Wow this may seem way too early, as we don’t know which one of the 23 announced Democrats will make it to the be their party’s candidate. However, there are states which consistently vote for Republicans or Democrats, and this is the basis of political forecasting. And there is this incredible paradox in forecasting as the forecast itself can alter the actual results. If the consensus says that party ‘X’ will win, then supporters of Party “X’ are less motivated to vote, and the supporters of Party ‘Y’ are more motivated, meaning more will vote. Also the candidates will change their strategy based on which states are leading or lagging in their run.
I sum it up like this: – it is difficult to identify how a population feels about a candidate when the candidates are doing everything in their power to change people’s opinion.
Polls can be so wrong, for a long list of reasons. Of the eligible voters, 42% did not vote. Any poll which included the general public, may be representative of the popularity of a candidate, but this isn’t what counts in an election. In an evenly divided state such as Florida, obtaining a representative sample in a timely manner is difficult. The polling margins in Florida (% difference between candidates) were very small in 2016, so results were significantly affected by sampling errors. If this stuff interest you, then you have to know the state symbols and the websites.
There are many sites of which 270towin.com and Larry Sabato’s Crystal Ball are the best right now in providing non-partisan forecasts of the US Presidential elections. In my opinion, forget any website which says “It looks like a landslide for Candidate ‘X’ or Party ‘X,’ I can’t predict much, but a landslide is out of the question. We have been a very evenly divided country politically for at least 19 years.
To make a educated guess of who will win the election, the prognosticators generally subdivide the electoral states into 7 categories (Solid Dem, Likely Dems, Lean Dems, Toss up, Lean Rep, Likely Rep and Solid Rep). There’s a 5 category model with Likely and Lean combined into one category.
There are 5 big toss ups, as follows with the state symbol and electoral votes: AZ(11), FL(29), WI(10), MI(16), PA(20). This is the “Consensus view” per the 270toWin website. The total is 86 votes. We have in these 5 states something I will call “Hardcore uncertainty.” The tossup total is 87 (one extra comes from NE),
The biggest prize is Florida, carrying 29 EV, or about a third of the total tossups. The candidate which wins Florida, Michigan and Pennsylvania takes 75% of the tossups.
Now, not everyone is agreeing on this tossup category. Larry Sabato’s Crystal ball from the University of Virginia, considers PA, AZ, WI and NH as tossups, for a total of 46%. Sabato adds NH but the state only has 4 EV’s. So, the big prize is Pennsylvania. In Sabato’s forecast, Florida leans Republican and Michigan leans Democrat, Net it is a gain of 13 EV’s in favor of Republicans.
As I take a more broad view of the “solids” in the map, it seems that evident that Republicans are the majority in the US heartland with a solid red, from North Dakota and Montana to Oklahoma and Louisiana. Democrats are solid on the very populous east and west coasts. Texas was usually considered a solid Republican state, but is now considered in the “likely” group for Republicans. Our country is not as divided as the maps might seem to present. It is far mixed up politically with local elections in red states going to Democratic candidates and vice versa. The third link from Wikipedia has an excellent discussion on “Map Interpretation.:
This forecasting is interesting because everyone sees the same information at the same time. In fact each political party is very focused on the polling numbers. and generates their own forecasts to boot. Again, I come back to the point that whatever the consensus forecast is at any one time, the subsequent actions of the parties acts to drive the statistics in the opposite direction, reducing the forecast’s accuracy. For example, if Democrats believe they can turn Texas or Republicans can turn Michigan, they will race to the state, with a barge of arguments why only their candidate can run the country. A final point is the hypocrisy in all this is not lost, as each candidate will tell their supporters that they are more interested in issues rather than polling statistics, when their staffs and paid consultants are looking daily at the numbers.