Polls Are Always Wrong

campaign_2016-2So, remember all the way back to Monday, when Quinnipiac had Trump up by 7 over Cruz in Iowa? So how on earth did Cruz pull out a 4 point win? The specifics probably only matter to polling wonks and campaign strategists. What does matter to all voters across the country are the polling trends leading up to nearly all elections since 2010.

Back then, the nation just voted in a midterm election that really wasn’t supposed to mean anything based upon the polls, but it did. It shifted things sharply conservative and few pollsters saw it coming. The drumbeat continued in 2012 with gains in the House, Senate, state houses and governors’ mansions. This left MSNBC’s Chuck Todd wearing his “this is my shocked face” in spite of the electorate sweeping Obama back into the white house. Again in 2014, the GOP added to their majority in the house and purloined the senate from Harry Reid’s boney fingers. Polls back then foretold of GOP gains but the talking heads on both sides were shocked, SHOCKED at the volume of conservative wins.

Then last fall Matt Bevin was elected governor of Kentucky by double digits. Good news for those who love liberty and small government except that it wasn’t supposed to happen. The polls had Bevin down somewhere between 3 to 5 points just prior to the voting and he won by 9!

Wait, what? How did Matt win? Why were the polls so wrong?

There are a lot of bits and pieces that contribute to the numbers that polls produce and the details of how that happens is best left to those who know the mechanics of polling analysis. But what is clear is that polls underreport the conservative voters and have for years. It may sound odd but the “why” of this is actually much less important than the ongoing trend of underreporting.

Let’s posit two theories of why this happens and project what the results might be. In this video, Bill Whittle contends that the statist progressively biased mainstream media gives conservatives a 15 to 20 point “headwind” in elections. One might argue the numbers might be higher or lower but only a statist progressive would argue that bias doesn’t exist, but they wouldn’t back up the argument with any verifiable facts.

Second, polls have been undercounting conservative voters for years as evidenced by elections in 2010, 2012, 2014 and last November. One could make a strong argument that this is caused in part by Whittle’s contention or that traditional polling metrics are obsolete in this millennium. Regardless, the trend exists now and it doesn’t appear that things will change any time soon.

The Bevin win means that the polls got it wrong by about 14 points. This is, more or less, in line with polls in the past half dozen years. Where it gets interesting is in pairing that trend with the November 4, 2015 Quinnipiac head to head polls for president which has Clinton and Sanders (who will never, ever, ever win the nomination) only winning over Trump. Carson, Rubio, Cruz and Christie all top the Democrat hopefuls, Carson by double digits. Here’s the breakdown:

  • Clinton v Christie: Christie +5
  • Clinton v Carson: Carson +10
  • Clinton v Cruz” Cruz +3
  • Clinton v Rubio: Rubio +5
  • Clinton v Trump: Clinton +3

If you need to know, Trump also loses to Sanders by 2.

One day later, Elon University published similar results, and with Trump as the sole GOP loser and Carson and Rubio ending up on top in single digits. With the election so far away, does this even matter now. Yes, but not in the obvious way.

So let’s look at these numbers within the context of polling projections from previous elections and the final electoral outcome.

Most polls exaggerated non-existent Democrat strength. The amount can be argued but it’s not unreasonable to peg is at about 10 points, well short of Whittle’s contention of 15 to 20 points. Yes, the two items are unrelated but there is a relationship between MSM bias and left leaning polls from places like Quinnipiac, and the effect is cumulative. So, adding those ten points to the November Quinnipiac data and everyone would have beaten Clinton. The margins look like:

  • Carson: +20
  • Christie: +15
  • Cruz: +11
  • Rubio: +15
  • Trump: +7

They also all beat Sanders, as much as it makes no nevermind.

Now, there’s a lot of wood to chop between now and the 2016 election. We’ll hear a lot more from the candidates and the field will thin. Paul*, Bush, Carly, Kasich, Jindal, Huckabee*, Graham, and Santorum* are all statistically dead but have yet to exit the race. Expect Christie leave for New Jersey and eat cannolis. We can only hope that someone will soon steal Trump’s ball cap and he’ll go back to playing eminent-domain roulette. Yes, it could still happen. After New Hampshire next week, Trump, Cruz and Rubio might be the only ones left.

On the Democrat side, Sanders will soon go home and yell, “Get off my lawn” at neighborhood kids. There is probably no way the DNC would let an admitted socialist sit atop their ticket in spite of the fact they’re fine with Clinton who is also a socialist, just not yet out of the closet. Clinton will be the Democrat candidate barring unforeseen circumstance but there are more and more of those new unforeseen circumstances arriving in the news every day. While her numbers are slipping within her primary against Sanders, those numbers, and more importantly the declining trend against potential GOP candidates likely has her staff somewhere between scared and apoplectic.

Every head-to-head poll from November 2015 had her losing support with Carson, Christie, Cruz, Rubio and Trump all gaining. The RCP rolling average poll (combining polls from CNN/ORC, Fox, NBC/WSJ, PPP and Quinnipiac) shows similar trends for September through November in all races as this one pitting Clinton and Carson. Carson’s gains have reversed but Clinton’s numbers seem to be heading back to earth.

Combine this trend with the polls 10 point lean to the left and it bodes well for Republicans this election season. But before getting too self assured GOP, it’s abundantly clear that this election is unlike any that came before still in the memories of those alive to have witnessed them. Now, there are new rules of engagement and the spoils will go to those who most effectively exploit them.

There are a lot of primaries between here and the parties’ conventions and with the bombastic Trump saying anything and everything to shake things up it’s too early to place any bets. But it will be fun to watch.

*Have exited the race since the initial publication.

Leave a Reply

Your email address will not be published.