A national poll released Tuesday by CNN is getting a lot of media attention simply because it shows Donald Trump leading Hillary Clinton by two points – a statistical dead heat.
The survey has some liberals panicking, many Trump supporters celebrating and plenty of pundits claiming this race has shifted toward the spray-tanned GOP nominee.
All of these reactions are premature and fail to take into account some important factors.
First and foremost, the CNN survey included a +4 Republican sample, when the electorate in the most recent presidential election was +7 Democratic. According to Huffington Post’s Pollster, the Democrats continue to have an edge of seven points when it comes to party identification.
Add the fact that no election since 1980 has had an electorate with more Republicans than Democrats, and the CNN poll quickly seems like an outlier.
NBC News noticed these flaws and adjusted the CNN survey using 2012’s electorate – which is likely to resemble 2016 – and it gave Clinton a four-point edge over Trump, which is much more in line with other polling.
As Chuck Todd pointed out, the CNN poll “assumes an electorate we’ve never seen before.”
Todd noted that CNN assumes half the electorate will be whites without a college degree – a group Trump does quite well with – when this group only made up roughly a third of voters in 2012.
Predicting the outcome of presidential elections isn’t just about one poll, though.
An NBC News/Survey Monkey poll released on Tuesday with a sample size of over 32,000 registered voters – compare that to CNN’s sample of about 800 likely voters – showed Clinton with a solid six-point lead over Trump.
With all polls taken into account, including the flawed CNN survey, Clinton still leads Trump by an average of about six points.
As Nate Silver tweeted today, it’s never smart to make assumptions about an election based on a single poll:
Most bad polling analysis stems from people’s desire to take the most recent/most hyped poll as the gospel truth, instead of averaging data.
— Nate Silver (@NateSilver538) September 6, 2016
Not to mention the fact that all of this talk of national polling is essentially meaningless in the Electoral College system.
As Jason Easley wrote earlier today, presidential elections are decided on a state-by-state basis. On that score, Clinton is in pretty solid shape as she doesn’t just lead Trump in most battleground states, but she also has many more paths to victory than the Republican nominee.
The new 50-state poll released by the Washington Post/Survey Money on Tuesday underscores this fact:
— McKay Coppins (@mckaycoppins) September 6, 2016
According to the Post, Trump would need to run the table when it comes to the swing states if he wants to be president – a highly unlikely scenario.
Today’s lesson isn’t to ignore polling you don’t like. It’s to take into account the polling aggregate – not just a single survey.
One single poll, as CNN showed, may have flaws that present an inaccurate snapshot of the race. Considering a wide range of polling will always give a more accurate picture of where the election stands and weed out any outliers that pop up.
If we do that, Clinton is still the favorite as we head into the homestretch.