06 Appendix C – Methodologies

Appendix C - Methodologies



The data in this report originated from a series of research studies conducted by Barna Group of Ventura, California.

Protestant senior pastors, whether interviewed by phone or online, were recruited from publicly available church listings covering 90 percent of U.S. churches that have a physical address and a listed phone number or email address. Churches selected for inclusion were called up to five times at different times of the day to increase the probability of successful contact. Data were minimally weighted to match church characteristics from the National Congregation Study (by Association of Statisticians of American Religious Bodies) for denominational affiliation, church size and region.

Interviews with U.S. adults were conducted online and by telephone. All telephone interviews were conducted by Barna Group. All households were selected for inclusion in the sample using a random-digit-dial technique, which allows every telephone household in the nation to have an equal and known probability of selection. Households selected for inclusion in the survey sample received multiple callbacks to increase the probability of obtaining a representative distribution of adults. Between 20 and 40 percent of telephone interviewing was conducted on cell phones.

For pastor and U.S. adult surveys, regional quotas were used to ensure that sufficient population dispersion was achieved. There were also minimum and maximum ranges placed on the distribution of respondents within several demographic variables that were tracked during the field process to ensure that statistical weighting would not be excessive. When a particular attribute reached one of the parameters, the sampling selection process was varied to preclude individuals who did not meet the necessary demographic criterion, with the interviewer seeking a person from the same church or household who fit the desired criterion.

Online interviews were conducted using an online research panel called KnowledgePanel® based on probability sampling that covers both the online and offline populations in the U.S. The panel members are randomly recruited by telephone and by self-administered mail and web surveys. Households are provided with access to the Internet and hardware if needed. Unlike other Internet research that covers only individuals with Internet access who volunteer for research, this process uses a dual sampling frame that includes both listed and unlisted phone numbers, telephone and non-telephone households, and cell-phone-only households. The panel is not limited to current Web users or computer owners. All potential panelists are randomly selected to join the KnowledgePanel; unselected volunteers are not able to join.

Once data was collected, minimal statistical weights were applied to several demographic variables to more closely correspond to known national averages.

When researchers describe the accuracy of survey results, they usually provide the estimated amount of “sampling error.“ This refers to the degree of possible inaccuracy that could be attributed to interviewing a group of people that is not completely representative of the population from which they were drawn. See the table for maximum sampling error. There is a range of other errors that can influence survey results (e.g., biased question wording, question sequencing, inaccurate recording of responses, inaccurate data tabulation, etc.)—errors whose influence on the findings cannot be statistically estimated. Barna makes every effort to overcome these possible errors at every stage of research.

Previous Section

Appendix B - Definitions

Read Section
Next Section

Appendix D - Acknowledgements

Read Section