National censuses attempt to reach every individual, and thus the sample size offers the greatest confidence. One exception is that religious data from the Canadian census is sampled based on a "long form" sent to one in five households, although the statistical power of a 20% sample is still very robust. With response rates approaching 100% of a national population, national censuses offer the greatest possible methodological rigor with very little error margin.
National censuses and independent sociologic studies provide information that even internal church data cannot. Although church officials have access to greater internal data that is released to the public, even internal church statistics provide them with no way of knowing how many inactive members still consider themselves to be affiliated with the church, especially since so many of these members in the developing world are in the "address unknown file." Before the 2002 Chilean census, LDS Chilean public relations director Rolf Acevedo observed in an article in the LDS Church News that with the census data, "we will also be able to see exactly how many Chileans claim membership in the Church."
National census data suffers from two main disadvantages. First, census data on religion is only available in certain countries where the census includes a question on religious affiliation. Although it is fortunate that such data are available from the top countries for LDS membership in Latin America as well as Canada, Australia, New Zealand, and Austria, there are many other nations with a large LDS population where the census does not include a religious question, such as the United States and the Philippines. Second, when religious affiliation is included on a national census, the only questioned asked is generally that of one's religious preference. No questions are typically asked about participation, limiting the conclusions that can be drawn from such data.
National censuses, when available, therefore provide very precise information on a single religious question: the number of individuals who identify a religious group as their faith of preference. Official membership figures are defined by different denominations in widely different ways. For the LDS Church, official membership statistics have no obligatory correlation to member participation or even self-identified preferences. Self-identified religious affiliation statistics from national censuses represent a more meaningful indicator of actual member belief or activity than any of the Church's publicly released statistics. Although nominally identifying oneself as a Latter-day Saint does necessarily imply church activity, it would be difficult to claim that those who do not identify themselves as Latter-day Saints are active or contributing members.
Because the number of individuals who regularly attend church services is in all cases less than the number of individuals who identify that faith as their denomination of preference, census data provides a ceiling for the activity rate of Latter-day Saints among current membership in a country. For instance, if a census finds that approximately 100,000 individuals in a nation identify themselves as Latter-day Saints in a nation where the Church claims 500,000 members, we can be certain that not more than 100,000 of those members attend LDS services or have any ongoing connection to the Church, although the number actually attending regularly may be substantially less.
Religious questions on national censuses also have inclusion criteria which may differ from church membership criteria. For instance, the Chilean census reported religious affiliation for individuals age 15 and over, whereas LDS membership rolls also include both unbaptized "children of record" from birth to age eight and baptized members between eight and fourteen. Consideration of census inclusion criteria for the religion question is necessary for data to be correctly understood and analyzed.
 LDS Church News, December 1, 2001.