A newly released PRRI poll covering Aug 10-16 has Hillary leading Donald Trump by +13 (48-35). 2% volunteered for Johnson.
These are obviously very good numbers. Still, I had never heard of PRRI, and they’re not rated by Nate, so I checked them out on their website. Their full name is Public Religion Research Institute. From what I can glean, they are a nonprofit, nonpartisan—but from what I can see also pretty progressive—research organization polling public opinion on issues like LGBT rights, climate change, and racial justice. They’ve been used by The Atlantic, the Brookings Institute, and other venues. From what I can see, this may be their first-ever presidential poll.
PRRI’s background might imply a pro-Hillary slant to their numbers—which might help them get to +13. On the other hand, the mostly R leaning Q poll had Hillary +10 today. Moreover, much of what I’m seeing in their internals mirrors what we’re seeing elsewhere:
Trump holds a five-point lead over Clinton among white voters (44% vs. 39%, respectively). Black and Hispanic voters, in contrast, strongly prefer Clinton over Trump. Two-thirds (67%) of Hispanic voters and 85% of black voters say they would vote for Clinton if the presidential election were held today, while fewer than one in five (18%) Hispanic voters and just 4% of black voters say they would vote for Trump.
I guess they find a smaller Trump edge—just 5%— among white voters than in some other polls, which might account for the extent of Hillary’s overall lead. And they find that Trump actually trails among males by one point, while he’s getting clobbered among women.
Male voters are evenly divided between the two candidates: 43% support Clinton while 42% support Trump. Female voters, in contrast, prefer Clinton over Trump by a nearly two to one ratio. A majority (54%) of female voters express a preference for Clinton, compared to 28% who support Trump.
Bottom line: any new poll by a new pollster, even if the results are favorable, needs to be treated with caution. Moreover, I wish I could find percentage breakdowns of race or ethnicity, gender, and party among those being polled.
Still, this pollster’s methodology does have impressive features. One is that the sample is quite large: 1630 registered voters. And then there’s this:
All PRRI studies include bilingual (English and Spanish) interviewing. Telephone studies are conducted by professional interviewers and include a high proportion of cell phone interviewing.
So we have live professional bilingual telephone interviews, including cellphones. Good stuff. It could be that this poll is reaching some voters who might not be reached by more restricted methods.
Anyway, take this poll for what you will, but the numbers are too good not to report. I’d like to see more polls from PRRI. I’d also be glad to hear from anyone who knows more about PRRI than I do.