Analysis: New U. S. Senate Poll by PPP


Public-Policy-Pollingby Elliott Stonecipher

Yet another poll on the Louisiana Senate race has been released, this one by Public Policy Polling*, called “PPP,” among those well-versed in such things.

Once again, as has more or less been the case for quite awhile, the poll has the heads-up race between Senator Mary Landrieu and Congressman Bill Cassidy within the poll’s margin of error:

Landrieu-Cassidy-Manes…  48% Cassidy,
…  45% Landrieu
…    7% not sure.

In the question which includes other candidates,

…  42% Landrieu
…  34% Cassidy
…  12% Rob Maness
…    2% someone else
…  10% undecided.

Mary-I'm a WinnerImportantly, Landrieu is, as expected, maxed-out in terms of name recognition, with only 7% unable to offer an opinion of her job performance.  The other two named candidates have what is referred to as “room to grow,” with 23% too unaware of Cassidy to offer such an opinion, and 50% are unable to do so for Maness.

Among those who do offer such opinions, the net margin of approval for all three is negative.

…  Landrieu’s approval is net -10% (42% approve and 52% disapprove)
…  Cassidy’s approval is net -4% (37% / 41%)
…  Maness’ approval is also net -4% (23% / 27%).

Again, the numbers may improve for the latter two, but Landrieu is a known quantity among virtually all voters.

The Maness Effect

PPP offers the following take on the Maness candidacy, in context: 

“Maness continues to be no real threat to Cassidy’s chances of finishing in the top two.”

Palin-mmmainWhile that may be true, there is no poll measurement of the role Maness may play in cutting away Cassidy’s vote in televised debates, media advertising, or any other way.

As I discussed in a piece earlier this week, to the degree that also-Republican Maness goes after Cassidy between now and November 4th, he expands the path for an outright win by Senator Landrieu on November 4th.

Notably, a bit of data-diving shows that Maness now gets 17% of the vote among Republicans, 6% among Democrats, and 14% among Other voters.

Landrieu African-American Vote Fully Reflected

Often in such polling, especially for this campaign, Senator Landrieu’s “vote” is understated due either to too few black voters in the poll sample, or a “trial heat vote” response far lower than history proves she will most likely get on election day.  In this poll, however, such is not the case, meaning Landrieu’s epic black support is probably fully baked-in the results.

Specifically, 31.2% of all Louisiana voters are African-American, compared to 30% of the poll sample.  In the 2012 re-election voting for President Obama, 31.0% of those casting ballots that day were black, and in the most recent, and more directly comparable, mid-term election in 2010, 26.8% of voters were.

In this poll, with a possible over-representation of African-American voters in the survey sample, Senator Landrieu gets 89% of the black votes in the poll’s two-candidate question, and 83% in the question with Rob Maness added.

Given historical mid-term voter turnout percentages by race – such as the 26.8% black vote in 2010 – the survey sample of 30% is likely a bit high, with 28% a more likely target for her campaign.  If the Senator gets 95% of those votes, the net to her is 26.6-points (95% x 28%) of the total vote.  In the heads-up question in the poll, she is netting almost exactly the same, 26.7% (89% x 30%).

Question_WhoOther Notes

Other results of note from the PPP analyst are shown on the opening page of the poll report, linked above.  Additionally notable to me are these points:

(1)  Party cross-over between Landrieu and Cassidy favors Cassidy.  In the hypothetical heads-up vote between the two:

…  Among Democrats, Landrieu gets 77%, and Cassidy gets 18%.
…  Among Republicans, Cassidy gets 82% and Landrieu 13%.
…  Among Other registrants, Cassidy polls 46%, Landrieu 37%, and 17% are unsure.

(2)  Also in the heads-up trial heat question responses, age results are notable.

…  As expected, Landrieu leads 56% to 36% among youngest voters (18-29).
…  Among the next two age groups the vote is even:  46% Landrieu / 45% Cassidy among those 30-45 years old, and 47% Cassidy / 46% Landrieu among those aged 46-65.
…  Then, among those Over 65, Cassidy leads Landrieu 57% to 37%.

Generally, voter turnout, especially in such a mid-term election, increases with age, thus particularly favoring Cassidy.

(3)  As noted by PPP, Cassidy leads Landrieu 68% to 24% among white voters, with 8% undecided.

a.  If the undecideds are added proportionately to the votes of each candidate, for example, adding about 5-points (68% x 8%) to Cassidy’s vote, the Congressman polls a bit above 73% of the white vote.  Such would not be at all unusual by a Republican candidate in a Louisiana statewide race against a Democrat.

b.  In the most recent mid-term elections in 2010, white voters were 70.3% of the total.

So, a key target of the Cassidy campaign results:  if the Republican gets 73% of the “white vote,” and white voters are 70% of the total vote (as was their turnout in 2010), he would win with 51.1% of the vote (73% x 70% = 51.1%).

Edwards at Baton Rouge Press ClubEWE Beats Jindal for Governor

No, that’s not a mistake.  When asked, “Who would you rather have as Governor:  Edwin Edwards (or) Bobby Jindal.”  The result is 47% Edwards, 43% Jindal and 10% undecided.  We might guess the candidates for the Republican vice-presidential nod in 2016 have that data-point from this poll if nothing else.

Elliott Stonecipher

Elliott Stonecipher’s reports and commentaries are written strictly in the public interest, with no compensation of any kind solicited or accepted.  Appropriate credit to Mr. Stonecipher in the sharing – unedited only, please – of his work is appreciated.


* PPP has a very good reputation for accuracy, and, arguably, turned in the best performance in the 2012 presidential race.  Though it is referred to by many commentators as a “Democrat leaning” polling firm, the truth is a bit more direct:  it is a Democrat polling firm, noticeable in such factoids as identities of a couple of its 2012-cycle clients, liberal blogger Daily Kos and labor union SEIU.

The poll was conducted September 25th through 28th, with a total sample of 1,141 likely voters.  Given that sample size, the poll results are accurate within + / – 2.9% when results are calculated among that total sample of respondents.  Subgroup results such as demographic breakdowns, have larger margins of error.