Regulator: ONS should have ‘elaborated on limitations’ of controversial teacher Covid analysis

The UK’s statistics regulator has advised the Office for National Statistics (ONS) it should have been clearer about the “strengths and limitations” of controversial analysis which suggested teachers were not more at risk of catching Covid than other key workers.

Ed Humpherson, director general for regulation at the Office for Statistics Regulation (OSR) has written to the ONS about its Covid-19 infections survey from November.

The study claimed there was “no evidence of differences in the positivity rate between primary and secondary school teachers, other key workers and other professions”.

But Humpherson said the ONS “should have elaborated on the strengths and limitations of the analysis, what conclusions can be drawn and what these conclusions mean – for example, in relation to the impact of the rate of prevalence on the findings and the ability of the survey to determine differences between groups”.

The claims in the study were disputed by University of Cambridge academic Sarah Rasmussen, who initially raised concerns via Twitter before lodging an official complaint. She said the report’s accompanying data did not support its conclusions and was “unacceptably misleading”.

It comes amid a growing debate over whether school staff are more likely to catch Covid-19 than other professions.

Analysis this week by the National Education Union (NEU) of staff absence data suggested there were “much higher” infection rates in school staff than the population at large, though experts have urged caution as the methodology involved two sets of data that “aren’t fully comparable”.

And new analysis published today by the Education Policy Institute (EPI) concluded it was “highly likely” the number of infections among teachers is higher than in the rest of the population, but due to a lack of data it is not clear by how much.

On Monday, the ONS is set to release its first report on Covid related deaths by occupation since June – which may shed some new light on how teachers have fared in comparison with other professions.

In his letter to the ONS, Humpherson acknowledged that while there was a “challenge in clearly communicating technical information to users” it was “important to clearly explain findings such as ‘no evidence of a difference’ to avoid misinterpretation by others”.

The letter went on to advise steps ONS should take in the future when communicating “similar analysis”.

ONS should “engage and collaborate with expert users” prior to releasing ad hoc analysis in order to improve the final analysis, said Humpherson.

It should also be clear about the ability of the survey to detect differences between groups and should explain the decision behind different methods of analysis.

“For example, in relation to the new occupational risk analysis, what are the criteria for an occupation being classed as high risk?”

Humpherson added that the ONS “must clearly communicate its findings, including appropriate commentary and caveats”.

Finally he recommended that the ONS offer “as much insight as possible into how interventions that were introduced for some workers at different stages of the pandemic impact the findings”.

Your thoughts

Leave a Reply

Your email address will not be published.


  1. Dominic Wall

    While the ONS statistics published last Monday captured critical information, it took the NEU to draw out the impacts hidden in the data; most striking of which was that support staff in special schools are most at risk of catching Covid at work.
    This is no surprise as they provide personal care up close to disabled children every day, some of whom will have asymptomatic infection.
    The answer is already here, last week dozens of LAs took the step of including these staff in the priority group of ‘care workers’, and some have already received the vaccine.
    This now needs to be rolled out across the country to protect all of these staff, and the children they care for, so that we can get children and staff back in school, and take the pressure off the hard-pressed parents of children with SEND.

  2. The data appears to have cover a time period of 9 months. 5 of which were when schools were closed to most children. I assume an average of the 9 months were looked at. A more reliable data set would surely be September to December when schools and businesses were open more normally