STEM Dopers, how Do you View Social Sciences?

In My Humble Opinion
21

Absolutely. It’s possible to do actual science in every field I mentioned. But some of those fields are dominated by ideas that were not arrived at scientifically or which have testable predictions.

For example, there is good science being done in macroeconomics, but the field itself rests on a very unscientific foundation. The same is true with psychology. There are actual, scientific, falsifiable experiments going on in psychology, but the field itself is full of unscientific beliefs and methods passing as science.

A string of raw facts; a little gossip and wrangle about opinions; a little classification and generalization on the mere descriptive level; a strong prejudice that we have states of mind, and that our brain conditions them: but not a single law in the sense in which physics shows us laws, not a single proposition from which any consequence can causally he deduced. This is no science, it is only the hope of a science.

— William James, “Father of American psychology”, 1892

22

I totally agree with this. However, falsifiability can mean (at least) two rather different abilities.

  1. Accurately predicting future behaviour, events, or occurences ;
  2. Precisely and consistently describing reality.

In my view, some social sciences definitely meet both criteria . I’ll stick to linguistics since it’s the one I know best.

Criteria 1 - The Laryngeal Theory

In the late 19th century linguists were confronted with some puzzling irregular sound changes in Proto-Indo-European. Saussure suggested that these were in fact regular, and based on the influence of consonants that had in meantime disappeared in all surviving languages. Needless to say, this idea of “ghost sounds” was controversial. Yet, in 1915 Hittite was demonstrated to be an Indo-European language and, not only had it retained laryngeals, but they were found in the places where Saussure had posited them.

Criteria 2 - Linguistic universals

They are features and correlations that are found either across all languages or within well-defined subsets of languages such as :

  • all languages have consonants and vowels ;
  • languages with dominant Verb-Subject-Object order are always prepositional ;
  • if a language has inflection, it always has derivation ;
  • If a language has trial grammatical number, it also has dual grammatical number.

It is very possible that one day, we’ll stumble upon some language that contradicts one or several of those universals, so they are very clearly falsifiable. Until we do (and if we do) they are accurate descriptions of patterns that are consistently found in languages all around the world.

Which is why I posted it in IMHO and not Great Debates.

23

I don’t feel the Scientific Method accurately describes how many, maybe most, scientists spend most of their time. My laboratory days were filled with stirring stuff and describing what happened. Hypotheses in papers, when there are any, are often added post hoc. Most of my papers are “we tried a bunch of ways to make stuff and here’s the one that sucked the least.” Maybe I’m just a bad scientist :smiley:.

My undergrad diploma also has Linguistics printed on it and I have no qualms calling that a science.

24

I see the sciences on a spectrum. Mathematics, then physics, then chemistry, then biology, then sociology, something like that. Each seems a bit predicated on the item to its left being absolute and literal, and reluctant to consider the item on its right.

They deal with messy things, which I think we collectively have to try to deal with.

But I think they all can aspire to being scientific, and nobody can claim what they’re doing is perfectly scientific. Yes, let’s call them sciences, and accommodate the fact that the rightmost realms are hard to reduce to tidy and accurate rules.

I don’t even really believe in the scientific method, at least not as a typical method. I think play comes closer to describing what many scientists mostly do. We (I’m a physicist) get interested and curious, and we kind of play with the topic, always some mix of theory and experiment (I have School of Athens hanging in my office). When we want to write about it, making a formal input to the literature, we cast it in terms of hypothesis, experiment, and so forth. And if we have to propose something big, much funding, many people, we have to do more of that early on. But as a description of what we’ve been doing minute by minute, I think the scientific method is a bit, perhaps, revisionist.

One thing I really like about physics is the degree to which a mechanistic description can (sometimes) become so cleanly predictive. But it might be that I’m just scared to try to operate in the murkier worlds further right on my spectrum.

25

I tend to agree with this.

I’m in a STEM field. At work I use the concept of the null hypothesis. And then I perform tests that will either reject or not reject the null hypothesis.

26

It’s not by broad subject. It’s bits and pieces scattered here and there across all of the social sciences. For instance, Arrow’s Theorem is a rigorous result in political science, and the Law of Supply and Demand is a rigorous result in economics, and classical conditioning is a rigorous technique in psychology.

The definition you’ll see for the “Scientific Method” in most elementary-school science textbooks is far too rigid and specific. A more accurate definition would be something like “Try stuff, and see what happens”.

27

For what it is worth, engineers, chemists and physicists spend very little time discussing Popper. Despite studying a lot of the former, he is much more prominent in economics, business and philosophy texts.

It is wrong to reduce science only to hypotheses and the scientific method, though it is obviously of great relevance. Social scientists often show great ingenuity in finding natural experiments when hard or impossible to conduct. But many studies are underpowered, hard to reproduce, too subject to subtle errors or manipulation. This does not make the subjects less important, the real world is complicated, but it makes defining something as scientific less accurate. Is it really important social sciences be considered sciences? Will this change their relevance or perception now? These subjects are more mature than in their infancy and times change.

28
29

Is Philosophy science?

30

Would it be out of line to question the inherent privileging of the ways of knowing that are called “scientific” in these kinds of discussion? Sciences, whether “real” or not, make claims of authority based on specific models of evidence and conclusion. Science is a way of getting to truth. But it’s far from the only way. For some kinds of truth, I would absolutely 100% always want to rely on science and scientific knowledge. But there are more things in Heaven and Earth, Horatio…

The way I see it (this is overly broad, of course, and stereotypical, but it often shakes out this way), scientists value a certain kind of description, and base their conclusions on certain kinds of observations. Those are “empirical” and “reproducible” and claim to be “objective.” This way of approaching questions about the universe has, historically, produced some amazing successes. We know and can do things now, because of science and the scientific way of approaching questions, that are spectacular. And save millions of lives and give us comfort and joy and the love of our families.

So social scientists, quite naturally, crave that kind of success, that kind of authority, and that kind of practical, applicable and (apparently) comprehensive understanding of their subjects. So they want very much to have their disciplines be seen as science.

But I think that those of us in the humanities (as well as, probably, a large number of scientists and social scientists, if you really press the issue with them) would say: wait a minute. So our way of observing and describing and analyzing and coming to conclusions about the universe is not science. So what? For many kinds of questions, for many kinds of subjects, science does not give the most true, or the most useful, or the most communicable kind of information. There are some truths that are actually better understood, better explained, better shared, and better used through poetry, or art, or journalism, and so on. We don’t have to claim the accuracy of science, because we have a different kind of accuracy.

In anthropology, at least, for example, it’s been pretty widely accepted for some time, I think, that ethnography is a genre of literature. And for that reason, it’s a lot better at understanding and analyzing and communicating about human culture than anything that had to force itself to be quantitative and objective could be.

31

I’m not sure if I’d consider economics a science, it is certainly not a behavioral science. Behavioral economics, which involves repeatable experiments, is a lot closer. In fact it showed some of the assumptions of traditional economics were erroneous.
In our tutorial we conducted an anchoring experiment for three classes of experienced engineers. They all were subject to it - to such an extent that the results were significant for n = 20.
I’m sure you don’t say that genetics and evolution are bogus because of Lamarck. His work was shown to be false. So why denigrate psychology because of Freud? Maybe some psychiatrists are still into him, but I don’t think a lot of psychologists are. Not the ones I’m talking about at least. Clinical is a different story.
Obviously different branches of science have different protocols. You don’t need to worry about getting a good sample of electrons. My point was that all kinds of science can be botched if you aren’t careful. That doesn’t invalidate a field.

32

I’m a fan of behvioural economics, but it also lacks predictive power.

BTW, I’m not trying to say that non-sciences are all junk. I think there are lots of non-falsifiable things that are worth studying and from which we can learn things. But there is a difference between fields that have testable, falsifiable knowledge and those are are built on something else.

The reason is right in your sentence: Lamarck was dismissed because evidence showed his theories to be wrong. Freud persisted (and still persists with some) because his theories were not falsifiable. It’s not Freud’s fault - it’s the nature of his field.

Hard sciences with falsifiable theories have a built-in error correction mechanism. Social sciences, not so much. They govern themselves more by committee, by peer review, by status, and sometimes by fads. But ultimately many of the propositions in those fields have nothing behind them but the opinions of gatekeepers rather than tests for truth.

Sure, any science can be done poorly. But wrong falsifiable claims eventually get falsified. In fields without falsifiability, bad claims can stick around for a long time, and even be used as foundations for more claims until you have a pn entire body of knowledge built on false premises, with no way to test them for truth.

33

Like what? What is an example of a ‘truth’ that wasn’t arrived at scientifically? What other ways of knowing are you thinking about? Your examples of literature and such aren’t truth-seeking so much as they are illuminating feelings, describing shared experiences, engaging emotions, etc.

34

Aren’t those some of the most important truths in human existence? Aren’t they part of what makes us human? In terms living in this universe (the one we live in, the one that includes human beings), and fully experiencing it, knowing how it works and understanding it, predicting how it will work and how its parts interact, there are few things more important, more true, than beauty, goodness, honor, rage, awe, eternity, death, evil, passion, dedication, color, hatred, bigotry, patriotism, loyalty, terror, joy, etc., etc.

Science has little or nothing to say about those kinds of truths (nor is it designed to). But that doesn’t make them less true, and certainly not less important.

35

I would simply argue that none of those things qualify as ‘truth’. They are by their nature subjective and subject to change, and are not universally applicable. There’s no such thing as ‘my truth’. There’s only what’s true, and what isn’t.

36

Are you claiming that “none of those things qualify as ‘truth’” is a truth?

37

Yeah, this is a tough one. On the one hand, philosophy should be beyond mathematics on my spectrum. It is not lost on me that if you click on the first nontrivial link on nearly any Wikipedia page, and keep clicking on the first nontrivial link in each successive Wikipedia page, you wind up on the Philosophy page. On the other hand, it is tempting to then put love before philosophy, and oneness before love, and then a succession of turtles before oneness. I almost want to call the process on the basis of how soon we are willing to step down into the mere sciences.

38

Predict what? The entire economy? Psychohistory? No, and I’ve never seen that claimed. The behavior of people in controlled situations? That it predicts well. We predicted the results of our experiments, and we got confirmation we were right.

Evolution is still dismissed by some (many) but that doesn’t mean it’s not science and not falsifiable. Am I calling anyone who is still a Freudian the equivalent of a creationist? Sure am. I don’t think Freud is no longer even close to mainstream because of fashion. Remember the big fight a few decades ago about Freud’s contention that childhood sexual abuse was fantasy? That was a prediction which got falsified.
I don’t know about psychiatry and clinical psychology, but Freud seems to be absent from analytical psychology.

My daughter wishes psychology isn’t falsifiable. She has many ideas which she does studies about and finds they get falsified. Of course there are no papers about these. Just like in the hard sciences, you don’t publish failures unless perhaps you spent a lot of money on them.
I perceive you have some thought that our psychology is purely random, so that one study does not predict what we’d find in a similar study. But we share characteristics, some social, some evolutionary. They’ve detected the endowment effect in gorillas, for instance.

Yeah, bad claims that don’t get examined can stick around forever. Most papers are write-only. The more prestigious and specialized a journal the fewer people read it. I know someone whose PhD in physics was delayed because he found that something everyone thought was true wasn’t. Not significant enough for a dissertation topic, unluckily for him. Due to availability (another behavioral economics idea) we think that many if most papers have significant results that get repeated by others. Actually, they are mostly forgotten except for a few other specialists who read the paper just enough to say why their paper is better in the previous work section.
I had a survey paper that got a ton of cites because people could read and reference it and not bother going to the papers I surveyed. That was in computer architecture, not psychology. I was kind of a pre-Google Google.

39

Now this I totally agree with. Look at the “truths” of any kind of woo position. I think most of them are falsifiable, but the pitchers of this stuff are allergic to testing them.
I was in a critique group once with a woman who was writing one of these woo books. I asked if she had any evidence for her claims outside of anecdotes. I could have been calling her nasty names. I don’t think she knew she was full of it and was lying, I think she just didn’t think in terms of falsification and testing of claims.
Yeah, truths need to be tested. Certain types of psychology do test them. And lots of ideas fail. It is especially interesting when public policy methods get shown to be useless or counterproductive through experimentation.

40

Absolutely right. There are no hypotheses to falsify. Only formal proof is acceptable. Sure we make conjectures, but no one pretends they are true until refuted.

I think there is some real science in linguistics and lots in psychology. I am dubious about economics. As for sociology, my attitude is summarized in the following story.

The deans of the faculties of Arts and of Sciences are out for a stroll. Science says there is one department he really likes. Mathematicians need no fancy equipment, just a desk, a pencil, paper, and a wastebasket and a mathematician happily produces paper after paper. Arts says that he can top that. A sociologist needs only a desk, pencil, and paper to produce paper after paper.