RT Article
T1 Assessing the Validity of Data Synthesis Methods to Estimate Religious Populations
JF Journal for the scientific study of religion
VO 57
IS 2
SP 206
OP 220
A1 de Kramer, Raquel Magidin
A1 Parmer, Daniel
A1 Saxe, Leonard 1947-
A1 Tighe, Elizabeth
A2 Parmer, Daniel
A2 Saxe, Leonard 1947-
A2 Tighe, Elizabeth
LA English
YR 2018
UL https://ixtheo.de/Record/1669597911
AB The present study tests the validity of a data synthesis approach to population estimates of religiously defined groups. This is particularly important in places like the United States, where there is no definitive source of official data on its population's religious composition, and researchers must rely on costly, large-scale surveys, or congregational membership studies. Each approach has limitations, especially for estimation of small religious groups and for estimation within small geographic areas. Without official statistics, the degree of bias in estimates is unknown. Data synthesis, specifically Bayesian multilevel estimation with poststratification, offers a useful alternative that maximizes the utility of data across all sources to estimate multiple groups from the same sources of data. This method also facilitates comparison of groups. This study provides evidence of the validity of the approach by synthesizing data from Canada, a country that includes questions about religious identification in its national census.
K1 Bayesian estimation with poststratification
K1 Jewish Population
K1 Measurement
K1 Religion
DO 10.1111/jssr.12513