Syndicate content

Annals of Misleading Statistics: Literacy 83

David McKenzie's picture

On the World Bank’s today page today I saw the following:

This seemed really high to me, and a strange way of presenting statistics. Following the link, it directs you to this World Bank Data Viz Tumblir which has a bunch of statistics all presented in the form, if the World had only 100 people, then…

The first thing that bugged me was this way of presenting statistics: what would it mean if the World had only 100 people – would they all be in one country, or distributed throughout the world? If they were throughout the world, then how would they be taught to read without anyone else around? Basically this seems a bad way to present percentages – why not say instead 83 out of every 100 people in the world can read…

But as I said, this seemed way too high – think of all the babies who can’t read for a start. So I followed the sources further (I commend the site for at least documenting where their numbers come from), and found a link to this UNESCO report which reports that in fact the total adult literacy rate in 2009 was 83.7%. So a rounding error, on top of using a statistic for a subpopulation (adults) and presenting it as if it were for the world. Now we all know these global estimates on top of that involve substantial uncertainty and estimation error, so I won’t even hazard a guess on what the confidence interval around this 83 is for adults.

So while I applaud efforts to try and make development data more accessible, let’s make sure it is used accurately, and I don’t think much of this way of presenting percentages.


Submitted by Charlie on
That is the tradeoff isn't it? Difficult to make information accessible, interesting and afford it the nuance and context that it deserves at the same time. Thanks for pointing this out!

Submitted by Stuart on
Shows the need to simplify the right things when communicating data; I suspect most people likely to come across that blog can understand percentages pretty easily (though some might struggle with percentage changes and changes in percentages, etc.) so there's no need for all this 'If the world had...' stuff, which also comes across to me as rather patronising. The UNESCO Global Monitoring Report on Education for All has a lot more detail on literacy - (see p. 95). This uses the same data and gives a global figure of 84% for adult literacy (data from 2005-2010 depending on country), but admits that the data is based on self-reported literacy in household surveys, so is very likely to overestimate the true level.

Add new comment