Hacker News Reader

The reproducibility problems of psychology general - of which this is only a small fragment - is showing a few things:

1) Psychology is harder than almost anyone thought.

2) We are only now starting to have enough data to know just how hard psychology really is.

3) There is too much of an onus to publish, "Publish or perish" in science in general.

4) Negative results are very valuable for the public and all other researchers, but generally negative for the researchers career. Even scientists need food.

5) People will use whatever that can pass for facts to further their own agenda, whatever that may be. Left, right, and center. Incidentally, psychology as a science has now taught us that both through its successes in that area, and through how people outside of psychology believed the authorities of psychology to be implicitly right whenever it fit their point of view.

Psychology is a science, no need for the quotes, it's just happens to be incredibly hard to turn the experimental results into reliable facts in that particular field, so shortcomings of both people and systems become magnified. The high difficulty is probably also why some authorities might have gotten an unusually large role in the social networks underlying any area of research, and the public eye. People want answers, even if there isn't really enough data, and even when researchers says: We really don't know.

Science is always only our best guess, in some areas it pretty darn good, but in others not so much, yet.

> If anything, this whole debacle has revealed to me how pathetic the whole "science" of psychology is in its current state.

What? How an academic being defensive of his study, conducted in 1971, an invalidation of "current" psychology? That's 40 years ago.

> I mean, has the field of psychology given us anything of value? I'm asking genuinely.

What an extreme leap, and an odd generalization for me to have to argue against. How about Milgram, Asch, Cognitive Dissonance? Studies of memory, vision, neurons can inform the development of AI. Things like IQ tests have helped detect and prove that leaded gasoline was damaging the intellect of a generation. Psychology is THE field to give clarity to question such as what gender dysphoria is about (research shows it is usually correlated with a brain that is more structurally similar to the gender one identifies as [1]).

So I'm not sure where a wholesale indictment of a field could really come from.

[1] https://en.wikipedia.org/wiki/Causes_of_transsexuality

You're mixing up psychology with psychiatry and neuroscience... These are intertwined to about 80s. Not anymore, hopefully.

The hardest of these is probably neuroscience and produces very micro results, but at least they are known to be valid.

I think the evidence begs to differ:

https://www.nature.com/articles/nrn3475

https://www.sciencedirect.com/science/article/pii/S009130571...

That wiki link, did you read the section on psychology in it? It pretty much reinforces my opionion. Seems like most of the real understanding of it is coming from biological factors.

Now it seems that those initial psychological theories were all argued with similarly low sample and methodologically wonky studies. Just paraphrasing the wikipedia link. In effect, that probably hurt more then it helped and that's why we're starting to see things like Denmark declassifying it as no longer a mental disorder.

https://www.scientificamerican.com/article/where-transgender...

Edit: I know I'm making a big blanket statement. As someone who does not read about all the latest psychological breakthrough, I'm saying in honesty that with my limited knowledge, the impression I get from psychological research is that its not very thorough. So I'm open to hearing that in the last 40 years this has changed, an example would be great, so that I can start maybe changing my opinion.

1. I provided 6 separate examples of the use of psychology (3 experiments, 3 real-world use-cases). You responded to one.

2. Something being biological is not contrary to it being psychological.

Your response is disingenuous. You certainly are well-aware that your blanket statement was a rash, emotional inaccuracy. Rather than concede that point, you're choosing to try to nit-pick at one sixth of the refutation.

This is HN, best just to admit your mistake.

> I mean, has the field of psychology given us anything of value?

The interactive computer system, thanks to the vision of J.C.R. Licklider (who was trained as a psychologist).

Good one, though it's not really related to the field of psychology and its research.

I guess maybe one could argue his vision was influenced by his studies of psychology, but its hard to know.

The problem, from where I sit, is that certain types of experiments have a built error detection or safety rails and psychology doesn't have this. For instance if a physics experiment has a result which appears to violate conservation of momentum or has information being sent faster than the speed of life, then it is likely that there is a flaw in the experiment. Unfortunately many of the social sciences don't have these easily detected error conditions. Thus, it becomes much harder to determine is an experiment is flawed or not. Making this situation even worse social sciences are very hard to control for and replicate because you can't just order "one standard human" for the experiment. Two experiments may not replicate each other due to no fault of the experiment. What we need is a model of human psychology that provides clear boundary conditions.

But there are methods to deal with more or less all of these problems, that is why social scientists learn research design/methodology separately so much. In my opinion being able to replicate results should still be required to accept a theory. My problem with this case is that things that I consider basics for the scientific method (replication or even a control group) are obviously missing.

To the people downvoting my comment I'd be genuinely interested in hearing why people see my comment as damaging to the discussion.

>But there are methods to deal with more or less all of these problems

There are methods to deal with designing experiments poorly, but even the best experiment could have some unknown unknown issue that skews the results. Physics has an inherent advantage that experimental mistakes often lead to results which contradict theory to such an extent that the experimenter can be fairly sure the experiment is in error.

> In my opinion being able to replicate results should still be required to accept a theory.

Strongly agree, however replication often becomes harder to achieve as the complexity of the system you are measuring increases. Photons are simple in ways that humans just aren't. For example various mental disorders only appear in certain cultures, the culture of a photon isn't a problem a physics experimenter needs to worry about. With studies on humans there are so many more places for failure to occur both in the original study and in the replication. Training matters, but the task is significantly more difficult.

I am not going to describe the multitude of advances due to psychological sciences. I'd just like to point out that while the Stanford Prison Experiments got a lot of press, these experiments occurred a long time ago in a particular subfield of pschology, and has very little to do with the validity of psychological sciences as a whole.

With so many to choose from coming up with one you thought would survive scrutiny here should have been as easy as giving the blithe dismissive answer you did.

As blithely dismissive as grandparent who dismissed a whole field of science based on one scientific study (and hand waves at unnamed others)? Are you serious?

If you want a 'practical' set of findings, just take a look at all the psychological tricks learned and now applied to increase social media addiction, gambling addiction, etcetera.

Cognitive biases research have often proven robust under repeated tests and paradigms. https://en.m.wikipedia.org/wiki/List_of_cognitive_biases

I have in my own research examined and confirmed a number of the memory error biases listed on that page.

I'm not dismissive without cause: https://www.scientificamerican.com/article/psychology-s-cred...

As a bystander, I'm not a psychologist, the last few years have been article after article affecting the credibility of psychology.

Then, I paused and thought, what has psychology done to improve our lives? And I can't think of anything.

So I'm calling out to others to show me the flip side. Every wikipedia page I go to has a huge section in criticism that basically says no one knows, some people say this is true and works, an equal amount of others say its not and doesn't. Like the page on CBT: https://en.m.wikipedia.org/wiki/Cognitive_behavioral_therapy

Cognitive Bias is my first showcase of something that seems mostly proven without controversy. Can I ask what has been the next step in using our understanding of inherent cognitive biases? You mentioned they're used to increase addiction to products. That's a shame, has it helped in other areas, maybe to more positive outcomes?

Whatever problems there are, this doesn’t reflect them. As you say, he is the emblem of the field. How do you think he got there? This is someone who has made their career on one surprising result, and is probably quite used to the conference circuit and interviews now. Lots of personal reasons to fight the truth.

Ya, but the fact that his research is still thought at school, his textbooks are still used as material, and that I'm not seeing an openly acknowledging crowd of newer psychologist going: "Oh ya, we know about the methodoligical flaws in Zimbardo's research, and today's research is completly different, way more rigorous, and no one uses his data points for anything serious anymore..." makes me worried there's a pathological trend in the field.

Which is why I openly asked for people to reassure me I'm wrong to believe that the science of psychology is bogus. I've yet to see any data point or argument to reassure me by the way, I've only seen people be dismissive or point that I shouldn't generalise like that. The latter is true, but also doesn't mean my hypothesis is not correct.

Meta-analysis as we know it today is largely as it is because of psychologists using it to study therapy effects.

Referring to Zimbardo and the Stanford Prison Experiment as the "emblem of its field" is a bit strange in my opinion. It's like referring to Snapchat as the emblem of computer science.