New poll shows Roy Moore tied with Dem in Alabama? Not so fast …
hands manipulating data

New poll shows Roy Moore tied with Dem in Alabama? Not so fast …

Posted October 18, 2017 10:52 AM by Rob Eno hands manipulating data
Inok | Getty Images
    • Font Size
    • A
    • A
    • A


On Tuesday evening, Fox News dropped what many are terming a shocking poll on the special election to fill Jeff Sessions’ former Alabama Senate seat. The poll showed Republican Roy Moore and Democrat Doug Jones tied at 42 percent of the vote each. While the race is tighter than recent Alabama elections, there is one major problem with this poll: A poll of “registered voters” in any election is close to meaningless, and it is doubly so in a special election.

The Alabama Senate special election is less than two months away, on December 12, 2017. Registered voter polls are rarely used that close to an election, even when it’s not a special election. The reason is that not all registered voters actually vote. In the last non-presidential-year general election, 2014, only 36.4 percent of eligible registered voters turned out to the polls. You can start to see why one must take a poll of all eligible voters with a grain of salt.

We’ve already had several 2017 special elections for federal office. In the Georgia 6th District election, 58 percent of voters turned out for the election. In South Carolina’s 5th District, fewer than 90,000 voters voted in the special election, significantly fewer than the 170,000 voters in 2014.

The fact is that only a small percentage of “registered voters” actually cast ballots in a special election.

Using a “registered voter poll” and reporting it as a measure of where the race stands at this point is absurd.

In addition to being the wrong type of poll at this point in the race, in 2014, FiveThirtyEight, the data journalism project of Nate Silver, reported that registered voter polls skew predominately towards the Democrat.

That was a long windup, so here’s the pitch: Polls of so-called likely voters are almost always more favorable to Republicans than those that survey the broader sample of all registered voters or all American adults. Likely voter polls also tend to provide more reliable predictions of election results, especially in midterm years. Whereas polls of all registered voters or all adults usually overstate the performance of Democratic candidates, polls of likely voters have had almost no long-term bias.

The author at FiveThirtyEight went on to outline that the organization almost always “corrects” registered voter polls to account for their Democratic lean.

In other words? Long-term data analysis shows that registered voter polls should not be trusted.

Is this particular race close? Yes, other likely voter polling shows it much closer than recent GOP statewide races. Is it tied? I’ll wait until another likely voter poll comes out showing that before I’d wager that it is.

Remember, not all polls are as they seem. Especially when used by a media organization to push a narrative that one of the candidates is “too extreme.” That seems to be what Fox News is doing here.


Let’s FIGHT BACK together …

… against the mainstream media's biased reporting, selective facts, and outright propaganda. Sign up now for the daily dose of sunlight you need to disinfect the media's lies. It's free!

* indicates required

Robert Eno is the director of research for Conservative Review. He is a conservative from deep blue Massachusetts but now lives in Greenville, SC. If you see something you’d like him to cover, tweet him @robeno.