Poll

 

Lawrence Research National Religion and Politics Survey Results | July 2011

     The data behind “Mormons Believe… What?!”


Note: In order to view PDF files, you must have Adobe Acrobat Reader installed on your computer.

 

National Survey Methodology:
Between July 6 and 13, 2011, Lawrence Research of Santa Ana, California, interviewed by telephone 1,000 randomly chosen American adults (18+). Opinionology/SSI conducted the interviewing in evening hours, except for Friday and Saturday evenings, and on Saturday morning to provide the best probabilities of reaching the broadest cross-section of the American public at home.
The sample was drawn proportionate to population density in each of the 50 states, and was stratified and controlled by geographic regions. Random-digit dialing was used so that adults with listed or unlisted telephones would have equal opportunities to be chosen to be interviewed, which is the central criterion of random-probability sampling.
Samples were drawn for both land line prefixes and for cell phone prefixes, with the goal of achieving an 80-20 mix, the current standard used by most pollsters. The sample in point achieved an 83-17 mix, weighted to 81-19.
The raw data were weighted on age and education to align with the parameters established by the 2010 census. For example, 18-24 year olds came in at 9.1 percent and were weighted to 12.8 percent. Seniors required minimal weighting: 16.5 percent raw and 16.8 percent weighted.
Education was also weighted because the sample was better educated than the adult population as a whole. Therefore, those with high school educations or less were weighted up from 23.9 percent to 44.7 percent, while those with graduate degrees were weighted down from 17.1 percent to 9.1 percent to conform to U.S. census figures.
The weighting design did not produce major changes in the data, but only marginal changes in the majority of variables:
  •  For example, belief in God moved from 87.5 percent to 88.3 percent.
  •  Those identifying themselves as very conservative moved from 17.9 percent to 19.1 percent, and the somewhat conservative from 31.0 percent to 32.1 percent.
  •  As for the definition of marriage question, the original split was 59.1 percent for “one man and one woman” versus 36.6 percent for “any two people.” The weighted numbers are 64.0 percent and 32.8 percent respectively, the difference being a matter of degree rather than substance.
Results for the total sample (N=1000) have a margin of error of ±3.1 percentage points for midrange results (between 35 percent and 65 percent) 95 times out of 100. The margin of error for non-midrange results is correspondingly smaller. Results for split-sampled questions (N=500) have a margin of error of ±4.5 percentage points.
Principal Investigator
Gary Lawrence received a B.A. degree in political science from Brigham Young University in 1967 and a Ph.D. from Stanford University in 1972 in a program that focused on the measurement and analysis of attitudes.
Before starting his own firm in 1986, Dr. Lawrence was a vice president of the research company that did Ronald Reagan’s polling work from governor through president. He has conducted over 2,500 research studies in his 40-year career.