18/08/2024
Trolling Polling
Is the quiet Tory lurking?
In 2016 I was asked by the U.S. State Department to spend a week in the Nordic countries: Sweden, Finland and Iceland, with Larry Larocco, a former Democratic Congressman from Boise and close friend, to explain the dynamics in the upcoming Trump-Clinton election.
Our presentation to college classes, editorial boards and business groups like the Swedish-American Chamber of Commerce was Political Science 101. First, most Europeans canât wrap their arms around the fact that the American system is really 50 separate state elections all of which then funnel into the puzzling mathematics of the electoral college. Along the way we were also peppered with questions about Trumpâs economic policies, Clintonâs approach to the Middle East and just who in the hell was going to win.
Iâve saved my PowerPoint from that trip because over the years I have often reused some of the slides. I just dug through it last month to do similar presentations from the RNC for the State Department, to remote audiences gathered at the U.S. Embassies in Lithuania, Croatia and Tunisia.
Attached with this column is the last slide in that 2016 deck. It is a late October projection on who would win the 2016 election. Look at the odds. Hillary in a landslide. I grabbed it from one of the nationally respected polling sites. Note they had her winning the blue wall: Wisconsin, Michigan and Pennslyvania. She lost all three.
On Friday, the well-respected Cook Political Report now has Harris ahead or tied with Trump in six of seven battleground states. On the eve of the convention, she has pulled ahead in the key swing states of Pennsylvania, Michigan, Wisconsin, North Carolina and Arizona, and tied with Trump in Georgia. Nevada is the only battleground state that Trump still leads, the poll found.
As subscribers to my podcast and fans of my blog on the âLost Political Middleâ (both found at Scottklug.substack.com) I am not thrilled with either candidate. As a former Republican Congressman, I cannot abide Trumpâs personal conduct and outrageous behavior and Harrisâ proposals for rent control and policing grocery store pricing sounds like Romanian economic policy from the 1970âs.
I understand why Harrisâ fans are over the moon about her candidacy. In just a few weeksâ time she has erased Trumpâs polling lead, and her crowds are a sign of an enormous wave of enthusiasm. The contrast in energy between the two camps is palpable. But like an early lead in an NFL play-off game (I am looking at you my Green Bay Packers) The last half of the 4th quarter often wins games as the 49ers proved. And in politics, so do the last few weeks.
Remember in 2016 Sam Wang, a neuroscience professor at Princeton University and co-founder of the blog Princeton Election Consortium, which analyzes election polling, called the race for Clinton. He was so confident that he made a bet to eat an insect if Trump won more than 240 electoral votes. Nothing like a grasshopper meal live on CNN.
Pollsters will claim they have fixed their models since the â16 races, but I not so sure. Traditional phone calls to land lines havenât worked for decades. Now they must use a combination of online polls, cell calls, text messages and landlines and then bake in some assumptions to come up with an updated projection.
But whatever the secret sauce experts concocted in some mad scientistâs polling lab, the projections have remained flawed. Early election night in 2020 Democrats had the champagne chilled early because polls showed Biden likely winning North Carolina and Florida which meant game over. Instead, the nation stayed up late at night waiting for the usual slate of true battleground states to trickle in. Errors in seven key Senate races were even more egregious which caused political campaigns in both parties to misallocate resources to states they were doomed to lose, or not add enough firepower in states they could have won.
As Pew Research concluded âMost preelection polls in 2020 overstated Joe Bidenâs lead over Donald Trump in the national vote for president, and in some states incorrectly indicated that Biden would likely win or that the race would be close when it was notâ.
I keep looking at this weekâs polls of Michigan, Wisconsin and Pennsylvania and still thinking about the train wreck in election polling in the UK. Pollsters are still trying to cope with whatâs called the âShy Toryâ factor, the conservative voters who wonât respond to polls or lie about who they will vote for as a middle finger to the establishment. The elections in 1992 first got it wrong where the projections of a one-point Labor Party victory were in reality an eight-point Conservative Party win. Same flaws in 2015 and even this summer which wildly underestimated the votes for the Liberal Democrats and Reform parties. Looking back in recent years they were also wrong about Brexit, the vote to leave the EU and half of the pollsters were dead wrong on the Scottish vote to break away from the UK.
The race will tighten in October. It always does. The Trump folks rolled out of Milwaukee with confidence that was sky high. The Harris camp will do the same later this week in Chicago.
Today itâs all seashells and balloons for the Harris camp. Her exuberant supporters on one side of the political divide should take a cleansing breath as should the jittery Republicans on the other.
Ten weeks is a lifetime in politics. I keep that PPT slide to remind myself that polling is ephemeral. Never as good as it looks for the leaders, never as bad as it looks for todayâs losers.