Great movie for pastors: Lars and the Real Girl
Working bibliography of biblical studies books on ecclesiology

Weekly U.S.A. Church Attendance: The Sociologists Weigh In

How many people attend church on a weekly basis?

There are two ways of figuring out the answer to this question.  You could ask a representative sample of adults how often they attend church.  Or you could go find a representative number of congregations and ask them to provide an attendance count.  Both methods are used but neither is perfectly reliable.  The problem with the first method is that people might lie or "misremember": Do they really attend as much as they say they do?  There is evidence that people answer that they attend more than they actually do.  The problem with the second method is that it is very difficult to get a representative number of congregations and to get them to count attendance in the same way.  Need I mention that this is a difficult task?  "There are more than 300,000 congregations in the United States." Mark Chaves, Congregations in America (Cambridge: Harvard University Press, 2004), 3.

As you will see below, most sociologists stress that (a) self-reporting of church attendance has been relatively stable (around 35%?); (b) that the actual attendance is probably lower than what is reported; and (c) that it is difficult to declare based on the evidence how much church attendance has decreased if at all. 

Here are two contemporary analogies which hint at the enormity of the task of detecting weekly church attendance rates:

(1) How do we really know that the U.S. economy is doing poorly?  It is a massive thing and so every day different indicators are trotted out--some positive and some negative.  Similarly, it is very difficult to tell whether the religious attendance is up because there are so many people and institutions involved.    

(2) The second analogy is this: How do we know whether the U.S. population wants to elect a Democrat or Republican for president?  Let's just look at the popular vote:

In 2008, 53% Democrat (Obama) to 46% Republican (McCain).

In 2004, 48% Democrat (Kerry) to 51% Republican (Bush). 

In 2000, 48% Democrat (Gore) to 48% Republican (Bush).

As a country, there has been a small shift in the last eight years from Republican to Democratic presidential candidates.  But we need to have a national election to precisely determine to what degree the country has shifted (or not shifted).  Up until the last days, the result is in doubt. 

Similarly, there are people who are stopping and starting attending church regularly in the United States but it is difficult to precisely determine how many.  Like political beliefs, people's religious beliefs and practices do change but it is difficult  to easily discern how much the entire country is shifting. 

The best we can do is to

(a) look to national polls where there is a margin of error acknowledging their limitations (Gallup, Barna, Wuthnow, Stark); or

(b) check the attendance levels at some churches and try to make reasonable deductions (Olson, Chaves, Ammerman, Thumma, Stetzer).     

Without further ado, here is what the sociologists say. 

Most researchers are familiar with the following article that argues that reported numbers may not reflect the actual number of people who show up. 

C. K. Hadaway,  P. L. Marler, and Mark Chaves, “What the Polls Don't Show: A Closer Look at U.S. Church Attendance,” American Sociological Review 58 (1993): 741-752.  (It is available as a PDF at the link and will only take you 5 minutes to read.  It is easy to read and fascinating).

There is also a 2005 article which only Scott Thumma  and Randy Travis and David Olson seem to have noticed: it strengthens the argument of the 1993 article that church attendance may be below 22% of the population. 

C. Kirk Hadaway, and Penny Long Marler, "How Many Americans Attend Worship Each Week?  An Alternative Approach to Measurement," Journal for the Scientific Study of Religion 44 (2005): 307-322.  

Here is the abstract:

Opinion polls indicate that over 40 percent of Americans attend worship services each week. However, attendance counts in several North American counties and Roman Catholic dioceses suggest that worship attendance may be much lower. In this article a new measurement strategy is used to estimate total weekly worship attendance. First, using a variety of resources we develop an estimate of the total number of religious congregations in the United States by religious family. Contrary to many published sources, the total number of congregations is estimated at just over 330,000. Second, using known population values and sample-based attendance counts we develop estimates of average weekly worship attendance for religious congregations by religious family. The resulting totals suggest that fewer than 22 percent of Americans attend worship services each week. This lower level of attendance provides further evidence that Americans tend to overreport worship participation and are less religiously active than the polls show.

Nancy Ammerman writes,

"somewhere between 25 and 40 percent of the population (depending on which surveys you believe) showed up for services on any given weekend."

Nancy Ammerman, Pillars of Faith: American Congregations and their Partners (Berkeley: University of California Press, 2005), 1. 

Jackson Carroll cites the Gallup information (though he footnotes that it has its detractors),

"In tracking U.S. poll responses over time to questions about the importance of religion, church attendance, and self-reported church membership, we can see considerable stability over time.  Although ups and downs in the trend lines are evident, they are, in general, remarkably stable."

Jackson Carroll, God's Potters: Pastoral Leadership and the Shaping of Congregations (Grand Rapids: Eerdmans, 2006), 38.

Scott Thumma and Dave Travis write,

"In addition, when phone pollsters ask self-identified Christians, 'Have you attended a religious service in the past seven days?' the figure is consistently around 40 percent, with a recent poll showing a quite optimistic figure of 47 percent.  Recently, academic researchers Kirk Hadaway and Penny Marler have made very credible claims that this stated behavior doesn't match the reality found in actual congregations around the country.  We generally agree with these researchers' findings that show actual attendance on any given week to be around 20 percent of the adult population.  Nevertheless, we feel the larger percentage is still a significant measure of those in the population with a strong commitment to Christian religious practice, even if this practice doesn't include worship attendance every week."

Scott Thumma and Dave Travis, Beyond Megachurch Myths (San Francisco: Jossey-Bass, 2007), 4.

Stanley Presser and Mark Chaves write,

"Yet, existing evidence does not definitively establish whether attendance at religious services declined in American society from the 1950s to the present. We examine the trend in religious service attendance between 1990 and 2006. Evidence from several sources converges on the same answer: weekly attendance at religious services has been stable since 1990. However one reads the evidence about trends between World War II and 1990, the recent past has been a time of stability."

Stanley Presser and Mark Chaves, “Is Religious Service Attendance Declining?” Journal for the Scientific Study of Religion 46 (2007): 417.

Robert Wuthnow writes,

"A great debate has emerged around the topic of religious participation in recent years.  Some observers believe it is declining.  Others do not.  Some argue that it is being replaced by more private expressions of faith--belief without belonging, as some writers describe it.  Still others question whether religious participation can be measureed accurately at all.  Perhaps people somehow want to be 'religious' badly enough to lie to pollsters about how often they attend. 

Much of this debate is fairly easily resolved once we move from speculation to actual research.  Overall, polls and surveys suggest that attendance at religious services among the full adult population of the United States has remained fairly constant for at least thirty years.  Constant, but with perhaps some erosion in recent years.  The reason some observers think religious participation may have declined more than polls and surveys suggest is that research based diaries suggests that it has.  Diary studies ask people to record carefully what they do day by day and hour by hour.  Thus, they require more thoughtful answers about what someone did, say, on a recent Sunday morning than is typically the case with a brief question survey. The trouble with these diary studies, though, is that they have not been conducted among very many people, and it is unclear whether the methods used over the past several decades provide comparable evidence.  However, nobody claims that religious participation has been rising in recent decades."

Robert Wuthnow, After the Baby Boomers: How Twenty- and Thirty-Somethings Are Shaping the Future of American Religion (Princeton: Princeton University Press, 2007), 51. 

Wuthnow's endnotes on page 260:

An excellent overview of the recent research is available in Chaves and Stephens, "Church Attendance in the United States," see also Chaves, "Abiding Faith."

Hadaway, Marler, and Chaves, "What the Polls Don't Show" is the source of much of the recent debate about the accuracy of survey data on church attendance; on diary evidence, see especially Presser, "Data Collection Mode and Social Desirability in Self-Reported Religious Attendance"; other useful contributions to the debate include Hadaway, Marler, and Chaves, "Overreporting Church Attendance in America," Hout and Greeley, "The Center Doesn't Hold," Hout and Greeley, "What Church Officials' Reports Don't Show," Smith, "A Review of Church Attendance Measures," and Marcum, "Measuring Church Attendance."

Rodney Stark writes,

In The Churching of America, 1776-1990, the authors [including Stark] reconstructed church membership rates, beginning in 1776 and ending in 1990 . . . The reason the book's title refers to the churching of America is because through the past two centuries Americans have become far more churched--an increasingly higher percentage actually belong to a local congregation.

Church attendance has held rock steady, except for the entirely understandable decline in Catholic attendance [after Vatican II relaxed its rules about attending Mass].

In recent years there have been a number of attempts made to minimize the actual rate of attendance in America, some of them remarkably impassioned . . . No one disputes that a significant number of Americans who tell pollsters that they go to church every Sunday, probably only go most Sundays and therefore that the consistent finding that about 35 percent are in church any given Sunday is a bit high . . . More careful studies by more qualified researchers suggest that the overstatement of church attendance is quite small, being only about 1.1 times the actual rate of attendance, so the average rate probably is around 31 or 32 percent."

Stark's approving reference at the end of the last quote refers to the article:  

Brian M. D'Onofrio, Lindon J.  Eaves, Lenn Murrelle, Hermine H.  Maes,  and Bernard  Spilka, "Understanding biological and social influences on religious affiliation, attitudes, and behaviors: a behavior genetic perspective," Journal of Personality (67) (1999): 953-84.

Rodney Stark, What Americans Really Believe (Waco: Baylor, 2008), 12, 14, 17, 204.

David Olson writes,

"From 2002 to 2005, the yes answer to Gallup polls ranged from 40 to 44 percent.  During that same period, the Barna Research Group reported similar findings: 43 to 47 percent of American adults surveyed said they attended church on the past weekend (Olson 26)."

Olson concludes based on his own research that

"The research of the American Church Research Project shows that 17.5 percent of the population attended an orthodox Christian church on any given weekend in 2005" (Olson 28).  He continues, "By using statistical modelings to calculate the frequency of attenders in the typical American congregation, the results show that 23 percent of Americans are 'regular participants'" (Olson 29).

David T. Olson, The American Church in Crisis (Grand Rapids: Zondervan, 2008), 26-29.

I have difficulty understanding Olson's methodology.  See his research methodology page online and the one paragraph description of his methodology on page 16 of the book.  I will reproduce here that paragraph and the paragraph before it  which dramatizes the statistics so you can judge his methodology.  At the end of each chapter, Olson notes that there is more information online about each chapter (e.g. chapter 1) but I did not find this to be the case; nor the statistics about birth rates (page 230, note 18 )--I got Page Not Found notices.

Here is the description of his methodology on page 16:

"In reality the church in America is not booming.  It is in crisis.  On any given Sunday, the vast majority of Americans are absent from church.  Even more troublesome, as the American population continues to grow, the church falls further and further behind.  If trends continue, by 2050 the percentage of Americans attending church will be half the 1990 figure. 

For the past twenty years, the American Church Research Project (TACRP) has compiled comprehensive data on the state of the church in the United States.  This research provides reliable attendance numbers for each of the 3,141 U.S. counties, for each state, and for the nation as a whole.  The database includes attendance figures for more than 200,000 individual Christian churches, as reported from head counts conducted by each church.  It also includes zip code and founding date of most of these churches.  In addition, for the 60,000 churches that do not report attendance, their membership was multiplied by the denominations attendance-to-membership ratio.  For the remaining 45,000 Christian churches that report neither membership nor attendance, a realistic estimate was developed by using a statistical model that factored in the population, growth rate, and average attendance of the churches' counties.  The research includes all historically orthodox Christian churches but does not include nonorthodox Christian churches or non-Christian religions" (Olson 16).  

What Olson seems to be saying is that: "Attendance is down in the churches for which I have records."  This says something but one needs to be careful about generalizing to the entire American church. 

D. Michael Lindsay, assistant professor of sociology at Rice University, notes in response to Olson's research:

"Counting heads to estimate weekly worship service attendance is far less reliable than estimates based on survey responses . . . For researchers to generalize head counts to the entire adult population, they must be conducted as an exhaustive consensus or a representative sample."

D. Michael Lindsay, "Gallup's Research Remains More Reliable Than Counting Heads," Rev. Magazine (Mar/Apr 2008): 59.

See also:

Mark Chaves and Shawna Anderson, "Continuity and Change in American Religion, 1972-2006," in Social Trends in the United States, 1972-2006: Evidence from the General Social Survey (ed. by Peter V. Marsden; Princeton, Princeton University Press, forthcoming).

E. Brooks Holifield,  “Towards a History of American Congregations,” in American Congregations, Volume 2: New Perspectives in the Study of Congregations (ed. by James P. Wind and James W. Lewis; Chicago: University of Chicago Press, 1994), 23-53.

See my post:

October 16, 2008

On the use of surveys by church leaders

and other posts within the Sociology category