Correa Train Engineers Desperate December Fundraising Stunt: Picking Apart Lou’s New ‘I Lead Dunn 43% to 5%’ Poll

.

.

[UPDATED 12/29:  SEE BELOW]

“ALL ABOARD!”, the candidate pleaded. “NOW, PLEASE, ALL ABOARD!”

1. Polls, Transparency, and Spin

“Poll” is a funny term.  Do the first five people commenting here under their own names agree?

See what I just did there?  I just did a poll!

And that is why, when you read a poll, you want to pay attention to the presence or absence of words like “scientific,” “representative,” “random sample,” and references to the specific population being polled.  Because, if you don’t see them, that “poll” you’re looking at may have been cooked up to show a specific result.  You give me the choice of who to poll and how to phrase the questions and I could have shown you a result giving me an 75% lead over Tony Rackauckas in last year’s race for District Attorney.  (Note: neither of us finished with as much as 75% of the vote.)  (Note 2: See, I just spun you!  Beware!)

The best survey research firms will release the wording of their questions, describe what population was being sampled and how, and some will even show their internal analysis to help readers better understand the data.

And then there are polls done by candidates desperate to bring in money at the end of a fiscal quarter after which reports are due.  Their polls may be especially dodgy.  I’m not saying that they are, I’m saying that they may be.  It’s worth taking a look.

2. Lou’s Big New Stunning Announcement

Here’s what Lou Correa’s campaign put out last Friday — puzzling timing for a favorable bit of campaign news, unless you want to make sure that no one is going to test it over the next few days.

LOU CORREA LEADS BY HUGE MARGIN IN NEW POLL

Correa leads closest Democrat by nearly 30 points

SANTA ANA, CA—Today, the Lou Correa campaign for Congress released a poll memo showing Correa leading his nearest Democratic opponent, Bao Nguyen, by nearly 30 points. The poll, conducted by Fairbank, Maslin, Maullin, Metz & Associates, is fresh out of the field, and was conducted from December 13th through December 16th. The initial results show Correa with a decisive lead over all other candidates:

• 37% Lou Correa
• 5%   Joe Dunn
• 4%* Jordan Brandman
• 8%   Bao Nguyen
• 12% Adam Nick
• 1%    Other
• 33% Undecided

After messaging, Correa’s support grows to 43%, while the other candidates’ totals remain nearly unchanged. The memo showed Correa with a net favorability rating more than double that of any other candidate in the race. To read the polling memo, click here. “It’s remarkable for a candidate in an open seat to start out with such high name recognition and such a big lead, especially this early,” said John Shallman, chief strategist for Lou Correa’s campaign. “This result shows that voters have taken note of Lou’s tremendous record of accomplishment throughout his career for the communities of the 46th district.”

For more information, please visit www.LouCorrea.com. *Brandman announced he was not running for Congress during the final night of interviewing— the public news of his decision did not break until after the survey was completed.

###

Wow!  Why bother even having the race, huh?  Sure, it’s all about name recognition at this point, but still — what an advantage!  Guess you front-running donors had better jump onto the victory train!  Unless … unless something is not necessarily what it seems!

Maybe we should check that link to the polling memo, written by the firm of Fairbank, Maslin, Maullin, Metz & Associates:

http://www.shallmancommunications.com/wp-content/uploads/2015/12/Correa-memo1.pdf

Wow-wow!  Looks — graphic!  And in fact it took me a few minutes of poking around to realize that … some of it doesn’t exactly … make sense.  Panicked potential donors, you may want to stand down for a moment before you give into a high-pressure donation demand.

3. Legendary Failure and Its Possible Implications

Normally, I’d start my analysis from the top to the bottom, but there’s something sort of stark in that bottom chart that I should ask you about first: what does the red part of the bar graphs represent?

The chart is labeled “favorability of candidates.”  The scale at the bottom goes all of the way to 60%.

The legend labels the blue portion as “favorable.”  Just above that, it mentions that a manipulation was performed where positive information is given about each candidate.  Do you get the impression that blue is the “first pressing” — favorable ratings before the prompt — and the red is the “second pressing” — additional favorable responses after the prompt?

Did you get the impression that you’re supposed to add those figures together to get a total favorability rating:

Correa: 62%
Dunn: 29%
Brandman: 23%
Nguyen: 22%

The graph seems to invite you to do so.

It would also be possible in some charts that the first part of the bar would be “strongly favorable” and the second part of the bar would be “somewhat favorable.”  That would still invite you to, more or less, sum the total length of the bar to assess the strength of candidates.

But — I hate to bring this up — is it possible that the blue is “favorable” and the unlabeled red is “unfavorable”?  That would change your sense of the graph, wouldn’t it?  Now the first thing we can note is that Correa’s unfavorables — as you would see if they were labeled as such — are almost twice as high as those of any other candidate.

You could also note that by subtracting the red “unfavorable” number from the blue “favorable” number, you could derive a “net favorability” rating, which would look like this:

Correa: 32%
Dunn: 13%
Brandman: 7%
Nguyen: 10%

That starts to look a little more plausible, given Correa’s more recent involvement than Dunn’s (and … well … what happens to Brandman.)

But even that is not necessarily the number you would want to know before being stampeded to empty your pockets.  How do we account for the fact that both Dunn and Bao have more “room to grow” as they become better known?  I’m not saying that this is the right answer, but: we could look at projected maximum favorability — essentially an estimate of what those with neither a favorable nor unfavorable current attitude would do if they develop attitudes like those who have already settled on a judgment.  For this, we take the ratio of favorable to unfavorable responses and apply it to the currently neutral portion of the population — something that on the academic side of the field has sometimes been called “saturation”:

Correa:  47/62 x 38% uncommitted = additional 29% favorable = 66% potential max with saturation of population
Dunn: 21/29 x 71% uncommitted = additional 51% favorable = 72% potential max with saturation of population
Brandman: 15/23 x 77% uncommitted = additional 50% favorable = 65% potential max with saturation of population
Nguyen: 16/22 x 78% uncommitted = additional 57% favorable = 73% potential max with saturation of population

Now don’t make a snap decision based on those numbers!  Maybe the unlabeled red part of the bar chart isn’t unfavorability!

We’d need to know, for example, more about the sample itself — was it composed of people who were especially likely or unlikely to already have a favorable or negative attitude towards each candidate?  (For example, if the sample were all police officers, prison guards, building trade members, conservative Latino 100 members, and Correa staffers, they’d be expected to have their opinions set.)

More importantly, we’d need to know how likely it is that candidates will have the means to extend their current favorability ratio into the portion of the population that hasn’t expressed opinions — which means looking at fundraising (not just for campaigns, but also for independent expenditure committees.)  We won’t know those figures until, I believe, the end of January.

And, finally, we’d want to know how these numbers would change in relation to the composition of the field.  For example — where are Brandman’s votes going to go?  (We’ll address that in Part 2 of the series that I started last week, on which I haven’t had the time to work.)  More significantly, what’s going to happen when a prominent Republican enters the race?  (I’m told that this is likely — and that it will be someone who lives within the district (that is, not Adam Nick, not Lynn Schott)– but mum’s the word until it happens.  Someone could be trying to set me up!)

What else don’t we know here?  Let’s go back to the top of the press release — or, actually, what matters is the polling memo.

4. “Likely June 2016 Primary Election Voters”

This will eventually be exactly the population surveys should assess.  (Well, that presumes that it is a random sample of a representative population of such voters — a presumption I do not make here.)  The question is: should it be done now — and how did they determine who belongs in that population?

For example: did they ask them if they plan on voting?

That’s the easiest way to measure this.  And at some point it will probably be the right way to do it, when people are aware of the election and the candidates and there intentions become more set.  Now, it probably leads to a substantial underestimate of the voting pool.

(How could one tell?  Well, we could ask what proportion of the sample indicated that they plan to vote and compare it to the percentage that usually does vote in Presidential primary elections.  If they disqualified 75% of the voters, this obtained sample is probably weighted heavily towards people who are either “higher information” voters or those aligned with groups — unions, groups like SOAR, certain churches — that really turn out the vote.  My guess is that the polling firm won’t tell us — but maybe I’d be in for a pleasant surprise!)

What else they could do?  They could apply a model.

The easiest and most common model would be a “likely voter” model, which asks: “did you vote in previous elections?”  (Actually, they don’t have to ask.  Campaigns or their consultants have purchased the records of who did and who did not vote when.)  This sort of screen is highly open the manipulation, though.  If you know that your candidate is likely to be more heavily favored by people who voted previously — and the most simple effect of this is that it excludes people who weren’t able to vote in June 2012 — or at the latest November 2014 — then to make your candidate look good you narrow the criterion for inclusion in the survey.  If your candidate is sort of an insurgent (like Dunn) or someone depending on the youth vote (like Bao), you’d probably use a looser criterion.

How can we tell if manipulation has taken place?  A good first approximation is to know this: what percentage of those people who were successfully contacted qualified for the survey?  The lower the percentage, the tighter the screen.

Or, at least that’s true all other things being equal.  A pollster trying to come up with a different result has more tricks that they can use.  For example — what percentage of those people for whom contact was attempted at all qualified for the survey?  This gets at the trick of, say, calling people solely during the weekday and weekend evenings — like the Saturday night before Christmas, although something that brazen is best left to the professionals at the parties’ national committees — when you will tend to get voters who are older, more retired, and more conservative.  If your candidates’ base involves such people, you can thus arrange to get a disproportionate number of them into your obtained sample without leaving fingerprints.  (Well, not unless you open up your books and show everyone the recipe for the “secret sauce.”)

There’s more; you’re bored; we’re moving on.

5. Evident Across Most Demographic Groups

There’s an important and welcome sentence after the first graph:

“This support is evident across most of the District’s demographic groups including men, women, union members, non-union members, whites, non-whites, all age groups, and in … Anaheim and Santa Ana.”

Good!  That’s important to know.  Of course, it doesn’t rule out whether it means “of those members who picked up a landline phone at 11:00 a.m. on a Tuesday morning or 7:30 p.m. on a Saturday night,” but we’ve already covered that.  Now it’s time for me to take off my survey and experimental social psychology Ph.D. hat and put on my Law School hat — because now we’re talking about weasel words rather than statistical artistry.

“Evident” — if, say, Correa’s positives exceeded those of Dunn’s and Bao’s by 1% among union members, then — discounting negatives (see above) the support was “evident.”  If you think that “evident” means anything like “comparable,” you flunk law school.  It just means that there is some sign of it — it doesn’t tell you how much.

“Most” — now you see a list after that of six binomial (“yes or no”) demographic categories along with age, which might have five categories within it.  The lawyer asks: what’s the weakest claim that the author could actually be making here without actually lying?  Six out of eleven.  Now when you read this, you probably read that list of demographic groups to apply to the phrase “most of the District’s demographic groups” — as in, “these are the eleven in which this support is evident.”  Instead, it may modify only the phrase “the District’s demographic groups” — in which case that little word “most” cuts the assertion down from a statement about 11 groups to one about as few as six.

6. A Telling Footnote

Alight briefly on the footnote just below that, which starts with the word “Methodology.”  Three things to note:

1. It doesn’t rule out a damned thing that I speculate above.  And if it had wanted to, it could have.
2. It’s written like a press release itself, without bothering to strive for scientific impartiality.
3. That “margin of error of +/- 4.9% just refers to the sample size.  It presumes — without evidence — that none of my other concerns apply.

I’d love to talk more about this, but….

7. “Positive Paragraphs”

There are some pollsters that I’d presumptively trust to construct and convey fair “positive paragraphs” to their respondents.  Given how this press release — from the polling firm itself, mind you, not from the consultant — I’m not going to grant them that consideration.  I’d love to know exactly what these “positive paragraphs” included — but there I go again, wanting evidence of the “secret sauce”!

Look, I know that unless those paragraphs are really innocuous — and they could be (especially if the game was otherwise rigged) — they’re not going to release them and that they’ll note that the standards of the field don’t require it.  That’s fine.  All it means is that you don’t have to credit everything that they’re saying — because some political organizations do tend to stretch the boundaries of the truth and — based on the first non-PR link that I found about Correa’s consultant in a Google search — the law as well.  No need to “rush to commit to the winner” — because this press release isn’t that great for Correa once one discounts it for bluster, spin, and possible worse.

8.  Contact Information

I don’t usually try to contact sources who I believe are simply going to try to bamboozle me.  If anyone else — and I’ll do a good turn here to Liberal OC in exchange for them posting those very helpful letters to the Anaheim City Council last week (when they were not yet publicly available) — the contact information for the polling firm is in faint gray letters at the bottom of page one of the polling memo.  Go ahead and ask them whatever you want!  I’m not the least afraid of a proxy debate with the average political pollster.

9. So Where Do We Stand in This Race Now?

My intuition is that the red bars do — though inexplicably unlabeled (which lent much to the tone of this piece, because COME ON!) — refer to “unfavorability” ratings.  If so, then the net favorability figures I calculate above (because they di’int) reflect something reasonably close to current standing the of each candidate.  (I just wouldn’t put much weight onto that hunch — at least without knowing much of the other information noted above.)  But given that Lou Correa was a State Senator for eight years until a year ago and just ran for Supervisor earlier this year, I’m not surprised that lots of people have an opinion of him — nor that, absent an effective opposition campaign that would go to attitudes rather than brute turnout, as Janet Nguyen did for Andrew Do while edging Correa in the Supes race — he has good favorables in the part of the district that he won.

But this is a different ball game — and among Democrats Joe Dunn and Bao Nguyen both have call on important groups of likely voters.  The nice thing is that one or both of them will be able to point to this survey as their numbers inevitably rise and point to gathering momentum.  That’s why this would have been a really stupid move for a candidate who was focused primarily on what would happen in May or June — although it was probably a necessary move for one who is desperately trying to make it out of December with a war chest that isn’t testament to failure.  I’ll be very interested in seeing how many of his contributions — compared the those of Dunn and Bao — came in the immediate wake of this press release.  I presume that he’ll stay in the race regardless — unless Loretta re-enters it and gives him an out — but that is a rebuttable presumption.

[Full disclosure: My daughter is the Campaign Treasurer (a compliance rather than a political position) for Bao Nguyen, who would I think be a very welcome addition to Congress.  That doesn’t affect my writing here.  I’ve written favorably about Joe Dunn over the years and I would be very very happy to see a brilliant investigator like him in Congress.  I consider Correa’s needless and cynical blocking of better medical care for the poor and working- and middle-classes, and his repugnant record as a drug warrior, to have contributed greatly to needless death and suffering within his district and beyond.  But he’s done some good things too.  None of that affected the analysis that I did, for which I am uncompensated, once I decided to take a good cold hard look at his press release.]

UPDATE 12/29: No, Bao’s Not Desperate, Just Slightly Misguided

As I noted in a comment below from December 24:

Bao is now fundraising off of this poll, pitching himself on its basis as “the #1 Progressive Candidate” in the race.

Dude….

100+ hours later, Chumley — and I don’t mean to insult the one-man-band over there by not using his name; that choice is now permanently fixed in the OJ Style Guide and I have no remaining choice in the matter — posts a trackback to this post (which I’ll take apart paragraph by paragraph as fair use for criticism allows) that says this:

Remember that poll in the CD-46 race issued by Lou Correa’s campaign? The one that showed the former State Senator well ahead of all other challengers for Loretta Sanchez’s congressional seat (including the 2014 losing Republican candidate)? And remember how our friends at the OrangeJuice blog offered a long, hard and largely unreadable explanation of why it was “bull” and was really “a desperate fundraising stunt?”

Most of the people commenting on (and we’ll guess reading at all) Liberal OC are by now anonymous commenters directly aligned with Chumley and people associated with this blog, so we’ll guess that they do remember it.  But what I want to know is: what of the above is “unreadable”?  Critics (mostly of the anonymous variety, but also some doofuses) lob that brick at my writing repeatedly — but without specifics.  Well, this post is broken down into relatively small subsections!  So if Chumely can just indicate by number what part he was not able to comprehend — or, as the case may be if one takes him literally, even read — I will be happy to explain it to him.  Thus we will move closer to Peace on Earth and Goodwill Towards All.

Well, CD-46 candidate and Garden Grove Mayor Bao Nguyen is using his status in this poll to tout himself as the leading Progressive candidate in the race for an end-of-the-year fundraising pitch for his own campaign. Is it brilliant? is it stupid? Is it cheap? Is it thrifty? Depends on who you ask.

In case it’s unclear, I have no problem with anyone giving Bao money!  He’s a good guy.  To answer the questions: it is neither “brilliant” nor “stupid” — mostly because it is extremely cheap and thrifty!  More on that below.

We sent a note to Mayor Nguyen asking which poll this was, thinking Mayor Nguyen had commissioned his own. Sources confirmed Nguyen used Correa’s poll.

Oh. My. God.  You know, even if I had REALLY THOUGHT that Bao had “commissioned his own” poll and then neglected to, you know, REPORT ITS RESULTS I would like to think that I retain enough sense not to say so in print.  Really, Chumster?  You really thought that he had done this?  Dammit, I guess I’ll have to include Bao’s pitch here after all: double indented in bright blazing blue:

I’m excited to share some great news with you.  A recent field poll named me as the #1 Progressive candidate in the race.  The poll reaffirms my place as the true progressive candidate.

I’ve supported a progressive agenda my entire career, not because it’s expedient but because these are the issues I’m passionate about and have fought for:

–  An economy that supports workers not bankers. A minimum wage indexed to inflation. Rebuilding the nation’s crumbling infrastructure.

–  A woman’s right to chose. Opposing the attempts to defund Planned Parenthood. Equal pay for equal work. True equality.

–  Immigration reform, a path to citizenship, a fair shot at the American dream.

–  Fighting against climate change, and fighting for a clean energy future.

–  Creating a just system of education that prepares our students for the 21st century economy.

–  An end to all housing and employment discrimination against LGBT people, and full equality for the transgender community.

I care deeply about these issues, as I care deeply about the 46th District.  Please support my campaign during this end the year appeal.  The end of the next reporting period is December 31, and it is important to show as much support as possible.  I need your help.

To support my campaign and make a contribution, please [donate].

Nice pitch — and very nice agenda!  I see little in that agenda to distinguish Bao from Dunn, so I wouldn’t go along with his claim to be “the true” progressive candidate (we clearly have two of them), but still — a good substantive pitch.

I only have two other quibbles: neither Bao nor anyone else in California should use the term “field poll” because it invites readers to believe that it is a Field Poll — generally the gold standard of California pollsters — rather than a “pay for the results you want” piece of crap.  Second and related, let’s reserve the term “great news” for news that is, well, at least tepidly good, which this is not.  Until we see what was in those “positive paragraphs” for Bao and for Dunn, it’s “mildly nice” at best.  OK, back to Chum.

If the poll was a “desperate fundraising stunt” by Correa, doesn’t this logic apply to Nguyen’s campaign as well? To borrow a line from the guys with too much time on their hands at the OJ Blog, “See what I just did there? I just did a poll!” The Mayor hasn’t gotten back to us on which poll demonstrated his progressive leadership status.

No, it’s not desperate!  Do you know why?  BECAUSE IT DIDN’T COST BAO ANYTHING OF NOTE TO DO IT!  It’s just an aside, an attention-getting introduction to a substantive fundraising pitch — unlike Correa’s.

To analogize: “DESPERATE” is something like selling a kidney to raise money.  If we set what Correa did — spending presumably multi-thousand of dollars on a lousy and slanted poll — as the equivalent of “selling a kidney,” then I’m tempted to say that what Bao did is more like “selling blood plasma to raise a few bucks.”  But even that comparison is unfair to Bao.  It’s more like “selling a kidney” versus “shaking hands with someone you dislike because they still might give you money” sort of comparison.

Nguyen’s 8% showing in Correa’s poll suggests he, and not former State Senator Joe Dunn (who came in at 5%), is the best hope for progressives in CD-46 and tries to brand Correa []as Democratic yet not progressive. To have to admire Nguyen for using another candidate’s, allegedly flawed, poll for a “desperate fundraising stunt.” He saves money by not having to do his own. But in using Correa’s poll, Nguyen lends credibility to it.

That Chumley is trying to use this to pump up Bao as the “best hope of progressives” to the detriment of Dunn is deliciously telling: it shows that he (and whoever feeds him) is really scared of Dunn and not so scared of Bao in the general election.  (That seems to be the general consensus outside of Bao’s circle, which is admirable loyal and optimistic.  I think that one can make a case that Correa should be more scared of Bao, al though I don’t find that case compelling; what’s interesting here is what Team Correa seems to think when it urges on an attack.  And that makes me REALLY want to look at those positive paragraphs for each candidate that Correa paid for!)

As for Bao, yes, this does save him money and it gives him a fundraising hook.  But the problem with it is right there in Chumley’s uncharacteristically sensible final sentence: “in using Correa’s poll, Nguyen lends credibility to it.”  And that is why, while Bao didn’t earn derisive laughter and name-calling, he did earn the mild “come off it” of a younger generation: a single, unadorned, “Dude….”

We hope that Liberal OC readers will find the portions of this argument that don’t peel their eyeballs and fry their brains to be useful.  To simplify: to pay thousands for lousy poll is absurd.  To take advantage of an doeky and desperate competitor’s profligacy is merely slightly distasteful — simply because of the possibility that some lummox will use it to promote the value of the lousy poll.  Hope that helps.