Not only is it often unclear how accurately self-reported survey responses reflect the respondents actual behaviors, but it is often even more unclear how many individuals claimed as members by the LDS Church are actually found in a given survey population. Because many disengaged and inactive Latter-day Saints do not identify themselves as church members, population-based surveys consistently undercount the number of Latter-day Saints claimed on church membership rolls. In the United States, sociologic research suggests between 57% and 65% of individuals claimed as members by the Church identify themselves as such. National censuses in Latin America have demonstrated a vast gap between the number of members claimed by the LDS Church and the number of individuals who identify the LDS Church as their faith of preference, with correlation ratios of only 20-27%. When the Chilean census (27%) and Mexican census (24%) are corrected for the exclusion of under-15 and under-4 year olds from religious affiliation questions, respectively, the real correlation between church-reported membership and self-identified membership in these nations is just 27%, as I will document later in the analysis of individual census data. For the reader's ease of calculation, I will use 25% as a round estimate in the following example.
The implications of the problem of the "unknown denominator" upon sociologic research are profound, especially in areas where the correlation between official church membership numbers and self-identified religious preference is low. For instance, in a random sample of 10,000 people in Mexico, we would expect from official church membership to find approximately 100 Latter-day Saints, or 1% of the population. Yet based on the 25% correlation between official membership claims and self-identified religious preferences expressed on the national census, researchers would likely only find approximately 25 people identifying themselves as Latter-day Saints in a representative sample of 10,000 Mexicans. 15 of these 25 people (60%) who identify themselves as Latter-day Saints might report attending church weekly. A researcher might therefore extrapolate the 60% self-reported weekly church attendance rate among this sample to the more than 1,000,000 Mexicans claimed as members by the LDS Church, leading him to conclude that some 600,000 Mexicans attend LDS congregations weekly.
Such a conclusion would be fallacious because the study methodology picked up only self-identified Latter-day Saints, who represent only one-quarter of the number of Mexicans claimed as members by the Church. Put another way, in addition to the 25 individuals from the sample who identified themselves as the Latter-day Saints, the sample of 10,000 Mexicans would have contained approximately 75 individuals claimed by the Church on official membership rolls, but who did not identify the LDS Church as their faith of preference, and may not even be aware that they are still on LDS membership rolls. These individuals are included in official church membership figures, but are not included in the denominator of the researcher's sociologic study, as the researchers had no way to identify them. When these missed 75 inactive Latter-day Saints are added back into the denominator, one would arrive at an activity rate of not 60% (15 out of 25), but of 15% (15 out of 100). This is a drastically different result which lies far beyond any survey's calculated margin of error.
The population of individuals who identify themselves as members of the LDS Church is vastly different in belief and behavior from the much larger population of individuals worldwide claimed by the Church as members but who do not identify themselves as such. As a result, such studies uniformly overestimate LDS member belief, activity, and participation by a modest 30 to 40% in the United States up to a factor of three or four times in areas of the developing world where LDS activity rates are very low.
There is no adequate means of calculating in a self-reported random survey how many people were really on church rolls but failed to acknowledge such. Census data may provide some guide to expectations, even if an accurate denominator of Latter-day Saints within a sampled population cannot be established with complete confidence. It is methodologically impossible for independent sociologic population-based research, with the exception of national censuses reaching every household, to accurately count the denominator of Latter-day Saints in a sample.
Because inactive and disaffiliated respondents most frequently do not identify current membership in the LDS church, no study methodology short of a full national census which reaches every household can firmly establish the correct relationship between the numerator of self-identified LDS members and the denominator of official membership statistics. Even large, methodologically robust samples of national populations are confounded by non-random geographic and social distribution of LDS members, significantly increasing the potential for error.
Most major religious surveys demonstrate little awareness of this problem. Although some discrepancies exist between self identified membership and official membership claims for other denominations, the magnitude of this challenge in studying LDS membership appears to be unique. Other missionary-oriented denominations, such as the Jehovah's Witnesses (100%+) and Seventh-day Adventists (75-100%), have demonstrated an excellent correlation between self-reported religious affiliation and denominational membership claims on national censuses around the world. Even lower-demand faiths such as the Catholic Church demonstrate good correspondence, even if nominal self identification does not imply active participation. The outlying low correlation between official membership claims and self-identified religious preference for the LDS Church, especially outside of the US and Canada, creates severe difficulties when religious researchers unfamiliar with these issues assume that the number of respondents who report the LDS Church as their faith of preference is an accurate measure of LDS membership account in the study population. It is not. This problem is further aggravated by the lack of objective information from the Church regarding member activity.
Studies which have made valiant attempts to address this issue by considering past as well as current religious affiliation have still fallen short. For instance, the Pew Foundation Survey inquired about religion and of birth as well as current religious affiliation. Although this did add some additional information, this study methodology still misses most of the "churn" of the LDS faith of individuals not born in the church who attended only a for a brief period around the time of their baptism before returning to old ways. Such individuals, who represent the vast majority of LDS inactives worldwide, are not picked up either by questions of religion of birth or of current religious affiliation. The only possible remedy to this would be to add an additional question to inquire whether the respondent was ever baptized in the LDS Church, but the esoteric nature of this question makes it unlikely that it will ever be included in general religious surveys among populations in which Latter-day Saints are a minority. It has also never been demonstrated that the question would be answered accurately even if asked.
The lack of adequate information regarding the denominator of Latter-day Saints in a sample population is almost certainly the most difficult problem faced by sociologists and religious researchers in the population-based study of the LDS faith, because the problem is both very large and completely invisible from within study methodology. An underestimate of the denominator skews all downstream data. Unlike respondents who can be queried more rigorously to minimize methodological problems, non-self-identifying LDS members by definition typically cannot be identified. Although there is no adequate way to address the matter, conscientious researchers should disclose and discuss the existence of the problem of the unknown "real" LDS denominator in population-based samples in their discussion and analysis.
 Pew American Religious Landscape Survey. religions.pewforum.org.