Academic Urban Legends | Hacker News Reader

Academic Urban Legends

sss.sagepub.com

129 pointsdougmccune57 comments

There are an amazing number of scientific urban legends in the lab too that are passed down orally from senior to junior members. My personal favorite is that drinking MilliQ (ultra pure water) will cause the cells in your gut to explode [1]. How such a stupid idea ever got taken seriously by scientists who should know better has always amazed me.

[1] On the topic of drinking MilliQ water I have always encourage all the members of my lab to drink it. The reason why (apart from the taste which I quite like), is that it allows you to detect when you have bacterial contamination of your MilliQ carboy that will ruin your experiments.

Yup, I heard the gut-exploding-ddH2O in a lab and a girl in the class eventually did drink a little bit of it and was absolutely fine. A science champion, if you ask me.

I have had people try to physically stop me from drinking MilliQ water as they thought I was going to die in front of them!

I have always thought it would be a good business idea to sell MilliQ branded water for human consumption. The complete lack of taste is an interesting taste in itself - the closest I can describe it as is it is like drinking liquid air.

I've heard that handling ultra pure water, can cause acid like burns due to water drawing out chemicals from the skin. I heard it from a bio PHD, and can't find any evidence googling it now.

This one is new to me - even more stupid than the drinking myth if that is even possible.

Well, it might have an element of truth. It is Hydric Acid, after all.

It is fairly well known that water will dry out the skin by leeching the oils.

(edit: and when I say element I mean microscopic element)

Actually most water is a dilute solution of carbolic acid with a pH around 5.5 because of the dissolved CO2 in the water.

This sometimes catches people out when they use pure water to store something like DNA which will breakdown over time at this pH. It is in theory why you should use a buffer like Tris/ETDA to store DNA. I have found that storing DNA for a short time in water is normally OK, but if you want to keep the DNA for more than a month it is a good idea to use a buffer.

You probably mean carbonic acid. Carbolic acid is an organic compound also known as phenol.

Do you mean you detect contamination by taste? Is MilliQ that pure?

Yes you can detect contamination in MilliQ by taste. I don’t know the threshold, but I do know that if it has a taste then the water is contaminated enough to ruin your experiments.

Are the contaminants dangerous? Or just normal microbial flora that you'd get on kitchen glassware?

No they are not dangerous. Pathogens are adapted to living in nutrient rich environments (i.e. our bodies) where their problem is avoiding our immune system. The bacteria that can grow in MilliQ water are those adapted to living in low nutrient environments and these bacteria are always harmless - except that is to your experiments :)

Caulobacter vibroides

Wouldn't the very act of drinking it contaminate that source with bacteria from your lips?

I don't work in a lab, but I assume you'd pour it into a cup/beaker and drink out of that

Don't drink it directly out of the container...

I've heard people in my lab make similar claims about regular distilled water -_-. Even when I drink it in front of them they still have doubts.

Don't they sell distilled water in the supermarket? How do they figure supermarkets can get away with selling something that dangerous?

Supermarkets also sell things like salt, tobacco and alcohol which will kill you if consumed in sufficient quantity

And Hot Pockets.

That particular legend is way beyond the lab now:

http://www.fastcompany.com/1750612/dangerously-clean-water-u...

I love the “informed comments” on this article too. How do these people not melt when it rains?

Based on the fact that you didn't mention doing it, I am assuming you're missing out on a brilliant opportunity for hijinks here. Suggestion - take an alka-seltzer and some red food dye (I'm not sure if you can drip food coloring directly onto an alka-seltzer slowly enough that the dye is absorbed without causing the alka-seltzer to foam up, but probably you can experiment with something like this) and hide it in your mouth ahead of time. Loudly proclaim that it's harmless to drink MilliQ. Take a big beaker of MilliQ and drink it, allowing plenty of it to reach the alka-seltzer. Pause for half a second, start to look uneasy, grab your stomach, then start allowing blood-red foam to spill out of your mouth. Fall to the ground shaking and moaning.

Bonus points if it'a "demo" you do while teaching a lab class.

Sounds like fun :)

I has a colleague who used to like putting dry ice bombs in people’s tip waste containers. You would be sitting there pipetting away and them boom - tips all over the bench.

I guess the original publication which listed the high iron content of spinach was:

Bunge, G.. 'Weitere Untersuchungen über die Aufnahme des Eisens in den Organismus des Säuglings'. Zeitschrift für Physiologische Chemie, 16 (1892)

http://echo.mpiwg-berlin.mpg.de/ECHOdocuView?url=/permanent/...

If that's true, then indeed, it was not a misplaced decimal point. Instead, Bunge listed the iron contents of dried food.

It is interesting that the paper does not actually cite the original paper. I guess it is not critical to the paper, but it does complete the story.

This does bring up a important point in that is old scientific publications in languages other than english. I have many times run into references in German, French or Russian that are only available as low quality scanned pdfs. As a monolingual English speaker these are completely inaccessible to me. I would be great to have all of them OCRed so at least we can run them through Google Translate.

It seems more interesting to me that Rekdal doesn't cite this article:

http://www.bestthinking.com/articles/science/chemistry/bioch...

It's a 2012 followup by Mike Sutton to the two publications from 2010. It contains many new details, including a reference to Bunge (1892).

Perhaps, one should not only try to cite the original source, but also look for recent research on the same topic?

>Perhaps, one should not only try to cite the original source, but also look for recent research on the same topic?

This is (or should be) done standardly by all scientists. It is actually the best way to learn about a field. Find the key paper (or at least the paper the field thinks is key) and then read all the papers that cite that key paper.

I am sure this has been done by somebody, but it would be great to have this citation link structure as a graph where you could click on any node and read the paper.

The problem is that key papers are cited by a LOT of other papers, these with variable quality.

Yes they are, but there is no better way of finding those papers that can’t be found using keyword searches. I have found hundreds of critical papers over the years by using this technique that I would never have found in a thousand years using any other search method.

Take a closer look, Jannic. I did cite that page, created in 2010. The version I referred to was from 20 March 2014 (date also found in the list of references, in line with the requirements of this particular journal)

A small anecdote on this point. I first sent the article to a journal where the editors insisted that I say something about Bunge and the other German scientists involved in the early history. I refused, because it would have taken attention away from the main topic of the article, and sent the article to SSS instead.

Wow. This academic paper we just read on academic urban legends has an urban legend payload - I now believed there was a "misplaced decimal point" (i.e. typo) in a single source, that then became the source of an academic urban legend about spinach.

It'd be interesting if we could break down academic papers into stand-alone facts, which could be cited more succinctly than the paragraphs of prose that we do now. Ideally, this would look more like academic proofs, and it'd be easier to argue and discuss components of a given paper. I'm sure that a good amount of beautiful subtext would be lost, but might it be worth it to have a more accurate way of linking knowledge?

This could now be accomplished by providing an offset (say number of characters from the start) to where the information cited can be found.

What I find more of an issue is when trying to track down the source of some “fact” is going three or four papers up the citation chain and finding the paper at the end doesn’t actually say what the starting paper claims.

Thanks for summarizing the OP.

This sounds like a problem that hyper-links should have solved.

> It'd be interesting if we could break down academic papers into stand-alone facts, which could be cited more succinctly than the paragraphs of prose that we do now.

Perhaps there could be some kind of semantic syntax that would group together the factual claim and supporting information (data and references). For example,

<"e=mc^2",cite1,cite2,data1,data2>

The claims still could be included in prose, but also easily extracted, analyzed, manipulated, etc.

With a little work the claims could form a semantic web, with all that entails, and the validity of cites could be checked automatically when the document is opened or on a regular basis, updating the claim appropriately. For example, the UI could display the claim differently based on whether the cited source has been withdrawn, partially invalidated, superseded, confirmed, rarely cited, heavily cited, based on impact factor, if the source's own citations have been invalidated (e.g., claimC cites claimB which cites claimA; if claimA is invalidated, there should be some sort of indication in claimC -- consider that the invalidation of a seminal claimA would set off a massive chain reaction, as it should), etc.

But that all seems obvious and I suspect I'm not the first to think of it.

It is indeed interesting, and as you suspect it has attracted some attention. Check out "nanopubs,"(http://nanopub.org), or Simon Buckingham Shum's very cool work on ScholOnto (http://projects.kmi.open.ac.uk/scholonto/) among others.

Microformats were popular back when blogging was new. [1]

http://microformats.org/wiki/citation

That approach has been tried and in the bioinformatics community years ago. Basically scanning published papers that mentioned genes relationship and extracting them in order to build a graph. Say someone publishes "T3SS1 induces cell death" so you can now add an edge to a graph "T3SS1->autofaghy" or something like that.

Please, no. The twitterification of academic papers would be terrible - all subtext and nuance would be lost. This is seen already with 'science journalism': a scientist reports an effect, couched with appropriate caveats, but a journalist then rips that effect free of any nuance or context and states it as fact. It can be highly (and sometimes intentionally) misleading.

Exactly, and it's not like papers are collections of 'facts' that can be extracted anyway. Papers are long arguments based on some data choices, some analysis choices, and followed by interpretations --- none of which are 'facts', and all of which are debatable.

I am thinking in the exact opposite direction: stop this publication influence, citations SEO and plagiarism madness.

Get shit done and put it on your blog.

Normal people hear ideas, forget them, internalize them and regurgitate them as their own all the time. While you're scouring databases and calling libraries to try to find who's got access to the paper you're looking for because it looks like a primary source of what you're working on, you could be experimenting from secondary sources or even from textbooks if the thing is now mainstream enough.

OK what if the experimental setup is hard to replicate and I just make up attractive sounding claims for for my blog with flashy graphics and a folksy charming writing style?

what's the difference with today? You didn't fight for the author's order and you didn't spent months on publication issue, but your bullshit is still in the wild.

If you want to be in Nature today, you need attractive claims and flashy graphics already. And the peer review is just not working, they are favoring already famous people and ideas.

> If you want to be in Nature today, you need attractive claims and flashy graphics already. And the peer review is just not working, they are favoring already famous people and ideas.

Citation?

While I can imagine that being amazing if that could be done well, but it seems like a really difficult problem to get right.

I would prefer an approach like PaperBricks caught on instead: http://arxiv.org/abs/1102.3523 -- where a single section of a current paper could be published by itself.

I think it would be both amazingly useful and could speed up the iteration rate of science substantially.

Be sure to read to the end, or, after you read enough to grasp the idea then skip down to the section "The narrow path has its temptations".

Yeah, my first thought was, "wow, I always thought that about spinach!" But, that wasn't the point of the article at all. This was an engagingly written article that kind of read like a mystery whodoneit?

And the most concerningly satisfying is that everybodydoneit, apparently. Including me, who didn't bother to read any one single paper cited.

Even better, read the section "An avalanche of low-quality research?" near the end. Even this notion is put to the test, and found wanting :)

Excellent article!

Thanks a lot! :)

This article made me think of The Leprechauns of Software Engineering[1], a good book about the myths that are being passed as facts because of bad citations in Computer Science papers.

[1] https://leanpub.com/leprechauns

> The explanation for this phenomenon is usually that authors have lazily, sloppily, or fraudulently employed sources, and peer reviewers and editors have not discovered these weaknesses in the manuscripts during evaluation.

I've seen this sort of thing in the literature I'm familiar with (HF Radar). I don't think it's laziness or fraud, but a version of the telephone game played as a kid. A 'fact' will get stated, and then repeated over and over until some of the original meaning is lost and it becomes gospel. This can happen because the fact is not crucial to the science in the paper but always part of the introductions or statements about how the system works.

The statement I'm thinking about in particular is "if you have N receive antennas, you can only find (simultaneously) N-1 signals". This is true if you are using covariance-based methods such as Multiple Signal Classification (MUSIC). It got repeated and repeated until the caveats (simultaneous, and MUSIC) are longer included. Of course, if you have two orthogonal antennas you can distinguish two directions, perhaps not in complex situations!

So that meams you can't get anything useful out of one antenna?

If it's directional, you can get one direction. Which disproves the N-1 meme that is in so many papers.

Reply on news.ycombinator.com