A Review of Church Attendance Measures

Estimates of church attendance rates in the United States differ by the measure used : headcount approaches generate rates about half those of survey-based approaches. Based on a series of experiments conducted on the 1996 General Social Survey (GSS) and a review of other studies and experiments, these differences in church attendance rates seem to result from several factors. First, because of both cognitive processes and social desirability effects, standard survey items yield overreports. Second, people often define attending religious services more broadly than what is tallied by headcounts. Finally, headcounts and survey-based estimates yield different results because of various other factors, such as the difficulty headcounts have in separating adult attendance from total attendance, and their frequent exclusion of non-weekend services