Early voting is already underway, and different media outlets are taking a crack at predicting the outcome based on what has already happened.
Among the biggest projects to attempt such a feat, is the Reuters/Ipsos States of the Nation project, which attempts a model showing the different outcomes based on voter demographic turnouts.
The tracking poll attempts to make contact with 15,000, who have already participated in early voting and it quizzes them on where their allegiances lie.
Currently, the poll is projecting Hillary Clinton to capture a minimum of 278 electoral votes, more than enough to secure the presidency.
However, there is something of a danger when relying on such a model, and it comes from that word “tracking.”
— Rick R Wells (@RickRWells) October 31, 2016
Essentially, tracking polls track the same individuals over a period of time instead of resampling new groups of people each week or month. If there are initial errors with the sample, it will be built into each projection.
That’s why, on the pro-Trump side, there has been a great deal of criticism for the L.A. Times poll that has consistently shown the GOP nominee ahead even as he has faltered in other polls throughout the cycle.
So does that make a tracking poll useless? Not necessarily. They can be good at spotting trends from sampling errors over a longer period, but that initial sample is everything.
In the case of Reuters/Ipsos early voting, the polls are among likely voters, and there is no clear distinction how the samples were taken.
What was the percentage of Democrats to Republicans to Independents? The methodology keeps that information pretty cryptic, but here is the official explanation.
“All three data streams are combined once a week into an aggregate file. This combined data is adjusted at the national and state level by age, gender, race and ethnicity, geographic region, and educational attainment, to match the US Census Bureau’s June 2016 Current Population Survey. Each respondent is given a weight based on how representative they are of the broader population.”
“Because the States of the Nation project is so large, state-wide polling data is reportable for almost all of the nation’s 50 states plus Washington DC.”
Essentially, you would have to know how many likely voters identify as Democrats versus Republicans versus Independents.
While it’s true that over half the country identifies with one of the two major parties, that number is now at a historic low (about 55 percent, according to this look from Gallup earlier this year).
Within that 55 percent, Democrats had the edge at 29 percent compared to 26 percent from Republicans.
The remaining 45 percent could go either way and judging from the two major candidates’ historic lows — neither have hit 50 percent in any reputable polls this election cycle and most ceiling out at 45 percent — they are.
What this means, and it is something the Reuters/Ipsos results freely admit if you read into the methodology, is that early voting doesn’t tell much when you can’t tell what voter turnout will be like on Nov. 8.
According to this report from Yahoo, an estimated 19 million Americans have participated in early voting so far.
Statistic Brain reports that 218.959 million Americans are eligible to vote, so 19 million is approximately 8.7 percent of that. Reuters/Ipsos is estimating a 60 percent turnout rate, which means approximately 112 million additional people will vote.
More than 126 million voted in 2012, so given the primary turnout and the deep divisions on both sides of the political aisle, actual turnout could be closer to the 146 million, who are registered.
That would be a historic high for certain, and early voting has no real way of telling who it would favor at this point.
That said, the Reuters/Ipsos model does have a scenario where Trump wins, thus bucking the current early voting model.
“Trump has strong support among whites,” the site explains, adding that he “leads Clinton by 44 percent to 39 percent among all white voters. But he is supported by just 20 percent of minority voters. So to win, he would have to see his core supporters come out in large numbers while others stay home.”
If turnout is higher than expected for Republicans and older, white Democrats, the site continues, “and lower than expected for Democratic minorities and younger, white Democrats,” then Trump’s likelihood of victory is 75 percent (about 26 Electoral College votes).
— ABC7 Eyewitness News (@ABC7) October 31, 2016
But what do you think, readers?
Will the early voting forecast of a Clinton victory hold true, or is turnout going to break in Trump’s favor? Sound off in the comments section below.