# 11/11/20 - All The Losers and Bigger Losers From Democracy's Night of Nights

Well, it looks like Joe Biden is going to be the next president of the United States and I for one have never been more ambivalent! Despite Trump declaring victory already, dog-whistling about voter fraud and demanding recounts in several states, it seems that the rats (Republicans) are beginning to flee the sinking ship and it's finally Biden Time ©. Yes one day soon, Donny will wake up in the White House for the last time, put on his marquee-sized suit, pack his belongings into a cardboard box, hand in his access cards and be escorted off the premises by security. As he's standing in the car park pensively surveying his old residence, he will look down at a framed photo of Ivanka and exactly none of his other children and a single tear will dribble down his cheek as he says to himself, "thank Christ that's over".

Aside from Trump, there are a couple of other big losers from this election:

  1. The Democrats - Technically they won but talk about falling over the finish line
  2. Nate Silver - Whooops 😬


Image by Nathan Saad.

# The Democrats

Don't listen to the establishment Democrat talking heads that have been in rationalisation mode ever since the disappointing numbers started rolling in, this is an awful result for progressives and many on the left are devastated. Yes, Joe Biden has won, but lest we forget his opponent was an odious caricature of a baseball mascot impersonating a clown, who's key policy achievements include an economic recession, the genesis of potential civil war and the death of ~250k Americans from a virus he failed to control (and that was all just this year). There is no question that the electoral college system is flawed, voter suppression is effective and that Biden's principled decision not to hold dangerous mass-gathering events (unlike Trump) hampered his campaign, but if ever an election was a gimme it was this one. Not only did the Democrats fail to win the Senate (not confirmed but likely), they also lost seats in the House and these results, combined with the stacked conservative Supreme Court, will render it near impossible to pass any meaningful legislation over the next 4 years.

I suspect Biden won't be too dismayed by this arrangement. As much as Fox News has tried to paint him as a communist (lol), in reality Comrade Joseph's greatest hits include:

  • architecting the 1994 crime bill that led to the mass incarceration of Black Americans
  • support for the Iraq War
  • a cosy relationship with and legislative leniency for the banking industry (earning him the nickname "The Senator From MBNA")

While Joe Biden was pushed to the left during the Primaries and ultimately did run on a RELATIVELY progressive platform, many of these positions appeared to be the result of political calculus rather than personal convictions. Legislative paralysis may prove to be a convenient excuse for not following through on campaign promises. Across the aisle Republicans will no doubt be disappointed with last week's result, but perhaps some solace will be found in the knowledge that the Democrats have inherited a busted economy, a worsening health crisis and a divided nation, and will need to juggle all three of these hot potatoes with one arm tied behind their back.

Sign Up To My Mailing List

Hey you are reading this thing! If you are enjoying the article then you should definitely subscribe below. I may be a bot but I will never spam you.

* indicates required

# Pollsters

Stop me if you've heard this one before... Despite assurances that they had dealt with the errors of previous campaigns, the pollsters got this election very wrong. Many convoluted theories are floating around about why the polls are so inaccurate but it's actually not that complicated; people don't answer the phone anymore.

In statistics, if you want to know something about a population you can take a random sample from that population and if your sample size is high enough, the measured value in your sample should roughly align with the true population value. For example, consider a school of 1000 students. If you put the names of every pupil in a hat and pulled out 50 names at random, the average height for those 50 students will be close to the average height for the whole school. Unfortunately, there are a couple of practical problems with this technique when it comes to election polling.

# Just because people say they are going to vote a certain way doesn't mean they will

Unlike in the school example where we measured the true mean height in our sample, we can't record the actual voting numbers for our sample (since the polls are done prior to election day). All we can do is measure the way people say they intend to vote and extrapolate from there. An individual voter may intend to vote for Candidate A when the poll is conducted but swap to Candidate B by election day, or they may just decide not to vote at all. Another controversial hypothesis (sometimes referred to as the "Shy Trump Voter Theory") is that some voters are dishonest and say they will vote one way when they intend to do the opposite.

# The samples used for polling aren't random (not even close)

Most pre-election polling is conducted by randomly dialling phone numbers to obtain a sample but this is not a very effective way to get a representative slice of the population. Even if everybody in the population who owns a phone is equally likely to be called, they are not equally likely to pick up the phone OR equally likely to want to participate in the survey. To use a simple example, people that answer their home phone are far more likely to be older and therefore the views of older Americans will be over-represented in the raw results.

Polling companies are well aware of these inherent biases and do a lot of work to account for them. Many record demographic information from participants and then use census data to make adjustments. For example, if 1000 people are polled in a state and the proportion of those that are Hispanic is 5% but we know from census data that the true proportion of Hispanic voters in that state is 9%, then they will weight the results of the Hispanic participants to account for their under-representation. This approach does somewhat attenuate the effect of sample bias, but it is only a partial solution. While it may be possible to correct for broad demographic trends, some biases are more complex and/or harder to pick up. Consider trust in institutions. People who have low trust in institutions are more likely to vote Republican but they are also less likely to want to participate in a phone survey. In this recent election, polling companies have tried to use education level as a proxy for this effect but this is a crude generalisation and clearly it didn't work. To ape the ignominious Don Rumsfeld, in polling there are known unknowns and unknown unknowns. By definition, there is very little that can be done about the latter.

None of this is to say that we should stop polling, but we certainly need to manage expectations better and pollsters (and certain poll aggregators) should hesitate before marketing their products as unbiased and infallible prophecies. Much like economic modelling and COVID-19 forecasting, polls are an invaluable tool BUT they are not scientific and rely on subjective assumptions. To the average person, polling looks deceptively similar to machine learning in that it involves maths, data and making predictions, but in truth they are entirely different fields. Perhaps some of the hype around data science has unfairly bled into public perceptions of survey research when in reality it might be more helpful to think of it like cricket's "WinViz" or the NRL's "VB Hard Work Index": entertaining, somewhat informative and best consumed with copious amounts of salt.


# Where To From Here?

Despite all the cynicism in this piece, I am of course glad to see the back of Donald Trump (even if "Trumpism" endures). Moving forward, I hope that progressives in America and Australia take a lesson from this election cycle because if the left continues to serve up "well-qualified", milquetoast candidates with contentious records (think Hilary Clinton and Bill Shorten), they will keep on getting trounced by populist, charismatic, policy lightweights (think Donald Trump and Scott Morrison). While Biden has prevailed this time, he was on track for a huge loss if not for the pandemic.

Another take away (and it's an uncomfortable truth) is that policy wonkery and "getting epic" on Twitter does not translate to electoral success. For the most part, the public couldn't care less about 10 point plans and find the contrived "zingers" ministerial staffers slave over off-putting. It's not a coincidence that a game show host and a literal adman are the leaders of our two countries. The public is seemingly so starved of authenticity that they will accept simulacra even in these respective hideous forms.

If you enjoyed reading this post why not sign up to my MAILING LIST? (opens new window) You'll love it.