I have been poking around with some data for a little bit now and I think it might be time for some of it to be discussed here…

One of the things that has interested me recently in the vast multitude of data that that the Presbyterian Church (U.S.A.) office of Research Services puts out is the size distributions of congregations.  Undoubtedly, one of the reasons that I have been looking at it is because in my professional work with earthquakes there is a very well-established and useful relationship for magnitude distribution of quakes.  I have found that a similar, though proportionately different, relationship seems to hold for congregation size distribution in churches.  More on the details of that another time, but here I want to take a first look at congregation size in the PC(USA) over time.

As I said, Research Services puts out a whole bunch of information about the church every year and for this post I want to focus on two particular data sets — Table 2: Distribution of PC(USA) Congregations by Membership Size and Synod (2009 data ) and Table 6: Fifteen Largest PC(USA) Congregations Based on Membership Size (2009 data ).

Unfortunately, with the revamped PC(USA) web site it appears that a lot of this older data is no longer on-line from the church, but thanks to the Internet Archive and the Way Back Machine we can get data back to 1994. The equivalent of Table 6 is there, while Table 2 goes back to 1995.

First, as usual, a couple of technical notes about the data.  Concerning Table 2:  1) For my own purposes it would be helpful to have the distributions reported in equal size ranges (e.g. 0-50, 51-100,… 301-350, etc.) rather than the uneven ones in the table (0-50, 51-100, … 301-500…). (They are pretty evenly spaced on a logrithmic scale which does work well for some of my calculations.)  2) It is interesting that the largest range used is >1600 when going to >2000 would correspond to the usual definition of a megachurch.  But it is also important to note that the PC(USA) tables are based on membership while the megachurch definition is based on worship attendance.

Now, I have taken the sixteen available annual reports for Table 6 and put together a spreadsheet covering 1994-2009. For those years when a church was not one of the top fifteen there is no entry in the rank column for that year and the membership number is taken from the on-line statistical report for that church.  Since that is only a 10 year report numbers were directly reported for 1999-2009 and 1998 was calculated from the 1999 membership and the gains and losses reported that year.  Those numbers are not included before 1998.  And yes, I did graph up all the data but most of the lines are so tightly clustered that I did not think it added anything to include here.  At a future date I may present it as groups of churches and include graphs of the subsets then.

OK, I think that does it for the obligatory introduction and geeky details — on to the data.

Turning first to the large churches over the course of these 16 annual reports (1994 – 2009) nineteen different churches have appeared on the list of 15 largest.  Of those, nine have been on the list all 16 years, five were on the list at the beginning and have dropped off, one left the list and later returned, and four have been added.  A couple of the congregations have held fairly steady positions on the list, only moving up or down four places or less in 16 years, and Peachtree (Atlanta) has held the number one spot for all 16 reports.  All 19 churches were part of the PC(USA) in 2009.

Looking at membership numbers, five of the original 15 are larger than they were in 1996, a couple of them significantly larger.  For the other ten, one has a minimal (<5%) decrease, but looking at the churches whole histories over this time most show fluctuations and it is not unusual to see periods of several years with very stable membership numbers.  Over the whole time range one church, Fourth (Chicago) showed no declining years and two churches that joined the list had no declines in the time they were on the list – Second (Indianapolis, 99-09) and Christ (Edina, MN, 02-09).  While some of the churches showed significant declines from 1994 to 2009, no church had declines all 15 intervals.

One of the most interesting properties of this list over time is that for the 15 largest churches there is a fairly constant total membership.  The combined membership begins at 73,689, increases to a peak of 75,872 in 1998, generally decreases to 71,368 in 2004, and then moves up and down again until finishing at 71,722.  This represents a 2.7% decline over 15 years and a 5.9% total variation.

But when looking at this pattern it is clear that the variation is less a function of the general decline of the PC(USA) and more a reflection of significant membership changes in individual congregations occasioned by some event or transition in the congregation, usually a change in the senior pastor.  For instance, the 1998 peak marks the year just before changes at both Peachtree and Menlo Park.

In fact, maybe the thing I find most interesting in this analysis is the response of membership at these churches to changes, particularly in pastoral leadership.  While I won’t explore this in depth now, here is the graph of church membership for three churches, Peachtree, First (Orlando), and First (Nashville).  I have normalized the membership numbers to the peak just before the transition and placed that peak at Year 5 on the graph.  The similarity of the growth-transition/drop-growth pattern is strikingly similar and I’ll be looking at it in more detail in the future.

But for our purposes today, what this analysis does show is that for the largest churches in the denomination over the last 15 years the decline as a group is nowhere near what it was for the denomination as a whole and factors that are usually cited for decline in the denomination are subordinate to local influences when it comes to changes in the size of the membership at these churches.  It is not so much that the 15 largest churches at any given time are necessarily declining, but that there is rotation in the members of the list and the total size of the churches on the list remain relatively constant, or at least fluctuate within a certain narrow (+/- 3%) range.

Turning now to the other data set, it is important to note that for the largest churches in the denomination there has been a decline when viewed as the number of churches with membership >1600.  I have compiled the data for the whole PC(USA) from Table 2 from 1995 to 2009 into another spreadsheet and looking at the top category we can see that the number of large churches held fairly steady from 1995 to 2003 (in the range of 113 to 124 churches) and then from 2003 has steadily declined to 91 churches in 2009.

Looking at all the data ranges we see that only the lowest two ranges, churches with memberships of 0-50 and 51-100 have increased over the range of the chart.

Because of the large scale differences between the lower ranges and the largest ranges I also plotted the distributions normalized to their size in 1995.

Now it is easier to see that the number of churches with memberships of 50 or less increased almost 30% and the next higher range (51-100 members) increased slightly (5%).  All other ranges showed a decrease of between 13% and 36%.  And while the largest churches showed significant decrease in numbers, the greatest percentage decrease was with the slightly smaller churches in the range of 301-1200 members.

Clearly what is happening is that as members have left the PC(USA) individual congregations have remained active and the churches have been slipping from the larger size ranges to the smaller ranges.  This is not an unexpected conclusion since membership has declined 22.1% in this time period, but the number of churches has only declined 6.2%.  With this change the mean size of congregations has dropped from 235 to 195 and the median size has gone from 128 to 97.  This increasing concentration in the lower size ranges is a reflection of the Presbyterian tendency to let a congregation continue until the membership drops to a point where the members themselves realize that the church can no longer sustain itself.

That is probably enough data for today and by now you have probably come up with some of your own applications from this exercise.  Let me mention two of concern to me: 

1) Are we training our seminary students for this world of lots and lots of very small churches?  If more than one-quarter are 50 members or less and half have less than 100 members what should seminary students know about the world they will be stepping into.  Taking this a step further, what are the best models for a pastor in small churches?  Yoked ministry?  Tent-making? Commissioned Lay Pastor? House churches? Something else?  If the future is full of lots of very-small churches what should pastoral leadership look like?

2) Should the future be full of lots of very small churches?  Should presbyteries be considering what is the best model for congregations?  Should the number of congregations decrease in proportion to the decrease in total membership?  Should governing bodies at all levels be more aggressive about counseling and shepherding congregations into a new reality?  I don’t know, but these are questions Presbyterians around the world are asking.

So two ends of the size spectra – two differing behaviors in membership variation.  But what does each suggest to us about ministry at that end of the distribution?