In the debate over requiring health insurance coverage of contraceptives, much weight has been placed on a poll showing that most Catholics side with the Obama administration's position. The poll by Public Religion Research Institute did indeed find that 58 percent of Catholics "believe that employers should be required to provide their employees with health care plans that cover contraception."But a second question the pollster asked - whether religiously affiliated colleges and hospitals should be required to provide employees with health plans that cover birth control at no cost - "is really what the debate is about," poll director Daniel Cox told me.On that, the general public supported availability of free contraceptives, 49-46. For Catholics, the figure was 52-45. This, too, has been widely reported as evidence that most Americans, including the majority of Catholics, disagree with the bishops' position on the issue.The problem, though, is that there is a pretty large margin of error for Catholics - 6.5 percentage points in either direction - because the poll surveyed just 219 Catholics. So we see that the poll's director was much more cautious with the way he characterized the result."There we actually find public opinion is very divided," Cox said. "Catholics were also roughly evenly divided." It's textbook journalism to consider the margin of error when reporting on a poll. But that was not often done on this story, especially for the data on Catholics. For example, MediaMatters, always eager to condemn bad journalism, quoted extensively from the poll on its Web site but cut out the information on methodology, including margin of error.Cox said that when the poll first was released on Feb. 7, news media coverage focused on the first question, the one showing 58 percent of Catholics "completely agree" or "mostly agree" that all employers should provide their workers with health plans that cover birth control. Later in the week, he said, some coverage began to cite the second question - which, I would add, may shed some light on what Catholics who "mostly agree" with the mandate had in mind.Forty-one percent of white Catholics surveyed supported requiring religious hospitals and colleges to comply with the mandate and 58 percent were opposed. Cox found that significant because white Catholics are "a politically important group" that Barack Obama had lost to John McCain in 2008 by 5 percentage points. It's an indication that the Obama administration's rule was opposed by a significant number of the white Catholics who supported Obama in '08. (Cox said the poll did not survey enough Latino Catholics to break out their data.)This helps explain why Obama moved quickly to offer a compromise aimed at pleasing at least those Catholics who voted for him in the past.I thought this poll was fair, but there are a lot of bad polls, and there is a lot of bad reporting on polls. With the political campaign season progressing, my suggestion is to disregard polls and news reports that fail to disclose a poll's methodology; to be alert to the way questions are phrased and ordered (the questions should be disclosed on the poll's Web site); and to be wary of those who cherry-pick from a poll's sometimes complex or even contradictory results.