Wednesday, May 27, 2015

IFU Science

Note: This essay is entirely opinion. There's nothing scientific about it. I'm going to ramble a bit. AND I'm going to use bad words. F-words. Lots and lots of F-words. And it will probably change as I think about it some more

Today's label isn't for anyone in
particular. It's for all the researchers
who authored all of the studies with
all of the bogus results as reported
in The Lancet.
Getting Slashdotted today is the following question: Can Bad Scientific Practice Be Fixed?

Let's start off answering this directly: Duh. Of course it can.

Bad practices can always be corrected. The key to doing that is examining and teaching good practices. Of course, what we'd like to do instead is ask how the bad practices came about in the first place. To do that, we need to know what the problem is; and that's that as many as half of the 'scientific' studies out there are simply wrong. The British Journal The Lancet posits that 'something has gone fundamentally wrong with science' [PDF].

This is, to put it bluntly, bullshit.
(Before I move on let me point out that The Lancet is a medical journal, and the observations made in the linked paper presumably refer to the practices of medical researchers. However, I'm looking at it in a broader context, as I've noticed some trends that have made me uncomfortable for a while now. In fact, several previously drafted topics found their way into the essay you're now reading.)
What has happened is that something is fundamentally wrong with the scientists that we churn out. Scientific practices aren't wrong; rather, they're not followed. The Lancet points out that the scientists are not 'incentivised' to be right; as if a scientist should need incentive to strive to be right. Instead, it appears, these people we call 'scientists' are being paid to produce sensationalist popular fiction.

And there's the key word: 'popular'.  Not as in 'widely accepted'. No... more like 'entertaining' and 'socially relevant'. Well, any true sociologist will tell you that sociology has very little to do with hard science. Popularization has nothing to do with valid results. But we're caught up in a "science bubble" where innovation is aggrandized and overvalued, not terribly unlike any other "bubble" that eventually leads to hardship and correction.

The Cult of Science

To me, this is well illustrated by the phrase "I Fucking Love Science".

I really don't want to trash on the website, because it's their mission to report on the latest in science. To some degree they play the hand they're dealt. But the way they and others do it... at least part of the time... is part of the problem.

via Wikimedia Commons
Let me give you an example. I picked this story off their home page today because I've spent a lot of time armchair-researching pyramids, so it interested me. Amazing Drone Footage Of Sudan's Mysterious & Forgotten Pyramids. The first thing you should notice is that this isn't really about the pyramids. That's something that National Geographic will talk about at a later date. No, it's about flying RC helicopters with cameras on them. They're called 'drones' now out of a desire for coolness, but people who are older than 12 should know that they are simply radio controlled helicopters. And that's the story... that 'drones' are being used to take pictures. Now, if we're very lucky, the more serious people at National Geographic will ignore the tools and focus on what they're trying to do with them.

And yes, I noticed that there is a little smidgeon of info about the pyramids so you know where on the globe they might be found. But even this is in horrible English. Observe: "Standing tall for over 3,000 years, National Geographic has used drones to newly explore the tops of these little-known pyramids..."  Sorry, but National Geographic has not been standing tall for over 3,000 years. And yes, I'm going to be a grammar nazi at the moment, because grammatical errors (and I've made some whoppers) are more easily forgiven if there's some message beyond "l have a really cool camera".

Actually, that's part of my point. If something is recorded with a drone, the story is automatically about the drone. It's not just a problem in science, either. Engineering suffers from the same failing. For instance, if something is made with a 3D printer, no matter how trivial it would have been to simply sculpt it, then the story is about the 3D printed widget, emphasis on the 3D printing. 3D printing has become a magic phrase. This has come to such a head that NASA has announced a competition to award a prize for the best 3D printed lunar habitat. Now, if it were me, I'd want the best, most innovative habitat. Full stop. This would allow teams to get truly creative and come up with something that's economical, utilitarian, and opens the door to 'outside the box' thinking. Not here. NASA simply defines the box and forces the would-be competitors into it. Pretty sad. It's actually sadder when you realize how the love of a buzzword can broaden a definition into near-unrecognizability. By current standards a popcorn ceiling might be considered '3D printed', just as an RC helicopter is now a 'drone'.

3D printed lunar habitat. 3D 'printer' shown on right
illustration by Jeremy Liu via

Actually, engineering suffers from the further failing of being constantly mistaken for science. A lot of what people 'fucking love' is stuff, not science.

Mammals and Cows

People who 'do science' are scientists. Now, here's a word that carries almost zero information, so I'm increasingly wary when I see it. In various contexts with various accuracy the word 'scientist' can refer to an agronomist, an anthropologist, an archaeologist, an astronomer, a botanist, a chemist, a climatologist, a crackpot, a criminologist, a cytologist, an economist, an engineer, an epidemiologist, an ethologist, a geneticist, a geologist, a geographer, a marine biologist, a mathematician, a meteorologist, a microbiologist, a paleontologist, a physicist, a phytologist, a politician (seriously), a programmer, a psychologist, a roboticist, a 'rocket scientist', a seismologist, a sociologist, a zoologist; among many, many more.

The thing is, consulting an astronomer is pretty useless if the topic is cytology. Nevertheless, that's exactly what the media do. This is pretty much guaranteed to confuse a lay audience about the concept of 'domain'. An expert is only an expert within his domain. Pick a popular science program and you'll see the same handful of famous faces picked for name recognition rather than expertise in the subject at hand: Neil DeGrasse Tyson, Michio Kaku, etc.. These are highly respected in their fields; but are often called upon to opine on subjects well outside of their fields. While their broad experience gives them familiarity with a number of disciplines, it's still necessary to differentiate those times when they are speaking as experts from those times when they're speaking as people who fucking love science. This is rarely done. Isn't it possible that they set the bar for that budding scientist who grows up believing that the road to superstardom and fortune is that sensational discovery that puts his name in the paper and his face in front of a camera?

To further muddy the waters, the Science Council, a UK organization, lists 10 kinds of scientist as follows:
  1. Explorer
  2. Investigator
  3. Developer/Translational
  4. Service provider/operational
  5. Monitor/regulator
  6. Entrepreneur
  7. Communicator
  8. Teacher
  9. Business/Marketing
  10. Policy maker
I think they're improperly expanding a label that only applies to the 'S' in 'STEM' with all of STEM even more egregiously than did the list I gave earlier. From the point of view of communicating the scientist's actual credentials, this is useless.

Using the generic term 'scientist' is like using the word 'mammal' when what you mean is 'cow'. It's technically correct, but it doesn't tell you a damned thing about the milk.

In popular culture we cultivate a Gilligan's Island stereotype of the 'scientist' as inaccurate as any 1950s B-movie portrayal. And this is how scientists portray scientists. Is it any wonder that the science itself has gotten sloppy?

If you ask me, we need fewer people who fucking love science, and more people who fucking understand science.

Putting it in practice

We do have, in my completely unsubstantiated opinion, throngs who just fucking love science, who don't properly employ the scientific method, and who are competing with actual scientists for funding. Since their sensationalism actually works, scientists are forced to modify their tactics to compete. As The Lancet notes, "...scientists are incentivised to be productive and innovative" as opposed to being right. The words "productive and innovative" give me pause.

It seems intuitively obvious to me that at some point in every topic our knowledge will have been optimized. At that point, every innovation can only make things worse. Furthermore, the closer we are to optimized knowledge, the fewer useful innovations remain. The upshot of this is that the more we know about something, the more a new idea about it is likely to be a bad idea. This is one of the reasons why 'revolutionary' ideas have such steep obstacles to acceptance: they have replace all of the knowledge we've already amassed.
Example: Tic-tac-toe (noughts and crosses) is a solved game. We know everything there is to know about it. On his opening move X must play the corner. Everything else is a bad idea. Furthermore, between two strong players there is never a winner. There can be no useful innovation in gameplay according to these rules. 
 Our knowledge doesn't have to be complete to be optimal. Optimal is simply "the best understanding given the facts at our disposal". That which is optimal may change, given new facts. Optimal can also mean that for all practical purposes our knowledge is complete even when it factually isn't. Thus, even though Relativity was a revolutionary revision in our understanding of Gravity, it has no practical effect on the study of ballistics, where Newtonian formulae are still applicable.


Simply having a bad idea isn't unscientific. The scientific method is supposed to catch them. The problem lies in publishing results that are unsubstantiated or irreproducible. You should have been very careful in constructing your tests so as to ensure this doesn't happen. If your carefully constructed tests don't yield the results you were depending on, then shame on you. You shouldn't have been depending on specific results. By this I mean that there's no such thing as a scientific experiment that's a failure. A scientific experiment tests a hypothesis; it yields a result; that result is what it is. You take the result and look for explanatory conditions in the experiment. Finding none, you revise the hypothesis. Having a hypothesis and carefully disproving it is science.

With the scientific method, properly applied, you can't lose. You can spend an entire lifetime testing hypotheses, never proving a single one correct, and you'll have done the world a service by having disproven a thousand bad ideas. Sounds pretty useful to me. But this isn't "productive" in the usual economic sense. That's because the usual economic sense attempts to mis-apply performance indicators to Science that are more appropriate to Engineering. It's another side-effect of lumping all of STEM together.

Perhaps we shouldn't do that anymore. Science and Engineering, to me, have a relationship analogous to that of Artistry and Craftsmanship. They can be combined, but they are distinct. And it is no slight to the craftsman if you do not call him an artist. Indeed, there are plenty of artists who are terrible craftsmen. Likewise, being a scientist doesn't make you an engineer; and there's nothing wrong with being an engineer, so why call them scientists? By maintaining this distinction we can divorce ourselves of the notion that scientists must produce in the manner of engineers. And with patronage, not percentages, we can remove pressure from those whose focus should be pure research.

The Lancet asks "Would a Hippocratic Oath for science help?" and I think they're onto something. Absolutely there should be an indoctrination into an ethic. The scientific method itself is that ethic. Somehow it should be impressed upon those budding scientists until they live it and breathe it; and that has to be done in school by constant reinforcement, and in professional life by the strictest standards and consequences for failure. It's what will keep science, science.

No comments:

Post a Comment