Saturday, May 06, 2006

Lying with statistics -- again

Haven't we been over this one before? When you report on public opinion surveys, you may only report on what is surveyed -- not what you think the questions should have said or might mean. Making stuff up is not permitted. Copyeds, that means you are allowed -- nay, required -- to challenge stories that violate Rule A up there and to hold them up until they are brought into conformity with reality.

Would this be inspired by one of the Carolinas' most irksome Saturday features? Funny you should ask:

Carolinas among most religious states
A Gallup Poll analysis found that South Carolina and North Carolina are among the eight most religiously active states in the nation.

Analyzing 68,000 interviews conducted during the past two years, Gallup found that 58 percent of those surveyed in South Carolina, Alabama and Louisiana say they attend church or synagogue once a week or nearly every week. That's tops in the nation. North Carolina isn't far behind, with 53 percent saying they worship weekly or almost weekly.

No, no, no, no, no. Neither the lede nor the hed is true. This analysis does not measure how "religious" a state is. It does not measure how "religiously active" a state is. It does not even measure how often people go to services. It measures self-reported attendance at services. Even Gallup, which has its shameless moments, refers to the analysis as measuring "reported attendance."

It really seems that this paper owes its readers some answers. Why do you continue to allow this particular department to fictionalize news? And what other statistics are you lying with?

0 Comments:

Post a Comment

<< Home