How different countries are reacting to the COVID-19 risk and their governments’ responses

When governments really need to get an urgent message across to the whole of society, and take what might be very unpopular measures, it’s important to get it right.

The Behavioural Insights team in the UK have released data on testing they’ve done to see which posters are most effective at changing people’s hand-washing behaviour.

We at the Winton Centre do not work on changing people’s behaviour, but we are interested in how well the content of different messages gets across, and how that affects people’s understanding, worry, and ability to make decisions.

So, as the pandemic spreads across the globe and governments take different approaches to tackling it, we’ve spent the weekend collecting data from several countries.

At the end of last week we collected data from 700 people in each of the UK and US on how they were responding to the risk of the coronavirus, and their governments’ reactions.

Over the weekend, we’ve collected the same data from more countries: Australia, Mexico, Spain, Germany and Italy (with the same caveats about the sampling — see footnote).

The comparisons make fascinating reading. Perhaps the most striking thing is the similarities across countries (with Mexico as the most different), but there are also some subtle differences hidden in there…

All countries felt that their governments’ response in the past few weeks had not been firm enough — even Italy, which has had strict curtailment of social gatherings for at least 10 days before the survey and Spain, under lock-down for a week. Mexico, with the fewest restrictions at the moment, is most united that the response is definitely ‘not firm enough’.

One thing that we thought might affect people’s attitude to their government’s strategy was how much they felt that we as individuals should give things up for the benefit of society, and here the cultural differences are, perhaps, surprising. Italy seems incredibly pro-social whilst Germany — and the UK — much more individualistic.

What we don’t know from this data is how much these feelings have been affected already by the virus: Italy and Spain, for example, have been under extreme measures for over a week and people could already be feeling the need for everyone to take action for the greater good. These are things that we at the Winton Centre will be interested to look into further.

Worry about the virus is very high in all countries. And the Mexican data sadly shows that worry about the virus has not displaced worry about other topics: in fact Mexico is clearly a place where worry is high across a lot of topics. Maybe it’s good as a society that worry about one thing doesn’t stop us worrying about another, but high levels of worry as an individual might not be healthy. This is going to be interesting to investigate further, and we look forward to digging into the data more.

To see how the above graph compares to survey responses on issues such as climate change, immigration, unemployment, and terrorism, visit the post on the Winton Centre's Medium channel.

Different countries vary a lot in how much they understand about their government’s strategy. Germany seems particularly confident about their understanding of the government’s strategy, whilst the UK and US much less so (this data was collected at the end of last week, before the UK prime minister’s broadcast on the 23rd March which started a lock-down):

And trust them to deal with it (again, Germany seems to have more trust than others — how has the communication strategy been different there?):

Although trust in their scientific advisors is generally high in all countries:

And it’s clear where they are getting their trusted information. Not social media, but WHO:

And trust in government information is variable:

This is only the tip of the iceberg in the data. Some of it is following on from our work on communicating uncertainty (such as the work we’ve just published in this paper here) and we will be analysing and publishing the results as quickly as we can.

For all the non-experimental measures such as those we’ve quickly graphed above, we’ll be putting all the survey data into our Open Science repository for anyone to use: https://osf.io/jnu74/ just as soon as we’ve got the last few data points from each country in.

The Winton Centre will be doing much more analysis on the data over the coming weeks, and we hope to run it in a couple more, contrasting countries, as well as surveying the UK several times over a few months.

The survey is of around 700 people per country, representative by age and gender for each country, but of course respondents are all online so that means that generally they’re from more-educated, higher socioeconomic status groups (and that can bring a liberal bias, although we have the data on their politics so we’ll be checking that). It’s the best we can do for a long survey in a short time. But do handle the data with care. 700 participants means that any proportion has a margin of error of +/- 4%, assuming the sample is representative. A few countries had slightly fewer than 700 participants and we are continuing sampling tonight to get the numbers up.

The Winton team working on the survey and results are: Sarah Dryhurst, Gabriel Recchia, Claudia Schneider, Anne Marthe van der Bles, Sander van der Linden, John Kerr, Alex Freeman.

Translation: Maria Climent Palmer and Giulia Luoni.

For more information and media enquires, please contact:

Dr Sander van der Linden: sander.vanderlinden@psychol.cam.ac.uk

Prof Sir David Spiegelhalter: d.spiegelhalter@statslab.cam.ac.uk