Corrected national polling data shows Trump in the lead

The Guardian and SurveyUSA have released a new poll on the general election matchup among Donald Trump, Hillary Clinton, Gary Johnson, and Jill Stein.  The polling results claim that Trump trails Clinton by 3 percentage points, 39% to 36%, with Johnson at 6% and Stein at 4%.

Then we look at the demographic composition of the poll, and we find a massive liberal bias.

A full one third (33%) of respondents were Democrats, which is a 7% bias over the number of Republicans surveyed (26%), even though we know that on a national scale, the percentage of Democrats and Republicans is about equal.

Better yet in terms of showing gross bias, 53% of those surveyed apparently voted for Barack Obama in 2012, and just 30% voted for Mitt Romney.  This polling ratio of 2012 Obama:Romney voters is absurdly biased toward liberal voting preferences, given how the actual 2012 election results saw just a 51.1%-to-47.2% split in favor of Obama.

Thus, the poll is biased about 20% toward Obama over Romney supporters – otherwise roughly known as liberals versus conservatives.  Among respondents who did vote for Romney, Trump has dominance (74%) compared to Clinton (just 6%).  Thus, had this poll been weighted to reflect the actual 2012 voting patterns, Trump would emerge with a large lead over Clinton rather than a 3% deficit.

A whopping 47% of those surveyed apparently had a bachelor's degree, compared to U.S. Census Bureau data showing that just 32.5% of the population 25 years and older has this level of education.

Despite these anti-Trump biases, there is some good news in the polling data showing that the standard-model liberal narrative against the GOP presidential candidate is entirely false.

Clinton currently has just a 10% edge among all women (regardless of race) over Trump, which is less than the 11% advantage Obama had over Romney in 2012 and far less than the 13% deficit McCain had compared to Obama in 2008.  Repeated polling data for several months shows Trump clearly winning the white female vote over Clinton.

Trump is getting 17% of the black vote, far higher than the 4% McCain received and 6% that Romney was able to obtain.  Among Hispanics, 29% say they will choose Trump, more than voted for Romney (27%), and Trump is also on track to pick up more of the Asian vote (29%) than did Romney (26%).

Keep in mind that these are clearly underestimates of actual support from women and non-whites for Trump, as the poll is glaringly skewed away from Trump in its demographic composition.  But if we correct for these biases, it becomes clear that Trump is likely out in front of Clinton by at least several percentage points and may be on pace to a solid electoral showing among women and non-whites.

All this points to a potential Trump landslide come November.

The Guardian and SurveyUSA have released a new poll on the general election matchup among Donald Trump, Hillary Clinton, Gary Johnson, and Jill Stein.  The polling results claim that Trump trails Clinton by 3 percentage points, 39% to 36%, with Johnson at 6% and Stein at 4%.

Then we look at the demographic composition of the poll, and we find a massive liberal bias.

A full one third (33%) of respondents were Democrats, which is a 7% bias over the number of Republicans surveyed (26%), even though we know that on a national scale, the percentage of Democrats and Republicans is about equal.

Better yet in terms of showing gross bias, 53% of those surveyed apparently voted for Barack Obama in 2012, and just 30% voted for Mitt Romney.  This polling ratio of 2012 Obama:Romney voters is absurdly biased toward liberal voting preferences, given how the actual 2012 election results saw just a 51.1%-to-47.2% split in favor of Obama.

Thus, the poll is biased about 20% toward Obama over Romney supporters – otherwise roughly known as liberals versus conservatives.  Among respondents who did vote for Romney, Trump has dominance (74%) compared to Clinton (just 6%).  Thus, had this poll been weighted to reflect the actual 2012 voting patterns, Trump would emerge with a large lead over Clinton rather than a 3% deficit.

A whopping 47% of those surveyed apparently had a bachelor's degree, compared to U.S. Census Bureau data showing that just 32.5% of the population 25 years and older has this level of education.

Despite these anti-Trump biases, there is some good news in the polling data showing that the standard-model liberal narrative against the GOP presidential candidate is entirely false.

Clinton currently has just a 10% edge among all women (regardless of race) over Trump, which is less than the 11% advantage Obama had over Romney in 2012 and far less than the 13% deficit McCain had compared to Obama in 2008.  Repeated polling data for several months shows Trump clearly winning the white female vote over Clinton.

Trump is getting 17% of the black vote, far higher than the 4% McCain received and 6% that Romney was able to obtain.  Among Hispanics, 29% say they will choose Trump, more than voted for Romney (27%), and Trump is also on track to pick up more of the Asian vote (29%) than did Romney (26%).

Keep in mind that these are clearly underestimates of actual support from women and non-whites for Trump, as the poll is glaringly skewed away from Trump in its demographic composition.  But if we correct for these biases, it becomes clear that Trump is likely out in front of Clinton by at least several percentage points and may be on pace to a solid electoral showing among women and non-whites.

All this points to a potential Trump landslide come November.