WashPost Folks Wrong About Lack of Trump’s Big Wave
Earlier this month, three writers for the Washington Post eyed both the U.S. Census Bureau Current Population Survey and David Leip’s Election Atlas to come up with an analysis that led them to suspect that Trump never received the predicted wave of working-class voters on Election Day (linked here).
In their analysis, they used the Census survey of about 100,000 people and then compared this with the “turnout rate” of whites, blacks, Hispanics, and Asians, as well as turnout rates in six swing states, comparing 2012 and 2016.
Instead, though, I looked at the actual vote numbers – ballots cast – not turnout rates, and the perception is actually different. The analysts also seemed to be missing the correlation between increased actual ballots, drops in some categories, but shifts from Obama to Trump and independent candidates (white or otherwise).
Here’s what I found:
In Michigan, as an example, the chart shows a slight dip in the turnout rate in the state. But in actuality, more people voted in 2016 – 4,790,917 – than in 2012 – 4,717,728 (about 73,000 more).
The chart shows drops in all turnout rates except Asians who rose in turnout but are a small portion of the Michigan population – 2 percent. So, even a tiny bump is pretty insignificant as it relates to the Asian population there. The white turnout rate drop is smaller than the black turnout rate but it doesn’t give a percentage for each or an actual estimated number of voters based on the actual turnout that was higher in physical voters than 2012.
So, when the analysts say this: “There are many factors that may have contributed to a victory that was decided by a mere 78,000 votes. But now we know one thing it wasn’t: spiking turnout among Trump’s core constituency,” this conclusion doesn’t match the data (and they didn’t show the exact numbers to show that it matches).
While Clinton received 293,000 fewer votes in 2016 in Michigan than Obama in 2012, Trump received more than 185,000 more votes than Romney’s – that would be considered a surge. At the same time, in 2016, four indies received more than 242,000 votes – this would also be considered a surge when compared to 2012, where three indies received about 43,000 votes.
There’s the shift: Trump won the state because more people voted in 2016 than in 2012 and the Obama coalition – of which white, working-class people played a huge role in that state – didn’t support Clinton; they voted for Trump and the indies.
The analysts concluded that Florida was the only state that showed a jump in white non-college-educated turnout. But what about Ohio? It showed a slight uptick in turnout rate but actual voters appear to be close to even: 5,368,334 in 2012; 5,325,395 in 2016, a difference of nearly 43,000 less or 0.8 percent. In 2016 though, Trump received more votes than Romney and Obama in 2012: 2,771,984 (Obama had 2,697,260; Romney 2,593,779). This along with higher white voter turnout goes completely against the analysts’ perception. Clinton was far behind with 2,317,001 and indies received nearly 160,000 more votes in 2016 than they did in 2012.
Let’s move to Pennsylvania where the analysts show a slight turnout rate increase but it was actually more than 370,000 votes when compared to 2012, with Trump receiving more votes than Obama and Romney did in 2012 – again, a surge. Clinton also received more votes than Romney did in 2012 – about 220,000 – but 63,000 shy of Obama’s numbers. Indies pulled in about 140,000 more votes than 2012.
In Iowa, the chart shows a pretty big turnout rate dip. However, in actual voters, it was less than 1.7 percent: In 2012, 1,563,943 people voted; in 2016, it was 1,537,673, a difference of about 26,000 – the size of a single, large town. The shift though was pretty dynamic; Trump received nearly 799,000 votes – 71,000 more than Romney but about 18,000 less than Obama – again, that’s a surge. Clinton, however, received 166,000 less than Obama. Indies received 68,000 more votes in 2016 when compared to 2012.
In Wisconsin, the chart shows a dip in the turnout rate but nearly the same amount of actual voters cast ballots, according to the data: 2012 – 3,055,484 to five candidates; 2016 – 2,943,059 to six candidates or about 112,000 less (a 3.7 percent drop). Here, the presumption of a surge to Trump doesn’t hold up – he received about 800 more votes than Romney did; Clinton received 231,000 fewer votes than Obama. Indies, however, received nearly 119,000 more votes in 2016 than 2012.
I requested in the comment section that someone post the raw Census and Election Atlas data but it never was posted. Either way, the analysis is flawed.
Washington Post chart screenshot.