Trump will take Utah in November

There is a lot of chatter in the #NeverTrump echo chamber that the presidential candidate they love to hate will lose Utah to Hillary Clinton, or at the least it will be a nail-biting race.

It won't be close.  Utah will easily go to Donald Trump in the general election, and polling trends are headed in that direction.

The poll that started off the hysteria was released March 20 by Deseret News/KSL, showing Clinton narrowly ahead of Trump by 38% to 36%.  But a more recent poll released May 16 in the same media outlet had Trump with a large lead, 13% over Clinton at 43% to 30%.

Over the last two weeks, a couple more polls have been released supposedly showing a tight race again.  The most recent, which came out Saturday from the Salt Lake Tribune, has Trump and Clinton tied at 35% in a three-way race with Gary Johnson, who has 13% of the respondents' support.  A Gravis Marketing poll conducted on June 1 puts Trump over Clinton by 36% to 29% in the head-to-head, and by 29% to 26% when Johnson (who gets 16%) is included.

The problem, as discussed many times previously in this column, is that demographics is destiny when it comes to polling accuracy.  What we see time and again when it comes to polling data involving Trump versus the liberals is that there is a strong liberal bias in the demographic composition of those being surveyed.  Rarely, if ever, is there bias that would favor Trump, and certainly not to the extent regularly seen against Trump.

Only the Gravis poll appears to provide some limited demographic data, and it suggests a problem.

A full 70% of those surveyed were Mormons, many of whom don't like Trump because of Mitt Romney's continued opposition, while just 55% of all adults in the state are Mormon.  Some may argue that 2012 exit polls suggested that 75% of Utah voters in the last presidential election were Mormon, so we should aim for this percentage in polling data this time around.

Wrong.

Exit polling data is notoriously incorrect, to such a degree and in such unpredictable ways that it is effectively useless for developing high-quality demographic profiles of the likely voting base in future elections four years down the road.  Polling demographics should always be matched to the potential voter base, and in this case, that would have involved just 55% Mormons.

Another problem involves the level of education of the polling respondents.  When asked about their highest level of education completed, 54% said bachelor's degree or higher.  But according to the U.S. Census Bureau, just 30.6% of those 25 years and older in the state have completed a bachelor's degree or more.  Only 2% of those surveyed did not complete high school, whereas the census data has this at about 9%.

The age demographics in the poll also appear to be way off.  Just 17% of respondents were 65 years and older, while 72% were between 30 and 64 years of age, for a ratio of 30-64:65+ of more than 4:1.  By comparison, the actual ratio between these two age groups is about 2:1.

Probably of greatest importance is the political orientation of those polled: 51% were Republicans, and 21% were Democrats.  Work by Brigham Young University's Center for the Study of Elections and Democracy suggests that while the percentage of Democrats in the survey matches with reality (20%), the percentage of Republicans is likely far too low (should be 65%, not 51%) if this 2012 data still holds.

So it looks as though the poll may be biased strongly in favor of Mormons, toward those with high levels of education, against older voters, and against Republicans in favor of Democrats.  Put all these together, and one might reasonably suggest that the survey base is stacked heavily against a fair assessment of Trump's real level of support in the state.

With no detailed demographic data for the other polls, we can't make a first-cut assessment of their reliability, but if we attempt to correct the Gravis poll for possible bias, Trump's lead looks to be in the range of 15%, which is roughly consistent with the May 16 Deseret News poll.

Unless another poll backs up the Salt Lake Tribune outlier suggesting a Clinton-Trump tie, Trump's support is probably stable with a healthy advantage over Clinton, and this is likely to hold until the general election.

There is a lot of chatter in the #NeverTrump echo chamber that the presidential candidate they love to hate will lose Utah to Hillary Clinton, or at the least it will be a nail-biting race.

It won't be close.  Utah will easily go to Donald Trump in the general election, and polling trends are headed in that direction.

The poll that started off the hysteria was released March 20 by Deseret News/KSL, showing Clinton narrowly ahead of Trump by 38% to 36%.  But a more recent poll released May 16 in the same media outlet had Trump with a large lead, 13% over Clinton at 43% to 30%.

Over the last two weeks, a couple more polls have been released supposedly showing a tight race again.  The most recent, which came out Saturday from the Salt Lake Tribune, has Trump and Clinton tied at 35% in a three-way race with Gary Johnson, who has 13% of the respondents' support.  A Gravis Marketing poll conducted on June 1 puts Trump over Clinton by 36% to 29% in the head-to-head, and by 29% to 26% when Johnson (who gets 16%) is included.

The problem, as discussed many times previously in this column, is that demographics is destiny when it comes to polling accuracy.  What we see time and again when it comes to polling data involving Trump versus the liberals is that there is a strong liberal bias in the demographic composition of those being surveyed.  Rarely, if ever, is there bias that would favor Trump, and certainly not to the extent regularly seen against Trump.

Only the Gravis poll appears to provide some limited demographic data, and it suggests a problem.

A full 70% of those surveyed were Mormons, many of whom don't like Trump because of Mitt Romney's continued opposition, while just 55% of all adults in the state are Mormon.  Some may argue that 2012 exit polls suggested that 75% of Utah voters in the last presidential election were Mormon, so we should aim for this percentage in polling data this time around.

Wrong.

Exit polling data is notoriously incorrect, to such a degree and in such unpredictable ways that it is effectively useless for developing high-quality demographic profiles of the likely voting base in future elections four years down the road.  Polling demographics should always be matched to the potential voter base, and in this case, that would have involved just 55% Mormons.

Another problem involves the level of education of the polling respondents.  When asked about their highest level of education completed, 54% said bachelor's degree or higher.  But according to the U.S. Census Bureau, just 30.6% of those 25 years and older in the state have completed a bachelor's degree or more.  Only 2% of those surveyed did not complete high school, whereas the census data has this at about 9%.

The age demographics in the poll also appear to be way off.  Just 17% of respondents were 65 years and older, while 72% were between 30 and 64 years of age, for a ratio of 30-64:65+ of more than 4:1.  By comparison, the actual ratio between these two age groups is about 2:1.

Probably of greatest importance is the political orientation of those polled: 51% were Republicans, and 21% were Democrats.  Work by Brigham Young University's Center for the Study of Elections and Democracy suggests that while the percentage of Democrats in the survey matches with reality (20%), the percentage of Republicans is likely far too low (should be 65%, not 51%) if this 2012 data still holds.

So it looks as though the poll may be biased strongly in favor of Mormons, toward those with high levels of education, against older voters, and against Republicans in favor of Democrats.  Put all these together, and one might reasonably suggest that the survey base is stacked heavily against a fair assessment of Trump's real level of support in the state.

With no detailed demographic data for the other polls, we can't make a first-cut assessment of their reliability, but if we attempt to correct the Gravis poll for possible bias, Trump's lead looks to be in the range of 15%, which is roughly consistent with the May 16 Deseret News poll.

Unless another poll backs up the Salt Lake Tribune outlier suggesting a Clinton-Trump tie, Trump's support is probably stable with a healthy advantage over Clinton, and this is likely to hold until the general election.