Very small changes in the Dem / GOP mix make a YUGE difference.
A recent WaPo poll had Clinton over Trump by 2 points.
Naturally, voter preference varied by self-proclaimed party affiliation:
Clinton got 86% of of the Dem vote; Trump got 88% of the GOP vote.
Independents went 47% to Trump and 40% to Clinton.
WaPo weighted their sample 37% Dems, 30% GOP and 29% Independents.
In concept that weighting should represent WaPo’s best guess as to the mix of voters on election day.
Some pundits argue that WaPo’s aggregate tally over-weights towards Dems.
Maybe it does, maybe it doesn’t. We’ll know in a couple of days.
For today, let’s just play around with the numbers to demonstrate how critical the turnout assumptions are …
Let’s start by checking the WaPo calculations…
The WaPo weightings and the voter preferences actually roll up to slightly different numbers than WaPo puts in the headline.
Not a big deal: Clinton 45.2%, Trump 43% … 2.2 points difference.
Why are the numbers different?
Sharp eyes may have noticed that 37% + 30% + 29% doesn’t add up to 100% … it only adds up to 96%.
Again, that’s not a big deal … just means that the numbers need to be massaged a bit.
To make things simple, we’ll compare to these ‘actual’ calculations from this point forward.
Now, let’s see what relatively small changes in the turnout weightings do to the overall result.
Still using the WaPo voter preferences by party affiliation, we can solve for a turnout mix that puts Clinton and Trump in a dead heat.
Note that it only takes a 1.4% shift in the Dem / GOP weightings to make the race a tie.
Said differently, if Dem turnout is 35.6% (instead of 37%) and GOP turnout is 31.4% (instead of 30%) … then the race is tried.
And, that’s still using the WaPo party-affiliated voter preferences – 86% of Dems for Clinton, 88% of GOP for Trump.
Let’s do one more scenario ….
Let’s assume – as some pundits do — that an equal number of Democrats, Republicans and Independents show up at the polls.
That is, Dems are 32% of the mix (instead of 37%), GOP are also 32% (up from 30%) and Independents go from 29% to to 32%
Look what happens:
Through the miracle of simple arithmetic, Trump goes from a 2.2% deficit to a 3.5% lead … all while holding the WaPo party-affiliated voter preferences constant.
Bottom line: The headline numbers that are being reported shouldn’t be taken too literally … they are way too sensitive to the the polls’ turnout assumptions.
More practically, the conclusion to be drawn is that — since both candidates are preferred by roughly the same percentage of their own-party voters – the election will hinge on relatively small spreads in voter turnout.
It’ll be Clinton’s ground game against Trump’s momentum.
We’ll know in a couple of days.