Religious Attendance Measurement

Hadaway, Marler, and Chaves found that actual attendance was roughly half of what polls reported. Where Gallup-style surveys put weekly attendance near 40 percent, head counts implied that about 20 percent of Protestants and 28 percent of Catholics were actually attending in a given week. The gap reflects a large, systematic upward bias in self-reports, which is why the authors argued that attendance levels should be anchored to behavioral benchmarks rather than to the survey question alone.

In a time diary, respondents reconstruct everything they did on a day without being asked specifically about religion, so attending a service must surface on its own among all activities. In a direct attendance question, by contrast, respondents are explicitly cued to report churchgoing, which invites them to answer with their religious identity rather than a precise recollection. Because the diary removes both the cue and the social-desirability pressure, far fewer people record attendance, corroborating that the direct question inflates it.

Not useless, but it cannot be read as a true rate. Self-reports systematically overstate attendance, so headline levels are too high and trends or comparisons drawn from them must be interpreted cautiously. The remedy is to validate and correct self-reports against independent benchmarks - head counts or time diaries - and to report levels anchored to those measures. Self-reports remain informative about relative differences if the bias is roughly constant, but the overreporting must be acknowledged.