Is sociology really a "science"?

A few thoughts on the field from a 6th year sociology PhD student:

There are quite a few sociologists who would absolutely object to labeling sociology a science. This is generally because of anti-positivist, post-modern critiques of the social sciences. As someone said above, sociological data is all interpreted subjectively. You can’t really conduct objective experiments. Your subjective biases influence everything about your work, including the research questions you ask. Claiming to do objective, scientific research may do more harm than good. Better to be upfront about your subjective position and work with it. Etc.

A second issue is that there is a real disjuncture between the folks who do sociological theory and the folks who do sociological research. The theorists who tend to be most influential also tend to be anti-positivist and anti-quantitative research. Meanwhile, researchers increasingly borrow their theoretical underpinnings from other fields. In my specialty area, I see a lot of borrowing from economics and psychology, with some biology thrown in for good measure. I’m sure other areas draw from other fields.

There have been some comments upthread about sociology continually using “dusty” theories, but I just don’t see it. I think it’s important to read the early sociologists to understand where sociology came from and how it’s gotten to where it is, but I can’t remember the last time I saw a journal article trying to test Weber, Durkheim, or Simmel’s theories. I suppose there are some neo-Marxists out there, who I’m sure are enjoying the current global economic situation (at least from an academic standpoint).

I have the most recent issue of the American Sociological Review on my desk. (This is probably the premier general sociology journal.) It includes the following articles:

Types and Trajectories of Music Genres
Cultural Globalization and Arts Journalism
Hispanic Segregation in Metropolitan America
Patterns and Determinants of Racial Segregation at Multiple Geographic Scales
Spatial Dynamics of White Flight
Sibship Size and Educational Attainment in China
Immigrant Children’s Education Achievement in Western Countries
Changing Relationships between Education and Fertility

This is a fairly “science-y” group of articles. The first two are sort of soft, but the three on segregation and the three on education all use secondary, demographic datasets (along the lines of Census data). They all develop hypotheses and test them with the data that they have available. The last article is a good example. The general assumption - since the demographic transition - has been that more education means fewer children. The authors of this article test that assumption for men and women born between 1940 and 1964 in Norway (using Norway’s excellent vital statistics data) and find that it isn’t actually true. There’s no relationship between education level and number of children for women (when controlling for age at first birth). For men, there’s actually a positive relationship between education and fertility. These are important findings because much of Europe is in a panic about low fertility. The authors speculate that Norway’s family-friendly policies have contributed to the relatively high birth rate among highly educated people. I’d say that sounds like science, wouldn’t you??

And now I must get back to that pesky dissertation…

Take a look at BetsQ’s reference to the article about education and birth rate in Norway. Where’s the prediction there?

Much as I hate to use House as an example, it’s a very good one. Nobody’s making predictions - they’re guessing. Then revising their hypotheses (or throwing them out) after getting results from the actions they take based on that guess. Predictions have a very limited usefulness in the sciences, and I’d argue they’re most useful only when a theory has been elevated to the status of scientific law. Outside of that predictions are misleading at best.

Not disputing that. My argument is that what AHunter3 presented as a sociological prediction is not sociology, and my question is: “How would someone presenting himself as a sociologist explain the failure of this particular prediction, should it actually fail?”

Uhm, I don’t have access to the article, but from the way he described it, the prediction is that there will be a negative correlation between education level and fertility. In this case, the conclusion should be that the hypothesis is not supported and other things need to be looked at to either explain why this is so (e.g. the theory that increasing education decreases birth rates, or that data presented is an anomaly).

If that is correct, then it is indeed scientific. I am not a scientist, but I have taken a number of physics classes and I have an interest in science in general. Any scientific activity has two parts, data collection and interpretation. In some cases, you are testing a hypothesis (note I am not using the word theory, which means something else in formal science), and in others you are just gathering data to see if patterns emerge. That is true in sociology and in physics. I see the path like this:
[ol]
[li]Data[/li][li]Analysis[/li][li]Conclusion[/li][li]Prediction[/li][li]Data[/li][li]Analysis[/li][li]Conclusion[/li][li]Repeat[/li][/ol]
After enough times you eventually have enough information to develop a theory that can then be used to make wider predictions.

Sociology, psychology, and anthropology all have problems that may make this more difficult than physics or chemistry. First they fields overlap with each other and other fields (such as biology). Second, as said above, the researchers are also the subjects, at least in the sense that they are human and part of human societies. There is a greater awareness of these issues, and if you look at modern textbooks, you will see that there are things being attempted to help compensate.

Jonathan

That number of children is inversely proportional with education levels. The Norwegian study disputes that. Sounds like science to me!

Just off the top of my head, epidemiology seems to be a subfield of both medicine and biology whose whole raison d’etre is the development of models of how diseases spread.

You do know that there’s more to medicine than clinical practice, right? Disease models make predictions regarding how a particular disease will respond to a particular treatment, and other models predict how those treatments will otherwise affect the patient’s health. Still other models predict how sleep and the lack thereof interacts with an individual’s health and cognitive state.

Regarding the silly notion that biology doesn’t make predictions, I suggest you read up on the evidence for common descent, which includes specific predictions made by the theory of evolution.

It is science, but that’s not a prediction! It’s a hypotheses. There’s a big difference. A prediction would be more along the lines of “In a country with the same level of family support policies as Norway, the birth rate will be correspondingly high.” Given that the birth rate of a country depends on many more factors than domestic family policy alone, I would be very surprised if that prediction held true.

Explanation, not prediction. You might be able to make predictions based on those models if you’ve accounted for enough factors, but developing the model itself is not a prediction, nor are predictions inherent in development of models.

Well, that’s not science, of course. The thing is, a lot of people will tell you that homeopathy works or that vaccines cause autism. That doesn’t mean biology isn’t a science, does it?

Indeed, it’s entirely possible that there are more snake oil salesmen selling chiropractic nonsense and homeopathy and ear candling and all that idiocy than there are honest to God medical researchers. Biology would still be a science, would it not?

You can play this game with any science. Idiocy about astronomy is widespread; shit, we still have people on the SDMB, on something of a regular basis, who sheepishly admit that the moon’s phases are caused by the Earth’s shadow, and then you have moon hoaxers and people who think space aliens stuck things in their asses and media reports that Mars will look as big as the Moon next week and all that stuff. Astronomy remains a science.

What people see on “Oprah” and read in the papers is rarely science because that sort of thing just isn’t conducive to science. Science is a long, hard slog. The people who win Nobel Prizes win them after years and years or work and usually years and years after their work is done, when science has had a chance to examine their work. That’s not something you can distill into an eight-minute conversation with Larry King. When someone says economics (to use another common whipping boy) isn’t a science, their grasp of economics is usually limited to an article they read in the paper that was political opinion, or some flappery they see on MSNBC, not science. If you actually read economics research, though, and meet economists and take economics, you quickly realize it’s very much a science. Sociology’s the same thing. For every pop fool you see on a talk show there’s a dozen scientists you’ve never heard of and will never meet, at a university somewhere or in the field, working with data and evidence to try to get sociology moving forward.

Interesting article, thanks! I think in some places the author is playing fast and loose with the term, but in others there are clearly predictions. I retract the statement that science doesn’t make predictions.

Nevertheless I still disagree that an academic discipline must make predictions in order to be considered a science. The goal of science is to explain - clearly without a rational explanation, predictions are impossible; the existence of a rational explanation, however, does not require that predictions be made from it.

I think the problem here is semantics. You say a hypothesis is not a prediction, I say - sure it is! It’s a kind of prediction. It’s not a definite, this-will-happen prediction, it’s not a law of physics, but it’s a tentative prediction. And that kind of prediction - I think this will happen - is necessary to the scientific method. That’s how models get refined - by predicting, testing, seeing how the results differed from expectations, and revising the hypothesis. So in that sense, all science does predict.

Again, it’s not so much that prediction must be a goal (although I beg to differ that there are any sciences where it’s not), but that the process of explanation involves making predictions. Any explanation you can come up without predicting and testing it isn’t scientific. That doesn’t necessarily make it invalid - history may not do a whole lot of experimentation, but it’s hardly useless - but it does make it a non-science.

Yes, sorry, I’m not implying that a field becomes non-scientific if it includes quacks - in fact, I’m not arguing that sociology isn’t a science at all. I was just trying to clarify where the question, ‘Is sociology a science’ comes from. All I’ve seen of sociology is the astrology side, so it seems like a pertinent question. Please, fight my ignorance!

If you’re asking me to teach you sociology, it should be borne in mind that I am a dimwit. I know what science is, but God forbid I should try to explain a particular one.

On the other hand, some disciplines like academic history are plagued by the opposite problem. In history, everyone wants to be radical, unusual, and new. Everyone tears apart everyone else’s work, and the most innovative people are the most praised - very briefly. They quickly get attacked a few years later.

I’ve heard someone say that while it’s a “flock” of sheep and a “herd” of cattle, it should be a “malice” of historians.

Not at all. As you yourself noted, predictions deal with what will happen (or what we think will happen), but not every hypothesis focuses on that. Evolutionary theory looked at the fossil record, asked “How did this happen?” and came up with hypotheses that weren’t predictions. Same with the Norway birth rate/education article mentioned upthread - there was no prediction, merely the question “Is there a negative correlation between these two social factors?” In both cases hypotheses have been tested and found to be true or false but there isn’t a hint of a prediction about them.

It is entirely possible to formulate hypotheses without making any predictions at all, whereas you can’t make a prediction without at least a hypothesis at hand. This relationship - in which prediction requires hypothesis, but not the other way around - clearly indicates that it is possible to have science where the goal is not prediction. OTOH, you can’t have a science where the goal isn’t explanation.

The prediction was what would be found. For example, if I had hypothesis that all SDMB posters with user names containing the letter z were guests, I could make the prediction that I search of the user data base would show no members with a letter z in their names. The fact that it is a prediction about what I will find when I review data about the past does not make it less of a prediction. Sociologist can and do make predictions about the future, they just can’t publish a paper about it yet. Papers are published after results are in. They also cannot control the experiment very well. For example, they could predict that if the current policies of both countries stay the same, Norway will have a higher average birth rate than Denmark. So, what happens when both countries change their policies? Or what if they go to war? Even if their conditions stay stable, it would take decades to acquire sufficient data. Since it needs to be repeatable, you would have to do a second study which would mean each sociologist would be lucky to fully test even one prediction.

Jonathan

Ok, I’ll grant that you can perform an experiment without having a hypothesis first, and come up with your explanation afterwards. In this case, there has been no prediction, because you performed your experiment without trying to guess which way it would go. But nobody stops there! The next step is to take that hypothesis and test it. Otherwise, you’re not doing science, you’re doing natural philosophy.

How exactly is an explanation different from a prediction? You don’t think that anyone expects their explanations to be true in the future? Take a look at medicine - if medical science didn’t make predictions, how would doctors know what to do? How can you prescribe a medicine without some idea of what it will do? Even if you state your conclusion as ‘morphine causes drowsiness’, you’re making an implicit prediction that if I take some morphine, I will get drowsy.

Evolution has made and continues to make many testable predictions, including species extinction and microevolution. A prediction does not have to be a position on a future event, just an out of sample event.

I’ve done some more reading on the subject to try to bolster my position and I’ve found I can’t. I misread the first article I consulted in response to Capt Ridley’s original post upthread regarding what predictions sociology can make about human society. I took an argument that explanatory power should be given more weight than predictive ability in determining whether scientific theories should be accepted or rejected to mean that predictive power was completely irrelevant.

I still think there are better examples of sociology’s predictive power than the near-pure statistical query AHunter posted, but I’d have to get the sociology text I own out of storage before elaborating further. As it is, I haven’t much more to say right now.

Ultimately, questions over the epistemological status of sociology will remain until those sociologists who are following the scientific method make a concerted effort to break completely away from those who are not. IMO, that means running those who don’t follow the scientific method out of university departments, lobbying for funding cuts, changing the name of the discipline etc. (This may be happening, I don’t know.)

It’s possible to play games of tu quoque with physics, and other objective sciences, but the fact remains, quacks in physics are marginalized and not considered physicists by other working physicists, they receive no funding, they do not get positions in physics departments, etc.

Are those researchers, who, as AHunter describes, hold certain “truths” as axiomatic, considered sociologists by other mainstream sociologists? Do they consider themselves sociologists? (I have no idea.) If so, isn’t it a little disingenuous to compare the situation in physics with that of sociology?

I’m pretty sure there’s room for both in the field. It makes for interesting conversation during coffee hour, if nothing else. Most people are pretty up front about their epistemological positions. I just don’t see the variety of epistemologies as a problem. Hell, my own work is mostly pretty positivist, but I’ve done some pomo work as well.