Thread by @MaxCRoser

Why do people – including journalists – rely on worldometers.info for data on the pandemic?

It has big errors. It has been like this for weeks. And it's not getting better.

I especially don't understand why people rely on Worldometers because there are much better alternatives.

E.g. the @ECDC_EU does very thorough work every single day ecdc.europa.eu/en/geographica…

These are researchers who know what they are doing and who care to get the data right.

@google's top search result for 'mortality rate covid-19' is this Worldometer page worldometers.info/coronavirus/co…
Worldometer is just copy-pasting whatever crap they find.
It'd be funny if it wasn't serious.

Metrics are dangerously mislabeled.

At the very top (worldometers.info/coronavirus) they claim that there there are 1,621,348 Coronavirus Cases.

The big problem with the pandemic is that we don't know the number of cases.

We know *confirmed cases*. The total is surely much higher.

Below that is this table that lists the world's country.

Worldometer claims that Spain did 355,000 tests. No source provided.

That test count for Spain is an unverified number that was released *20 days ago*.

elindependiente.com/vida-sana/2020…

This is the top of their 'Age' page (
worldometers.info/coronavirus/co…)

Their definition claims that this table reports the infection fatality rate (IFR) by age.

If it was indeed the IFR it would be terrible what this table reports here.
But they are wrong and it is actually the CFR.

If you want to check that I'm not making this up, these numbers are reported in this paper in Table 1.
cdn.onb.it/2020/03/COVID-…

These numbers are clearly not what worldometer claims they are.

The problem is not to make mistakes. Every publication is making mistakes and then corrects them when they had time to look into it.

The problem that I have with Worldometer is that mistakes that were there weeks ago are still there.

I never check the site myself and don't see much value in how others are using their site – checking a table for daily updates.

(Because day to day variation is driven by reporting as much as by changes of the actual situation in the country.)

I'm annoyed by worldometer because it wastes so much of my and my team's time.

For weeks we get messages of people asking why do we not show this or that – 'worldometer has the data'.

And too often when you look into it, they provide no source or it is wrong.

It's disrespectful to researchers because they are getting under pressure when they can't do what worldometer claims to do.

And it is disrespectful to their own readers who want to know how the pandemic is impacting their lives and are getting misinformed by them.

My point is not UX. The UX of worldometer is surely better than the UX of the ECDC or us.

My point is that worldometer is used as a data source by journalists and I don't think it should, because it has too many mistakes and better data sources are available.

I didn't know they have a Twitter account –– I'd be interested in what @Worldometers thinks.

Do you plan to correct mistakes on the site?
Do you plan to rely on better data sources?