Is Psychology Science?

[Cirion Spellbinder]

brimstone, if you don't think soft sciences are a reliable source of truth, then why have you cited the golem effect as a reason to support higher standards in education before? Is some soft science not as soft as it seems? Also, what makes economics a hard science?

[brimstoneSalad]

brimstone, if you don't think soft sciences are a reliable source of truth, then why have you cited the golem effect as a reason to support higher standards in education before?

Psychology is highly influenced by expectation, it's the basis of placebo effect. This and the effects of cognitive dissonance are some of the most established and well founded in the field.

It's mostly the details and unfortunately sometimes grand claims that are softer science. Creating expectation itself is more of an art, and not that well studied beyond being able to make very broad generalizations.

Is some soft science not as soft as it seems? Also, what makes economics a hard science?

Psychology is actually more divided between soft and hard. There are serious psychologists who use scientific methodology and proper controls to experiment with behavior, human and non-human, and learn about the workings of the mind (neuropsychology/cognitive psychology are particularly hard, but some others are too), and then there are others who are less rigorous and deal with very soft and squishy ends like social psychology, gender studies, etc. There's a lot of pseudoscience in it all too. The trouble is that people are SO influenced by expectation, doing placebo controls is essentially impossible in many cases, and often it's inherently impossible to do because the situation is impossible to create a placebo for.

I don't fully agree with these articles, but it can give you some insight into the discussion:
https://www.psychologytoday.com/blog/under-the-influence/201308/the-psychology-the-psychology-isnt-science-argument

Here's a soft scientist complaining about criticism (the article is terrible and indicates pretty much total ignorance of science from the writer):
https://blogs.scientificamerican.com/psysociety/psychology-8217-s-brilliant-beautiful-scientific-messiness/
This was kind of funny, though:

From what I can gather, there are plenty of phenomena in the "hard sciences" -- most notably, in physics -- that are not observable. String theory? Quantum mechanics?

That guy is sort of ignorant and out of touch. Quantum mechanics is observable, he may be talking about the models, and those are different; they use deductive logic to fit what's happening and create a theory that yields predictions; these are VERY simple models. String theory is different, and is widely criticized for not being a theory (and often for being bad science), probably more widely than is psychology within Physics.
Actually, it gets worse. Serious face desks when she starts talking about how hard it is to measure love and compares it again to quantum physics.

Based on reading a bit of that, I probably agree more with the criticism:
http://articles.latimes.com/2012/jul/13/news/la-ol-blowback-pscyhology-science-20120713
But I haven't read that one yet (I'm just skimming it, not sure if I have time to read it fully since I'll probably agree with everything in it but you should read it if you can).
EDIT: Didn't realize how short it was, I read it all. It was fine, but the Scientific American blog article actually does a better job of demonstrating why Psychology isn't science through a social psychologist clearly not understanding science.


Economics is hard when you deal with rational agents, as in game theory. It deals extensively with computer modeling and mathematics; although it mostly only works when it's prescriptive. E.g. do this and things will work better. It can't necessarily predict the economic weather (which is what I think some people want it to do).

[Cirion Spellbinder]

Thanks for the thorough reply brimstone, I feel like I understand this divide more closely now. However, I'm a bit confused by this:

The trouble is that people are SO influenced by expectation, doing placebo controls is essentially impossible in many cases

Could you elaborate on this? It's been a while, but I thought placebo / nocebo effect could be controlled (at least in medicine) by giving a portion of the subjects a fake treatment and the other portion the real treatment, and if both experience the same effects you can conclude that your drug doesn't work. Could you give me an example of an experiment where placebo would need to be controlled for, but couldn't be?

and often it's inherently impossible to do because the situation is impossible to create a placebo for

I'm not sure what you're saying here.

[brimstoneSalad]

Could you elaborate on this? It's been a while, but I thought placebo / nocebo effect could be controlled (at least in medicine) by giving a portion of the subjects a fake treatment and the other portion the real treatment, and if both experience the same effects you can conclude that your drug doesn't work. Could you give me an example of an experiment where placebo would need to be controlled for, but couldn't be?

In psychology. Let's say you want to show that when teachers have low expectations of students, the students don't perform as well (golem effect).
You can tell the teacher "these are poor performing students", and look at the result. You can compare that result to a control where you told the teacher nothing. But you can't always compare that result to the equivalent of a sugar pill. You just have intervention and non-intervention. You can compare different interventions, but none of them really look like a placebo.

This is OK, actually. But you don't really have the fake intervention that looks like a classical placebo, and some people have problems with this. The important thing is having a control.

The need for placebo controls in certain fields is actually overblown:
https://www.ncbi.nlm.nih.gov/pubmed/12535498

Things that don't respond to placebo, like cancer, don't NEED a placebo control.
Ethically, particularly in medicine, it's often much better to compare two interventions:

https://www.ncbi.nlm.nih.gov/pubmed/14986774

Discussion on this is ongoing, but the point is that concern is often overblown.
Psychology can't always use placebo controls because it's the study OF placebos, in effect; it's a study of expectation and the effects of that.
That doesn't make it unscientific, as long as the methodology is sound and the terms are well defined -- unfortunately, they often aren't.