Color Bars

The Common Error in Reporting the “Margin of Error”

Election 2014 With the November elections fast approaching, we are sure to see more news about polls. These horse race stories receive a lot of attention and – depending on the outcome of the poll – can provide a needed boost or unexpected kick in the gut to a political campaign. But too often we see headlines that fail to accurately reflect the results of a poll – specifically when it comes to the margin of error.

Why is that?

Journalists will often say a race is tied or virtually tied if a candidate has a 5 – point lead and the poll has a 5 percent margin of error. Seems to be a reasonably conclusion, right? Actually it’s not, because the margin of error represents the likelihood of the expected result.

To show what I mean, let’s look at a hypothetical example.

Jane Doe is running for City Council. A recent poll shows 50 percent of voters plan to vote for Jane Doe over her opponent Bob Smith.

The same poll, which has a 5 percent margin of error, found Bob Smith has 45 percent support among voters. Does that mean the two are tied because they are within the margin of error?

No, because the margin of error affects both candidates.

With a 5 percent margin of error, Jane Doe might have as much as 55 percent support among voters, or as little as 45 percent.

Bob Smith, meanwhile, could have as much as 50 percent support, or as little as 40 percent.

That means the gap between the two candidates could be as wide as 15 points, or it could be a dead heat. But most likely, Doe and Smith are five points apart.

So, the next time you hear that a race is a virtual dead heat, dig into the numbers and see if the reporter has fallen into this common “margin of error” trap.

« | »