March 6, 2013 www.catholicleague.org
Bill Donohue comments on four recent surveys of Catholics:
The larger the sample size of any population, the more costly the survey, but the more accurate the findings. Catholics make up anywhere between 70 and 78 million Americans, but even a sample of 1,500 can yield relatively accurate results (the margin of error in such a survey would generally be 3 percentage points). Two surveys of Catholics published today did not come close to this baseline sample.
The New York Times poll sampled 580 Catholics; it had a 4 percentage point margin of error. The Pew Research Center sampled 184 Catholics, allowing a margin of error of 8.2 percentage points.
Such samples are inadequate. By contrast, a Rasmussen poll of Catholics published last month had a sample size of 1,000; the margin of error was 3 percentage points. Ten days ago, Gallup did a survey of just one subset of Catholics, Hispanics, and had a sample size of 28,607; its margin of error was only 1 percentage point.
There are lots of problems with even the best surveys, but when the sample size is ridiculously small, the findings cannot be taken seriously.
The larger the sample size of any population, the more costly the survey, but the more accurate the findings. Catholics make up anywhere between 70 and 78 million Americans, but even a sample of 1,500 can yield relatively accurate results (the margin of error in such a survey would generally be 3 percentage points). Two surveys of Catholics published today did not come close to this baseline sample.
The New York Times poll sampled 580 Catholics; it had a 4 percentage point margin of error. The Pew Research Center sampled 184 Catholics, allowing a margin of error of 8.2 percentage points.
Such samples are inadequate. By contrast, a Rasmussen poll of Catholics published last month had a sample size of 1,000; the margin of error was 3 percentage points. Ten days ago, Gallup did a survey of just one subset of Catholics, Hispanics, and had a sample size of 28,607; its margin of error was only 1 percentage point.
There are lots of problems with even the best surveys, but when the sample size is ridiculously small, the findings cannot be taken seriously.
No comments:
Post a Comment