In the British Social Attitudes survey, NatCen have asked about public attitudes to official statistics and the Office for National Statistics (ONS).
The latest survey finds that trust in official statistics remains high. This article highlights the importance of official statistics, trustworthiness, and public discussion.
Trust remains high: Trust in official statistics remains high, with an estimated 88% of those giving an opinion saying they trust the ONS.
Building higher: Statisticians and data analysts need to stand up for statistics in debates and help the ultimate recipient of official statistics: the public.
In 2018, half of the people answering NatCen’s British Social Attitudes (BSA) survey were asked about official statistics. These questions were funded by the UK Statistics Authority.
Public attitudes to official statistics in 2018 are similar to those in 2016. Trust in official statistics remains high, with agreement of general accuracy. Of those expressing an opinion, an estimated 88% agreed they trusted the ONS, and 85% trusted its produced statistics.
The latest BSA estimated 93% of British adults believe official statistics are important for understanding Britain.
Occasionally, when discussing statistics publicly, I am accused of enjoying statistics for its own sake, or for placing (somehow abstract) statistics in a privileged position above ‘real lives’ or experiences. We need the ONS and official statistics to help us understand Britain.
For example, understanding changes in the UK labour market is incredibly important for government policy. If some people cannot find work or struggle financially, then greater comprehension is needed for a proper response.
The Labour Force Survey contains, within its estimates, the experiences of tens of thousands of people. Those people do not cease being real because they answered a survey.
Telling the story
In a subsequent question, NatCen asked people who trusted and distrusted ONS statistics about why they held that belief.
In the 2018 BSA, the top three reasons for distrust were misrepresentation by politicians or the media (24%), the government having a vested interested in results (23%), or that figures alone not telling the whole story (23%).
The misrepresentation of official statistics should be taken very seriously. Fighting back against that misrepresentation is one of the reasons I became a Statistical Ambassador for the Royal Statistical Society.
When journalists, fact-checkers and statisticians respond to these misrepresentations, stress should be placed on the ONS’ trustworthiness. This is about what the ONS (and other producers of statistics) does in order to earn trust, not merely about whether it is trusted.
It should be demonstrable that the ONS is independent from government. That independence, along with the trustworthiness generated through its Code of Practice, should help alleviate these concerns.
The perception in accuracy of statistics series varies, from 87% estimated to believe the Census to 70% believing the crime statistics are accurate.
As Hetan Shah, the RSS Chief Executive, suggests:
One reason may be that people do not see their own lives reflected in the reported figures.
It would be difficult for regular reports by statistics institute to fully answer all questions people may have about employment, crime, inflation or other matters. It is true within organisations too — my current role is focused on that extra analysis, beyond what regular reports can describe.
The ONS needs to take the lead, producing statistics relevant to people’s areas or regions, as well as proactively conducting analysis on its own data to help answer wider questions in society.
The communication problem
The ultimate recipients of official statistics are the public. Nevertheless, it can be difficult for statisticians (and other experts) to be heard in public debate.
Statisticians and data analysts have a natural, and wholly admirable, tendency to prefer accuracy over excitement. The problem is that websites want interesting articles, television studios want engaging guests, and radio likes people without soporific voices.
The answer is to be accurate and also compelling to read, listen to and watch.
We need to stand up for accuracy, for statistics, and for evidence in public debate. If official statistics are distrusted, then governments go unchecked.
The British Social Attitudes survey is a random probability survey of Great British households, in which the postcode region, addresses and the specific person at each address are all randomly selected. For the 2018 BSA, interviews took place between July and November 2018. A total of 1,968 adults (aged 18 or over) were interviewed. The response rate was 42.4%.