Biased Polling is Good News for Corey Stewart

As usual, the newspapers say one thing and the polling data says another.  Today, I decided to take a look at the Real Clear Politics Average of Polling and saw that Virginia Democrat Senator Tim Kaine is beating Stewart by about 19 points (avg. 51% to 32%).  At first glance, this looks like a dead race – Kaine will beat his Republican competitor by a landslide.  But, being the nerd I am, I looked into the actual polling, which thanks to my education, I actually know how to read.  I chose to look at the most recent poll, the Roanoke Poll (August 22nd).

When you look at the data, Stewart is really running neck-in-neck with Kaine and may actually be winning.  Even better? Stewart is highly unknown.  This is a VERY winnable race for Stewart and I see many of the same markings I saw for Trump on the National race when I predicted the President would win in 2016 with about 45% of the popular vote in Western Journalism. Let me explain.

Before I begin, let me address the pollsters’ bias.  In Question 14 it asks the poll respondents if Tim Kaine’s policies are “too liberal” for Virginia.  The natural follow-up regarding his competitor in Question 15 (Q15) should have asked if Corey Stewart was “too conservative.”  It did not.  Instead, it asked if Stewart was “too extreme.”  Such language tends to act as a subtle depressing message to factual polling data collection.  Respondents pick up on those clues.  This was about as bad as one could get and the professor leading the polling, Dr. Harry Wilson, should be professionally ashamed of himself.

The poll further dug into issues such as gun control and Donald Trump.  In this Roanoke poll, a majority of Virginians favored gun control (48 to 44%).  This seemed odd since it was inconsistent with previous polling on the subject itself.  Regarding the President, more than half had an unfavorable view.  Whereas that may or may not be possible in Virginia, Trump’s approval rating was at 32% (Q5).  That number is well below the President’s fairly consistent 40 – 45% approval rating in Virginia.  Other bizarre facts in the polling data showed Northam with a (Q10) 17% disapproval rating; this is highly unlikely in a state that is pretty polarized along political lines.  Thus, it was obvious that the pollsters were oversampling Democrat voters and that is clearer upon greater scrutiny.  Of course, ask about a Republican candidate’s “extremism” and you are likely to get dishonest answers on other hot button topics, such as gun control and President Trump – thus, no surprise.  But let’s dig into this poll which should really give Stewart’s camp some hope:

  • Q8. The economy is the number one issue at 25% of respondents.  Given the strength of Trump’s economy, this should have factored more favorably in the polls and it shows a clear path to a political strategy for Stewart.
  • Q20. Northern Virginia and Richmond were represented in this poll by a combined 52% (31% NoVA; 21% Richmond). These two locations dwarfed the largest concentration of registered Virginia voters, Tidewater – to which the pollsters sampled only 24% of this typically conservative voting region. The remainder of the state received a representation of 24%.  Consequently, the geographic breakdowns for Virginia’s VOTING population in this poll defied reality.  The numbers are closer to 25% NoVA, 25% Tidewater, 15% Richmond, and 35% the rest of the state. Again, voting populations are different than simple numbers of people.
  • Q22. The education breakdown also seemed to favor “non-Trump” voters. Only 15% of the respondents had either less than High School or High School educations.  An astonishing 85% of the poll had anywhere from some college to an advanced degree.  While Virginia is a highly educated state, it is not THAT educated and college graduates tend to vote as their professors – with Leftist biases.
  • Q27, Q28, and Q28a. These three questions tried to gauge liberal, moderate, and conservative biases as well as Republican and Democrat biases.  28a claims to combine 27 and 28 to determine a mix and the pollsters determine that the poll is 34% Democrat and 33% Republican.  But the logic falls flat because Q27 shows a “moderate” bias of  32% and yet Q28 indicates a Democrat v Republican bias of 33% to 25%.  Oddly, Q27 showed a 37% conservative plurality bias – higher than the results for Corey Stewart.  Naturally, one would assume that Corey Stewart would at least capture that voter constituency, with some Libertarian defectors replaced by some Independents.  That did not occur, which makes this entire study highly suspect.  More likely, the pollsters defined moderate to mean what they wanted it to mean and adjusted the R/D quotient to fit a narrative of fairness.  Given the leading Q15 on “extremism,” it is safe to say the pollsters are fairly incapable of an unbiased position.
  • Q24 and Q25. The racial demographics were off as well. Blacks were at about where they should be as a percentage of the Virginia population (19%).  But the number of Hispanics was under-represented (5% vs an actual number closer to 13%).  “Other/multi-racial” demographics were slightly high, but fairly close to accurate (poll 10%; state 8%).  Non-Hispanic Whites were consequently over represented at 65% in this poll (if we subtract the Hispanic numbers from the data as posited in this poll).  That is nearly ten percentage points higher than their statewide total of the electorate.
  • Q26. Nearly one-third of the respondents had a family income over $100,000. An additional 15% had a family income between $75,000 and $100,000.  That means 45% of the polling respondents had an income higher than the Virginia median family income of approximately $70,000 annually.

So, what does this all mean?  This poll heavily favored educated, wealthy, White Liberal voters from Northern Virginia … NOT Stewart’s base.  In fact, not any Republican’s base in Virginia. Even with that significant polling bias, Kaine could not secure a majority of respondents.  Worse for Kaine? In this poll 43% either did not know or refused to answer about their feelings toward Stewart (Q12) AND 14% of the voters are undecided (Q7) with 79% of those voters unsure how they will vote and the remaining 21% splitting their vote evenly (11% Kaine; 10% Stewart).  Those are astounding numbers coming against an incumbent Senator with all of the advantages of a demographically biased poll!

  • Q7. If the election were held today, Kaine would get 49% of the vote, Stewart 33%, the Libertarian gets 4%, with 14% undecided.

This particular poll claims a 4.3% error rate.  That is farcical.  I believe, like so many other polls in 2016, the biases of the pollsters are skewing their perspective of accuracy, and the error rate is so high, these polls should be considered invalid.  That said, Stewart has a genuine shot at winning based on my interpretation of the data.  He needs to keep his nose down, hammer home the economy and conservative values, drive up that Trump voter base, and explain who Corey Stewart “is.”  Polling like this will help suppress Democrat voter turnout – as it did with Hillary Clinton’s constituency – because after all, “Kaine already won.”  Meanwhile, Republicans and Conservatives need to do more studies like this to better understand a consistent and deceitful pattern of behavior by pollsters and ensure their base shows up to vote on November 6th.

Based on what I just read – if Corey can get 75% of the 2016 Republican turn-out, he will unseat Kaine.