Donald Trump has significantly cut into Hillary Clinton’s sizable lead in Michigan, but still is facing an uphill battle with voters between now and Election Day on Nov. 8 if he is to take a state that has been a key part of his strategy for winning the White House, a new Free Press/WXYZ-TV poll shows.
The exclusive poll done for the Free Press, WXYZ-TV (Channel 7) and their outstate partners by EPIC-MRA of Lansing showed Clinton, the former secretary of state and Democratic nominee, leading Trump 41%-34% in a four-way race that also includes Libertarian Gary Johnson and Green Party nominee Jill Stein.
That 7-point edge for Clinton over Trump, a Republican businessman, casino developer and reality TV star, is down from what had been an 11-point margin earlier this month. But it still is slightly higher than the 6-point advantage President Barack Obama had over former Massachusetts Gov. Mitt Romney -- a native of metro Detroit -- at this point in Michigan four years ago. Obama went on to win by 9 points.
This is a much different race than that one, however, given that both candidates continue to have high unfavorable ratings -- 53% for Clinton, 61% for Trump -- in this most recent survey. The number of undecided voters -- 13% -- also is more than double what it was toward the end of October four years ago.
That uncertainty could allow Trump’s supporters to argue that Michigan is still up for grabs. But with less than two weeks remaining in the campaign, there is little direct evidence to prove so in a state where no Republican nominee has won since 1988, and where Trump has trailed in virtually every poll taken for months.
“He has seen some improvement from the last survey, but seven points is still significant,” said EPIC-MRA pollster Bernie Porn, who said younger, less active voters may come off the sidelines and serve to build Clinton’s lead. A turnaround for Trump, he said, “would be one of the most remarkable comebacks in history.”
For the poll, EPIC-MRA surveyed 600 active and likely voters using live interviewers and a random sample that included 30% cell phones between Saturday and Monday. The poll has a margin of error of 4 percentage points, though that would be higher for subsets of specific voter demographics such as gender, race or regions.
Michigan’s polls appear to be following a trend across the U.S. with Trump narrowing Clinton’s advantage in Pennsylvania, Florida, New Hampshire and other battleground states, but with the Democrat maintaining what appears to be enough of an edge to win barring a collapse in support. Trump, meanwhile, has taken to characterizing polls showing him behind as "phony,” though he trumpeted them when they have showed him ahead.
At present, Real Clear Politics -- a website which aggregates and averages polling data -- shows Clinton with a 9-point edge on Trump in Michigan ahead of this most recent survey and has taken it out of the list of tossup states, despite Trump’s five visits since the Republican convention in July and a strategy that called for winning in Rust Belt states including Michigan, Pennsylvania and Wisconsin. Nationally, the site has Clinton leading Trump by 5 percentage points.
If that trend holds, Trump’s chances could be slim: Early voting has already begun in many states and the campaign has had to spend much of its time working to claw back leads in battleground states like Ohio and Florida, while trying to stave off potential losses in traditionally Republican states like Georgia, Arizona and Utah.
Still, the Free Press/WXYZ poll has positive indications for Trump in Michigan: In several regions across the state, Trump has taken clear leads -- including in northern Michigan, where he leads Clinton 49%-32% and the undecided vote is virtually nil -- while Clinton’s sizable lead in metro Detroit has shrunk slightly from 55%-25% earlier this month to 50%-26% now. Trump also leads in central Michigan and out west, though his 35%-32% edge in the latter -- a traditionally Republican stronghold -- should concern his campaign.
The poll also showed that Trump had retaken a lead among white voters -- 41%-33% -- since early October, and held a slight advantage -- 39%-36% -- among men, where Clinton had led among men and women in the state before. Trump also had a 45%-32% lead in Macomb County -- the home of the so-called white, working-class “Reagan Democrats” who helped Ronald Reagan win in 1980 and 1984 -- though the number of people surveyed in the county was small enough that the margin of error for that sample would be large.
While Trump had improved his performance among many voters responding to the poll, he still faced stiff headwinds in Michigan among African Americans, women, college educated voters and others.
Among women, Clinton saw her huge gender gap over Trump shrink somewhat -- it was 20 points earlier in the month -- to 45%-30% in this poll. But considering Trump’s much smaller lead among men, that continuing advantage could be decisive. And where Trump had a lead among whites -- not uncommon for a Republican nominee -- Clinton had the support of 88% of African Americans surveyed, which is enough to make up the gap.
Trump's support among blacks was so small it didn't register in the poll. And Trump’s lead in Macomb was more than offset by larger ones for Clinton in Oakland (51%-25%) and Wayne (58%-18%) counties, both of which are more populated than Macomb and are the only counties of the state's 83 with populations over 1 million. Overall, the southeastern part of Michigan accounts for the lion’s share of the state’s vote. In Detroit, Clinton leads with 93% while Trump’s support in the city didn’t register in the poll despite several visits.
Clinton continued to have a lead among all education levels as well -- including not only a 42%-33% lead among college graduates, but also among those with a high school diploma or less, 44%-35%, despite an expectation that those voters would break for Trump. She also scored better with the 26% of voters who consider job creation the most important issue and the 14% who say education is most important. Trump did better with the 9% who cited national security as their top concerns and 7% who said trade deals were most important.
There also appeared to be an enthusiasm gap: Clinton continued to get nearly half of voters who scored their enthusiasm for voting at a 7 or higher on a scale of 1-to-10, compared to less than 40% of Trump’s supporters who said their enthusiasm was so high. And while 34% of all respondents said they would “definitely” vote for Clinton, 27% gave that response for Trump. Four percent of respondents said they had already voted by absentee ballot in Michigan for Clinton; 2% said they had already voted for Trump.
Self-described independent voters -- accounting for 19% of the survey pool -- continued to be deeply ambivalent about the election, however, with 26% supporting Trump and 25% supporting Clinton, a virtual tie. Johnson got 15% of their support. A whopping 32% said they were undecided.
Porn said that with undecided support falling at the low end of the enthusiasm spectrum, however, he wouldn’t be surprised if a lot of undecided voters end up staying home on Election Day -- a prediction which, if correct, could make it difficult for Trump to overtake Clinton.
As for the polls being wrong, Porn said it’s unlikely that so many polls predicting the same result are all off-base, though it does happen. The Free Press, EPIC-MRA and others got the Michigan Democratic primary wrong this year, for example, predicting Clinton would win by double-digits when in the end she lost to U.S. Sen. Bernie Sanders of Vermont.
General elections tend to be much easier to poll than primary elections, however, with much greater turnout and a much larger pool to survey. But Trump and his supporters have maintained that if enough new voters go to the polls -- especially in areas with high numbers of disaffected white, working-class voters -- he will win.
“There are thousands upon thousands of interviews that would have to be wrong for that to be the case,” Porn said about the outcome in Michigan and the polling there and across the U.S. “As a pollster, I have a hard time believing that many polls could be wrong or that many folks are withholding their judgment.”