Imza wrote:Is psychology a science? Working on my PhD in psychology and doing scientific research, to me this question seems ludricious but I still see it being asked frequently, both from graduate students I meet in other fields and on internet discussion boards. As a result, I wanted to open this topic up with some questions I had regarding this issue to see what everyone else thinks. I thought it might be better to start this in the general science section but I'll start here and maybe if it's relevant, it can be moved.

So the questions I have or at least feel are important to ask on this topic are the following...

Why do people think psychology is not a science? Is it simply lack of current knowledge on psychology as a field (thinking psychology is still Freudian based) or do people have legitimate criticisms of the methodology used in psychology that disqualifies it from being considered a real science? What exactly would psychology have to do to meet this "supposed" criteria to be considered a science?

Some arguments I have seen also differentiate psychology by calling it social, soft or behavioral sciences. What exactly is the distinction here that separates it from hard, physical, and natural sciences? For the label "social" it seems at least logical in terms of the subject matter but I think it implies more than just that.

Also, just to note, I don't necessarily support the implications of the questions above, I just think this is an interesting topic that is often discussed by people that are not actually in the field of psychology and would be interesting to here the perspective of those people here who work in the field or are interested in it. I would also be interested to see any research or professional material on this topic, I have searched and found very few academic papers that cover this topic.

I'm currently trying to write an article on this topic because I think it's an interesting issue. The gist of what I'm hoping to write is that yes it's a science, but the long answer is: yes, with caveats. Overall, I think the main hurdles for most people can be summed up as difficulties in understanding: 1) what science is, and 2) what psychology is. The first can be answered without getting into any real controversial area of the philosophy of science by pointing out the fundamental aspects of what makes something a science which most people would probably agree with: i.e. testable ideas, experimentation, objective data, conclusions based on balancing the evidence, results that can be replicated etc. When we understand that science doesn't need to produce perfect results, or amazing conclusions, because science is the method not the conclusion, then it becomes much easier to understand how psychology can be a science.

The second hurdle is how people perceive psychology. As you mention, there is the pervasive trope through our culture that psychology is based on Freud. More to the point, I was talking with a friend yesterday and I told him that I did psychology in university, and his response was: "So, can you like read my mind now?". So there is a significant PR problem with psychology and very few people seem to understand what exactly it is that we do. After explaining how Freudian analysis has been rejected from mainstream psychology for at least half a century, and realistically, far longer than that in experimental psychology, there still comes the idea that psychology = clinical psychology. People ask how applying therapeutic techniques to the "human mind" can be considered a science, you just "give them drugs and talk to them for a bit". This conflation is obviously mistaken, as it's akin to confusing biology with medicine - where nobody would deny the claim that medicine is not a science, it's an applied field or an "evidence-based practice". This is what clinical psychology is.

This leads us to the next problem in understanding what psychology is: the belief that it is a "social science" or part of the "humanities". This stems from the idea that psychologists do what they do to understand humans and the society they live in. For some psychologists this of course can be the case, but calling the field as a whole (which is largely unconcerned with human subjects or the implications their findings have on humans and their society) is highly inaccurate. It is similar to finding an AI researcher whose work impacts how we think about the behavior of humans, and suggesting that AI or computer science is thus part of the social sciences of humanities. Now, there is nothing inherently wrong with social sciences or the humanities (and the question of whether they are "real" sciences is another interesting one), but the reason why it is problematic to apply it to psychology is because it produces the assumption psychology cannot be a science because it cannot accurately control for all the variables (due to the unethical nature of such experiments being performed on humans). And to a degree, this is true for the areas using human subjects - they have to resort to quasi-experimental methods, instead of fully controlled studies. But such an objection to psychology as a science assumes that if humans cannot be studied experimentally, then it cannot be a science. Which of course is untrue. (This also leads to a debate about the necessity of randomised-controlled trials which, as I'm sure you're well aware, are not essential for scientific study when another statistical design is used in its place; e.g. single-subject designs).

Leading into the next problem is the idea that human behavior is "random" or "unpredictable". I think this stems from our belief in free will, and the apparent agency behind our actions. However, from studying psychology, we know that humans are incredibly predictable; essentially they are nothing more than large meaty billiard balls. The apparent "randomness" of human behavior comes about due to our ignorance of the environmental variables controlling our behavior at any given point in time, and once we start controlling for even just some of them, it all suddenly makes sense. It's like observing falling objects in a number of different places in the universe and concluding their actions are "random", before realising that controlling for the variable 'gravity' results in their behaviors suddenly become uniform and structured.

The major problem with viewing psychology as the study of humans, is that it misses precisely why psychology (at least in part) is a natural science. It is the study of behavior; what makes organisms tick. Using objective data, repeatable experiments, etc, we generate universal and fundamental laws which can account for and predict with great accuracy the behaviors of a range of organisms in any given situation. The organism itself is irrelevant, as the subject matter that the field is interested in is behavior. Behavior, as a phenomenon to be studied, is a natural part of the universe that responds in predictable ways to variables - making psychology, necessarily, a science.

With all that said, there are some "more valid" concerns with the state of psychology. For example, the Kuhnian objection that psychology has no underlying theory/paradigm, or the objection to the use of statistical inferences in place of experimental manipulations for the more social aspects of psychology, but these are just the caveats I mentioned earlier. Some areas of psychology are more scientific than others; neuropsychology, cognitive psychology and behavioral psychology obviously are largely no different to physics or chemistry, whereas social psychology, personality psychology, and the medical research behind clinical psychology are less so, due to their restrictions on what variables they can and cannot manipulate.

As for academic sources on this, I like Schlinger's discussion here: Why Psychology Hasn't Kept Its Promises

ABSTRACT: This essay posits that psychology’s general lack of respect as a science stems from two related problems: the continued focus on conceptually vague mentalistic constructs and the adherence to a methodology that emphasizes statistical inference over experimental analysis. The lack of a thoroughgoing experimental analysis has so far prevented psychologists from discovering a set of foundational principles thus inhibiting them from being able to predict and control individual behavior. Psychologists can remake their conceptual and methodological foundations by focusing on the relationship between observed behavior and its context and by adopting methods of experimentation that would aid in the discovery of orderly functional relationships. This knowledge could be used to more parsimoniously explain complex behavior, including cognitive phenomena, and to more effectively solve practical problems

There are areas where I disagree with him, like I think he makes the same mistake of strawmanning cognitive psychology that Skinner made, but he also makes some really good points. The paper also links to (and discusses) some other good criticisms of (parts of) psychology and discussions on its state as a science:

Holth, P. (2001). The persistence of category mistakes in psychology. Behavior and Philosophy, 29, 203–219

Kimble, G.A. (1996). Psychology: The hope of a science. Cambridge, Massachusetts: The MIT Press.

Cohen, J. (1994). The earth is round (p < .05). American Psychologist, 49, 997–1003

Loftus, G.R. (1996). Psychology will be a much better science when we change the way we analyze data. Current Directions in Psychological Science, 5, 161–171.

Machado, A., Lourenço, O., and Silva, F.J. (2000). Facts, concepts, and theories: The shape of psychology’s epistemic triangle. Behavior and Philosophy, 28, 1–40.

Viney, W. (1996). Disunity in psychology and other sciences: The network or the block universe. Journal of Mind and Behavior, 17, 31–44.

twistor59 wrote:Didn't Feynman once make some provocative comments about psychology ? Or was he only talking about psychoanalysis - Freud and all that stuff ?

Yeah his comments about it being a cargo cult are largely concerned with the relevance of psychoanalysis:

All experiments in psychology are not of this [cargo cult] type, however. For example there have been many experiments running rats through all kinds of mazes, and so on — with little clear result. But in 1937 a man named Young did a very interesting one. He had a long corridor with doors all along one side where the rats came in, and doors along the other side where the food was. He wanted to see if he could train rats to go to the third door down from wherever he started them off. No. The rats went immediately to the door where the food had been the time before.
The question was, how did the rats know, because the corridor was so beautifully built and so uniform, that this was the same door as before? Obviously there was something about the door that was different from the other doors. So he painted the doors very carefully, arranging the textures on the faces of the doors exactly the same. Still the rats could tell. Then he thought maybe they were smelling the food, so he used chemicals to change the smell after each run. Still the rats could tell. Then he realized the rats might be able to tell by seeing the lights and the arrangement in the laboratory like any commonsense person. So he covered the corridor, and still the rats could tell.
He finally found that they could tell by the way the floor sounded when they ran over it. And he could only fix that by putting his corridor in sand. So he covered one after another of all possible clues and finally was able to fool the rats so that they had to learn to go to the third door. If he relaxed any of his conditions, the rats could tell.
Now, from a scientific standpoint, that is an A-number-one experiment. That is the experiment that makes rat-running experiments sensible, because it uncovers the clues that the rat is really using — not what you think it's using. And that is the experiment that tells exactly what conditions you have to use in order to be careful and control everything in an experiment with rat-running.
I looked into the subsequent history of this research. The next experiment, and the one after that, never referred to Mr. Young. They never used any of his criteria of putting the corridor on sand, or of being very careful. They just went right on running rats in the same old way, and paid no attention to the great discoveries of Mr. Young, and his papers are not referred to, because he didn't discover anything about rats. In fact, he discovered all the things you have to do to discover something about rats. But not paying attention to experiments like that is a characteristic of cargo cult science.

Essentially, the exception to the 'cargo cult' rule is how psychology is actually done (and was mostly done before he made these comments), and the likely explanation is that he simply didn't know much about psychology. He was basing his opinions largely on his impression on the field, rather than the actual research being done within it. The reason why Young isn't often cited is because his finding wasn't overly novel - researchers like Guthrie, Hull and Tolman had already discovered that rats use a number of different methods to determine the correct route.

Efilzeo wrote:It is not a real science because its object is not objectively misurable.

Behavior isn't measurable?

Efilzeo wrote:

Mr.Samsa wrote:

Behavior isn't measurable?

Not objectively, I think. It is more like a collection of experiences, where you try to find some common rules. But you do that a posteriori, by the consequences.

This is how science is always done. We observe, and generate possible general laws. Then we turn these laws into mathematical laws that are used to accurately predict and control phenomena. If they don't accurately predict phenomena, then we adjust them or reject them completely.

Efilzeo wrote:Even if there are many experiments, they are not based on a stable object, they're based on humans,

Why focus on humans? Behavior is applicable to all living things, which allows us to create general laws.

Efilzeo wrote:...who can be mutated by so many factors, non-physical factors.

I'm not sure what you mean by "non-physical" factors that change behavior? Given what we know about behavior, finding the causes of behaviors are fairly easy and these causes appear to be physical. There doesn't seem to be any room for non-physical causes to change behavior.

Efilzeo wrote:In fact also medicine is not a certain science, but the difference is that in medicine you have a physical object.

I agree that medicine isn't a science, because it's an applied evidence-based practice, but how can disease not be a "physical object"? And given your comparison of psychology with medicine, is it possible that you've confused psychology with clinical psychology?

Efilzeo wrote:Anyway I don't think that there is a complete difference between 'science' and 'non-science', I think there is a graduality of 'science' on many disciplines.

I agree with this. However, certain areas of psychology are of course natural sciences since they are just as rigorous as chemistry or physics.

Efilzeo wrote:

Mr.Samsa wrote:

This is how science is always done. We observe, and generate possible general laws. Then we turn these laws into mathematical laws that are used to accurately predict and control phenomena. If they don't accurately predict phenomena, then we adjust them or reject them completely.

I don't know to which kind of psychology you are referring to, but which areas of it has mathematical laws and can be predictive?

Most areas of experimental psychology involve predicting phenomena, and the easiest way to do this is through mathematical modelling. For example, behavioral psychology has mathematical laws for choice (matching law), self control (Rachlin and Green's model), etc; psychophysics and cognitive psychology have laws of timing (Weber's law) and discriminability factors (contingency discriminability model), etc; and of course neuropsychology includes laws which measure chemical reactions, or the effects that damage on particular parts of the brain has on behavior. I'm sure that even the fluffier sides of psychology have mathematical laws, like social and personality psychology, but I'm not well-versed enough in those areas to list any.

Since humans are such predictable objects, most of these laws are highly predictive. Look at the matching law, which predicts for how organisms choose between alternatives, and accounts for (usually) between 95-100% of the variance.

Efilzeo wrote:

Mr.Samsa wrote:Why focus on humans? Behavior is applicable to all living things, which allows us to create general laws.

Yes and also with other animals you can just see what they do and try to describe it. But you're not certain if your explanation is correct, even if you do experiments on a sample, and this problem is much more relavant with humans because we are more influenced by mind factors.

I'm not sure what you mean by "you're not certain if your explanation is correct". If your explanation is parsimonious, falsifiable, and highly predictive of novel situations, then it is scientifically correct. We can apply your argument to gravity - sure, we can describe how objects fall using Newton's laws of motion, but we don't know if our explanation is "correct" because invisible gremlins might be pulling objects down to the ground in a fashion that is perfectly consistent with our notion of a relationship between mass and falling objects.

Efilzeo wrote:

Mr.Samsa wrote:I'm not sure what you mean by "non-physical" factors that change behavior? Given what we know about behavior, finding the causes of behaviors are fairly easy and these causes appear to be physical. There doesn't seem to be any room for non-physical causes to change behavior.

You're right, the word 'non-phisical' has actually no sense, but I used it meaning the psychological aspect of our mind.

Sure, but psychological factors are able to be studied. We know what causes thoughts and beliefs, etc, and that's how we can accurately manipulate them. They're just parts of a causal chain.

Efilzeo wrote:

Mr.Samsa wrote:I agree that medicine isn't a science, because it's an applied evidence-based practice, but how can disease not be a "physical object"? And given your comparison of psychology with medicine, is it possible that you've confused psychology with clinical psychology?

Yes is possible, but what do you mean with 'psychology'? The applicated one isn't just a collection of techniques? Phisiological pshycology, social psychology, work psychology etc., isn't it just a collection of experiences based on a common context?

Clinical psychology is the application of what we know about the way people think and behave to help treat mental disorders. Psychology, on the other hand, is the study of behavior. When we study behavior we do so using experimental techniques, predictive laws, and we identify causal relationships between phenomena. Clinical psychology thus isn't "scientific" as it's not trying to be scientific, it's trying to apply scientific knowledge to create change in the world.

Efilzeo wrote:

Mr.Samsa wrote:I agree with this. However, certain areas of psychology are of course natural sciences since they are just as rigorous as chemistry or physics.

If we exclude the neuro-psychology, which one is comparable with natural sciences according to you?

Behavioral psychology is the obvious one, given that we've discovered a number of universal laws that apply to all organisms.

Pebble wrote:

Mr.Samsa wrote: This conflation is obviously mistaken, as it's akin to confusing biology with medicine - where nobody would deny the claim that medicine is not a science, it's an applied field or an "evidence-based practice". This is what clinical psychology is.

Seems like the correct place to address the issue raised in the other thread. Medicine is certainly not a science, it is however, moving in the direction of a solid evidence based foundation. This has come at a price, learning from previous mistakes, improving the quality of evidence that is accepted as evidence and being considerably more coy about asserting the validity of ideas that are subsequently likely to be shown to be baseless.

Unlike standard science the tools are necessarily somewhat different - since the results of experimental methods are not always generalisable or indeed fully applicable. However the benefits of rigorous trial construction, recognition of the sources of bias and high quality audit and registries have been incorporated to determine if the results of changing practice lead to predictible outcomes.

Agreed, definitely. In case I need to clarify, by pointing out that medicine is not a science, I was not attempting to disparage medicine or to suggest that its conclusions are suddenly questionable. It's simply a comment on the obvious limitations of treating patients, and the impossibility of rounding up patients on a ward, dividing them into randomised groups and giving them different medicines as part of standard treatment. Instead, as you say, they rely on the evidence gained from solid scientific studies where the treatment chosen is decided to be the best option for a given patient, disease and situation.

In other words, being a "science" isn't an automatically good thing for any given field. If medicine were a science, then I guarantee that the treatment and health of patients would necessarily fall. (And, of course, medical research is a science as it differs from the day-to-day work of doctors and surgeons etc).

Pebble wrote:

Mr.Samsa wrote:
Some areas of psychology are more scientific than others; neuropsychology, cognitive psychology and behavioral psychology obviously are largely no different to physics or chemistry, whereas social psychology, personality psychology, and the medical research behind clinical psychology are less so, due to their restrictions on what variables they can and cannot manipulate.

Clinical psychology I do not keep fully upto date with, so you will need to point out to me whether any DBRCTS can ever be constructed in this field, if there are occasional or widespread examples of adequately powered RCTS, whether these are accepted as simply hypothesis generating and repeated by indepenent researchers in confirmatory trials before being accepted as valid, whether any large scale audits of the impact of implementing the results occur? I appreciate that CBT meets some of these requirements, but am unaware of other examples.

CBT, as you mentioned, has garnered popularity in the success of the RCTs in the area, but a number of other psychotherapies have as well. I think your difficulty in finding results here might be due to poor search terms; in the other thread you mentioned that you searched for "double blind" and "clinical psychology". This would be like searchint for "double blind" and "medicine" - nobody really publishes their paper as "A double blind study of clinical psychology" in the same way they don't publish "A double blind study of medicine". Searching for "psychotherapy" is more effective, if you (understandably) can't be bothered trying to hunt out the specific names for treatments and look at them in relation to RCTs.

The Cochrane reviews come up with quite a few hits, with each review containing numerous papers with RCT methodology. (Obviously some reviews rely on weaker evidence, i.e. no RCTs available, so not all of the reviews of the search will contain RCTs).

Generally, most psychological treatments take a while to become common practice, in the same way that medical treatments require vast amounts of evidence in order to be implemented. Look at CBT, a treatment that has a vast amount of support behind it, but studies are still being done to specify exactly where its strengths are and where its weaknesses are.

Also, as a side point, whilst RCTs are hugely useful statistical designs, they are not the only statistical tool we can use to reliably establish a causal link. For example, single-subject designs (or Small-N designs) can be used to achieve the same power as RCTs with only a handful of subjects. There are limitations to the interpretation of their results, but of course there are limitations to RCTs as well, which Small-N designs can avoid. As such, psychology has also a vast amount of research demonstrated the effectiveness of various treatments, like language therapies, anxieties, eating disorders, and of course in treating/curing autism (with most of the research using this design being done in the field of applied behavior analysis).

Pebble wrote:"Double blind' & "medicine" = 26,464. Unsurprising since double blind and the other term are used as 'and', not necessarily immediately linked and the search is of the entire abstract and all key words.

Interesting. I still think my point is valid though, just "medicine" wasn't a good example.

Pebble wrote:My point was that only where one can get as far as double blind is one dealing with truly unbiased information. RCT are an improvement other methodologies, what I have read of Cochrane reports in this area is largely negative (nothing is proven), with RCTs that are either inconsistent, or inadequate or not repeated in a standardised fashion to allow conclusions.

Hmm not from what I've seen. But let's assume that all of the RCTs done in clinical psychology turn up negative - so what? Science is a process, a methodology, not a conclusion. So even if they haven't been able to conclusively establish something, it doesn't change the fact that they are doing science. However, I disagree that they are largely negative. I'd say a lot of them present evidence from both sides, with the evidence weighing one way or the other, or they are largely overwhelmingly in favour of some treatment.

Pebble wrote:Small trials have inherent statistical limitations - in clinical field they never prove anthying, they are but hypothesis generating unless you have near 100% efficacy, which is clearly nonsense except with antibiotics.

I agree, but I wasn't discussing small trials, I was discussing single-subject designs. The difference being that you don't have, for example, a small control group and a small treatment group - you have a small group that experiences all conditions. This way they are more than just hypothesis generating, they are able to identify causal relationships.

Pebble wrote:1. My point really was that 'double blind' is a pretty well impossible design for interactive therapies since there is no indistinguishable placebo comparator - single blind can be managed but is rarely tried and has inherent design faults due to differences in patient experience - thus for example comparing surgery to medicine is inherently flawed because of the greater placebo effect associated with the more dramatic intervention (surgery). Surgery versus sham surgery produces markedly different results. Thus comparison of intensive therapy versus usual care is inherently flawed in psychological trials.

Indeed, that part can be difficult, in the same way that surgery versus sham surgery cannot be done double-blinded but it can be done in some cases (to a degree) in clinical psychology. Often psychotherapy is controlled with "education", where those implementing it are not sure which is the treatment group and which is the placebo, or the psychotherapy is done in conjunction with medication (and medication on its own), and the clinical psychologists implementing it cannot know which group the patient is in.

I admit that this part of research in clinical psychology is problematic, but it applies to a lot of other healthcare fields - like surgery, physiotherapy, etc. It's something that we need to be aware of, but it doesn't rule out its validity.

Pebble wrote:2. I have no requirement that all trials are negative or positive. Simply that each trial should be analysed for bias using "SIGN' or similar criteria and the over all contribution to any conclusion weighted based on the degree of freedom from bias. Second that where the available data are contradictory, the weight of evidence using relatively objective analytical approaches should be used. Meta-analysis are a minimum standard, though hugely over rated.

As far as I know, this is already how it's done.

Pebble wrote:3. Uncontrolled studies cannot lead to fair conclusions in respect of efficacy - the understanding of disease causation in the field of psychology is far too rudimentry to really give credit to such approaches. So I don't buy conclusions from small 'pathology' only uncontrolled studies of intervention - the history of medicine is littered with the erronous consequences of such methods of appraisal.

Edit: Cause and effect in uncontrolled studies - up to a point! One is really studying associations, plus or minus temporal relationship, one may get some insights into dose relationship if the study is large enough (generally we are into the thousands now). But even that is not enough - does removal of cause lead to regression? Now we need comparator groups. Is there an evidence based pathobiological relationship? Back to the neuropsycholgists now.

I agree that uncontrolled studies cannot lead to conclusions about causal relationships, and I don't think any psychologist would suggest that they can. And not all mental disorders will have a biological cause, so we won't usually have to come back to the neuropsychologists to discuss the cause of disorders.