![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Bad Science
Mar. 13th, 2009 11:49 amThere's a science story doing the rounds and the way that it is being reported is really pissing me off. Here's an example from the Telegraph Vegan Diet Increases risk of birth defects, scientists warn.
What's wrong with this story? Well, the headline and the fist paragraph are both flat out lies. You can tell this because when you actually read the news story you find that they never actually quote any scientists saying that a vegan diet increases the risk of birth defects. If any scientists had actually said that, and it's their headline, wouldn't they, ya know, actually provide any evidence that this were the case. Instead they quote that B12 deficiency in early pregnancy increases the risk of birth defects and a scientist recommending that B12 supplements be recommended to women who are likely to conceive in the same way that folic acid is. The closest the Telegraph gets to substantiating its headline is to point to that B12 is found in "meat, eggs and milk" conveniently forgetting to mention that it's also found in yeast extract and lots of foods marketed to vegan are fortified with it. In fact, B12 deficiency is very rare even among vegans, because you only need a tiny amount, and tends to effect people past childbearing age. Furthermore, when studies have been done on pregnant vegans they show no worse outcomes, and young babies tend to do a bit better than average because a higher proportion of vegans breastfeed compared to omnivorous mums.
This article is following one of the classic Bad Science Reporting formulas. You get a press release with some findings which had been indicated by previous studies and has some public health implications yawn yawn. Then you realise that you can turn it into a Story by connecting the implications to a convenient scapegoat. Got a story about sunlight exposure and vitamin D? Make it about burqas. Got a story about B12 deficiency? Make it about vegans. Write a sensationalist headline blaming your chosen scapegoat and then cunningly use sub-clauses to make your connection without quite putting words in the researchers mouths.
I'd usually blame this on journalists being scum but I'm a bit suspicious. This story was reported with the same spin in several places. I'm a bit suspicious the the sexing up may have occurred before it crossed the journo's desk. A couple months ago there was a news story everywhere about research which might lead to a prenatal test for autism and how we should have a discussion about the ethical dilemmas such a test might present. I friend of mine who works in medical research in Cambridge later told me that the actual research that the story was loosely based upon a) confirmed findings which had been sort of found before an b) only showed quite weak relationship.* However, if you're related to Ali G, weakly confirming a relationship which we already sort of knew about and has no clinical applications doesn't have to get in the way of being in the international press. So the research group called lots of media outlets, fed them the abortion spin and the press jumped on the story.
This kind of spin is annoying because a) it's dishonest and b) it misleads the public about information which might be important to their decision making. The Telegraph headline is irresponsible because it will lead to more pregnant vegans getting hassle about their diet and omnivorous women thinking that B12 efficiency is something that they're immune to. A much better headline would be "Scientists recommend women take B12" or "B12 deficiency associated with birth defects". Those headlines would have been honest and might have actually changed people's behaviour in a way that might improve their pregnancy outcomes.
Edit: Here's a link to the abstract of the study. Some interesting points to note:
It doesn't mention vegetarians or vegans at all, they measured B12 deficiency through blood samples during pregnancy rather than looking at diet.
It states that previous studies have established a link between B12 deficiency and birth defects and this study only aimed to quantify the already known relationship. So an honest headline would read "Link between B12 and birth defects measured more accurately than in previous studies".
*As they described it there was a scatter graph with a cloud of dots that were sort of upward sloping enough that you could fit a line but not enough to mean that any level of testosterone couldn't be seen with any level of autism.
What's wrong with this story? Well, the headline and the fist paragraph are both flat out lies. You can tell this because when you actually read the news story you find that they never actually quote any scientists saying that a vegan diet increases the risk of birth defects. If any scientists had actually said that, and it's their headline, wouldn't they, ya know, actually provide any evidence that this were the case. Instead they quote that B12 deficiency in early pregnancy increases the risk of birth defects and a scientist recommending that B12 supplements be recommended to women who are likely to conceive in the same way that folic acid is. The closest the Telegraph gets to substantiating its headline is to point to that B12 is found in "meat, eggs and milk" conveniently forgetting to mention that it's also found in yeast extract and lots of foods marketed to vegan are fortified with it. In fact, B12 deficiency is very rare even among vegans, because you only need a tiny amount, and tends to effect people past childbearing age. Furthermore, when studies have been done on pregnant vegans they show no worse outcomes, and young babies tend to do a bit better than average because a higher proportion of vegans breastfeed compared to omnivorous mums.
This article is following one of the classic Bad Science Reporting formulas. You get a press release with some findings which had been indicated by previous studies and has some public health implications yawn yawn. Then you realise that you can turn it into a Story by connecting the implications to a convenient scapegoat. Got a story about sunlight exposure and vitamin D? Make it about burqas. Got a story about B12 deficiency? Make it about vegans. Write a sensationalist headline blaming your chosen scapegoat and then cunningly use sub-clauses to make your connection without quite putting words in the researchers mouths.
I'd usually blame this on journalists being scum but I'm a bit suspicious. This story was reported with the same spin in several places. I'm a bit suspicious the the sexing up may have occurred before it crossed the journo's desk. A couple months ago there was a news story everywhere about research which might lead to a prenatal test for autism and how we should have a discussion about the ethical dilemmas such a test might present. I friend of mine who works in medical research in Cambridge later told me that the actual research that the story was loosely based upon a) confirmed findings which had been sort of found before an b) only showed quite weak relationship.* However, if you're related to Ali G, weakly confirming a relationship which we already sort of knew about and has no clinical applications doesn't have to get in the way of being in the international press. So the research group called lots of media outlets, fed them the abortion spin and the press jumped on the story.
This kind of spin is annoying because a) it's dishonest and b) it misleads the public about information which might be important to their decision making. The Telegraph headline is irresponsible because it will lead to more pregnant vegans getting hassle about their diet and omnivorous women thinking that B12 efficiency is something that they're immune to. A much better headline would be "Scientists recommend women take B12" or "B12 deficiency associated with birth defects". Those headlines would have been honest and might have actually changed people's behaviour in a way that might improve their pregnancy outcomes.
Edit: Here's a link to the abstract of the study. Some interesting points to note:
It doesn't mention vegetarians or vegans at all, they measured B12 deficiency through blood samples during pregnancy rather than looking at diet.
It states that previous studies have established a link between B12 deficiency and birth defects and this study only aimed to quantify the already known relationship. So an honest headline would read "Link between B12 and birth defects measured more accurately than in previous studies".
*As they described it there was a scatter graph with a cloud of dots that were sort of upward sloping enough that you could fit a line but not enough to mean that any level of testosterone couldn't be seen with any level of autism.
