The polls for this election cycle have been all over the map, calling into question whether voter sentiment in 2016 can be measured at all.

The latest Fox poll has Hillary Clinton up by 3 points over Donald Trump.  This flips the results of last month, when Trump was beating Clinton by 3. 

Clinton’s edge over Trump is due to a six-point drop in support for him rather than an increase for her.  Trump was up by 45-42 percent three weeks ago (May 14-17, 2016).  Since then, he lost three points among self-identified Republicans and 11 points among independents. 

Clinton is ahead among blacks (+76), unmarried women (+34), women (+18), lower-income households (+14), and voters under age 30 (+13).

Trump is preferred among white evangelicals (+42), whites without a college degree (+25), whites (+16), men (+15), and independents (+5).

Expect the race to remain tight, as people are pretty set with their choice. 

Voters were asked if there is a chance their candidate “could say or do something” that would make them change their mind before Election Day.  More than 8-in-10 Clinton backers say there is no chance at all (57 percent) or only a small chance (24 percent).  Sentiment is almost identical for Trump supporters: 57 percent say no chance and 23 percent say just a small chance. 

Voters may be committed to their candidate, but that doesn’t necessarily mean they’re happy about it.  Trump supporters are split between being happy to vote for him (51 percent) and holding their nose (48 percent).  Clinton backers are more upbeat:  60 percent happy vs. 37 percent holding their nose. 

For comparison, in March 2000, only one quarter of George W. Bush supporters (24 percent) and those backing Al Gore (25 percent) said they would have to hold their nose.

You don't have to be much of a pollster to figure out that the public are unhappy with their choice for president.  But is there something different this time that explains the volatility of the electorate?

Polling is based on models of the electorate that include turnout and weighting of responses by groups including whites, blacks, Hispanics, urban, rural, suburban, Republican, and Democratic, to name a few.  The more people who agree to be interviewed, the more accurate the poll – supposedly.  But none of it matters if the model is screwy, or doesn't reflect the reality of what's happening in the electorate.

I'm not buying the high percentage of Trump and Clinton supporters claiming that nothing their candidate does will force them to change their vote.  I think people are changing their vote often this election cycle, given the high negatives of both candidates and the near disgust with which voters view both Clinton and Trump.

In short, many voters are having a hard time determining which candidate is the lesser of two evils.

By the looks of things at this point, there are going to be a lot of red faces in the polling community the day after the election.