What Polls Can You Trust?
Mainstream media dismissed online polls of 100's of thousands of people that declared Bernie Sanders the winner of the first Democratic Debate. They instead declared Hillary Clinton the winner based on expert opinions and more traditional, supposedly scientific, polls.
Here are the common arguments against online polling:
…Instant online polls are informal and unscientific. The results rely on a self-selecting group of respondents with no regard to political affiliation, age, country, or even whether the person doing the responding actually watched the debate.
These arguments are legitimate problems with online, bit also with traditional “scientific” polls too.
The argument that the informality of online polls makes them less accurate is bullshit. If anything I’d expect informality to make a poll more accurate. I mean when are people being more honest? In an informal setting (like at home) than in a formal setting (like an interview)? People are generally more real in a casual setting. I think that’s pretty obvious.
On air CNN belittled their own Facebook poll, repeatedly referring to the voters as “kids” even though in reality about 95% of FB users are voting age adults. CNN dismissed the significance of the Facebook poll (that Bernie Sanders won with 80% of the vote) because they claim it over represented younger people. That’s probably true. But exactly how overrepresented the younger demographics were is well…debatable. Any kid today will tell you, Facebook is for old people (The kids use Instagram.) The kids are right, Facebook is getting older. The largest FB demographic is 35–54 yr olds.
While CNN did not write articles about their online Facebook poll or discuss it much on air, they did write about their phone poll that declared Hillary the winner. The problem is that poll was likely at least as skewed in terms of age as the online poll! More than half of those interviewed were over landlines, which of course over represents older age demographics. It is well documented that in recent CNN polls the older than 50 yrs old demographics are significantly overrepresented.
Did they even watch the debate?
You cannot tell if somebody taking an online poll actually watched the debate. This is absolutely a problem with online polls. But you know what? You can’t tell if people you’re interviewing over a phone actually watched the debate either.
Huffington post wrote “A 55 percent majority of registered Democratic voters who watched the debate said Clinton won”. But more than half the people counted in the HP poll did not watch the entire debate! Just “clips or highlights”.
Bernie Sanders won every single focus group (Fusion, CNN, FOX) but many dismiss this too as unscientific. And they’re right. No 20 person focus group can scientifically represent every American. But everyone in those focus groups actually watched the full debate. It seems to me that’s a more credible group to decide who won a debate than people who only watched a few clips.
The Mere-Exposure Effect
They did not ask about their position on any other candidate before moving onto questions about the debate. When you expose someone to a word or image they tend to develop a preference for it, merely because they are familiar with it (the mere-exposure effect).
So when answering the following questions about which candidate performed best, after the poll takers had been primed with a Hillary Clinton question, that could subconsciously make her appear more favorable than she would otherwise. It’s hard to believe we are that malleable. But we are. It’s a well researched phenomenon, and the reason advertisers are so eager to pay millions of dollars in order to flash their brand names and logos in front of our eyes.
(A legit study would have also asked the poll taker’s position on every other candidate and randomized the order these questions were asked.)
In fact, the mere-exposure effect is why there is such an argument online about who won the Democratic Debate, in the first place. Because simply seeing a headline declaring a candidate the winner can make the candidate more appealing. It shifts public opinion.
So what can you trust?
Do not trust political commentators and authors with any authority. Their opinion is no more legit than your own. You can’t trust polls either, at least not without analyzing how responses were gathered. The more you investigate polls, the more you realize how unscientific and misleading they can be.