The initial collection of RCP Swing State Polls presented here on August 3 has been revised, corrected, and extended. The group followed includes all 15 states Rick Wilson recommends, plus Texas, well just because…
As Rick Wilson, one smart ass political strategist, keeps saying: There is no national election. The Electoral College is All that Matters. There are 50 state elections this year, and only 15 make any difference in the outcome. 35 are already decided and baked in.
These 16 states we follow hold 225 electoral votes combined, or 83% of the total needed to prevail in November and elect a President. In 2016 Trump won 10 of them (179 electoral votes) and Clinton 6 (46 electoral votes). This set of state wins provided 58% of all Trump’s Electoral Vote total (304 votes).
So, in this time of trouble with other serious concerns like COVID-19 weighing on all of us, let’s have a simple dashboard of the state polls that matter starting 100 days out (the outer limit of relevance to the final result.)
Thanks to the number crunchers and aggregators at RealClearPolitics.com, we have compiled a list of all swingy state Presidential Polls reported each day starting on July 26, 2020. All polls are included. If there is more than one state poll reported on a given day, the results are averaged for this purpose.
The charts are presented in summary files of 10 consecutive days each. So, there will be ten deciles of results leading to Election Day 2020.
As of today, then we have 1 full decile and nine days of data to describe.
There are still lots of blanks, as you would expect in mid-August. Here is the final chart for the first 10 days of the 100 Day Election Countdown (Decile 01), ending on August 4, 2020.
These results by themselves don’t mean much in isolation, until a significant number of additional days are tabulated closer to the election. As Nate Silver reminds us, the variability is much greater the further out in time the measured event is scheduled to take place. What does matter is the trends observed as time passes and we get closer.
Here goes.
One additional note. For our purposes we are tracking polling done per state per day in sum. Sometimes more than one poll might be reported in a given state on a particular day. When that happens, we take a simple average of those polls to get a single state day result.
During the 10 days from July 26 through August 4, there were 21 State Polling Days reported by RCP in the 16 Swing States in our Inventory. Biden won 19; Trump 2. Trump’s two wins were by one percentage point in Ohio, and again in Georgia.
During this 10-day period there were 21 state polling days reported out of a possible total of 160, or 13%. No state polls were conducted during this time in four of the Swing 16: Colorado, Iowa, Nevada, and Texas (two Reds and two Blues from 2016).
Another analytic summary added to the RCP Polling results is information from the Cook Partisan Voting Index (PVI), which is the standard for assessing how a state tends to actually cast its votes on Election Day, regardless of poll indicators.
From the Wikipedia entry:
The Cook Partisan Voting Index (abbreviated CPVI or PVI) is a measurement of how strongly a United States congressional district or state leans toward the Democratic or Republican Party, compared to the nation as a whole. The index is updated after each election cycle.
The PVI has been in use for more than 20 years and proven its utility and reliability over that time.
Given the circumstances of a likely close election, the potential for “shy” Trump voters in polling responses, the complications of COVID-19 Pandemic raging in the US now, etc., we have adopted the following very conservative approach to assessing whether Trump or Biden seems to be leading in a particular state. When ever Biden is ahead on average for a state we deduct the full Cook Partisan Voter Index percentage from that lead where the state leans Red.
In states which normally lean Blue, we also adjust (deduct) the full PVI from any Biden average lead to try and avoid any complacency stay home factor that might demotivate potential Democratic voters this year. So, we have a lose-lose situation to diminish any apparent Biden polling advantage. Extra Conservative, belts and suspenders.
That said, we then present the Decile (10-day average) Biden State polling lead. If this number is negative, Trump is leading. In Decile 01, by this measure, Trump is ahead in 2 of the 12 Swing States reporting polls: Georgia by 4.0 points, and Ohio by 3.5 points.
There is lots of time for Trump to make up ground, and there are always October Surprises to spring, but these are not Break Out the Cigars results for the Trump Campaign. In 2016 Trump won Georgia by 5.1% and Ohio by 8.1% on Election Day.
Using this earliest iteration of the unfavorable Biden index adjusted lead, in three crucial Swing states of Michigan (5.5), Pennsylvania (4.5), and Wisconsin (3.0), Biden has small leads. A rule of thumb is that a Cook PVI of 10 or more is needed to make a race nearly out of reach for the opposing party.
No Biden cigars either, not by a long shot.
And it’s just too damn early.
Tomorrow, we will have the complete data for the second decile to report and summarize.
But here is the nine-day chart as of August 13 to set the table.
More later.
America, Do Your Job.
Vote November 3
Absentee or In-Person If You Can