Studying Religion, Making It Sociological 25
of congregations and religious movements, while in-depth interviews, speeches, pub-
lished texts, and archival materials provide insights into the nature of religious discourse
(Becker and Eiesland 1997).
Although these various methods all contribute valuable information, it is important
to acknowledge another common misunderstanding: that quantitative data in itself is
somehow inimical to the study of religion (perhaps because it simplifies a necessarily
complex topic). Quite the contrary. Polls and surveys about religion have become so
common in recent years that they are now intrinsic to our understanding of who we
are religiously: Let one poll show a slight upward trend in church attendance and
journalists announce a “religious awakening”; let another poll show a slight decrease,
and religion suffers from a “collapse.” The challenge is for all educated people, whatever
their discipline, to gain at least a rudimentary understanding of surveys, sampling, and
statistical analysis. Surprising at it may seem, especially with the amount of polling that
accompanies national elections, it is still possible to find graduate and undergraduate
students (often in the humanities) who do not understand how generalizations can
be made from a small sample to a large population, when or when not to use the
term “sample,” and how one might possibly “control for” the effects of race, gender,
or education level. Students who have not already done so, should consult one of the
many readable introductions to sociological methods (e.g., Babbie 1997).
Just as quantitative data require skill to collect and interpret, qualitative studies
also depend on specialized training. Here the difficulty arises from scholars not taking
seriously enough the particular training to which sociologists of religion are typically
exposed. Armed with an interesting topic and confidence that one is a good conversa-
tionalist, literary critics, theologians, and historians (perhaps with the encouragement
of a small research grant) set off to do qualitative interviews not realizing that the craft
of framing questions, asking them properly, and including the right follow-ups should
be as foreign to them as that of a sociologist examining rare manuscripts in an archive.
At minimum, scholars interested in utilizing qualitative methods should gain a rudi-
mentary understanding of the skills required (Burawoy 1991; Strauss and Corbin 1998;
Atkinson 1998).
Whether quantitative or qualitative data are used, an additional misunderstanding
is that work is somehow more sociological if it employs explicit hypotheses than if
it does not. Hypothesis formulation is a valuable exercise in sociology, but there is
also a reason why it seems strange to the inquiring student: It frequently takes the
form of pitting one na ̈ıve view of the world against an equally naive view, instead
of recognizing that events typically have multiple causes and multiple explanations.
Where hypotheses are most helpful is determining whether or not one has an argument
at all. Sociological studies of religion, in this respect, are helped by having a clear, strong,
and compelling argument, just as work in other disciplines is.
As the sociology of religion has matured, the single methodological characteristic
that most often sets good work apart from mediocre work continues to be the strategic
use of comparisons. Quantitative research necessarily involves comparisons; qualita-
tive work should, too. Students of religion, too, often neglect this basic insight, either
because they want to examine one case intensively or because they refuse to consider
what an appropriate comparison might be. The intellectual challenge is to recognize
the rich possibilities that are always present for comparisons, including temporal and
spatial comparisons, as well as ones based on gender, religion, or ethnicity.