Attendance and giving for every recorded Sunday since September 2014, in one place. Attendance answers who came; giving answers what happened next — and the two turn out to be far less connected than anyone expects.
Average Sunday attendance by month, the year-over-year overlay, and how each year grew or shrank.
One point per month: the average of that month’s Sundays. The line takes each year’s color as it crosses into it. The dashed line is the least-squares trend across the entire record — it is pulled down by the 2020–2023 valley, so the rolling average is the better read on where things stand now.
Each line is one calendar year, January through December. Click a year to show or hide it. 2026 runs through August.
Percent change in average weekly attendance against the prior year. Toggle to the January–August basis to compare 2026 fairly against the same stretch of every earlier year.
| Year | Avg. Sunday | Growth | Relative size | Sundays |
|---|
One card per year, then every month of every year in a single grid.
One card per year: the average Sunday, how it moved, the shape of the twelve months, and the moments that defined it. All sparklines share one scale, so a flat line low on the card means a flat, low year.
Average Sunday attendance for each month of each year. Darker means fuller. Read across a row for one year’s shape; read down a column to see how a given month has changed.
The shape a normal year takes, and whether the Easter crowd ever stays.
This panel combines all years, so it uses neutral bars rather than year colors. Each month indexed against its own year’s average, so 100 means “a typical Sunday for that year.” Averaged across 2015–2019 and 2022–2026; 2020 and 2021 are left out because the pandemic pattern was not seasonal. The bars are strikingly consistent year to year — a July dip is the calendar, not a decline.
Each year keeps its own color; the shade deepens from the weeks before, to Easter, to the weeks after. Easter is the largest crowd of most years. The question that matters is whether it holds: each year compares the four Sundays before Easter, Easter itself, and the four Sundays after.
| Year | Easter | Attendance | 4 wks before | 4 wks after | Held vs. before | Easter lift |
|---|
Easter 2017 (Apr 16) is missing from the export, so only its before/after weeks are shown. 2020 fell on the second Sunday of the shutdown.
Seven Sundays whose counts were far out of line with the weeks around them have been removed as missed or mis-logged counts: . 2014 holds a single Sunday (Sep 28) and is excluded from growth math. A few other Sundays are simply absent from the export.
Every Sunday’s offering, where it arrived from, and how the four quarters and twelve months compare.
Giving as bars, average attendance as a line. Start with months — a single Sunday swings too wildly to read — then widen to quarters or drop to individual weeks.
Box contributions are cash and checks passed on Sunday. Online is the giving platform. Benevity and Zeffy are employer-match and third-party platforms. Shown as a share of each year’s total.
Left: each month indexed against a normal week of its own year, giving beside attendance. Right: how each year’s giving divided across the four quarters.
Quarters with fewer than eleven recorded Sundays are left out, so 2026 shows only Q1 and Q2.
The single strongest pattern in the giving data, and the holiday effects that dissolve once you account for it.
Sort every Sunday by where it falls in the month and one pattern outweighs everything else in this data. The bars are what a typical Sunday actually brings in; the dots are how many people were there. The money moves; the crowd does not.
| Week of the month | Typical offering | Usual range | Typical attendance | vs. an average Sunday | Sundays |
|---|
Each occasion measured against an ordinary Sunday. Raw compares it to any normal week; adjusted compares it to normal weeks in the same slot of the month, which is the fair comparison. The gap between the two columns is the pay-calendar effect being stripped out.
| Occasion | Type | Giving (adj.) | Giving (raw) | Attendance | Median given | Median there | Years |
|---|
The connection between the two datasets — and the reason it is weaker than it looks.
Each dot is one Sunday: how many were there, and what came in. If attendance drove giving, the cloud would tilt upward.
Total giving divided by total attendance for the year, excluding the two major one-time gifts. This is the number that connects the two datasets — and the one that has moved most.
| Year | Total given | One-time gifts | Regular giving | Median Sunday | Avg. attendance | Per person | Online share |
|---|
All figures above, gathered as sortable tables. Click any column heading to reorder.