YouGov’s Numbers Don’t Add Up

Our next post comes from a Lockdown Sceptics reader who has scrutinised a recent YouGov study about the numbers of people who know someone who died from COVID-19, and found that it doesn’t really make sense.

YouGov have published a study of how many people know someone who has died of COVID-19. They have asked people in 16 different countries.

At first glance the results look coherent – countries with higher death rates are at the top, ones with notoriously low death rates at the bottom. So 19% of Spaniards claim to know someone who has died of COVID-19 while only 2% of Chinese and Singaporeans do.

However, 2% of Chinese is 28,000,000 people. And China claims that only 4,600 people have died of COVID-19. That would mean that every Chinese individual who died of COVID-19 was acquainted with over 6,000 people. Is that a reasonable number?

Based on those poll numbers, how many acquaintances did COVID-19 victims in other countries have?

Well, here is the answer:

What to make of the results? There seems to be a big disparity between countries.

Are the Chinese the most sociable people on earth? Does a typical Australian or Indonesian have five times more acquaintances that your typical European? Are Mexicans the most personally affected by COVID-19 in the world ?

Or perhaps some countries are hiding deaths? China is always under suspicion of lying about its data. But what about Singapore or Australia? Are those governments hiding deaths too?

The most plausible explanation is something that many Lockdown Sceptics probably already suspect: that YouGov polls, many of which rely on panels of people to fill them out regularly, aren’t very reliable.

You might be tempted to go further.

If you click on the “See Full Results” link you will see some fascinating stats. YouGov will have you believe that 23% of Mexicans claim to have lost a family member.

But it seems YouGov doesn’t put much faith in the Mexican data so it has left it out altogether from the summary they publish and which is circulating in Twitter.

We can safely conclude that this particular YouGov poll should be taken with a pinch of salt.


2 Responses to “YouGov’s Numbers Don’t Add Up”

  1. Derek says:

    Died “of” covid is not a definitive. That’s one spanner in the works. It should be ‘with’ or ‘of’.

    Now compare two countries with the same population; Singapore & Denmark, or the UAE & Sweden. Massive difference in how many people each victim knew. This will be down to density and with density comes greater ‘hearsay’, not necessarily greater knowledge. Which pretty much blows the whole YouGov chart out the window for any useful comparison.

    I personally might know 86 people, but does a Chinese person ‘know’ 6,000 – or just know ‘of’ 6,000?

    Just a thought.

    • Derek says:

      Correction: my first line should have stated:
      “it should be ‘with’ or ‘from’.