The figures in your report can do counterintuitive things when you add attendance as a criteria, and select different attendance ranges. There are two main reasons why.
Adding Attendance Decreases Sample Size
Adding attendance to the criteria for your report makes the sample size 2-4 times smaller on average. In some cases, adding attendance can make the number of “similar employees serving at similar churches” as much as 20 times smaller.
Example: ChurchSalary’s database currently contains:
Position 7,237 Senior Pastors
+ Status 6,977 Full-Time Senior Pastors (-4%)
+ Budget 991 Full-Time Senior Pastors at $1-2M churches (-76%)
+ Attendance 54 Full-Time Senior Pastors at $1-2M / 200-300 person churches (-78%)
In the example above, adding attendance made the sample size 18 times smaller, from 991 to only 54 senior pastors!
This is why ChurchSalary recommends that churches start by creating "budget-only" reports, which do not include attendance. A larger set of employees will almost always give you more accurate data.
You Are Filtering for More Affluent Congregations
In our research, adding attendance does not increase the accuracy of reports because it is already practically baked into the church’s total budget as a variable.
Note: Church income is primarily a result of giving x people. And most church budgets are primarily driven by congregational giving. Thus, if a church has more people, their income/budget will almost always be larger; fewer people and it will be smaller.
For example:
200 people x $2,000 per person = $400,000
400 people x $2,000 per person = $800,000
Adding attendance merely narrows down your search to congregations with a specific ratio of “money to people.” ChurchSalary calls this ratio "per person giving."
The formula to calculate per person giving is:
Total Budget (or Giving) ÷ Attendance = Per Person Giving
If you decrease the attendance range in your report, without also decreasing the budget range, you are sampling for more affluent congregations (higher per person giving). Conversely, if you increase the attendance range without increasing the budget range, you are sampling for less affluent congregations.
Example: If you generate a report based on a $1-2M budget range and vary the attendance range, you will sample employees serving at churches with different ranges of per person giving.
$1-2M ÷ 200-300 = Average $6,000 PPG (Range = $10,000 to $3,333)
$1-2M ÷ 300-500 = Average $3,750 PPG (Range = $6,666 to $2,000)
$1-2M ÷ 500-750 = Average $2,400 PPG (Range = $4,000 to $1,333)
This is why changing the variables in your report can impact the salary figures in counterintuitive ways.
Increasing the attendance range without changing the budget range will sample less affluent congregations.
Decreasing the attendance range without changing the budget range will sample more affluent congregations.
Both of these decisions to add attendance will decrease the sample size–which increases the likelihood that the distribution of salaries starts to look chunky and weird.