It’s cool to have some mathematical rules of thumb to describe crowds and so on if it is backed up by experimental data.
I did an economics undergrad. One thing that really frustrated me about the whole subject was how they kept calling it a social “science” and yet there was no experiment or predictions. Everything was a retroactive explanation for something, e.g. the financial crash or the affect of a minimum wage. Where’s the predictions? Another thing that bothered me was that often it felt like they were introducing maths to make it all seem a bit more legit: like solving integrals for second-price bid auctions or describing utility functions and so on. But that was always just a cute model with no experimental evidence to back it up.
Basically, I think social sciences should stop calling itself a science unless they start making predictions. And they should make it more like statistics and psychology classes. Evidence is king.
Here we go, another HN comment that's awfully quick to dismiss the history of economic methodology and economics as science because they took one undergraduate class in Neoclassical Microeconomics 101.
Please, this is exactly like a first year psychology student trying to psychoanalyse you: Absolute bullshit. Economics has a lot more to offer than supply-demand curves, and it has been evolving greatly in many different directions and sub-fields like behavioural economics (especially living its golden age right now). Different schools of thought exist, and not everybody in the field is solving derivatives.
Well, I have first class degree in economics to be fair. I got the highest score across the university. I didn't just take one class and make my mind up. But I agree that doesn't make me an expert in the field. I also agree behavioral economics is a better direction. However, I'm just saying many strands of it need to use evidence more. A lot of models were kinda cute but also pretty silly. Game theory is a bit of a kafuffle -- Ariel Rubenstein agrees. I didn't see any predictions or studying of real-world data in my whole time studying it. The least we can agree on is that there should be way more of a focus on that for university level.
It depends on the university that you have gone to, mine was very clear on the assumptions made and we were always critical about the content and its implications. I took several philosophy of economics classes throughout my degree as well. I wish you had a better introduction to the field though, and agree that universities should better educate their undergrads on the real-world aspects of the models at hand.
What sort of university did you go to where they don't do econometrics or macroeconomics, or international trade, or... all things where you can not open a textbook without being confronted with real life data?
The large majority of articles in top journals feature both mathematical theory and statistical&causal inference on real data.
>Here we go, another HN comment that's awfully quick to dismiss the history of economic methodology and economics as science because they took one undergraduate class in Neoclassical Microeconomics 101.
HN comment? There are lots of books by respected authors written on the subject.
Just because the field exists and gets grants and gets to dictate policies and so on doesn't mean its scientific. Anything scientific in economics is just the math part, which are orthogonal to economic behavior and markets.
>* Economics has a lot more to offer than supply-demand curves, and it has been evolving greatly in many different directions and sub-fields like behavioural economics (especially living its golden age right now).*
The main things it has to offer is salaries for economists and favorable policy suggestions to politicians.
"Anything scientific in economics is just the math part, which are orthogonal to economic behavior and markets."
You're just proving my point here, this is outright dismissive of all forms of experimentation that goes on in many fields of economics. Do not blame the study for your lack of deeper understanding of it; it has more to offer than just salaried academics and macroeconomic policycrafting.
Here we go, another HN comment that is quick to insult and dismiss someone else in order to feel smug. If you took as much time to read their comment as you took to insult them, you'd know that they majored in economics, they didn't just take 1 class.
I have an undergraduate degree in economics and a master's degree in finance, I cannot believe someone can go through a whole degree without understanding what economics as a science entails... Regardless, see OP's response and my reply to that.
> 2. Geologists don't run experiments with n earths and randomized double-blind trials. Do you think they're not scientists either?
Although they can't travel back in time to do controlled experiments, computer simulations are pretty routine for the study of the deep time of the earth, for example, the neoproterozoic "snowball earth" simulations [1].
These guys should take some clues from quantum mechanics, where observing or trying to control the system observed screws it up.
“Physics found we can take advantage of Facebook to spread opinions and control people’s minds” - “oh shit, we polarized the people, and now we’re on the verge of WWIII”
"These guys" have been considering biases introduced by injecting an observer into a social space for a long time now. I think you might have missed the point of the article and the social sciences.
https://en.wikipedia.org/wiki/Lucas_critique <-- Treating relationships between variables as structural and changing one to influence the other will affect the supposed underlying structural relationship.
Humans adapt their behavior by observing it. Looking at sth. doesn't affect the outcome, but instead the observation adds to the set of exogenous variables that will determine future behavior. We're not an ant colony that acts in an entirely exogenous setting, we change the setting both through the (un)intended consequences of the observed behavior and our attitudes toward them.
The allocation of scarce ressources did neither begin with capitalism nor economics. Many formal descriptions of foundational economic principles can (edit: be traced back to) ancient Mythology [1]. How societies applied these principles has differed greatly. Trade has been an important factor of human behavior since the dawns of civilization. In a framework of linear progress one might be tempted to search for deterministic laws in human behavior, but it might make more sense to just interpret phenomena case-by-case with a rather loose set of theoretical beliefs.
Paul Krugman is a Nobel laureate, but so is F.A. Hayek:
"The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design."
The work described in the article can be interesting as far as it goes, but it seems to me that a lack of humility coupled with insufficiently accurate models can be a dangerous combination.
Hayek got his Nobel in 1974 Krugman in 2008. to illustrate the point further: who do you think knew more about quantum mechanics when they died Niehls Bohr or Richard Feynman?
They have proved they can and will be corrupt as needed. Politics should not validate science, but people treat it so when we care about the Nobel prizes.
I agree! My point isn't that we should listen to people because they won an award, because people who've won the same award have come to different conclusions.
Applying physics to understand human behavior is same as applying human intuition to explain physics. Any correlations found, which are anyway "fitted", are bound to breakdown.
Firstly we don't understand physics fully. We have only more breadcrumbs but no answer to the ultimate "why", and are basically unable to answer a pesky kid who keeps repeating that. Now, we are applying this to explain another equally intractable problem, that of understanding human behavior. It is a strange case of reductionism. I wonder why people engage in such futile activities. May be this line in the article is the answer.
> Results like these have obvious value for corporate marketers.
A bunch of maths equations and charts presented in the form of a paper that promises to crack the human behavior with further research is too good for the corporates to let go. Ergo funding. Well done. I won't be surprised if all this started out as a joke that the corporates lapped up.
To be clear, there is nothing wrong with trying to quantify an observation or a phenomenon, and in fact is always welcome, but bringing in incongruent things into the mix reduces the gravity.
Fun fact: If they cut you open to remove your appendix and find out that it isn't inflamed or infected
then they still remove it since the scar left after surgery is universally known that you already had an appendectomy
and so doctors, on seeing the scar, will rule out anything to do with the appendix as a possible cause for your
harrowing abdominal pain.
So to eliminate this possible confusion surgeons - all around the world - always follow this unwritten rule.
We've had econophysics for quite a while. It has yet to provide substantial insights.
The issue is twofold.
As others have remarked, human beings react to each other and the intervention you do strategically.
Econophysics models often use things like particle behavior based on probabilistic choice rules to predict behavior. This may describe a situation if fitted, but it also tends to break down with interaction and intervention. Furthermore, we know that in such interdependent situations, it can be highly misleading and wrong to use aggregate functions and non-behavioral approaches. For example, if networks of externalities influence choices.
In many areas, such as auction behavior or anonymous trade, the game-theory based models fit empirical behavior extremely well. In other situations, it seems we start to understand better why people react differently.
Econophysics feels like a step back to old style economics. I believe the current research agenda, ie. giving people strategic agency and trying to understand their bounds to understanding and rationality, seems to be more useful to me.
Secondly, econophysics claim to fame is also party based on mathematical tools. I think this is misguided. While the math is different, and new tools are always useful, the economic models are actually based on quite deep foundations.
Those reaction curves you learn in Game-Theory 101 seem simple, but that is more a matter of representation. They can also be derived in much more general mathematical spaces. For example, instead of simply finding optimal functions in continous spaces, the same can be done in lattices based purely on orderings...
Economic theory has always been dominated by pure mathematicians.
and for sociology - I mean seriously. Some of the biggest names in formal sociological theory are physicians, such as White.
In my research I have personally found that physics style models, which are not based on strategic behavior, can be highly misleading as they allow, for example, for aggregation and inference that would not actually correspond to strategic agents making individual interdependent decisions.
Mostly, this is a failed research program and not at all the direction the social sciences are going. Economists at least will draw one lesson from physics and respond to econophysics, by following the "shut-up and calculate approach" - the result, econophysics will be ignored.
35 comments
[ 4.4 ms ] story [ 90.1 ms ] threadThe "phase transition" around a level of extreme opinion in a population is interesting, and I'd certainly like to find out more about this research
I did an economics undergrad. One thing that really frustrated me about the whole subject was how they kept calling it a social “science” and yet there was no experiment or predictions. Everything was a retroactive explanation for something, e.g. the financial crash or the affect of a minimum wage. Where’s the predictions? Another thing that bothered me was that often it felt like they were introducing maths to make it all seem a bit more legit: like solving integrals for second-price bid auctions or describing utility functions and so on. But that was always just a cute model with no experimental evidence to back it up.
Basically, I think social sciences should stop calling itself a science unless they start making predictions. And they should make it more like statistics and psychology classes. Evidence is king.
Please, this is exactly like a first year psychology student trying to psychoanalyse you: Absolute bullshit. Economics has a lot more to offer than supply-demand curves, and it has been evolving greatly in many different directions and sub-fields like behavioural economics (especially living its golden age right now). Different schools of thought exist, and not everybody in the field is solving derivatives.
Econometrics is a major part of modern economics and attempts to do just that
The large majority of articles in top journals feature both mathematical theory and statistical&causal inference on real data.
HN comment? There are lots of books by respected authors written on the subject.
Just because the field exists and gets grants and gets to dictate policies and so on doesn't mean its scientific. Anything scientific in economics is just the math part, which are orthogonal to economic behavior and markets.
>* Economics has a lot more to offer than supply-demand curves, and it has been evolving greatly in many different directions and sub-fields like behavioural economics (especially living its golden age right now).*
The main things it has to offer is salaries for economists and favorable policy suggestions to politicians.
You're just proving my point here, this is outright dismissive of all forms of experimentation that goes on in many fields of economics. Do not blame the study for your lack of deeper understanding of it; it has more to offer than just salaried academics and macroeconomic policycrafting.
Mock-experimentation has been going on since forever in all kind of pseudo-sciences of the "soft" variety.
Even Freud did clinical "experiments".
1. Economists make predictions.
2. Geologists don't run experiments with n earths and randomized double-blind trials. Do you think they're not scientists either?
Although they can't travel back in time to do controlled experiments, computer simulations are pretty routine for the study of the deep time of the earth, for example, the neoproterozoic "snowball earth" simulations [1].
[1] https://www.researchgate.net/profile/W_Peltier/publication/1...
“Physics found we can take advantage of Facebook to spread opinions and control people’s minds” - “oh shit, we polarized the people, and now we’re on the verge of WWIII”
https://en.wikipedia.org/wiki/Hawthorne_effect <-- Observing behaviour changes said behaviour, leading to invalid observations.
https://en.wikipedia.org/wiki/Lucas_critique <-- Treating relationships between variables as structural and changing one to influence the other will affect the supposed underlying structural relationship.
The allocation of scarce ressources did neither begin with capitalism nor economics. Many formal descriptions of foundational economic principles can (edit: be traced back to) ancient Mythology [1]. How societies applied these principles has differed greatly. Trade has been an important factor of human behavior since the dawns of civilization. In a framework of linear progress one might be tempted to search for deterministic laws in human behavior, but it might make more sense to just interpret phenomena case-by-case with a rather loose set of theoretical beliefs.
[1] Tomas Sedlacek - Economics of Good and Evil
"The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design."
The work described in the article can be interesting as far as it goes, but it seems to me that a lack of humility coupled with insufficiently accurate models can be a dangerous combination.
They have proved they can and will be corrupt as needed. Politics should not validate science, but people treat it so when we care about the Nobel prizes.
Firstly we don't understand physics fully. We have only more breadcrumbs but no answer to the ultimate "why", and are basically unable to answer a pesky kid who keeps repeating that. Now, we are applying this to explain another equally intractable problem, that of understanding human behavior. It is a strange case of reductionism. I wonder why people engage in such futile activities. May be this line in the article is the answer.
> Results like these have obvious value for corporate marketers.
A bunch of maths equations and charts presented in the form of a paper that promises to crack the human behavior with further research is too good for the corporates to let go. Ergo funding. Well done. I won't be surprised if all this started out as a joke that the corporates lapped up.
To be clear, there is nothing wrong with trying to quantify an observation or a phenomenon, and in fact is always welcome, but bringing in incongruent things into the mix reduces the gravity.
So to eliminate this possible confusion surgeons - all around the world - always follow this unwritten rule.
The issue is twofold. As others have remarked, human beings react to each other and the intervention you do strategically. Econophysics models often use things like particle behavior based on probabilistic choice rules to predict behavior. This may describe a situation if fitted, but it also tends to break down with interaction and intervention. Furthermore, we know that in such interdependent situations, it can be highly misleading and wrong to use aggregate functions and non-behavioral approaches. For example, if networks of externalities influence choices. In many areas, such as auction behavior or anonymous trade, the game-theory based models fit empirical behavior extremely well. In other situations, it seems we start to understand better why people react differently. Econophysics feels like a step back to old style economics. I believe the current research agenda, ie. giving people strategic agency and trying to understand their bounds to understanding and rationality, seems to be more useful to me.
Secondly, econophysics claim to fame is also party based on mathematical tools. I think this is misguided. While the math is different, and new tools are always useful, the economic models are actually based on quite deep foundations. Those reaction curves you learn in Game-Theory 101 seem simple, but that is more a matter of representation. They can also be derived in much more general mathematical spaces. For example, instead of simply finding optimal functions in continous spaces, the same can be done in lattices based purely on orderings... Economic theory has always been dominated by pure mathematicians.
and for sociology - I mean seriously. Some of the biggest names in formal sociological theory are physicians, such as White.
In my research I have personally found that physics style models, which are not based on strategic behavior, can be highly misleading as they allow, for example, for aggregation and inference that would not actually correspond to strategic agents making individual interdependent decisions.