This paper examines the overreporting of attendance at a large evangelical
church using a poll of church members conducted during the seven days follo
wing a specific Sunday morning worship service. It provides a direct (tempo
ral and individual) test of the connection between self-reported church att
endance and actual, observed attendance. Overreporting was found for worshi
p attendance even when the attendance rate was adjusted by removing non-wor
ship events. More importantly, by comparing poll responses to attendance re
cords kept by the church, it was possible to determine both the rate of Sun
day school attendance overreporting and which adult members misreported the
ir attendance. Most of those who said they attended Sunday school, but who
in fact did not, were active church members who claim to attend church "eve
ry week.".