On Christmas Day, a Twitter user claimed BBC “viewing figures have plummeted since the election”.
This claim — shared over 7,000 times — is false. Estimates from the Broadcasters’ Audience Research Board (BARB) suggest little change in the BBC’s weekly reach.
We can check this
A Twitter user gave supposed second-hand testimony of a BBC employee:
Spoke to someone at The BBC yesterday, this person told me they are SHITTING themselves right now, as viewing figures have plummeted since the election. I mean, REALLY [plummeted]
They now realise their propaganda bullshit has undone them, they fear there’s no way back
This claim can be checked, using viewing estimates from the BARB household panel. These are survey estimates, subject to multiple sources of potential error.
There are different measures to consider. A channel’s weekly reach is the proportion of individuals aged four or over estimated to have watched that channel at any point during a given week.
For the main BBC channels (BBC 1, BBC 2, BBC 4, and BBC News) there has barely been any change in estimates from the week ending 15th December — which included the election on Thursday 12th — and afterwards.
Another measure is the estimated share of total television minutes watched during the week.
On 9–15 December, the spike upwards in the BBC 1 viewing share was due to the final episode of Strictly Come Dancing. BARB centrally estimates that 12.6m people watched this two-hour programme.
BBC 1's estimated share of weekly viewing time in 16–22 December was 20.9% — about average for 2019. It is false to suggest that the BBC’s viewing figures have “plummeted” since the 2019 General Election.
The methodology of BARB
The Broadcasters’ Audience Research Board is responsible for measurement of television viewing in the United Kingdom. The BARB panel is a sample of households which represent television viewing across the nation. Producing these estimates takes the continual efforts of three different social research organisations.
Ipsos MORI conduct an establishment survey of 1,000 households each week — used for panel recruitment and weighting. The target is maintain a representative panel of around 5,300 homes, which contain around 12,000 people.
Once recruited, Kantar Media fits every TV set with a meter. Handsets and tablet applications are used to register when people are watching a particular screen. Audio samples are used to measure what is being watched. Also, Sky allows access to service information codes to precisely identifying what programmes people are watching.
Additionally, Kantar Media collect device-based viewing census data, by embedding software code into different digital television platforms.
RSMB Ltd then provide data processing and quality checking of the information gathered by Ipsos MORI and Kantar Media. This combined work means BARB can publish estimates on how many people have watched programmes, channels, and broadcasters.
These estimates are used by the broadcasters and journalists. Suppose a programme was estimated a half-hour rating of 10 (that is, 10% of the population watched this programme). The confidence interval means the true rating could plausibly be as small as 9.1 or as large as 10.9. In Wales, that interval runs from 6.7 to 13.3.