Table of Contents | Cover Page  |  Editors  |  Contributors  |  Introduction  |  Web Version

The principal basis for advancing knowledge in the social scientific study of religion is observation of the empirical world. How to conduct these observations in ways that will give good information is the task of the methodologist. The research methodologist is concerned not so much with whether to look at the causes or the consequences of religious actions, characteristics, and orientations but with how to look at them. Within the social sciences, various techniques have been developed to study virtually any aspect of empirical reality—in and out of the domain of religion. These techniques and how to employ them are the heart of research methods.

The social scientific study of religion began in France during a time when the principles of natural science were being applied wholesale to the study of society. The guiding philosophy of this period was positivism , a system of thought founded by Auguste Comte (1798-1857), which held that societies demonstrate regularities just the way that organisms do and that these regularities can be understood through empirical study. A number of assumptions operated behind the scenes in the creation of positivistic science. Among these are three important ones: a belief in the fundamental reality of being, the possibility of knowing the empirical world at all, and the separateness of the knower from that which is to be known. Each of these assumptions has been assailed by philosophers offering views of ontology, epistemology, and objectivity that differ from the views of positivists such as Comte.

Empirical Inquiry in the Social Scientific Study of Religion

The application of scientific principles derived from these assumptions was assumed to be the best way to understand reality. In fact, this model became so well entrenched in social science that it was considered by many the only way to do science. Thus the method of positivism has been construed as the scientific method; the assumptions behind the method, the guiding principles for the conduct of empirical research.

This scientific method includes a series of steps that will be familiar to any student of physical or social science: theoretical understanding of the object or process under study, development of hypotheses that guide one's study, direct observation of relevant aspects of the world, and derivation of empirical generalizations that lead to the refinement of theory. Science operates through the logical processes of deduction and induction to generate our understanding of the world and how it works (Wallace 1971). It is clear to sociologists today, however, in ways that were not clear to Auguste Comte, that there is no single or unitary scientific method. The Nobel prize winning physicist P. W. Bridgman denied the very existence of a special scientific method of inquiry. Philosopher Paul Feyerabend (1993:14) put it this way: "The only principle that does not inhibit progress is: anything goes ."

Positivistic social science, which once dominated the conduct of inquiry in the sociology of religion, now faces a number of competing paradigms, based upon competing assumptions. Among the critics of positivism are a number of alternative schools of thought. These include postpositivism, critical theory , and constructivism (Guba 1990). These alternative paradigms make different assumptions about the unity of science, about the universality of human experience, about the nature of knowledge, about the possibility and even the desirability of so-called objectivity (see Weber 1949). Consequently, they dictate alternative principles for the conduct of inquiry.

The thing that unites all these approaches, however—including traditional positivism—is a common desire to understand the social world. Understanding was and remains the major reason for doing research in the first place; but irrespective of how one proceeds in the conduct of inquiry, there are certain dilemmas that must be faced, certain questions that must be answered. It is the business of the research methodologist to help answer these questions.

The methodologist Joel Smith (1991) helps to cut through the competing assumptions offered by these and other schools of thought. He observes that regardless of which philosophical assumptions one adopts, the researcher's experience of the empirical world will always be partial, always be influenced to a great extent by whatever conceptions the researcher adopts. This fact suggests a series of dilemmas that must be resolved for research to go forward. Smith expresses his insight as a series of questions. Building upon his approach, I suggest addressing the following questions as the main tasks of methodology:

1. What do you want to know?

2. Why do you want to know it?

3. What will you observe so as to know it?

4. Which specific objects will you observe?

5. How many objects will you observe?

6. Within what time frame will you observe them?

7. How will the objects be observed?

8. How will the answers be decided?

9. What should you do?

10. To whom/what are you accountable?

These questions provide a useful framework for understanding the challenge before empirical researchers in all fields—including the social scientific study of religion.

What Do You Want to Know? Conceptualization

The first challenge for any researcher is to be clear about the focus of the study: to know what the research is to investigate. In the absence of an articulate research question, it is virtually impossible to conduct useful empirical inquiry. It is also important to realize that the ideas in the mind of the researcher are essentially conceptual. It is in the researcher's mind, first of all, that a coherent picture of the world develops. This means that, in a very real sense, what one wants to study does not exist apart from the mind of the researcher. It is in the researcher's mind that order emerges and systematic empirical inquiry becomes possible. Jacob Bronowski, the noted mathematician and philosopher of science, made the following observation about order in the empirical world:

Order does not display itself of itself; if it can be said to be there at all, it is not there for the mere looking. There is no way of pointing a finger or a camera at it; order must be discovered and, in a deep sense, it must be created. What we see, as we see it, is mere disorder. (1956:13 f)

The order that the researcher sees in the world is a product of an ordering of reality that begins with the development of concepts. Concepts are mental devices for ordering the world. Let us take religiosity as an example.

The question of why it is that some people are more religious than others has fascinated sociologists since the time of Émile Durkheim (1858-1917). It has been recognized for some time that there are different dimensions of religiosity. Charles Glock and Rodney Stark (1965) identified five dimensions: the experiential, ideological, ritualistic, intellectual, and consequential. Are there really five dimensions? Could religiosity not be seen to have more or fewer dimensions? Durkheim thought there were two. Such questions are essentially conceptual. The development of concepts and the identification of the dimensions that constitute them is a fundamental activity of the research process.

Why Do You Want to Know It? The Goals of Research

A second question is closely related to this first one. The researcher must be clear about why she or he is interested in the research question. This is another way of asking what the goals of the research are. Generally speaking, one conducts empirical research for one or more of three reasons: to explain a possibly causal relationship, to describe something that is of interest, or to explore possible relationships among variables. Choice about one's reasons for conducting research will facilitate responses to other questions one must ask as well, such as how to observe and how to analyze data.

In their study of racial differences in religiosity, Jacobson et al. (1990) sought to explain differences in the kinds of piety that characterize black and white Americans. Data for their study were derived from questionnaires administered to several hundred respondents sampled at two different points in time from the population of a well-known midwestern city. Their conclusion that the differences observed "may stem from basic differences in the functions that religion serves in the two communities" was based on the application of powerful statistical techniques (factor analysis and linear structural relationships analysis) to the quantitative data they examined. Their form of data collection, research design, sample, and analytical techniques all were determined in large measure by the researchers' intent to explain what they were examining. Had their goal been different, their approach also could have been different.

Consider, for example, the conduct of ethnographic research. One of the main purposes of ethnography often is to describe the social world. This sort of research operates in very different ways from the quantitative work exemplified by Jacobson and his colleagues. This is so in large part because the goals of qualitative research are frequently different from the goals of quantitative research. David Preston's study (1988) of two Southern California Zen communities may be taken as an example. The first few chapters of this ethnographic work constitute thick descriptions of various aspects of Zen practice: the social characteristics of members, the role of the Zen teacher and his interactions with his students, how and what one learns as a new practitioner, and so on. Preston's goal in these chapters was to achieve what Max Weber (1864-1920) called Verstehen , or understanding derived from the point of view of the other. Good description is essential to developing this sort of first-person understanding.

At the same time, while understanding is the goal of all research, in some cases it is not possible to know precisely how best to acquire understanding. Nor is it always possible at the outset of a study to know exactly what one should be looking for. Thus some research legitimately has exploration as a goal. My own research on Northern Ireland may be used as an example. I went to Antrim and Londonderry Counties as a field researcher with the intention of studying sermons in different Protestant and Catholic denominations, reasoning that if the conflict in Northern Ireland were about religion to any significant degree, I would hear about politics from the pulpits. I quickly discovered that politics is rarely mentioned in explicit terms in the churches.

I attended a number of church services for several weeks before becoming aware that the sermons should not be the main point of interest. Prayers, hymns, and announcements, along with theology, ecclesiology, and church history, were much more important indicators of the politico-religious intersection. This field research showed the importance of being open to the empirical world, the significance of exploration. One sometimes comes away from the field understanding that the less one knows about a subject, the more appropriate exploration is as a research goal.

What Will You Observe to Know It? Measurement

Once it is clear what th focus of the research is to be, the researcher must decide on what actions, characteristics, beliefs, and/or attitudes he or she will examine to locate indicators of the concepts that are of primary interest. As noted above, concepts exist in the mind of the researcher but the empirical world provides innumerable ways of operationalizing these concepts so that the empirical relationships anticipated from theory may be assessed for accuracy. Researchers must examine the world to test and develop their ideas.

It has been argued that anything that exists can be measured. At the same time, not every measurement of a concept is equally valid. Sociologists use the term validity to describe a concept that actually is measured by the researcher's attempt to do so. Let us consider the case of religiosity again. It may be argued that regular attendance at religious services is a valid measure of religiosity. Surely, most people would agree that it is a more valid measure than whether one was baptized at some point in one's life. At the same time, ritual attendance, while it may be one dimension of religiosity, would not be so valid a measure of religiosity as would a composite measure that incorporated several of the previously mentioned dimensions of religiosity.

What the researcher wants in the measurement of any concept of interest is as complete a sampling as possible of all dimensions that make up the concept. Adequate sampling of the content of the concept is a first step in the creation of valid measurement tools. These measurements eventually will involve such things as questions on a survey instrument, observations in a field setting, public records in an archive, or other empirical items that make visible the concepts in which the researcher is actually interested. Measurements are always approximations of the concepts. They can be good (valid) approximations that are manifested in a consistent manner through repeated measurement—that is, they are reliable —or they can be poor approximations. Once it has been determined what parts of the empirical world one will examine, and what parts of it one will ignore, then one must decide how many of the objects in the world one will examine. This is the next question to be decided.

Which Specific Objects Will You Observe? Units of Analysis

Most sociological studies focus on individuals as the source of their information. Through one or more of the data collection strategies used in the social sciences, information can be gathered about people (e.g., individual members of a congregation), social groups (e.g., Presbyterian and Episcopal clergy), organizations (e.g., Catholic parishes), and artifacts (e.g., hymnals). The phrase unit of analysis is widely used to identify the individuals, groups, organizations, and artifacts that supply the data for the study. The questions here are from whom or from what is the information to be collected? And about whom or what will the research be able to say anything?

One could, for example, research the same question using different units of analysis. Suppose, for example, that one is interested in the question of whether Jews or Catholics are more likely to marry someone outside their own religions. One could select a population of individuals from representative congregations to investigate the question. Or one could study marriage records (artifacts) in various communities to answer the question. Alternatively, if one were interested in divorce and remarriage among clergy, one could study individual rabbis, ministers, and priests, or one could look at the ways in which different religious organizations handle the question as a matter of policy. In any case, once it has been determined what will be observed, the next problem will be to decide how many observations will be sufficient. Will a study of one parish teach me about all parishes? Am I interested in all parishes? This suggests the problem of sampling.

How Many Cases Will You Examine? Sampling

In a totally homogeneous population, a sample as small as one case will yield an accurate representation of the population. But human populations are never totally homogeneous, and in religious matters the diversity of human expression is vast. This does not mean that a study of a single case cannot prove useful. This depends upon the purposes for which one is undertaking the research in the first place.

The key issue here is how many cases there should be and how these cases should be selected. The size of a sample should be dictated not by the size of the population in which one is interested but in the diversity of the elements making up the population. Two broad sets of techniques are available for selecting the sample. These are nonprobability and probability designs. In a probability sample (e.g., a simple random sample ; see Shepherd and Shepherd 1986), one knows with certainty how likely it is that any element in the population will be selected for inclusion, while in a nonprobability sample (e.g., a snowball sample; see Kaufman 1991) this knowledge is unavailable, sometimes because the actual size of the population is unknown or even unknowable.

Thus one might readily conduct a probability sample of the members of a particular parish because official membership lists are generally available for such populations. On the other hand, one would be hard pressed to select a probability sample from the population of clergy who have committed crimes. One could study clergy accused or convicted of such actions, but the total size of the population is unknown, thereby defeating an attempt to draw a full probability sample.

Within What Time Frame Will You Observe? Research Design

Some research questions—especially questions that implicitly or explicitly concern social change—call for a longitudinal dimension. This means that data must be collected minimally at two points in time to answer such questions. If, for example, one wants to investigate changes in religious commitment over time, as Smidt and Penning (1982) did, then data should represent at least two points in time. This sort of design is like a moving picture compared with a still photograph. For some research questions, a still photo is quite adequate. This sort of study, in contrast to a longitudinal study, is known as a cross-sectional study.

Once it has been decided how, how many, and when cases are to be examined, the next problem is to determine how the data for the empirical study actually will be gathered. Answers to this question often rely on a series of standard methods of inquiry, but new methods are constantly being developed. In general, there are four principal techniques in wide use. These are experiments, surveys, participant observation, and the use of existing records. Other techniques such as focus groups, depth interviews, unobtrusive measures, and content analysis are best considered within the context of these major approaches.

How Will You Examine the Cases? Observation

There are many different ways that one could design a study to answer a particular question. This entry focuses on four traditional techniques—some of which result in quantitative (or hard ) data and some of which result in qualitative (or soft ) data.

Experiments : Experimental and quasi-experimental designs are not the most common approaches to the empirical social world, but the logic underlying them informs many of the other approaches one sees more often. The logic of experimental designs is essentially the logic of positivism. It assumes a rational world in which everything that is has a prior cause. It assumes too that research can be designed that will make it possible to determine the contribution that each putative cause makes to the outcome. In such a model, factors either are independent variables that effect an outcome or they are dependent variables that are determined in some sense by these independent variables. The logic behind this model is pervasive in traditional science. It is shown in the following formula:

Y = (f)X

where Y is the dependent variable and X is the independent variable. Given this relationship, it may be possible to manipulate the condition of X to observe the effect that this manipulation has upon the dependent variable Y . Manipulation of the independent variable is at the heart of all experimental designs.

One way to manipulate a variable is through the use of experimental and control groups. Weldon T. Johnson (1971) used this approach to investigate whether exposure to religious revivalism would produce religious change in those attending. He selected two experimental groups and one control group comprising students. The first group was instructed about how to play an active role at a Billy Graham crusade meeting. The second experimental group was instructed to participate passively at the revival. The third group did not attend the revival but was a control or nontreatment group. The "religious commitment" of individuals in all three groups was assessed before the revival meeting and twice after at three-week intervals. The question for Johnson was whether attendance at the revival would affect students' level of religious commitment. Johnson concluded that attendance at the Graham meeting was ineffectual in changing beliefs, behaviors, or self-concepts. This technique of using experimental designs—even field experiments—is rare in the social scientific study of religion compared with the next type.

Survey research : Interviews and surveys are much more common approaches to the social world than are experiments. This technique may be implemented in a number of different ways: telephone interviews, mailed questionnaires, depth interviews, face-to-face interviews, and so on. At the heart of them all is the asking and answering of questions. Debra Kaufman, in her study of women who became Orthodox Jews, conducted long, semistructured interviews that allowed the 150 subjects of her study to "speak in their own voices" (Kaufman 1991). Other studies have relied upon more traditionally structured interviews for gathering the data for their studies. Kosmin and Lachman (1993) used a computer-generated sample of 113,000 people to conduct their National Survey of Religious Identification in 1990. The numbers of respondents they sought clearly militated against any sort of in-depth or unstructured interviews.

In any case, the questions in an interview constitute the operational definitions of the concepts in the study. Questions may be closed-ended or open-ended depending upon the nature of the research under investigation. The former refers to questions that present a fixed set of choices as responses, while the latter refers to questions that permit free responses to the questions asked. The type of questions asked, the way the questions are worded, whether others are present to focus the attention of the respondents, the amount of latitude the respondents have to answer the questions as asked, are up to the survey researcher. It is important to remember, however, that the interview is reflective of a relationship between the interviewer and the interviewees. As such, a sense of reciprocity is in order. Rapport with the respondents will almost always enhance the quality of the data collected. Rapport is always important in social research but nowhere more so than in the qualitative technique discussed next.

Participant observation : There are a number of different names for the next sort of observational approach; field research and qualitative research are two of the more prominent ones, but the most common name for it probably is participant observation (PO for short). As the name suggests, the researcher using this approach is both a participant and an observer. It quite literally may involve a detached form of "observation" of subjects with or without their explicit knowledge that they are being observed. It may involve a pattern of engaged interaction with the persons being observed. Babbie (1995) suggests a continuum of roles for the researcher to play, ranging from the complete participant at one end of the spectrum to the complete observer at the other end. Most of this research involves a mix of participation and observation.

A number of interesting research projects have been undertaken in the sociology of religion using PO techniques: a Billy Graham crusade, charismatic Catholics, the Cursillo movement, Quakers, the Divine Light Mission, Hasidism, meditation, Mennonites, Rajneeshpuram, Scientology, Synanon, the Unification Church, Yoga and Zen, as well as countless other groups, movements, and belief systems. PO has proved to be an indispensable technique for developing Verstehen (understanding).

Existing records : Public documents provide a rich source of empirical observations for the social scientist prepared to use them. They have the distinct advantage of virtually ruling out researcher-subject contamination and are therefore considered "unobtrusive" measures. In fact, such records are only part of the wealth of information that exists for social science use. Tombstones, court records, maps, manuscript census records, and many other sources have provided data for social scientists interested in historical and comparative research. Of course, the use of records and documents has been the métier of historians as long as there has been historical research, but sociologists also have come to appreciate the wealth that exists in archival data.

Sociological use of such data may be considered in two intersecting traditions: historical and comparative research. Numerous examples of each of these exist in the social scientific study of religion. Working in the historical tradition, Rodney Stark and Roger Finke developed a detailed statistical portrait of America for 1776. Their work depended in part upon an atlas of colonial maps resulting from "an effort to identify and locate every church congregation in America for the period January 1, 1775 to July 4, 1776" (Stark and Finke 1988:40).

Rosabeth Kanter used a comparative and historical approach in her 1972 book Commitment and Community . This work relies on archival data to study the characteristics of nineteenth-century communes and utopian communities. Her purpose was to see how those movements that endured differed from those that failed. Her research produced a theoretical framework for understanding the mechanisms of attachment and detachment that separate successful and unsuccessful communitarian movements.

Another way in which existing data have come to inform sociological research is through the technique known as content analysis , which may be applied to virtually any form of communication: written or oral, on film, tape, or video. Magazine articles, newspaper columns, popular songs, photographs, published speeches, every form of communication lends itself to content analysis. Shepherd and Shepherd (1986) used this technique to study modes of leadership within the Mormon community. They examined a 600-speech sample of between 9,000 and 12,000 addresses made to General Conferences during a 150-year span of Mormon history (1830-1979). They then classified the speeches according to their principal themes and analyzed their data to understand better the evolution of official rhetoric within Mormonism.

In an increasing number of cases, researchers are using multiple methods to develop a broader understanding of the phenomena they are studying. Michael Ducey used multiple approaches in his study of religious ritual in four churches in a Chicago neighborhood. In one parish where he conducted classic field research, he was known as "an inquisitive participant doing research on religious ritual." But he also used "interviews conducted with the pastors of all the churches and with many members of the congregations." In addition, he used documents "to construct a history of each church and of the community and to verify some of the information obtained in interviews" (1977:5). This multiplicity of methods for purposes of verification is known as triangulation .

How Will You Know? Analysis

Once the data are collected, one must decide what to make of them. What do the data tell the researcher? Contrary to the hackneyed expression, data never "speak for themselves." A crucial job of the researcher is to interpret what he or she has collected and analyzed. On the quantitative side, this may involve rudimentary comparison of numerical summaries and elementary descriptive statistics (as in Kosmin and Lachman 1993), or it may involve the use of powerful statistical techniques (as in Jacobson et al. 1990). Quantitative data most often are processed using one or more statistical and computational packages such as SPSS, MicroCase, or SAS. These tools of analysis allow the researcher to create graphs, construct tables, and conduct statistical analyses of quantitative data.

In terms of qualitative data, interpretation may not involve numerical analysis at all (Preston 1988, Kaufman 1991). Nonetheless, even in these cases, the use of computers for recording and analyzing data is becoming more common. Software has moved from standard word processors to text retrievers, text base managers, and even code-based theory builders such as ATLAS/ti, Hyper-Research, and NUDIST. The availability of code-and-retrieve capabilities promises to change the face of qualitative analysis.

It is at the point of analysis, of course, that the final product of inquiry—understanding—can emerge. In every case, quantitative as well as qualitative, there will be a necessary tentativeness in what we can say. Knowledge of the world is always tentative knowledge because our understanding of the world is, as Smith (1991) points out, always partial and mediated.

What Should You Do? Research Ethics

All the above considerations are essentially technical matters addressing questions of how to proceed. But there is another set of questions that also needs to be asked; this is the domain of right behavior—ethics. These issues address the more thorny questions of what the researcher should and should not do in the conduct of inquiry. There is hardly a more fundamental question for social science than that of right behavior.

Professional associations (e.g., the American Sociological Association) have developed comprehensive codes of conduct for their members. In the social scientific study of religion, as in other disciplines, these would normally revolve around concerns for the protection of human subjects. Concerns about human subjects generally are manifested in four domains: (1) harm, (2) consent, (3) privacy, and (4) deception. It is the first and foremost moral duty of any researcher to ensure that subjects will not be physically, psychologically, or emotionally endangered as a result of their participation in a study. This is best done by assuring that their informed consent to participate in the study at all is freely given. Insofar as possible, subjects must know what they are getting into when they agree to be studied.

Researchers also must be sure that their research is not overly intrusive, or at least that it will not reveal anything about the subjects that might cause them harm. The right to privacy of the subjects must be protected at all times. This protection of privacy is sometimes manifested as a confidentiality screen. Just as journalists are sworn to protect confidential sources, so too researchers must protect their subjects from unwanted exposure. In some cases, it is possible to collect the data without the researcher even knowing the identity of the respondents, so that the protection of respondents' confidentiality can be guaranteed—although this has the drawback of making follow-up studies impossible. Even when this is not feasible, there are ways to assure that subjects will remain anonymous to all but one or two key researchers working on a project.

Finally, subjects in research studies have the right not to be deceived. This last point is somewhat controversial because there may be good reasons that a researcher would not reveal details about the reasons for doing the research. Further, in some social psychological research, a degree of deception may be essential. In these cases, researchers must be careful to debrief subjects and be sure that no harm has been done to them.

To Whom/What Are You Accountable? Uses of Research

These issues do not fully exhaust the ethical concerns. Ethicists are interested in the question, "What should I do?" but they are also interested in the question of accountability. Feminist researchers have made a strong argument for the accountability of the researcher back to the respondents and the communities from which they come (Reinharz 1992). Many people have argued that one should feed back the findings of the study to the respondents not only as a matter of validity checking but also as a matter of reciprocity and social exchange.

The question of what happens with the findings of a study must be considered in the context of the researcher's values. At one point, people advocated "value-freeness" for social science, but that position is no longer tenable. Social scientists cannot simply hand over the results of their research and presume that policymakers will "do the right thing." The consequences of doing this in the physical sciences are plain. Alvin Gouldner (1963) reminds us that before Hiroshima, physicists used to talk about value-free science but that many of them are no longer so sure. The question of ethical neutrality , raised by Max Weber as long ago as 1917, remains a controversial issue today (see Weber 1949). How involved and how detached should the social scientist be?

Conclusion

In the social scientific study of religion, the role of social scientists as advocates for religious points of view has been one area of controversy; however, as Bryan Wilson has pointed out, to understand a religious group may require empathy, but "empathy need not lead to advocacy" (1983:184). The danger that awaits the empirical researcher—perhaps especially the participant observer—is that of going native . The researcher who goes native in essence becomes a member of the group he or she is studying and thereby loses whatever objectivity he or she might have had—assuming, as noted above, that objectivity is a genuine possibility.

Not everyone would see this as a bad thing, just as not everyone accepts objectivity as either real or desirable. Some researchers argue that advocacy is a very appropriate role for sociologists, noting that we are expressing our values whenever we decide what to study and what not to study. Our decisions on what is more and less important are expressions of our personal values—points about which we should be clear and explicit. Whether this should go so far as advocacy of a particular religious viewpoint is a question that has divided the discipline (see Horowitz 1983, Wilson 1983). How should the sociologist act in conducting empirical research?

The theologian Dietrich Bonhoeffer argued from his Nazi prison cell that it was the obligation of the modern believer to act etsi deus non daretur (as if there were no God). Such "atheism" is just the sort of position that sociologists of religion need to adopt. Peter Berger used the term methodological atheism to signify the sociological understanding of religion, an understanding that religion should be perceived "as a human projection, grounded in specific infrastructures of human history" (1967:180). The implication of this is that sociologists should treat religious ideas as they would treat any other ideas; that the subject concerns religion or God is irrelevant to the social and methodological issues involved. Others have used the expression methodological agnosticism to suggest that the existence of God should remain an open question to social research—neither true nor false but simply not subject to verification by social science. What, then, is the place of values in the study of religion? Berger thought personal valuations should be "kept strictly apart from the theoretical analysis of religion." Others would disagree, saying that it is important for the methodologist to be clear about her or his own values, to be explicit about them, and to act in accord with them.

This was the position advocated by former president of the Association for the Sociology of Religion, David Moberg. In his 1977 presidential address, Moberg argued that "making explicit the value commitments under which we operate can be very constructive for ourselves and for our discipline" (1978:2). The virtues he recommended at the time were integrity, humility, love, justice, vision, and transcendence. Any methodologist whose work takes cognizance of virtues such as these would likely advance the progress of the sociology of religion in particular and would likely advance empirical social science generally. Such progress is a suitable task for sociology as well as for methodology.

—Ronald J. McAllister

References

E. Babbie, The Practice of Social Research , 7th ed. (Belmont, Calif.: Wadsworth, 1995)

P. L. Berger, The Sacred Canopy (Garden City, N.Y.: Doubleday, 1967)

D. Bonhoeffer, Letters and Papers from Prison (New York: Macmillan, 1967)

J. Bronowski, Science and Human Values (New York: Harper, 1956)

M. H. Ducey, Sunday Morning (New York: Free Press, 1977)

P. Feyerabend, Against Method , 3rd ed. (Verso: London, 1993)

C. Y. Glock and R. Stark, Religion and Society in Tension (Chicago: Rand McNally, 1965)

A. Gouldner, "Anti-Minotaur," in Sociology on Trial , ed. M. Stein and A. Vidich (Englewood Cliffs, N.J.: Prentice Hall, 1963): 35-52

E. C. Guba, "The Alternative Paradigm Dialogue," in The Paradigm Dialogue (Newbury Park, Calif.: Sage, 1990): 17-27

I. L. Horowitz, "Universal Standards, Not Uniform Beliefs," Sociological Analysis 44(1983):179-182

C. K. Jacobson et al., "Black-White Differences in Religiosity," Sociological Analysis 51(1990):257-270

W. T. Johnson, "The Religious Crusade," American Journal of Sociology 76 (1971):873-890

R. M. Kanter, Commitment and Community (Cambridge: Harvard University Press, 1972)

D. Kaufman, Rachel's Daughters (New Brunswick, N.J.: Rutgers University Press, 1991)

B.A. Kosmin and S.P.Lachman, One Nation Under God (New York: Harmony, 1993)

D. Moberg, "Virtues for the Sociology of Religion," Sociological Analysis 39(1978):1-18: D. L. Preston, The Social Organization of Zen Practice (Cambridge: Cambridge University Press, 1988)

S. Reinharz, Feminist Methods in Social Research (New York: Oxford University Press, 1992)

G. Shepherd and G. Shepherd, "Modes of Leader Rhetoric in the Institutional Development of Mormonism," Sociological Analysis 47(1986):125-136

C. Smidt and J. M. Penning, "Religious Commitment, Political Conservatism, and Political Social Tolerance in the United States," Sociological Analysis 43(1982):231-246

J. Smith, "A Methodology for Twenty-First Century Sociology," Social Forces 70(1991):1-17

R. Stark and R. Finke, "American Religion in 1776," Sociological Analysis 49(1988):39-51

W. L. Wallace, The Logic of Science in Sociology (Chicago: Aldine-Atherton, 1971)

M. Weber, The Methodology of the Social Sciences (New York: Free Press, 1949)

B. R. Wilson, "Sympathetic Detachment and Disinterested Involvement," Sociological Analysis 44 (1983):183-188.

Hartford Institute for Religion Research   hirr@hartsem.edu
Hartford Seminary, 77 Sherman Street, Hartford, CT 06105  860-509-9500