Slashdot Mirror
Debunking the Lorentz System As a Framework For Human Emotions
New submitter Enokcc writes "In a series of research articles it was claimed that a famous system of nonlinear differential equations originally used to model atmospheric convection can also be used to model changes in human emotions over time. It took an amateur in psychology with a computer science background to notice how extraordinary these claims were, and with the help of experts on psychology he has now published a critique. The latest of the questionable research articles (with 360 citations) is now 'partially withdrawn.'" Notably, skeptic Nick Brown's paper is co-authored by Alan Sokal, famous for exposing nonsense by less diplomatic means.
Considering how poorly atmospheric conditions and climate are modeled, it's no wonder they can't model human emotions.
Having spent my career working on modeling various physical phenomena, I attest it's easy to fudge the results to produce any outcome you want, if you know how.
A noted psychological researcher (can't remember his) during a TED Talk said (to paraphrase) "ignore all psychological and neurological research in your lifetime because they more than likely got it wrong.'
For decades, we were taught that the brain doesn't grow new neurons and then neurogenesis was discovered.
At least some psychologists are trying to use math beyond statistics. It looks like they screwed up, by I give them credit for trying. Social scientists have historically sucked at using rigorous mathematics to describe the phenomena they observe. I for one, don't want more social scientists scared off by a backlash on this.
We have an uncited researcher giving a speech that discredits his own scientific field with a vague statement on the probability of the validity of research in his field. This is followed by a statement on how "we" were taught something about a field that was on the cutting edge of research (which of course reminds us on how "we" were taught that in the 15th century that everybody thought the Earth was flat), and then using that to draw conclusions on areas of the field that are considered established.
Mods, we have a winner! Mod this up to +5. The immense sarcasm in this post points to how many people talk about a subject they have no firsthand expertise in while drawing a broad conclusion with uncited or perhaps misinterpreted statements from the experts. It is a damning critique of kneejerk thinking. It is the most insightful post on Slashdot in months.
You can bet that if there had been a strong lobby or interest group invested in the results of this paper, there would be strong counter-claims and attacks on people trying to debunk it. That's the case in many papers in economics, for example: their data is shaky, their models arbitrary, and their conclusions absurd, but one or the other political party uses it to justify its economic policy, it acquires a lobby, and becomes unassailable.
But even in papers where merely a lot of scientific careers and reputations are at stake, you can't overturn established dogma until the proponents of that dogma have retired or died.
Debunking pointless papers like this, papers that don't do any harm, actually is itself harmful, because it gives the erroneous impression that "the system works" and errors get corrected. The only errors that get corrected in science are those that don't have a lobby.
So then what was this? A butterfly graph, the calling card of chaos theory mathematics, purporting to show the tipping point upon which individuals and groups âoeflourishâ or âoelanguish.â Not a metaphor, no poetic allusion, but an exact ratio: 2.9013 positive to 1 negative emotions. Cultivate a âoepositivity ratioâ of greater than 2.9-to-1 and sail smoothly through life; fall below it, and sink like a stone. ... ]
[
But Brown smelled bullshit. A universal constant predicting success and fulfillment, failure and discontent? "In what world could this be true?" he wondered.
One step closer to the tipping point where tipping points will become ... not so tipsy.
Psychology is not a science. It attempts to use methods and analysis from science but that is as far as it can go. Of course most of it can be debunked.
Whenever there's an enormous new "objective" trend in psychology or social science, I always think of the Rosenhan experiment.
In a nutshell, volunteers went to different psychiatric hospitals in the US, complaining that they all suffered from (made up) voices in their heads. They were all admitted under different psychological disorders. At this point, they all acted completely normally and told staff they no longer heard voices. In all cases, they were only released once they'd submitted to treatment, and "made better".
In a follow up after the original paper, psychiatric hospitals challenged Rosenhan to send more volunteers, and the hospitals asserted they would spot them easily. He agreed, and after three months the participating hospitals said that they had weeded out 42 imposters.
Rosenhan hadn't sent a single person to the hospitals.
It's a perfect example of how inaccurate psychology is once it relies on distinct catagories like "insane" and "sane". A "positivity ratio" as created by Fredrickson is absolutely no different.
Like in any field the "experts" are often anything but.
Insights are one thing, but constantly trying to hammer objectivity into something so complex as human behaviour is always going to be flawed.
The author is a graduate student in psychology, not an amateur. Graduate students are expected to write academic journal articles as part of their studies.
So they didn't just debunk the Lorentz-Lorenz law?
PlusFive Slashdot reader for Android. Can post comments.
Why not actually make a big difference if we're going to have a team of "expert" psychologists debunk something? You know, like get the bogus Duluth Model thrown out since it's used as the framework for almost all domestic abuse therapy / explanation. Since it presents abuse as gendered, but men and women are equally aggressive. Hell, there's plenty of evidence... So, Shouldn't be hard, eh?
Oh, that's right. It's fucking Psychiatry / Psychology -- Damn non-sciences. Whoopdee fucking do, let's debunk some shit everyone knows is bunk to begin with and doesn't fucking matter. I got a better idea: Let's throw out any and all existing shit about predicting why folks act certain ways and let the Neroscientists and Cyberneticians handle it -- You know, the actual sciences based on real evidence and repeatable observable physical phenomena linked to reality by more than uninformed guesses.
I guess everyone's scratching their own itches, but I mean, if we're not going to do the right thing and declassify psychology as science, then if I were looking at making a positive impact I'd start at the Duluth Model since abuse is largely cyclic: Most rapists and abusers were themselves abused. So, the current most wide spread approach to domestic violence counseling creates more female victims in the next cycle. I mean, if we're going to debunk shit, why not take your pick of other crap that's sticking out like a sore thumb, is obviously blatantly wrong, and just plain ol' sexist? Oh, I think you know why... Because you're not fucking scientists.
I saw that TED talk (and searching TED talks suck and can't find it, although found a wonderful talk by Russel Foster about sleep and our misconceptions about that and how researchers got that all wrong for centuries.) and the person was talking about lay people - lay people taking current studies as fact.
His point was that it takes about a generation for science to really get down to the truth - trying to duplicate results of studies, improved technology in research, more research, etc ...
How many people still believe the non-sense that we only use 10% of our brains?
Of course, someone IN the field would use previous research - even in his own generation. But a lay person shouldn't take too much stock in current findings until it has run it's scientific course.
Anyway, go back to insulting people and getting your silly mod points.
Nope, it's just that when you skim through the text very quickly, as usual, the name relativistically contracts.
Ezekiel 23:20
What's so sad for me about this whole story is that took an amateur and an outsider to debunk this research, and only after an ivy league school set up an entire institute for this snake oil. Now they're saying "oops, sorry, our bad for trusting the bunk we read in the peer-reviewed journals" but why weren't experts in psychology doing this debunking themselves? And why didn't it happen immediately upon the publication of this bunk? Why didn't UPenn take a second look at this crap before they devoted an institute to it? And why is the US government putting serious money into programs based on it?
All of this stuff will eventually get walked back in the coming backlash (one hopes), but the fact that psychologists themselves were not able to recognize the crap in their own journals should be a serious wake up call for that whole discipline. If a psychology department wants to have an elite faculty, I say that at least two should be highly skilled in data-analytic methods and devote most of their research activity to undercutting the work of others. Also, a lot more research money should go into replicating experiments that the field takes as significant. Unlike other people who post here, I do think that psychology is a real science, and one of the most valuable sciences we have. The fact that it's being done badly does not make it a pseudo-science. But it does highlight the urgency of drastic reform in the field. Like I said, this should be a wake-up call. Psychology departments of the world should all be resolved to never let this kind of disaster happen again.
The best criticism of Rosenhanäs experiment that I've heared so far:
If I were to drink a quart of blood and, concealing what I had done, come to the emergency room of any hospital vomiting blood, the behaviour of the staff would be quite predictable. If they labelled and treated me as having a peptic ulcer, I doubt I could argue convincingly that medical science does not know how to diagnose that condition.
You can't fake symptoms and then complain that diagnostics sucks.
Experiment would be legit if they'd faked improvements in a way matching symptoms.