Journalism’s “margin of error”

Big news yesterday when the LA Times reported that Bush's approval had gone up 5% in the past month.  Over at, the comprehensive evidence suggests that Bush has gone up over the past month, but a more modest 2-3% as can be seen here, courtesy of mystery pollster:

Of course the people writing the article for the LA Times, know their poll is a bit of an outlier on the high side.  In fact, though highly unlikely, given that these polls have a margin of error of +/-3%, Bush could have been as high as 43% in the earlier poll and at 42% in the latest.  So, his increase is actually within the margin of error.  Reporters at the LA Times (and elsewhere) know these things, but it gets in the way of an exciting headline like: “Bush and GOP Making Gains Among Voters.”  Unfortunately, it seems that reporters at every media organization treat the latest poll that comes from their own media organization as if it exists in a vacuum.  Sure, they want to highlight their own company's polls, but this is fundamentally dishonest journalism. 

%d bloggers like this: