The Epistemology of Method and Theory
Remember, we as religious scholars seek to verify our findings with observable data using specific methods. When the data are not directly observable, as in stories of metaphysical encounters, the reports of the experiences become the data. These data are then analyzed for their social meaning. But much more often, other sociologists largely avoid this conundrum by emphasizing phenomena that are more easily observable to begin with. We are generally unable to observe someone’s near-death experience or ghostly encounter—the report of the experience is the closest thing to it—but we can certainly observe dynamics such as religious socialization in process. With carefully trained eyes and observation, we can also see phenomena such as power dynamics, socialization, gender roles, ethnic differences, institutional behavior, historical precedents, and religious conflict at work.
The particular methods social scientists use help them answer important questions. So these methods are used to help reinforce the researchers’ social-scientific epistemological concerns. Epistemologically speaking, we are arguing that we know our results are accurate because we are basing our observations on data and adhering to established scientific methods. Remember as well, we are overlooking for now the possibility of an abductive practice—that is, a conclusion based on an incomplete view or assessment of the facts—and assuming our approach is fully inductive. That’s a can of worms we’ll save for just a bit, though if you have jumped a little ahead and mentally tied that can of worms to Saxe’s elephant, you are getting the right idea. 😊
A number of different research methods are used in the sociology of religion today, though a few are especially common. We’ll consider three here. For instance, those of you who have carefully read the American Grace excerpt have noted that the study resulted from amassing data from a series of surveys. Questionnaires were sent to many respondents, who answered the questions in the form of numerical scales. With the responses converted to numbers on a progressive scale, the results can be statistically analyzed. The same is also true of other quantitative (numerical-based) methods that convert responses to numbers. Internal statistical measures are also conducted to estimate many factors that interest social scientists. These include the probability of the reliability of measurement, also known as inter-respondent agreement. In other words, how often do the respondents agree on similar questions of interest to the researchers? Researchers also assess the correlation of particular measurements as well as the probability that the correlated relationships came about by chance. In terms of epistemology, these statistical measures form a logical basis for how we as researchers “know” that our measurements and the relationships we postulate are correct. We didn’t “prove” it—but we did gather evidence relevant to the possibility that it’s true.
Along the same lines, do the statistics generated in the methods of social science really guarantee that the researchers “know” that their results are true—or, as some argue, True? This is a very interesting question. Statistics do follow a general train of logic, so that the numerical results do have a particular meaning. The resulting methods are no doubt at all infinitely better than out-of-thin-air guesswork. Yet, as those with a sharp critical eye have no doubt already noticed, there is a process of interpretation that enters into statistical analysis. The numbers generated by quantitative methods definitely have meaning. But the researchers themselves infer that meaning—it is NOT simply a given result of the numbers. The numbers are real, but the researchers have interpreted the numbers. How do the researchers know their interpretation is correct? They have prior experience, they weigh their interpretation against comparable studies, and they read their results in light of what is already known on the subject. This results in a degree of knowledge and truth—though still, whether it is Absolute Knowledge and Truth are still debatable, even when statistics have been used. There’s that darn elephant again!
Other researchers use more qualitative methods such as personal observation. There, a researcher goes out and personally observes a given situation or social context, makes careful notes at opportune times, and examines the data that results. A close cousin, historical analysis, takes a similar approach, though the direct observation is of other historical sources, not people in face-to-face social settings. Immersion in part of an experience can be a powerful way of gaining insight into that experience. Whether it’s anthropologists who live with tribes in undeveloped nations or investigative journalists who spend weeks and even months talking with sources, checking facts, and carefully researching information to give us an in-depth analysis of a particular news story, the power of personal observation has a long tradition in social research (and related fields). Historical analysis—often done in quiet library collections or archives—likewise has been with us for decades, if not in some form for centuries.
Of course, these aren’t the only possible methods. Researchers in the sociology of religion may also use others such as content analysis or its close cousin, textual analysis, which involve analyzing existing texts for recurring words or themes. Content analysis tends to count recurrences (quantitative) while textual analysis examines their nature (qualitative). Other forms of secondary data analysis are possible as well, as are metastudies—essentially compiling information from a wide variety of studies. Those of you who will consult the socio-historical dynamics subtopic (AKA “Good, Bad, and Ugly”) in Module 6 will note the extensive Koenig 2012 metastudy analyzing religion and health outcomes, as one example. Targeted interviews and focus groups are also occasional possibilities. Please note, though, that as in the Assignment 2 instructions, actual experiments are extremely rare in the sociology of religion due to ethical concerns. So just a hint on Assignment 2: Please don’t discuss experiments as a common method in the sociology of religion, since they are NOT.
In any case, here again, epistemological questions arise. No doubt observation and analysis alike yield a degree of expertise, but questions related to the possibility of an abductive rather than inductive approach—notably the risk of incomplete or incorrect information—loom large. For instance, do we know for certain that a particular observer knows everything there is to know, even after ten, twenty, or even fifty years of observation or analytical comparison? Would factors such as the observer’s friends, interests, personality, biases, and other subjective matters influence the resulting observations and reports? No matter how well-trained or experienced the observer is, that observer is still human.
Even more broadly, how much did the observer or analyst fail to observe? Would another observer come in and immediately replicate the original observer’s observations? How does the observer know that his or her observations are accurate, and do the observations establish truth—much less Truth? Here again, careful and well-conducted observation (as with analysis) has its utility. It is indeed much better than none at all. But there are epistemological limitations. Careful, in-depth observation is good, but it doesn’t necessarily settle all questions, and observers and analysts don’t know everything. (Including about elephants. Too much?) 😊 We try to find truth by asking specific questions and using specific methods to answer them. This doesn’t “prove”—but it does examine and test.
OK, that’s good for methods for a bit; we’ll revisit that topic in a few pages . What about social theories? No, theories don’t “prove” either—but they do help explain. One caution before we begin to discuss theory: When you’re writing Assignment 2, please don’t confuse the general, broad intellectual traditions discussed in the Christiano and particularly the Johnstone readings—especially the anthropological and psychological—with the more specific sociological theories that I will inevitably accept as the best answers to the prompt on Assignment 2. In other words, within those broad intellectual traditions, I am looking for a discussion of the sociological theories as the best answer to that question.
