I agree with the above, but as the person in our church who returns these stats (for my sins, it's an awful job done properly) I find the advice quite surprising and certainly not what I do, as you would just get random answers. The data is supposed to generate Average or Regular Weekly Attendance and therefore any unusual attendance should be excluded. So 150 folk at a baptism one Sunday who never darken the church door otherwise from one year to the next clearly do not comprise Regular Attendance. It all depends on whether you want to produce statistically meaningful data or not.