Hacker News Reader

Comment by rossdavidh | Hacker News Reader

Whenever I hear discussion of whether we should defer to the judgement of scientists, I always want to ask, which field of science do you mean? In some, like physics, chemistry, evolutionary biology, geology, there is still much to be learned but what we have so far is pretty solid. You don't hear about the basics (matter is made of atoms, species emerge from natural selection, plate tectonics) getting overturned.

Then, there are other fields, like economics and nutrition, where the basic fundamentals of the field are a pile of mud. It is much like medicine in the era of balancing the four humours. In the meantime, we need to fund research, but not listen to any of the recommendations of the experts, until they achieve a basic level of understanding of their subject (which clearly hasn't happened yet).

View on HN

The fields of nutrition and economics do have the basics worked out. But the basics aren't things like "is eating too much salt bad for you?" The basics are about nutrient metabolism and the human body is a very complicated (/complex) system. The basics don't necessarily have to be "actionable", and they don't necessarily need to say how the system comes together as a whole. For instance, in your example of evolutionary biology, there is a lot of debate about whether natural selection or neutral evolution is the primary driver behind species diversity. We know that both exist and both happen, but we don't know how they came together to create the species diversity we observe.

Another aspect of this is that there are certain fields (namely economics and nutrition) where the popular press over-exaggerates every single paper as a groundbreaking discovery because it gets clicks. The media tries to present the idea that scientists as a whole have come to a brand new conclusion which should change the way everyone behaves, when in reality it's just a few people in a windowless room reporting on some observations they found and speculating on what those observations might mean.

> Then, there are other fields, like economics and nutrition, where the basic fundamentals of the field are a pile of mud.

That's incredibly dismissive of these fields. I can't speak much to nutrition, but economics is not a "pile of mud", definitely not its fundamentals.

Do you have any reason to think that it is other than (kind-of-literally) mud-slinging?

Can you elaborate on the 'pile of mud' fundamentals of economics? I've always been inclined to agree, but then I consider my biases in this regard, but also my own field, psychology, and pause.

If you subdivide 'science' into various fields, psychology is definitely on much shakier grounds than has often ben assumed. But if you subdivide 'psychology', the picture isn't as clear. I'd say marketing and the broader field of persuasion/propaganda, or something like group dynamics, has a much firmer basis in reality than, say, psychopathology or evolutionary psychology.

I'm very curious if people who are well-versed in economics can make similar distinctions, because I find it hard to believe it's 'all' mud.

People are very prejudiced against economics because economists don't predict well and have contradictory policy recommendations, but the basis of economics is quite solid.

(My favorite illustration: if there were no underlying laws to economics, then dictators would just issue edicts about currency and markets which would then just plain work rather than causing havoc -- and yet dictators very much suffer side effects of their dictates.)

This is very similar to weather prediction, which in some geographical areas is very untrustworthy, but not because its scientific foundation is lacking.

Psychology has subareas that are very rigorous and evidence based, but also major foundational areas that remain essentially speculative/opinion-based (what is mind/consciousness/free will?)

Evolutionary psychology is again firmly rooted, but is most famous for its speculative and unproven areas.

Marketing and persuasion/propaganda have subareas that incorporate statistics and/or psychology and/or sociology etc., but are not themselves actually sciences at all, so that's very much apples and oranges.

Really, I got downvoted for something where every word was inarguably true in the sense of objective fact not opinion, and informative about half a dozen subjects?

This place is as bad as reddit and slashdot. I haven't returned here very often over the years; this kind of experience makes one feel unwelcome.

My best advice is to quit caring about the downvotes.

People often come along eventually and correct unfair downvotes.

That's one reason the site guidelines ask you not to complain about getting downvoted; by the time most people read the complaint, it's obsolete.

https://news.ycombinator.com/newsguidelines.html

> Then, there are other fields, like economics and nutrition, where the basic fundamentals of the field are a pile of mud.

Practically all social sciences are suspect and really have no business calling themselves sciences.

What are your criteria?

I'm not OP, but Karl Popper's notion of falsifiability places a lot of the social sciences in the domain of myth vs actual science.

And yet, we want to know more about those subjects. We just have to make do with very limited falsifiability.

Good job Popper's falsifiability is flawed then; else evolution is not science.

> It does appear that some people think that I denied scientific character to the historical sciences, such as palaeontology, or the history of the evolution of life on Earth. This is a mistake, and I here wish to affirm that these and other historical sciences have in my opinion scientific character; their hypotheses can in many cases be tested. [Popper, 1981]

Certain things like 'carbon dating' are certainly scientific, and can be falsified. Evolution itself is the best historical explanation we have, but hard to set up an experiment to disprove.

I personally think 'falsifiability" is particularly useful in separating a soft 'fact' like Freud's theory of the id, and the Oedipus Complex from something like say, the theory of gravity.

Agreed, it is useful, certainly. But it's not to be wielded without subtlety: notably in your quote the use of 'imo' and 'in many cases'. Anyway, apologies for the terseness. My gripe is just when it is used as a magic omnibullet in discussions about 'what is science?'.

They are all cargo cult sciences. Fields dominated by theories entirely unproven because they are practically unprovable. Those theories remain even after they are often demonstratably wrong. Look up replication crisis.

I don't buy your premise, but supposing I did, what alternative do you propose? The purpose of social science is to understand human society. Do you have some concept of a way to do that which you feel would be more clearly accurate or scientific than what's currently done?

Bear in mind that the 'hard' sciences use theories which are partially or wholly incorrect, too; we still use Newtonian physics in many cases because they're easier to use as a model, even if they don't capture everything.

> what alternative do you propose?

That they don't call themselves science. Maybe call themselves studies and that they put more effort into what they call experiments.

> we still use Newtonian physics in many cases because they're easier to use as a model, even if they don't capture everything.

This is frankly a terrible example. Because it does not tell the whole story does not mean it is incorrect. There is no unifying theory of sociology either.

I believe there are a wide variety of methods used in social science. Are you specifically referring to small experiments conducted with undergrads in a classroom? Or do you have other targets in mind?

> Are you specifically referring to small experiments conducted with undergrads in a classroom?

No I'm talking about the entire field of social science being dominated by complete nonsense.

See the following documentary:

https://imgur.com/gallery/7EPPT

> That they don't call themselves science. Maybe call themselves studies and that they put more effort into what they call experiments.

I'm sorry, let me clarify: My question is not "what would you do with what we have now?", rather, it's "assuming we throw all of what we have now out, how would you do it in a way you consider to be science?" My assertion is that disregarding all our progress up until now in these fields on the grounds that they have not yet reached the level of maturity of physics or other hard sciences is unfair and will create a self-fulfilling cycle whereby we never gain any actual progress in these fields.

> This is frankly a terrible example. Because it does not tell the whole story does not mean it is incorrect. There is no unifying theory of sociology either.

Except that's exactly my point. Sociology doesn't tell the whole story, but your assertion that everything that comes out of it is tainted isn't right, either. You're comparing relatively nascent fields to relatively well-developed ones and expecting them to be exactly the same in terms of replicability, reliability and applicability, generally. That's unfair, and totally disregards history.

If we regressed even further, to before Newton, to the ancient Greeks or Romans or Egyptians, sure, they got a lot wrong - most things. But everything is a building block. I think that we are just now beginning to make real progress in fields like psychology, sociology, etc. You can call them "baby sciences," if you'd like, or invent whatever other word or term you prefer, but it doesn't change what they are, and it doesn't make them not worth proceeding with.

Further, we can't say that because some sciences progress quickly (e. g. computer science), all sciences should be expected to do so. As many in this thread have pointed out, good social science is harder to find and harder to do than good computer science. Progress will therefore be slower. It doesn't make said progress unimportant or impossible.

I'd love to read a book on the subject. What did it take for various scientific fields to become truly mature? Are their key ways to measure how far along we are?

Concrete, predictive theories would be great but I get that we're far from them in fields like nutrition short of, like, scurvy I guess? I wonder how much of this is due to the natural variation among individuals and how much is just because of ethical limits on scientific inquiry. Not that I'm advocating for human experiments but one has to wonder what we could discover.

Measurement. If you can’t measure it you can’t predict it.

I find a lot of the problems is people seem to have a job dealing with correlation not being causation. Maybe there's a correlation between salt and heart attacks but perhaps that's because the high salt folk eat a lot of junk food and its the junk food and lifestyle that's not the problem rather than the salt. You need further evidence to figure what's going on with any certainty but you almost never see an article saying that.

Reply on news.ycombinator.com