“To represent the religious history of America statistically and geographically is to generalize dangerously to court disaster openly.”
(Source: New Historical Atlas of Religion in America)
Statistics on religion, like most other statistics, must be interpreted with a critical eye. Data on religion may be reported at either the insitutional level or the individual level. That is, there are surveys of religious insitutions themselves and there are surveys of individuals and their self-reported attitudes and affilitiations. Keep in mind that surveys of religious institutions are also self-reporting, so over-reporting of membership is more likely that under-reporting. The diversity of religious organizations and varying definitions of membership can also have an effect on comparing membership numbers among different organizations. Some organizations may not share reports with the public, or even keep numbers. There has been no government-mandated count of religious organizations since the last Religious Bodies census taken in 1936.
Most of the resources that you encounter in this guide will have a disclaimer explaining the collection method and nature of the data. Be sure to read this section!