Sleep: the silent killer
"Too much sleep is deadly" proclaimed the newspaper headline. Is this true?
Age
16 to 18
Challenge level
Newpapers often report on scientific announcements, especially if they are health related.
The Daily Mirror proclaimed on its front page on 7th August 2018:
TOO MUCH SLEEP IS DEADLY
More than 8 hours may be sign of ill-health as study shows 'long sleepers' die early
The online version of the article begins with a similar (but more detailed) headline and subheading:
Based on these headlines alone, what do you think the findings were of the study reported in the article?
Here are two more extracts from the article:
Do these two extracts match up with what the headlines said?
The actual study is available here: it is a meta-analysis of other studies, which means the researchers have looked at all of the earlier studies they could find and collected all of the relevant results together to give a huge sample size with (one hopes) more reliable results. They summarise their conclusions (bottom of page 3):
("Higher risk of mortality" means "more likely to die", and "cardiovascular events" means "heart attacks and other nasty things to do with the heart or blood vessels".)
The researchers go on to give a clinical implication of their study (page 4):
This resource is part of the collection Statistics - Maths of Real Life
The Daily Mirror proclaimed on its front page on 7th August 2018:
Image
TOO MUCH SLEEP IS DEADLY
More than 8 hours may be sign of ill-health as study shows 'long sleepers' die early
The online version of the article begins with a similar (but more detailed) headline and subheading:
Getting too much sleep could be deadly as study warns more than ten hours a night increases stroke risk
Researchers have urged GPs to check the sleep patterns of patients after shocking study results
Researchers have urged GPs to check the sleep patterns of patients after shocking study results
Here are two more extracts from the article:
Researchers said underlying diseases could be the cause of regular lie ins - and they urged GPs to check the sleep patterns of patients during visits in a bid to save their lives.
“Our study has an important public health impact in that it shows that excessive sleep is a marker of elevated cardiovascular risk. Our findings have important implications as clinicians should have greater consideration for exploring sleep duration and quality in consultations."
Do these two extracts match up with what the headlines said?
The actual study is available here: it is a meta-analysis of other studies, which means the researchers have looked at all of the earlier studies they could find and collected all of the relevant results together to give a huge sample size with (one hopes) more reliable results. They summarise their conclusions (bottom of page 3):
Divergence from the recommended 7 to 8 hours of sleep is associated with a higher risk of mortality and cardiovascular events. Longer duration of sleep may be more associated with adverse outcomes compared with shorter sleep durations.
("Higher risk of mortality" means "more likely to die", and "cardiovascular events" means "heart attacks and other nasty things to do with the heart or blood vessels".)
- What does the word "associated" mean here?
- How does this conclusion compare to the newspaper headline?
- How does this conclusion compare to the rest of the newspaper story?
Clinical implications
The researchers go on to give a clinical implication of their study (page 4):
Our study suggests that abnormal sleep is a marker of elevated cardiovascular risk - and greater attention in consultations with patients to ask, and give advice about optimising the duration and quality of sleep may help reduce the burden of cardiovascular disease.
- What does this suggest the authors of the scientific paper believe? Does this fit better with the newspaper headline, with the report conclusion, or neither?
- If the data only shows that there is an association between a longer duration of sleep and cardiovascular disease, but not that a longer duration of sleep causes cardiovascular disease, is advising people who sleep more than 7 to 8 hours to sleep less a good thing, a bad thing or neither?
The online version of the article can be found on the Daily Mirror website.
This resource is part of the collection Statistics - Maths of Real Life
Here are some questions you could ask yourself to get started.
Consider just the headline and subheading.
Consider just the headline and subheading.
- Without reading more of the article, what does the headline make you think? How does it make you feel?
- What do you think the newspaper editor was trying to do by writing the headline like this?
- Read the rest of the newspaper article. Based on the article, what do you think the study is saying?
- To what extent does either headline fairly represent the rest of the newspaper article?
- What does the word "associated" mean here?
This does not, though, say that one causes the other, and the rest of the study (and some of the newpaper article) says this explicitly: the long amount of sleep could be indicative of some underlying health problem.
- How does this conclusion compare to the newspaper headline?
- How does this conclusion compare to the rest of the newspaper story? (You could analyse every paragraph if you have time to.)
- ("While...") This paragraph hints that this is causative: the language "loves a lie in" suggests that this is a conscious choice to sleep in late. The study seems more concerned with the average amount of sleep participants are getting rather than an occasional lie in. One would have to lie in a huge amount to average over 10 hours of sleep per night! With over 3 million people in the study, it is likely that the majority of those sleeping over 10 hours per night on average were not doing so by choice.
- ("A major study...") This explicitly says that long periods of sleep cause an increased mortality risk, which is not what the study says.
- ("Researchers said...") They did say this, though they weren't as dramatic as "in a bid to save their lives".
- ("The study...") This was not mentioned in the conclusion, but is in the rest of the study.
- ("Lead researcher...") This matches the study's conclusion, as long as we understand "is a marker" to mean something like "indicates" or "may indicate"; it is certainly associated with "elevated cardiovascular risk", so whether it means there is an increased risk or just that there may be an increased risk is somewhat nuanced.
- ("Our findings...") This is not mentioned in the study's conclusion, but is a more nuanced version of paragraph 3.
- ("If excessive...") This is a continuation of the previous paragraph.
- ("The researchers...") The phrase "got too much sleep" is making a value judgement, whereas the study's conclusion does not. Other than that, this matches the report's conclusion.
- ("One, over 11 years,...") This is a detail from the study which we cannot judge from the study's conclusion alone. It's somewhat imprecisely worded, though, as everyone will die eventually. Also, since that study has finished, the language of probability is not quite correct. It would be better to say something like "In one study considered by the researches, the proportion
of people who reported sleeping for more than nine hours on average who died during the 11 years of the study was 25% greater than for the rest of the participants." But that's a bit of a mouthful.
Also, what does 25% greater mean? If only 1% of the participants died during the 11 years, so 4 out of every 400 people, then this would mean that 5 out of every 400 long sleepers died during the 11 years, and that would not necessarily be a dramatic effect. - ("The review...") The conclusion said "Longer duration of sleep may be more associated with adverse outcomes compared with shorter sleep durations," so it suggests that there may have been adverse outcomes with shorter sleep durations too. But we would have to read more of the study to see whether this is what was actually meant.
- ("Dr Kwok...") This explains the research purpose, but does not give any results.
- ("We further...") As with the previous paragraph.
- ("There are cultural...") A probably uncontrovesial statement about sleep.
- ("These include...") A description of modern life, expanding on the previous paragraph; again, this doesn't directly relate to the results of the study.
- ("Some have...") This suggests that a longer duration of sleep causes problems, which is not what the study's conclusion says. It may be the case, as these people have previously suggested, that there is a causal link, but this paragraph ties back to the headline which says that this link is causal.
- ("Adults who...") This does seem to fit the wording of the study's conclusion, but contradicts paragraph 10 of the newspaper article. (And what does "at risk of dying" mean?)
- ("But until now...") This is not mentioned in the study's conclusion; it may or may not be the case.
- ("The review...") This is a correct statement: it was published here on 3rd August 2018.
Clinical implications
- What does this suggest the authors of the scientific paper believe? Does this fit better with the newspaper headline, with the report conclusion or neither?
- If the data only shows that there is an association between a longer duration of sleep and cardiovascular disease, but not that a longer duration of sleep causes cardiovascular disease, is advising people who sleep more than 7 to 8 hours to sleep less a good thing, a bad thing or neither?
Why do this problem?
This problem connects the technical side of doing statistics to the human side of communicating the results effectively. A very important citizenship skill is the ability to critically question reports and articles. In this problem, we take the research itself as solid (as the skills needed to evaluate that are far more technical in nature), but focus on whether the research itself is accurately represented or distorted.
Secondarily, this problem also develops literacy skills, by looking at both technical material (the research article) and non-technical material (the newspaper). It could be interesting to work with colleagues from the English department to think about how this issue relates to persuasive writing. For example, one might ask at what point persuasive writing crosses the line into dishonest writing. Alternatively, how could this research be presented in an engaging and accurate way?
The particular article discussed here was chosen because it appeared at the time this problem was being written, but there is nothing unusual about it: articles like this one appear in the press and online on a regular basis.
Possible approach
Students will either need access to the internet to read the online newspaper article which can be found here, or they will need the article or the excerpts from the problem printed out. A few quotes from the research article have been reproduced in the problem, but if students want to read them in context, they will again require access to it.
Students could be asked to look at the print headline and subheading and to write down what they think the key result of the study is. They could then be shown the online version and asked whether they still think the key result is the same or whether their ideas have changed. Then reading the rest of the article, this question can be posed a third time. How were their ideas different on each occasion, and what has caused this?
After reflecting on this, students should then consider the study's conclusion and reflect on how it relates to the way the study's results have been communicated.
A further question which might be considered is why the newspaper editors might have made these decisions.
If time permits, students could go on to the "Clinical implications" section and consider the further questions there.
Key questions
- How do newspapers or other news sources misrepresent scientific research?
- How can we check whether we should act on or be concerned about a significant news report?
Possible extension
Students could be asked to look at a newspaper or news website for a few days to find an interesting-sounding scientific article. Once they have found one, they should analyse it to find out how well the scientific research has been represented. If they can find the original research paper that the article is based on, if there is one, they will also be able to compare the report to the newspaper article.
Possible support
To help them get started, students could be asked scaffolded questions to help them consider the headline and subheading in detail, for example:
- Without reading more of the article, what does the headline make you think? How does it make you feel?
- What do you think the newspaper editor was trying to do by writing the headline like this?