the source of the Sun and the Express's figure.
When Full Fact asked where the extra 48,000 public sector workers in the estimate came from (who were not civil servants), the Cabinet Office were unable to be specific, saying that estimate would be the total of estimates from several different Government departments.
However, evidence from the Department of Health (DH) suggests that even the Cabinet Office's estimates may be imperfect. After the DH confirmed to Full fact that around 4,000 health workers could have been on strike, they conceded that this was simply based on the number of workers who failed to show up for work when they were expected to. Hence, the figures could include workers off ill rather than necessarily on strike.
Some news outlets however failed to bring out the distinction between civil servants in the Cabinet Office and public sector workers overall.
The Daily Mail fell into this trap:
"[Union leaders] were expecting around 400,000 public sector workers to join the latest wave of strikes and protests over pension reforms. But ministers said the real figure was dramatically lower at 100,000 — down from the 146,000 who took part in last year's walkout."
The Daily Mirror's reporting could also be regarded as confusing:
"But despite union leaders claiming a turnout of more than 400,000 workers, the Government said only 100,000 civil servants took action."
But why the wildly different estimates from the Government and the Unions?
It is clear up to now that the Government's estimates of the strike turnout are taken from departmental calculations - although it isn't clear whether these measure government workers who were on stike or merely off work for unknown reasons.
The Public and Commercial Services Union (PCS) claimed the 400,000 figure, in its press release yesterday, was based on estimates from the membership of the unions that took part in the strikes.
Full Fact contacted the PCS for more details on this and are awaiting a response. However some of the other unions who took part gave out some of their own estimates.
POA: Estimated 10,000 members took part
Unite: Estimated 'tens of thousands' based on feedback from regional offices
RMT: Estimated up to 500 members took part
UCU: Had taken no estimate, although said that 60,000 members had supported the action on paper
In spite of awaiting a response from the PCS, the estimates provided by the four unions who responded raise considerable concerns over the reliability of the 400,000 estimate.
Unite were unable to be more specific than 'tens of thousands', which could of course fall anywhere between 10 and 99 thousand, while the UCU had not taken any estimates of its own turnout.
Of course the PCS would most likely have been the largest contributor to the strikes and their figures should therefore be most significant to the overall estimate.
Neither the Government nor the Unions seem to provide especially reliable figures, at least at this stage. The Government's figures are taken from departmental staff levels which were not provided in one place nor were some necessarily recording 'strike turnout', merely a staff member being absent from work.
Meanwhile the Unions' estimates of how many of their members answered the call to strike vary considerably, with some failing to provide any estimates and others being vague about turnout levels.
While readers could have understood the figures in this morning's papers as 'fact', they are more accurately nothing more than competing estimates based upon as-yet unclear assumptions.
The PCS have now got back to us to explain their estimates.
The union claims that the total possible number of its own members who could have been on strike is 249,504 - that is the number of PCS members who are involved in the pensions dispute.
The PCS informed us that their estimated strike turnout was 71.82 per cent, based on union representatives counting in attendees to workplaces of PCS members. 37.6 per cent of PCS members (or 93,698 workers) were surveyed in this way - across every region of the UK - with the 72 per cent turnout from these extrapolated to apply to all 249,504 involved.
They also provided a specific example from HM Treasury and Customs:
"In the second largest department, for example, HM Revenue and Customs, we received reports from workplaces covering 87.12% of our 55,851 members. Of these, 37,246 were reported as being on strike - 76.54%. Which gives an estimate of more than 42,000 on strike in one department alone - well on the way to the government's claimed 102,000 in one hit."
All these calculations give an expected turnout of PCS members of around 180,000. A total of over 200,000 for the civil service is arrived at by assuming that the other participating unions would enjoy a similar turnout. However from the feedback that Full Fact has received from other unions, this does not appear possible to verify.
The source of the 400,000 figure remains unclear, and could simply refer to the total membership of the unions rather than the turnout figure. Full Fact will keep readers posted when we find out more.
*This update was revised to correctly reflect the PCS's calculations as they were explained to Full Fact