Without looking at any numbers, is this not easier to explain as "those persistently absent are also more likely to not turn up in the first week of term"?
Surely even a minister should be able to grasp that?
Sure, I think that's fine too. But I think drilling down to understand a sizeable fraction might be caused by health issues makes it more clear that it's not simply a case of getting people to drag themselves in during Week 1 to solve the problem
Also, having been slightly disparaging of the article, I would say that it's pretty obvious that the DfE are effectively asking the BBC to do an article like this as a public service, to encourage parents to take the absence of their children from school seriously. Absence from school is a driver of poor educational outcomes, particularly for disadvantaged kids, and one that post Covid is much worse and a lot of that is probably down to parent attitudes. So while I don't like the statistics, I applaud the public service intent of the article. So it seems a bit of a shame to pick this as the one among so many possible examples of bad BBC correlation = causation examples!
Ok, but you'll get no argument from me that missing school is bad - and to be honest I've still got the scars down my back from trying to make that case in June 2020 and March 2021 when it wasn't particularly fashionable.
And I think if there's one other lesson we've learned from COVID, it's that trying to do well-intentioned vibe-based things with numbers that don't really stand up to scrutiny isn't a good strategy in the long run, for some of the reasons I talked about here https://bristoliver.substack.com/p/anti-vaxxers-dont-always-lie
agree completely. they could have achieved the "public service announcement" intent AND upheld their use of stats guidelines and it would have been all the better for it, both in the immediate article and the long term authority of the BBC
I have worked in media before and found that the people who slapped a headline on a piece of copy were not the same people as the ones who wrote the article itself. Is it possible the well-meaning training and style guide missed out whoever writes the snappy headlines?
It’s a fair point that the subs writing headlines are often to “blame” for this sort of thing. But to me the article is just as bad in this instance: lots of wording which is likely to imply causation to many readers but which is loose enough that it’s hard to really pin down and say it’s wrong.
This is one example of reporting a correlation and continuing the article as if they had just demonstrated causation. But this is everywhere! Like around 50% of health & lifestyle in all news outlets including the BBC.
Admiring your thorough analysis. But yes, this sort of statistical illiteracy does seem pretty much universal among all broadcast and print media. And they would lose a lot of lifestyle type content if it weren't. "Studies show that people who drink 5 glasses of red wine a week live up to 4 years 9 months and 3 days longer compared with those who don't drink at all. Mind you, the estimated figure was actually 3.2 years, we've added on the 1.5 years of uncertainty to make it more interesting for you. This does lead us to wonder what might be behind this correlation, and an obvious suggestion is that a substantial group of seriously sick people are included among the non-drinkers. It's likely that this entirely explains the reduced life expectancy in this group. So there's no reason to suppose that moderate drinking will cause you to live longer. No reason at all. And now, some non-news about absence rates in the first week of school."
Without looking at any numbers, is this not easier to explain as "those persistently absent are also more likely to not turn up in the first week of term"?
Surely even a minister should be able to grasp that?
Sure, I think that's fine too. But I think drilling down to understand a sizeable fraction might be caused by health issues makes it more clear that it's not simply a case of getting people to drag themselves in during Week 1 to solve the problem
If you want some much more detailed and well grounded analysis than what the BBC can muster on pupil absence you should look at FFT Education Datalab e.g. https://ffteducationdatalab.org.uk/2025/04/pupil-absence-in-autumn-and-spring-2024-25/
Interesting, thanks for the pointer!
Also, having been slightly disparaging of the article, I would say that it's pretty obvious that the DfE are effectively asking the BBC to do an article like this as a public service, to encourage parents to take the absence of their children from school seriously. Absence from school is a driver of poor educational outcomes, particularly for disadvantaged kids, and one that post Covid is much worse and a lot of that is probably down to parent attitudes. So while I don't like the statistics, I applaud the public service intent of the article. So it seems a bit of a shame to pick this as the one among so many possible examples of bad BBC correlation = causation examples!
Ok, but you'll get no argument from me that missing school is bad - and to be honest I've still got the scars down my back from trying to make that case in June 2020 and March 2021 when it wasn't particularly fashionable.
And I think if there's one other lesson we've learned from COVID, it's that trying to do well-intentioned vibe-based things with numbers that don't really stand up to scrutiny isn't a good strategy in the long run, for some of the reasons I talked about here https://bristoliver.substack.com/p/anti-vaxxers-dont-always-lie
agree completely. they could have achieved the "public service announcement" intent AND upheld their use of stats guidelines and it would have been all the better for it, both in the immediate article and the long term authority of the BBC
Off to a flying start! Excellent piece. Thank you.
Thanks!
I have worked in media before and found that the people who slapped a headline on a piece of copy were not the same people as the ones who wrote the article itself. Is it possible the well-meaning training and style guide missed out whoever writes the snappy headlines?
It’s a fair point that the subs writing headlines are often to “blame” for this sort of thing. But to me the article is just as bad in this instance: lots of wording which is likely to imply causation to many readers but which is loose enough that it’s hard to really pin down and say it’s wrong.
This is one example of reporting a correlation and continuing the article as if they had just demonstrated causation. But this is everywhere! Like around 50% of health & lifestyle in all news outlets including the BBC.
Ok, I won't bother then
Admiring your thorough analysis. But yes, this sort of statistical illiteracy does seem pretty much universal among all broadcast and print media. And they would lose a lot of lifestyle type content if it weren't. "Studies show that people who drink 5 glasses of red wine a week live up to 4 years 9 months and 3 days longer compared with those who don't drink at all. Mind you, the estimated figure was actually 3.2 years, we've added on the 1.5 years of uncertainty to make it more interesting for you. This does lead us to wonder what might be behind this correlation, and an obvious suggestion is that a substantial group of seriously sick people are included among the non-drinkers. It's likely that this entirely explains the reduced life expectancy in this group. So there's no reason to suppose that moderate drinking will cause you to live longer. No reason at all. And now, some non-news about absence rates in the first week of school."
I mean, I agree it's tilting at windmills to some extent, but I think it's worth doing once every so often just to make the point
Too much bullscheist. Experts seem to be failed performers.
And we are burdened BY HAVING TO RELY ON EXPERTS. THESE PEOPLE ARE FULL OF THEMSELVES. AND MASSIVE DOSES OF OVER SELF-IMPORTANCE
Because the whole point is trying to decide whether it's coincidental or not