Counting Members of Faith Communities in America

I’ve been working on a review essay for the The Christian Century that looks at recent sociological studies of religion in America. One of the things I realized as I was working on the article is how complex the “membership” question is. It seems simple enough to ask how many Americans are affiliated with what religious groups? Google American religious group membership yields about 166,000 results, almost all of which will be wrong.

How might one answer the “counting members” question?

One of the most often-cited sources of information about organized religion in the U.S. is the Yearbook of American and Canadian Churches, published annually since 1952 by the National Council of Churches. The yearbook relies on voluntary reports by U.S. religious bodies (227 in 2011) and published annually. Its editors acknowledge that its data depends on definitions of membership that vary by religious group and that denominations count “members” differently.

Who, for example, is a Jew or a Muslim? Someone who belongs to an organized synagogue or temple or mosque? Or someone who was born into a Jewish or Muslim family? Or someone who self-identifies as a Jew or a Muslim? Do you count as “members” all who have been baptized or only those who have been confirmed? What about Uncle Harry, who never darkens the doors of his family’s Lutheran church, but will surely be buried from there?

Some religious bodies rely on membership rolls maintained at the congregational level but others use estimates at the regional or diocesan level. Some report all adherents, including children, while others include only adult members. And it is sometimes hard to take the denominational reports seriously; one national church body has reported membership at exactly 5 million members for at least twenty years! The 2011 edition reported 145.8 million members, down about one percent from the previous year, but it is difficult to know just what this means.

A second source of religious membership data is a decennial project conducted by the Association of Statisticians of American Religious Bodies (ASARB) that extends earlier work done by the National Council of Churches and the Glenmary Research center. (I’ve always had an appreciation for the work of ASARB, whose former president loved to refer to it as “the world’s weakest organization.”)

For 2000 the project reported on membership for 149 participating groups by states and counties using fairly rigorous methodologies for standardizing definitions of membership and also estimated membership for groups that do not normally report membership statistics (e.g., independent congregations, historic black churches, Muslims). Publication of Religious Congregations and Membership in the U.S. 2010 is anticipated in 2012. Data from the Religious Congregations project is available free at , which is a valuable source of statistical data on American and world religion.

A third and newer approach does not rely on reports from individual congregations and denominations but on polling data from individual Americans. This works well for larger religious groups who will normally be well-represented in a typical national sample of 1,500 or so people. It works less well for smaller groups. To gather data on smaller groups social scientists often combine several national surveys to explore members of smaller religious groups. This is the approach Wade Clark Roof and I used in our 1987 book American Mainline Religion. We relied on 11 years of sampling by the widely-respected General Social Survey.
In recent years religious affiliation data has become available from several very large survey projects.

One of the most comprehensive recent surveys of American religious participation was reported in 2008 by the Pew Forum on Religion & Public Life ( ). The survey included over 35,000 respondents, making it possible to present detailed demographic and attitudinal data on large and small religious communities. In addition to the report, individuals have access to an extensive website at Surveyors collected very detailed information on sample members’ religious affiliation. Sometimes the results are perplexing. For example: There are huge differences between Church of Christ and United Church of Christ members but surely some members of one group get recorded as members of the other, either because they are confused and give the wring response or a surveyor checks the wrong box. Members of a Unitarian Universalist congregation might give their affiliation as Unitarian Universalist, Unitarian or Universalist. These problems aside, in my judgment this is the best single contemporary resource for profiling large and small religious bodies in the U.S.

A somewhat different approach was developed at City University of New York and now based at Trinity College in Connecticut. Known as the American Religious Identification Survey (ARIS), it was conducted in 1990, 2001 and 2008 (see The study asks individuals the simple question, “What is your religion, if any?” and records the response e.g. “Christian,” “Protestant,” “Baptist,” “National Baptist,” etc. Those responding “Christian” or “Protestant” are asked “What denomination is that?”

Special challenges are introduced when surveyors use open-ended questions to ask individuals to identify their religious preference.

The advantage to the ARIS approach is that religion is reported in the respondents’ own words and not forced into predetermined categories. The disadvantage is that one cannot compare the survey results with reports from other sources that use different classification schemes. Some of the ARIS classifications are difficult to understand: Mainline Protestants include Episcopalians, Methodists and Presbyterians, but also Missouri Synod Lutherans, Moravians and Russian Orthodox.

So how many Americans are members of which religious bodies? Tell me what you mean by “member” and what you mean by “religious body” and I may be able to give you an answer. But it won’t be easy and probably won’t be accurate!