 |
| © Saul Loeb/AFP/Getty Images US President Donald Trump speaks during an event announcing the Roadmap to Empower Veterans and End a National Tragedy of Suicide (PREVENTS) Task Force in the East Room of the White House in Washington, DC, June 17, 2020. (Photo by SAUL LOEB / AFP) (Photo by SAUL LOEB/AFP via Getty Images) |
Analysis by Harry Enten, CNN
Poll of the week: A new national Ipsos/Reuters poll finds that former Vice President Joe Biden leads President Donald Trump 48% to 35%.
While Biden has led Trump in almost every Ipsos poll this year, his advantage this week is the largest in 2020.
What's the point: Even though the national polls were accurate in 2016, one of the complaints I hear most often about the polls is that Trump's supporters are either lying or won't talk to pollsters. Polls like Ipsos get around that argument because they use machines (e.g. they're done online) to conduct the interviews. There's no reason to lie to a machine. If Trump was doing significantly better in these non-live interview polls, then these critics of the polls may have a point.
The evidence indicates these detractors are, at least in this moment, wrong. There's no sign of shy Trump voters. Trump doesn't do any better in polls without a live interviewer.
The average of national surveys (accounting for the fact that some pollsters survey more often) this week from pollsters who didn't have a live interviewer put Biden up over Trump 50% to 39% (10 points unrounded). That's a huge advantage and very similar to the latest live interview poll average that has Biden up 51% to 41%.
Moreover, it's pretty clear that Biden's edge is growing in the non-live interview polls. In all the polls taken after the protests against racism and police brutality started in late May, the average has Biden up 48% to 41% (8 points unrounded). When you look at the polls in May before the protests started, Biden was ahead by a 47% to 42% margin.
In other words, Biden's lead over Trump this past week was double what it was a month ago in the non-live interview polls.
Importantly, Biden's lead is growing because his percentage of the vote is growing, not just because Trump's share of the pie is falling. Biden is up about 3 points, and Trump is down a proportional 3 points.
There's no indication over the last month in the non-live interview polls that Trump supporters last month are merely saying that they are "undecided" right now.
Nor is there any reason to think that the surveys are shifting because Democrats are more likely to answer polls these days. Many of these non-live interview pollsters weight by party identification, so they're less susceptible to fewer Republicans responding than have over the long term (like when the race for president was closer).
I also checked out the state polling done since the protests began. Averaging the non-live interview polls just like I did for the national polls, I found on average that Biden's margin was 9 points bigger than Hillary Clinton's was. A 9-point shift in Biden's direction nationally from 2016 would give him an 11-point lead nationally, which is around where the national polls have the race.
Now, none of this guarantees the race will stay this way. Trump might gain in the polls in the upcoming months.
Furthermore, it's conceivable some state polls could be off for other reasons (e.g. improper weighting). The high quality live interview polls I (and CNN as a whole) most often cite have been made over to try and correct for the widespread state polling errors that occurred in 2016. They weight for education, for example. Notably, most polling pitfalls won't matter if the race for president continues to favor Biden by a wide margin.
For now, the idea that Trump has some large hidden support is just an idea without a lot of proof.
No matter how you slice it, Biden is well ahead of Trump in the presidential race.
Read more at CNN
Poll of the week: A new national Ipsos/Reuters poll finds that former Vice President Joe Biden leads President Donald Trump 48% to 35%.
While Biden has led Trump in almost every Ipsos poll this year, his advantage this week is the largest in 2020.
What's the point: Even though the national polls were accurate in 2016, one of the complaints I hear most often about the polls is that Trump's supporters are either lying or won't talk to pollsters. Polls like Ipsos get around that argument because they use machines (e.g. they're done online) to conduct the interviews. There's no reason to lie to a machine. If Trump was doing significantly better in these non-live interview polls, then these critics of the polls may have a point.
The evidence indicates these detractors are, at least in this moment, wrong. There's no sign of shy Trump voters. Trump doesn't do any better in polls without a live interviewer.
The average of national surveys (accounting for the fact that some pollsters survey more often) this week from pollsters who didn't have a live interviewer put Biden up over Trump 50% to 39% (10 points unrounded). That's a huge advantage and very similar to the latest live interview poll average that has Biden up 51% to 41%.
Moreover, it's pretty clear that Biden's edge is growing in the non-live interview polls. In all the polls taken after the protests against racism and police brutality started in late May, the average has Biden up 48% to 41% (8 points unrounded). When you look at the polls in May before the protests started, Biden was ahead by a 47% to 42% margin.
In other words, Biden's lead over Trump this past week was double what it was a month ago in the non-live interview polls.
Importantly, Biden's lead is growing because his percentage of the vote is growing, not just because Trump's share of the pie is falling. Biden is up about 3 points, and Trump is down a proportional 3 points.
There's no indication over the last month in the non-live interview polls that Trump supporters last month are merely saying that they are "undecided" right now.
Nor is there any reason to think that the surveys are shifting because Democrats are more likely to answer polls these days. Many of these non-live interview pollsters weight by party identification, so they're less susceptible to fewer Republicans responding than have over the long term (like when the race for president was closer).
I also checked out the state polling done since the protests began. Averaging the non-live interview polls just like I did for the national polls, I found on average that Biden's margin was 9 points bigger than Hillary Clinton's was. A 9-point shift in Biden's direction nationally from 2016 would give him an 11-point lead nationally, which is around where the national polls have the race.
Now, none of this guarantees the race will stay this way. Trump might gain in the polls in the upcoming months.
Furthermore, it's conceivable some state polls could be off for other reasons (e.g. improper weighting). The high quality live interview polls I (and CNN as a whole) most often cite have been made over to try and correct for the widespread state polling errors that occurred in 2016. They weight for education, for example. Notably, most polling pitfalls won't matter if the race for president continues to favor Biden by a wide margin.
For now, the idea that Trump has some large hidden support is just an idea without a lot of proof.
No matter how you slice it, Biden is well ahead of Trump in the presidential race.
Read more at CNN
