Poll of Older, 25% African-American Electorate Finds Big GOP Leads

A new shocking poll from Monmouth University Polling might have been the biggest news of the morning had the NCAA not released their ruling on Todd Gurley’s eligibility. The poll found David Perdue leading Michelle Nunn by 8 points, 49-41, with Governor Nathan Deal leading Senator Jason Carter by 6 points, 48-42.

However, diving into the polls’ sample will likely make Democrats breathe easier and Republicans not smile quite as wide. Responding to questions on Twitter, pollster Patrick Murray shared the racial makeup of his poll’s 436 person sample (the smallest used so far this cycle in Georgia): 25% African-American, dramatically lower than 2010 turnout and the 29-30% that most other polls are now using.

Emory’s Dr. Alan Abramowitz had some critiques, too.

When comparing the poll’s age sample to Georgia’s 2010 turnout, the contrast is very apparent. Seniors, which predominantly cast Republican ballots in Georgia, make up an especially high percent of the electorate, while younger voters are under-sampled.

If it turns out that the electorate has taken a sharp turn towards an older crowd than the 2010 GOP wave election, with depressed African-American turnout, Democrats are definitely in trouble. However, I don’t think anyone, regardless of political affiliation in Georgia, would be willing to wager on that.


  1. And when a poll’s composition has methodological errors, you can’t do a simple unskewing, because those errors infect the other numbers. In 2010, 35% of voters under 50 where black, 58% white, 7% other. Assuming everyone turns out at the same rate as 2010 and applying those rates to the new registration totals, you’re looking at 36% black this year, 54% white, 10% other.

    In other words, it is almost impossible to believe that voters under 50 give Perdue a lead in a properly constructed LV sample. He would need to have 10% of the black vote, 50% of the other vote and 77% of the white vote among under 50’s to get to 50% with that group.

    • Chris,

      Explain something to me in layman’s terms, please. Go to the bottom of the press release in the Poll Demographics box. Where it has a column for Self Reported Party and Party of Registration… is that saying that 48% of the sample is still Republican on paper but 32% identify as Democrat in this poll?

        • We don’t have registration but I’m guessing they just looked at what primary respondents had voted in on their voterfile list. There’s obviously a huge Republican advantage on that these days, but you’ll notice that it did not translate into a ‘what party are you’ advantage on the question where they asked.

  2. I think one thing to keep in mind with many of the public polls that are released that have an electorate that looks more like 2002 or 2006 is that the last time the electorate actually looked like this in 2006, Sonny got something like 57% of the vote and Taylor was stuck at 38%.

    In these polls, even when Perdue/Deal lead, they aren’t getting anywhere close to Sonny’s number and Nunn/Carter are exceeding Taylor. So you take the “old Georgia”, which is relatively easy to poll – and Nunn/Carter are doing better than Democrats did back then with that electorate.

    And then you ask yourself, what are the net active voter registration changes since Nov 2006? Well, here they are (Nov 2006 to 9/1/2014):
    Black: +335,717 active registration
    White: +14,951
    Hispanic: +47,715
    Other/Unknown: +311,927.

    So gee, I wonder how the people that aren’t easily trapped by a “likely voter” screen will vote.

  3. Bobloblaw says:

    Rasmussen has a real good poll out today. Perdue 48, Nunn 48. I think it is legit because finally we are seeing undecideds choosing a candidate. I have been frustrated in all these races, not just GA, with high undecideds.

    BTW The Marquette University Poll, the most accurate poll in WI, has Walker up 7…yooohooo!!!! That is actually the most important race in the country, Senate or Gov. Sorry Georgians. If my choice was Walker win, Deal win or Walker win and Deal lose, ill take the latter any day.

  4. northside101 says:

    At this point—with basically five and a half days til the polls open next Tuesday—I’d be skeptical of any poll that shows Carter and Nunn less than 43 percent (especially Nunn). “43” is the percentage Roy Barnes got in 2010—and that was when the state was 62 percent white in voter registration (today it is 58%). Hard to imagine that at rock-bottom (and I mean “rock bottom”—both Carter and Nunn, win or lose, will exceed 43 percent next Tuesday), either would only be in the low 40s today.

    What would be interesting to see (but not holding my breath I’ll find out!) is how the candidates for governor and US Senate are polling in various legislative contests, say as some examples Buckhead’s House Dist 54 (old Ed Lindsey seat), the LaGrange-Newnan House Dist. 132 (seat of retiring Democrat Carl Von Epps), or the eastern Gwinnett House Dist. 105 (seat of incumbent Joyce Chandler). At the congressional level, CD 12 (Barrow-Allen). Results from polling in those districts would probably tell us more than a generic statewide poll.

    • Bobloblaw says:

      “”and that was when the state was 62 percent white in voter registration (today it is 58%). “”

      I think that number is population, not registration. Afterall GA is now nearly 10% latino but the electorate is probably 5% latino at most.

  5. jh says:

    With the sample issues aide, I thought the 5th question was pretty interesting. Only 49% of Perdue voters said Perdue was honest and trustworthy, compared to 71% of Nunn’s.

    Overall, 24% of their likely voters said Perdue was more honest and trustworthy, compared to 30% for Nunn. Given that posters have noted that the poll is skewed against her, the true margin is likely higher for Nunn.

  6. WeymanCWannamakerJr says:

    I don’t know all of the political insider stuff. Why is this one getting ink, well, pixels anyway?

Comments are closed.