Is sociology a science?

[Jerrymander]

I heard it said that sociology isn't a science because it makes theories that can't be falsified or that sociologists make unscientific claims.

I personally don't understand this. Questions like "is drug-addiction more prevalent among certain ethic groups?" or "Are there gender bias in hiring in banks?" are certainly things that can be tested and falsified.

And how would the anti-sociology say how we should know things about human behavior, cultures, institutions, ect?

 

[Loss Leader]

Advertising works, and that's probably the most concrete, practical application of sociology.

But the field does encompass a lot of social philosophy which is, at best, pre-scientific and, at worst, entirely untestable.

 

[Skeptic Ginger]

What philosophy?

I can see ideology getting confused with science, but there remains the science once you filter the non-science out.

 

[Dinwar]

Fundamentally, sociology is supposed to examine how humans behave. All systems obey rules, and human interaction--being a system--must necessariy obey rules. It's perfectly valid to look for those rigorously, and that makes sociology a science.

Sociology can, once the rules are identified, also provide insight on how socieities should operate to achieve some goal--in precisely the same way that evolutionary theory can be used to provide insight on agriculture. In as much as those rules are rigorously determined, sociology remains a science in this activity.

That said, sociology runs into some fairly unique issues, and many conclusions that are drawn are of limited in scope--it is entirely possible (and in fact trivially true) that a rule is limited only to an individual group. It is therefore extremely easy to over-state one's conclusions. Data are also problematic, in that there are innumerable complications with collection. Surveys are always suspect, experiments are often not possible, and natural experiments may not occur.

There's also the fact that because we're dealing with groups, questions quickly become politicized. At that point, it's no longer science. That's obviously a problem with individual researchers or research groups.

 

[Dancing David]

yes, done well it is.

 

[Dinwar]

yes, done well it is.

The devil's in the details.

 

[hgc]

Sociology covers a wide range of disciplines. A few that I know are very focused on evidence-based theories and heavy use of statistical analysis. At least some aspects of their work look a lot like hard science.

 

[fuelair]

sort of. Just not a hard science- not even close to Hari Seldon level

 

[Dinwar]

sort of. Just not a hard science- not even close to Hari Seldon level

I've never fully understood the difference between hard and soft sciences. From the layman's perspective the appear fairly different, but once you get into the details they often seem nearly identical.

I think a bigger issue is mature vs. imature science. An imature science will have a lot of nonsense floating around--see psychology in the Freud days, or physics in ancient Athens, or early geology. We don't know enough to know what ideas are nonsense, so we test everything. Mature sciences such as modern physics and geology, in contrast, have figured out the basic frameworks, and have established ways to determine what is a valid hypothesis vs. what is bat-crap crazy.

 

[dudalb]

sort of. Just not a hard science- not even close to Hari Seldon level

I love Isacc Asimov, but his idea that you can have a science dealing with Human Behavior that has all the mathmatical certainlty of physics or chemsitry is pretty silly. I think it's case that he was fascinated by history but by training was a biochemist,and dreamed about applying the discipline of one to the other. In the real world, it's not going to work except maybe for some areas of economics,and even that is questionable.

 

[Dinwar]

I love Isacc Asimov, but his idea that you can have a science dealing with Human Behavior that has all the mathmatical certainlty of physics or chemsitry is pretty silly. I think it's case that he was fascinated by history but by training was a biochemist,and dreamed about applying the discipline of one to the other. In the real world, it's not going to work except maybe for some areas of economics,and even that is questionable.

Biology as a whole is a series of exceptions with just enough rules to hold the whole thing together.

 

[smartcooky]

I heard it said that sociology isn't a science because it makes theories that can't be falsified or that sociologists make unscientific claims.

Well an awful lot of "science" isn't science using that criteria.

Cosmology for a start. Their are plenty of competing theories as to the nature of the beginning of the universe that cannot be proven or falsified.

Has anyone proven/falsified any of the following?

Inflation
Big Bang
Dark Matter
Dark Energy
String Theory
Multiple dimensions (beyond the three we know about)
Multiple universes
Branes

 

[blutoski]

I've never fully understood the difference between hard and soft sciences. From the layman's perspective the appear fairly different, but once you get into the details they often seem nearly identical.

I think there's a usable distinction, although I feel the nomenclature is reversed.

If the subject of study is simple and controls can isolate the factor being tested, it's a hard science. If it's complex and controls can only partly isolate the factor being tested, it's a soft science.

The high degree of complexity makes the soft sciences considerably more difficult to study, and ironically are the ones I consider 'hard'. Medicine, sociology, economics, psychology, political science &c.

 

[W.D.Clinger]

Well an awful lot of "science" isn't science using that criteria.

Cosmology for a start. Their are plenty of competing theories as to the nature of the beginning of the universe that cannot be proven or falsified.

Has anyone proven/falsified any of the following?

Inflation
Big Bang
Dark Matter
Dark Energy
String Theory
Multiple dimensions (beyond the three we know about)
Multiple universes
Branes

None of those have been falsified. No scientific theories are ever actually proven, but some are more thoroughly tested than others.

The four-dimensional nature of spacetime is supported by an impressive body of direct experimental evidence, so the physicality of multiple dimensions (beyond the three you know about) is about as well established as, for example, the germ theory of disease or the theory of evolution.

The evidence for a cosmological big bang is also quite impressive.

Inflation, dark matter, and dark energy are hypotheses that have been proposed to explain a variety of observations. These hypotheses can be tested by working out their consequences and making further observations. Although these hypotheses are not as well-established as the theory of relativity or big bang cosmology, the testing conducted so far suggests they are more likely than alternative hypotheses.

String theory and branes are broad families of proposed theories. From what I understand as a non-physicist, some variations of those theories have been tested and found wanting. Other variations of those theories remain viable.

"Multiple universes" can refer to a wide variety of ideas, most of them controversial. If they had been falsified, however, they would not be controversial.

ETA: Getting back to Jerrymander's statement, most of the items in smartcooky's list are indeed falsifiable (which is to say, they do make definite predictions that can be tested via experiment or observation). Multiple universes (possibly depending on what you you take that to mean) might not be falsifiable. Some specific string or brane theories might not make testable predictions, but the fact that some have already been falsified shows that unfalsifiability is not a charge that can be levelled against all such theories. All of the other items in smartcooky's list are falsifiable, but have passed the tests conducted thus far.

 

[fuelair]

I love Isacc Asimov, but his idea that you can have a science dealing with Human Behavior that has all the mathmatical certainlty of physics or chemsitry is pretty silly. I think it's case that he was fascinated by history but by training was a biochemist,and dreamed about applying the discipline of one to the other. In the real world, it's not going to work except maybe for some areas of economics,and even that is questionable.

Actually, that was the point I was headed for: ?thumbs up? image didn't come up on posting.....

 

[DarthFishy]

As someone who is sort of a sociologist I would say that yes it definitely can be a science. The biggest problem (as has been alluded to here), is the generalisation of results. Due to the large number of uncontrollable variables in the average sociological study it is difficult to to impossible to generalise about the results you get.

As an example I performed a recent study (which I still need to properly write up and get published) comparing my students results on aptitude tests to their final marks in University programming courses (at both first and second year). While I found some very interesting results (i.e. that there was no correlation between their results in second year and their results in the aptitude test) there are a large number of possible confounding factors that would make it difficult to proof or even repeat the results.

For example if I wanted to run the study again this year, I would get a new group of students, though I would have some repeaters so they would have to be removed from the study group. Furthermore I know the first year programming lecturer is constantly trying to improve her course so she might change things within the course. One thing we did decide to change was the programming language that they did at first year to better align with second year.

All these factors make it nearly impossible for me to claim that for all programming students aptitude tests are not significant in predicting future success.

So what can I do because this is still something that I strongly feel needs to be studied (for example if it can be shown that mathematical aptitude is not that important for programming then the entry requirements need to changed which could allow many more students into the field)? I try and study the phenomenon as best I can, trying to control for what I can, and when I publish my results I clearly state the limitations of my study.

If the study is replicated in similar contexts by enough other researchers we can start saying we suspect that the hypothesis might be true.

 

[DarthFishy]

I think a bigger issue is mature vs. imature science. An imature science will have a lot of nonsense floating around--see psychology in the Freud days, or physics in ancient Athens, or early geology. We don't know enough to know what ideas are nonsense, so we test everything. Mature sciences such as modern physics and geology, in contrast, have figured out the basic frameworks, and have established ways to determine what is a valid hypothesis vs. what is bat-crap crazy.

This is also very true, but it makes it an exciting time to be involved in a field. With some luck in 50 years time students within my field might by studying up the DarthFishy framework, or methodology for XYZ

 

[erwinl]

As someone who is sort of a sociologist I would say that yes it definitely can be a science. The biggest problem (as has been alluded to here), is the generalisation of results. Due to the large number of uncontrollable variables in the average sociological study it is difficult to to impossible to generalise about the results you get.

As an example I performed a recent study (which I still need to properly write up and get published) comparing my students results on aptitude tests to their final marks in University programming courses (at both first and second year). While I found some very interesting results (i.e. that there was no correlation between their results in second year and their results in the aptitude test) there are a large number of possible confounding factors that would make it difficult to proof or even repeat the results.

For example if I wanted to run the study again this year, I would get a new group of students, though I would have some repeaters so they would have to be removed from the study group. Furthermore I know the first year programming lecturer is constantly trying to improve her course so she might change things within the course. One thing we did decide to change was the programming language that they did at first year to better align with second year.

All these factors make it nearly impossible for me to claim that for all programming students aptitude tests are not significant in predicting future success.

So what can I do because this is still something that I strongly feel needs to be studied (for example if it can be shown that mathematical aptitude is not that important for programming then the entry requirements need to changed which could allow many more students into the field)? I try and study the phenomenon as best I can, trying to control for what I can, and when I publish my results I clearly state the limitations of my study.

If the study is replicated in similar contexts by enough other researchers we can start saying we suspect that the hypothesis might be true.

True.
You have to be very, very conscious about the restrictions of the study. Precisely because there are a lot more influences than is usual in a more mathematical oriented science.

And then you have what the professor of my wife said, way back. (parafrase) 'No matter how you try to control all the external influences, sometimes the rat just does as it pleases.'

 

[smartcooky]

I love Isacc Asimov, but his idea that you can have a science dealing with Human Behavior that has all the mathmatical certainlty of physics or chemsitry is pretty silly. I think it's case that he was fascinated by history but by training was a biochemist,and dreamed about applying the discipline of one to the other. In the real world, it's not going to work except maybe for some areas of economics,and even that is questionable.

To be fair to Isaac, his character, Hari Seldon, was using psycho-history to make only general predictions about the future behaviour of billions of people, and it went off the rails pretty quickly when the unpredictable occurred; the arrival of "The Mule".

If you want a good "real world" example of something akin to psycho-history at work, look no further than Nazi Germany; one man managed to lead a nation into war and it people to endorse genocide. His speeches played on the psychology of the German people, and they reacted predictably to that. It was mob psychology!

 

[Dinwar]

If you want a good "real world" example of something akin to psycho-history at work, look no further than Nazi Germany; one man managed to lead a nation into war and it people to endorse genocide. His speeches played on the psychology of the German people, and they reacted predictably to that. It was mob psychology!

I'm not sure you can blame Hitler for everything. The man was evil, no doubt, and the blood of millions is on his hands--but without a very strong organization supporting him, and the recent history of Germany, and likely the philosophies that Germany introduced in the previous century, he'd just have been another crazed lunatic.