Research methods in the study of U.S. Catholics

The 2005 survey of U.S. Catholics reported in the National Catholic Reporter, Sept. 30, is the fourth in a series that began in 1987. The purpose is to measure trends in Catholic behavior and opinion, and thus the survey methods were kept constant throughout. The first survey, in 1987, was commissioned on the occasion of a papal visit to the United States, and the others were done in following six-year intervals. All data were collected by the Gallup Organization, using telephone interviews based on identical random digit dialing sampling methods. The sample size in 1987 was 803, and it included persons 18 or older in noninstitutionalized settings who said she or he was Catholic.

The 1993 sample was 802, sampled in the identical manner. The 1999 sample was slightly larger -- 877 persons, of whom 12 percent were Hispanic. The 2005 sample was 875, of whom 15 percent were Hispanic.

The margin of error within any of the four samples is plus or minus 3 percentage points.

That is, 95 times out of 100 any reported figure is within 3 percentage points of the true figure in the Catholic population. When comparing across samples, the margin of error for the difference is plus or minus 4 percentage points.

--William D’Antonio and Dean Hoge

National Catholic Reporter, October 14, 2005