PSA: Beware of Bad Polls This Election Season


The reason why not all polls you see online are created equal in this election cycle!



This tweet was sent out today after Monday night’s presidential debate touting that Donald Trump was in fact the winner, cable news talking heads be damned!



So right now you’re probably a little confused, because most people agree tha Hillary Clinton had easily won their first presidential debate. Let’s explain.


The confusion comes from conflicting polls between Scientific Polls (CNN, YouGov, ect.) and “Internet Polls” (Drudge Report, Washington Times, Time, ect). While Scientific Polls had Hillary Clinton being the clear winner, Internet Polls had Donald Trump being the overall winner in their first presidential debate. So which is correct? Well, let’s not mince words here; Internet Polls are:




The reason they’re sketchy is due to their methodology. Sites like the Drudge Report and Washington Times aren’t trying to recreate a random sample size, which is so important in a legitimate political poll. A good example of this is the difference between the Drudge Report and YouGov polls.


Traditionally Scientific Polls are done with a random sample size of voters – preferably over 1,000 people – from various political beliefs and backgrounds. Essentially the randomized sample is supposed to represent the US electorate. While definitely not perfect, there’s still an honest effort to emulate the American populace in scientific polls.


In contrast, it’s important to realize the Drudge Report poll was more of an online reader survey than anything else, having the site’s readers opt-in to the poll. The Drudge Report is a site that specifically caters to those looking for news that has a conservative bent to it, in turn their readers tend to be more conservative in their political beliefs. So if you were to ask them “who won the presidential debate between Clinton and Trump”, they’re more likely to say Trump won due to their conservative bias. As you may have guessed, the Drudge Report’s poll doesn’t represent the opinions of average American, which makes its poll suspect. But that doesn’t necessarily mean “opt-in polls” can’t be scientific.


The best example of opt-in polls that are considered scientific/credible are the YouGov polls. While the YouGov polls are opt-in polls (ie not randomized surveys), they however also incorporate a pretty strict methodology of weighting their polls to create a representative sample of the US population. As we said before, the idea in Scientific Polls is to create a polling sample that emulates more-or-less the US electorate. This is what separates “good polls” (YouGov, CNN) from bad ones (Drudge Report, Time Magazine).


While news agencies should be more cognizant about the polls they’re reporting on, so far in this election cycle, that hasn’t necessarily been the case.


So when you’re online, just remember, not all polls are created equal.


(Photo Credit: Donald Trump Instagram)


Leave a Comment

Filed under News

Leave a Reply

Your email address will not be published.