Is one of those pairs of data not like the other two?
Gee, it looks to me like the pair on the right is different from the first two.
Clinton was cruising before her “coal” comment. In 14 state primaries before 3/13, she was running 8% ahead of her vote count in those same states in the 2008 primaries, while total turnout in the Democratic primaries was down 25% from 2008 in these states. Sample size was 5.5 million votes for Clinton and Sanders combined.
This excludes MI, one of the three “outliers” excluded from this comparison. TX and CA are the other two, voting POST-comment. These outliers had material differences from 2008 which rendered their turnout not comparable. In MI, the 2008 primary was unofficial and not contested by Obama; turnout in 2016 was up 97%.
In CA, the 2008 primary was the largest prize in a huge Super Tuesday which included NY, IL and NJ on February 5. Turnout was high as voters saw the chance to set the tone for the entire campaign and establish momentum for their candidate. In 2016, CA voted on June 7 when the race had already long since been decided and turnout was down 32%.
In TX, the 2008 primary came a month after the huge Super Tuesday, when the race was still neck-and-neck and TX was the largest prize remaining. Obama and Clinton heavily contested the TX primary, and turnout was high. In 2016, TX moved up to Super Tuesday while CA, IL, NY and NJ moved later, so TX was no longer the largest prize remaining. Clinton and Sanders did not contest TX as heavily as Obama and Clinton had in 2008, and turnout was down 51%.
So that’s the rationale for excluding the outliers. Now back to the chart. Clinton is cruising in the 14 “non-COAL” states before 3/13, and continues cruising in 21 more after 3/13, gaining 6% more votes than she did in 2008. Sample size was 12.8 million votes for Clinton and Sanders combined. In the 35 “non-COAL” states combined, Clinton gains 7% more votes than in 2008, while total DEM turnout is down 17%.
But the reaction in the five “COAL” states is drastically different, isn’t it? OH, PA, WV, KY and IN had a sample size of 4.1 million votes for Clinton and Sanders combined. Turnout was down 42% in the five states combined (-50%, -30%, -42%, -40% and -50% respectively), while Clinton received 43% fewer votes than in 2008 (-46%, -28%, -64%, -54% and -53% respectively).
Two takeaways seem really important here: first, the decline in support for Clinton was a direct response to her policies, and had already occurred by the time of the primaries in MAR-APR-MAY; second, therefore, James Comey and Wikileaks (and the Russians, if they were involved) had no effect on the decline in support. Those disclosures began in July.
Notice what has to be true, in order for Clinton’s whining that Comey re-opening the e-mail investigation in October cost her the election: these voters who had jumped ship in the primaries because of the “coal” comment had to have come back on board, but then changed their minds when Comey re-opened on October 28.
But that’s a slippery slope. If she grants that these voters will change their minds when they receive new information … then that same argument should hold that they will have come back on board when Comey gave the “all-clear” on November 6, and voted for her on November 8.
So the only impact would be if these voters had voted early, before Comey gave the all-clear. We should be able to see empirical evidence of that. PA would have had no impact from the reopening, because they had no early voting, so let’s take OH as an example. If the “true intent” of the voters on Election Day was to continue the Obama policies, then Clinton would have won the votes cast on Election Day by the same 50-47 margin that Obama had in 2012 in OH. That would require the margin in early voting to have been 61-28 for Trump, in order to produce the actual 2016 vote count. That would be easy enough to check, wouldn’t it?
And if the early vote had been 61-28 for Trump, would the news anchors have been confidently predicting a Clinton win early in the night? Wouldn’t the early returns have shown Trump with the huge lead from early voting (which gets announced first)?
Btw, if a bunch of voters who usually vote on Election Day had gotten disgusted when Comey reopened, and gone out and voted early to just get it over with while they were disillusioned … then we would see OH early voting totals about 400,000 higher in 2016 than in 2012. They were down a little, consistent with registered voters down a little from 2012, so I went with the calculation based on those who regularly vote early … doing so while they were disillusioned.
So the choices boil down to these two:
- Voters in OH, PA, WV, KY and IN withdrew their support for Clinton in the primaries because of her policies (the “coal” comment) … and never came back to her; or
- The early voting results would be 61-28 for Trump in OH, to reflect that these voters had come back to her, but jumped ship again when Comey re-opened, and then voted early while they were disillusioned.
Maybe the WaPo can send in their Judge Roy Moore team to get to the bottom of this! Those journalists have proven their ability to ferret out information vital to keeping the public informed about our electoral process, right? So do this great service to America, WaPo – send in your team and ferret out the early voting results in OH!
If it really was 61-28 Trump … while Election Day voting was 50-47 Clinton … then we’ll have to acknowledge that maybe Comey’s re-opening did have an impact. On the other hand, if you don’t find this split, then you’ll have to acknowledge that voters rejected Clinton because of her policies. And did so early, during the primaries, with no impact from Comey or Wikileaks.
We Deplorables are certainly willing to take this challenge and live by the data. Are you?
ERpundit – 07/01/18