Saturday, October 02, 2010

Yes and no

Here's a hed that tries to do a little too much and ends up as a pretty thorough failure. Let's look.
First, try to be accurate. According to the lede (and the longer version at the McClatchy site), this is three separate polls -- not "poll." They aren't the same thing. There isn't a substitute for being right. There's being right -- accurately reporting what went on -- and there's everything else. No exceptions.

The bigger deal is that the hed doesn't reflect what the polls do, which is kind of unfortunate, because it's interesting. As you might have gathered, this isn't a single survey of three states to see where "likely voters" are "calling shots." It's three surveys in three states that -- I kind of like this part -- survey registered voters and then break "likely voters" out of that pack. This requires a little wing-walking (and McClatchy ought to be more exact in reporting its confidence levels), but it's cool. But that isn't the point of the hed.

Are likely voters "calling shots" in these races? Yeah, to the extent they are in every race that's surveyed around this great land of ours. That's the problem. Nothing in the hed explains why anything in the story is worth reading. The story is about how likelies might be influential -- not whether they are, which is a given.

It's unusual for the McClatchy Washburo to write a poll story competently. Too bad McClatchy papers can't keep up with that occasional bit of good fortune.

Labels: ,


Post a Comment

Links to this post:

Create a Link

<< Home