How is scientific knowledge generated?

You've probably heard scientific knowledge referred to as one 'way of knowing' about the world. There are other ways, of course--our direct contact through the senses, trial and error, religious experience, etc. Science, ideally, produces 'new' or novel information about the world, and does so through a process of research, which involves testing, methodologies, peer-review, based on evidence and logic brought to bear on a subject. Take the example of evolution versus intelligent design. Evolution is based on the concept of differential reproduction--organisms with adaptive traits, traits that allow them to survive in the environment, may live long enough to produce offspring more likely to have those traits. So long as the environment 'selects' for those, they may stand a better chance of survival than those organisms without them. This is the general concept behind 'natural selection'--the environment selects for those individuals with adaptive traits that make them more 'fit,' more capable of surviving in their environment, amid predators, pathogens, and other hazards of life.

What evidence supports evolution? The fossil record. We know through the science of genetics as well just how close certain kinds of species are to others genetically, and infer that somehow, they may have been from the same genetic lines. Take the squirrel that learns to cross the street by going over the power line, versus the squirrel that gets caught in the middle during rush hour, runs one way, then the other, then in circles, jumps up and down, and pretty much flips out in a blaze of metabolic fury. Or is that fur? In any case, it could be that something about the line-crossing squirrel's brain, its structure, neural patterns, etc., is somehow different, and that squirrel may have a greater likelihood of surviving long enough to bear offspring, who are more likely to have that similar capacity hardwired into their tiny little brains.

With respect to intelligent design, it seems reasonable to ask whether the sort of complexity in life one observes could be the result of an evolutionary process starting with some pretty basic 'rules' and simple organisms. Or whether there was a 'ghost' in the machine, or a diety, some 'designer.' The designer of course could handle reconstructing a fossil record to make it appear the earth was 6 billion, rather than 6,000 years old. But if science generates knowledge based on logic and evidence, there may be some logic, but where is the evidence for the intelligent creator?

What is research?

Simply put, we can think of research as the ‘testing’ of ideas generated through intuition (Stock, 1986). We observe the social world, and we ask questions. Why are fertility rates higher in pre-industrial, agrarian societies, for instance? And we can hypothesize about what the answer might be, and ask questions that shed light on the reasons. Theory is not deterministic, though. We have a few laws, like the law of gravity, or thermodynamics, or electromagnetism. These are theories essentially for which scientists have found no evidence disproving them. Humans are pesky critters, and as soon as you think you've figured out some 'law' of social behavior, one of them is going to make a fool out of you.

So research generates scientific knowledge, and theory represents the structure of scientific knowledge. What you have learned in your sociology classes, whether it was presented to you as such or not, has been largely based on others' research and contributions to social theory. Karl Marx's critique of capitalism, Max Weber's concept of rationalization, Durkheim's study of suicide and the anomie induced by industrialization and urbanization. Their work may seem pedestrian over 100 years later, but it's important to remember they were swimming in unchartered and often choppy waters. And they were observing social life during a different time, a time of great change. The sociologists tended to look forward; the anthropologists tended to try to understand what was being lost from non-industrialized cultures.

How does knowledge in science accumulate?

The sanitized version of this is that scientists do research that contributes to theoretical understanding. They take a theory, hypothesize, test it, and based on their findings either add to, modify, or reject parts of the theory. Thus theoretical knowledge advances gradually. Or not. Thomas Kuhn contended that a substantial body of evidence that ran counter to the 'received view' of science could lead to a 'paradigm shift,' or more radically a revolution. We've come a long way from the Newtonian, mechanical view of the universe, for instance, to Einstein's theories of relativity, to current particle physics and string theory. And as Isaac Newton once said, 'if I have seen further than some, it is because I have stood on the shoulders of giants.' Well, if we know more about the social world, about bureaucracies, about capitalism, about suicide, about inequality, poverty, industrialization, mass media, social interaction, large organizations, politics and power, religion, the family, education, health, etc., it is because social scientists don't have to work from scratch every time they go to the office.

What form does scientific knowledge assume?

Science is in the business of generating novel knowledge, building on an accumulated body of work, using agreed upon methods of testing. Now sometimes the knowledge tears down the accumulated body of knowledge, or at least calls it into question, as in the case of Galileo. Interestingly, on March 15, 1990 Cardinal Joseph Ratzinger, who is now Pope Benedict XVI, said in a speech in Parma, Italy : "At the time of Galileo the Church remained much more faithful to reason than Galileo himself. The process against Galileo was reasonable and just." (from Wikipedia). Reminds one of Supreme Court Justice Antonin Scalia's famous quote: "Mere factual innocence is no reason not to carry out a death sentence properly reached." Translated into science, this would read something like "mere evidence is no reason to refute a theory scientists, perhaps including myself, have worked hard to develop."

Wheel of science: Induction and deduction

Some research builds on existing theory, and generates hypotheses designed to test that theory. For instance, if I wanted to test whether there was a difference between the road-crossing and line-crossing squirrels, and I thought that difference expressed itself in the brain, I might want to do some sort of visual examination, using MRIs or some other diagnostic tool, assessing multiple generations. Or I could observe squirrels, maybe mark them, even tag them or use radio collars, and track their movements to see if indeed there were some who crossed on roads, and others on lines, and what their offspring did. This would be a deductive approach if it were based on current theoretical understandings: Working from theory, asking a question, and devising a test, using accepted methods from my discipline, to provide results for that test.

Other theory, for instance grounded theory in the social sciences, begins with observations, and works toward generalizations. For instance, David Wagner's book 'Checkerboard Square,' lived among the homeless in a Northeastern mid-sized city for two years. From that experience he essentially rejected the middle class model that homelessness is somehow a social pathology to be remedied through housing, through 'mainstreaming' of the homeless. Instead what he observed was rejection of middle class values such as the work ethic, family, and the role of the government and the welfare state in providing 'assistance.' He said that many of the homeless he studied preferred independence and life on the streets to societal institutions that they felt had failed them. In other words, it was an inductive study, working from observations toward generalizations, and at some point, impacting the body of theory on homelessness and the homeless. Others could take Wagner's ideas, and test them in a more inductive approach (for instance, through interviews find out to what extent the homeless population reflects Wagner's findings). An inductive approach might attempt to elucidate other 'sub cultures' that might exist in different geographic settings.

Variables--expressing the language of science

Theory generally is written as sets of propositions, statements about some phenomenon. It represents an effort to explain, or if you're really lucky, to predict. For instance:

1. How can we explain the rise of capitalism?
2. Can we predict where we’ll find dictatorships or democracies?
3. Do out-of-wedlock births and lack of marriage cause poverty?
4. Can we predict who will become the next president in the U.S?
5. Where will the U.S. engage militarily?
6. What causes global inequalities?
7. What will happen to EOU in the next 10 years?
8. What are the employment opportunities for sociology majors?
9. Who will win the Super Bowl?
10. What will be the most promising investments in the next 5 years?
11. Will globalization lead to a clash of civilizations (and WWIII)?

Imagine answering any of these questions. The variable you're interested in understanding is the dependent variable (DV) . For instance, government type (#2), or rise of capitalism (#1), or poverty (#3). The independent variable (IV) is the one that you think, usually based on previous studies, research, and theoretical work, has a significant influence on the DV. A good exercise for you might be to go through this list, and see if you can figure out the DVs, and some possible IVs that you might hypothesize would affect the DV. For instance, the Super Bowl--maybe some key IVs include head coaching, defense, quarterback, and turnover ratio:

Okay, it's not really theory, but I've come up with a three-variable model that I could use either to explain who has won super bowls in the past, or who will win them in the future? Not important? Tell that to an oddsmaker in Las Vegas. Let's try something closer to sociology: Can we predict who the next president of the U.S. will be?

Now, this doesn't really tell us who will be the next president, does it? There are obviously other things to think about--potential as a fundraiser, charisma as a speaker, effectiveness of the campaign organization, physical attractiveness (like it or not, in a television age, this is important), etc. But it does tell us something about society, doesn't it, in terms of explaining who has been president? And we could draw on theories of racism, feminism, power, to help explain the trend to elect upper middle-class white males to run the country. Now, is it predetermined that all U.S. presidents will be white, upper middle class males? No. That would be a determinist view, and we can't say that it is predestined, because the next president could prove us wrong. But we can say there will be a high likelihood that the next president will be white upper middle class. Social theory is probabilistic, not deterministic. Theorists never claim something will happen for sure, or has always occurred in a certain way. But they will be able to say that there is a greater likelihood of B causing A (say, gender influencing the president), than D causing A (say, hair and eye color).

So, theory describes the relationships between variables. In some cases, it's ambitious--Marx essentially tried to explain the history of human societies' evolution from tribal to feudal to capitalist to, eventually, communist. Durkheim was trying to show that suicide was not a result of individual, personal mental illness, but was in large part determined by factors external to the individual--religious affiliation, marital status, gender, etc.

Now the most powerful theories are those that are parsimonious, that is, that can explain the most with the least. If you have a 17-variable model attempting to explain poverty, it's going to be tough to quantify the precise effect each of those IVs has on the DV. So theories generally try to explain the most they can, with the least number of variables. This is sometimes referred to as Occam's razor.

So, some general traits:

• Testability/falsifiability—theory needs to be testable. If it isn't, how do we use it to build knowledge systematically? How do we know that intelligent design isn't as valid as the theory of evolution? The most persuasive theories can marshal both evidence and logic to make their cases, and subject themselves to rigorous testing.
• Probabilistic--Some things are more likely than others to happen, some theories more persuasive than others.
• Variables--theory is often expressed as relationships between variables, both independent variables (the ones we think are causing change) and dependent variables (the ones affected by the changes). Keep in mind, there is a difference between correlation and cause and effect--just because two variables appear related, doesn't mean we've figured out how.
• Theory explains, predicts--it asks 'is there an explanation for this phenomenon?'
• Theory can be generated from observation (inductive approaches), or can be tested through more deductive methods (hypothesis testing).