“Predicting Election Outcomes via NextStage’s TargetTrack” or “Why Dean Led, Kerry was Droll and Lieberman Foundered in 2004”


Note: This post was originally published as “An Evolution Technology Prediction Markets Case Study”

Background

NextStage was approached late in September of 2003 by the Lieberman Presidential Campaign camp. At that time, Senator Joe Lieberman (D-CT) was running in the democratic Presidential primaries against a very large field, although the main contenders at the time boiled down to Senator Lieberman, Senator John Kerry (D-MA) and Howard Dean (ex-governor, D-VT).

The question we were asked was whether or not NextStage’s Evolution Technology™ could help the Lieberman camp improve their standing. The first part of meeting this request involved determining if the Lieberman camp’s current efforts would be successful in the long run, and I won’t keep you in suspense; we determined they would not be successful and history has borne us out.

However, to demonstrate the first and second stage predictive capabilities NextStage’s TargetTrack™ was able to provide [[(as documented in Working with Prediction Markets via NextStage’s Evolution Technology and Reading Virtual Minds Volume 1: Science and History. Readers who’d like the entire NextStage 2004 campaign analysis should contact NextStage directly)]], this case study shows a competitive analysis of similar web pages from the Dean, Kerry and Lieberman websites. These pages evaluated included the Home page, some Issues pages and pages profiling (or “About”) the candidates. The goal was to use Evolution Technology™ to determine which audiences the candidates were targeting and capturing via their online media, and to determine likely outcomes based on the material being used.

The content used was from Monday, 28 Sept 2003. The information presented on the following pages constitutes a summary of Evolution Technology‘s assessment of these website pages and was our opinion. While not a case for why one candidate’s website or candidate his or herself might be outperforming another, the information contained herein was useful for that type of analysis. No suggestions for modifying any sites were contained herein.

Gender Capture

First, we looked at what the targeted demographics for the web content were. The first category, “Gender Capture” shown below, indicates the Percentages of overall gender that the combined web pages targeted (Male v. Female).

Gender Capture Comparison for Lieberman, Kerry and Dean on 28 Sept 03

Lieberman was reaching out to a largely (63%) Male audience. Dean was almost 50/50. None of the candidates were targeting the “Women” voters. What this indicated was that Dean was doing an excellent job of getting his message across to both men and women.

Age Group Capture

Next we evaluated the age groups that would best respond to the web site messages over all web pages combined (shown below).

Age Capture for Lieberman, Kerry and Dean on 28 Sept 03

This chart shows that the material in Lieberman’s site was reaching out to a primarily middle-aged demographic. Kerry had a larger portion of the younger age groups, while Dean had the largest demographic coverage, including a larger part of the “senior citizens” demographic. Ages under 15 are not included in these charts because of their insignificance in the election process.

Based on just these two charts alone, NextStage was able to determine that on 28 Sept 03 Lieberman was going to remain third man if this was the race.

Age Group Comprehension

Thus far we determined that Dean was communicating best of all the candidates’ websites we were analyzing.

However, appealing to the broadest audience means nothing if that audience can’t understand what you’re telling them. Answering that was done by determining which age groups could best comprehend or understand each candidate’s message as presented on their websites.

The chart below shows which age groups would most likely understand and respond to the candidates messages as of 28 Sept 03. Kerry and Lieberman’s content was understandable by a younger age group (lower education level) while Dean’s messages were designed for both a more mature and much broader audience. One way to look at this is that Lieberman’s and Kerry’s messages were stated much more simply than Dean’s.

Which Age Groups could best understand each candidate's messages?

Generally, Comprehension Capture numbers should be close to Age Capture numbers, indicating that the messages are more likely to be hitting their targeted audience. In mass marketing efforts such as politics, however, the goal is to be understood by the largest audience possible and this often means aiming for a lower education requirement on the part of your target. There is a problem in aiming for a lower education requirement, however. Too simple a message can come across as off-putting and patronizing.

Dean’s message, while requiring a higher degree of sophistication to understand, was very well targeted to the age groups he was capturing. This could partially explain his ability as an internet fund-raiser.

Message Strengths

Whether you’re a political candidate or a business promoting a product, unless you can get that message out in a way that makes sense for what you’re selling you’re not getting things done.

All politicians are good orators; it’s part of their job. But the goal is to be good at talking about what’s on the minds of people when you meet them. Lieberman had many strengths but they weren’t strengths people were interested in during the 2003-2004 primary season.

Message Strengths – 1 (below) shows that Lieberman’s strengths were his ability to show (Visual) people Comparisons (his ability as a debater and in confronting issues). Kerry was good at communicating who he was (Identity) and that he has a Process in Place to solve the problems he talks about. Even if he didn’t communicate the process specifically, he was giving the message that he knew what the process was. If you add this to the fact that his messages require the least education to understand (Age Group Comprehension above ), this adds up to a good market penetration.

Liebeman's strengths were his ability to show comparisons. Kerry's strength was that he had a process in place.

Dean was a successful fundraiser because he grabbed the broadest age demographic (Age Group Capture on page ) and had the most even gender capture (Gender Capture above). He was also communicating his ability to provide Order and Structure (as shown in Message Strengths – 2 below) way above the others. This strongly appealed to voters’ economic and national security issues at that time.

Dean's strength was his ability to communicate structure and order to a concerned, anxious nation

Perhaps the final blow to Lieberman in this three horse race was that both Dean and Kerry both were indicating they had a greater ability to move Toward and Upward, which translates as their ability to get the nation to a better place (literally, “higher ground”) as shown in the Message Strengths – 3 chart below.

Both Kerry and Dean were better able to communicate their ability to improve things

Summary

NextStage’s Evolution Technology demonstrated its ability to function as both a stage 1 and stage 2 Prediction Market during the 2003-2004 Democratic Presidential Primaries and again during the 2004 Presidential elections [[ and we continue to do so via our Politics blog]].

The first stage of a true prediction market is to determine what a likely outcome will be. The second stage, and usually the more important stage — is being able to explain why the outcome will be what is predicted and to suggest ways to alter undesired outcomes into desired outcomes.


Posted in , , , , ,

The Fickle Vote (Politics and Your Mind 22 Oct 2012)


This post continues the Politics and Your Mind thread. You can read previous entries at

NextStage Political Reader determines political outcomes based on how people are thinking when they're on monitored websitesWe’re using The NextStage Political Reader to determine how people will vote in the 2012 US Presidential election and so far we’ve learned two distinct things:

  • The voting population that we’re able to sample is incredibly fickle. We’re sure their political leanings are driven by something but what, exactly, we’re not sure.
  • The first chart, Say, usually tracks the latest national polls for the same days within the margin of error.Pretty good, we thinks, as we’re not actually asking anybody anything, just observing how visitors behave on a variety of websites, the majority of which are not political at all.

The Gender Divide

Yes, we know it's suppose to be 'Alfred E. Smith'. We wanted to see who was paying attention.We first noticed the Gender Divide in Politics and the Gender Mind 10-12 Oct 2012 (With a Touch of Debate BS and PersonaScoping at the End). That Gender Divide continues…sort of. What we noticed the day after both Obama and Romney spoke at the Annual Alfred E. Neuman Dinner was how the genders responded to Obama’s self-deprecating humor. Men tend to like self-deprecating humor, women not so much and this showed up in the numbers the following day:

Humor and the Political Mind
Humor and the Political Mind - Men will favor men who laugh at themselves, women won't

The above finding amused us as this site’s most read post is Humor.

A Quick Read of the Charts

  • Say – What people would tell you in conversation. These numbers should be closest to national polls.
  • Think – Which party do people think will do better. This may or may not be who people would vote for or who they want to win, only who they think will do better.
  • Hearts – This is the number that most closely responds to the outcome of a secret ballot if held today.
  • End – If people could be magically transported to their real voting venue and knew for a fact that their vote would actually make a difference, how would they vote?

The Fickle Voter

We tracked daily visitor political leanings 19-22 Oct 2012. We’re not sure what happened over the weekend to cause the amazing Saturday Republican Surge nor the gradual climb of the Democrats (we don’t follow the news much unless we’re paid to). What we do know is that the Republican Surge corresponds to a surge in monitored visitor traffic (the % next to the days on the x-axis of the charts is the % of voting population we monitored each day). Saturday’s traffic volume was 3-9x what we monitored on the other days. If the increased traffic is the sole cause of the increased Republicanism (and recognizing we’re not necessarily monitoring the same visitors each day), this could be bad news for the Democrats.

Who Would They Say They’d Vote Right Now?
Would voters tell you they prefer Obama or Romney if you asked them right now?
But What Do They Think?
And What Are Their Hearts Telling Them?
Republican Support Seems to be Collapsing...Maybe
And In the End, Will Hearts or Minds Prevail?

A Final Word

The time for either party to strike is when voter hearts and minds are within a few points of each other. Hearts and minds with close numbers indicate voters’ higher and lower brains, conscious and non-conscious, are closely aligned. People act without thinking and without later remorse for their actions when conscious and non-conscious are aligned.

Come 6 Nov 2012, whoever synchronizes the most hearts and minds wins the election.


Politics and Your Mind 17 Oct 2012 (The Day After the Obama-Romney Hofstra Town Hall Style Debate)


NextStage Political Reader determines political outcomes based on how people are thinking when they're on monitored websitesThis post continues a thread started in Politics and Your Mind 9 Oct 2012 (Introducing the NextStage Political Reader) and continued in Politics and the Gender Mind 10-12 Oct 2012 (With a Touch of Debate BS and PersonaScoping at the End). We’re using The NextStage Political Reader to determine how people will vote in the 2012 US Presidential election.

The same caveats apply here as previously: NextStage is not monitoring all sites in the US nor was the entire voting population online during the time this data was collected.

A Quick Read of the Charts

  • Say – What people would tell you in conversation. These numbers should be closest to national polls.
  • Think – Which party do people think will do better. This may or may not be who people would vote for or who they want to win, only who they think will do better.
  • Hearts – This is the number that most closely responds to the outcome of a secret ballot if held today.
  • End – If people could be magically transported to their real voting venue and knew for a fact that their vote would actually make a difference, how would they vote?

The Gender Divide

The twelve or so hours since the debate started have been busy internet-wise. NextStage tracked over 10 million people (3.43% of the voter population) (over 7½ million women and close to 3 million men) and there is no doubt that President Obama carried the evening as far as women voters go. What President Obama also managed to do was make lots of men reconsider their voting decision. They didn’t go totally democratic and they did move away from republican by a substantial margin.

How Would They Say They’d Vote Right Now?
How did the Hofstra Debate Affect How Voters Say They'll Vote?

The real shift is in male consciousness, though. In their own minds, men believe President Obama won the debate.

But What Do They Think?

Heartwise, men still believe President Obama won the debate by a large margin although here we see that men still have a hankering for Governor Romney.

And What Are Their Hearts Telling Them?

But in the end and all things considered, the Democrats won the night.

And In the End, Will Hearts or Minds Prevail?

A Final Word

The Republicans may also believe President Obama and the Democrats carried the night. As I write this, it is slightly after 10amET and so far the only political calls to our offices have been by Republican operatives, and those to let us know how wrong President Obama was and what he wasn’t telling the debate audience. Normally there’s a 50/50 mix and the calls are issue oriented.


Politics and the Gender Mind 10-12 Oct 2012 (With a Touch of Debate BS and PersonaScoping at the End)


Following up with Politics and Your Mind 9 Oct 2012 (Introducing the NextStage Political Reader), we now supply a gender-based voter breakdown…

A Quick Read of the Charts

  • Say – What people would tell you in conversation. These numbers should be closest to national polls.
  • Think – Which party do people think will do better. This may or may not who people would vote for or who they want to win, only who they think will do better.
  • Hearts – This is the number that most closely responds to the outcome of a secret ballot if held today.
  • End – If people could be magically transported to their real voting venue and knew for a fact that their vote would actually make a difference, how would they vote?

What I’ll share today is probably why day by day evaluations are meaningless from an eventual outcome perspective while incredibly fascinating from a socio-political perspective.

What Changed Their Minds So Rapidly

The image below indicates an incredible shift in male voting patterns in a 24hr period.

Male voting change from 10-11 Oct 2012

Likewise, the following image indicates a similar shift among female voters in the same 24hr period.

Female voting change from 10-11 Oct 2012

What happened in 24hrs to cause that kind of shift? There may have been something from a news standpoint. NextStageologists spent a few hours looking at it and realized it was a shift in “who” was analyzed.

Or more correctly, how many of who was analyzed.

It’s All in the Numbers

The 10 Oct 2012 analysis involved 3,658,758 men (1.21% of all voters) and 931,852 females (0.31% of all voters). The 11 Oct 2012 analysis involved 124,916 men (0.04% of all voters) and 1,114,259 female (0.37% of all voters).

Of perhaps greater interest is that the 10 Oct 2012 results are from sites covering 19-75 year olds while the 11 Oct 2012 results sampled sites catering to 54-75 year olds. Evidently younger men and mature women favor the Democrats, older men and younger women favor the Republicans.

First, the 11 Oct 2012 sample is about 1/4 the size of the 10 Oct 2012 sample. Also, the 10 Oct 2012 sample was 3:1 male:female weighted, the 11 Oct 2012 sample was almost 10:1 female:male weighted.

And After the Biden-Ryan Debate…

Today, 12 Oct 2012, we did another run of the NextStage Political Reader and it seems all bets are off and the Democrats need to do some serious rethinking if they plan to win the election. Analyzing 1,708,074 visitors (0.57% of the voter population) we get the results below.

How Would They Say They’d Vote Right Now?
Men and Women Voting after the Biden-Ryan Debate

Again, the devil’s in the details. The above is primarily for a 12 hour stretch post debate, hence overnight traffic. Evidently Republican males don’t sleep much.

About that BS thing

There was much discussion of how well Romney did in the first Obama-Romney debate. Some of that might have been due to how much BS (BlueSky, a measure of how believable a person was) the audience detected in both candidates. Pretty much all their values were equal with the exception of BlueSky where the Democrats came off as less believable than the Republicans:

How Much BS Was There?
How Much BS Was There in the Debates?

Was Anybody Really Talking to the People?

The last question answered in this post is one of personal curiosity. I’ve always wondered (and several readers have contacted me with similar questions) if politicians talk more to each other or if they really talk to the voters.

Fortunately NextStage’s PersonaScope and SampleMatch tools can answer such questions via “{C,B/e,M}”s. {C,B/e,M}s are a shorthand for how people think, behave and what motivates them. Ever meet someone you just didn’t get along with? Chances are your {C,B/e,M} conflicted with the other person’s. Ever meet someone you hit it off with immediately? Chances are the two of you had highly complimentary {C,B/e,M}s. You can find out more about {C,B/e,M}s at Looking for Love? Now You Can Find All the Right Places! (On the Evolution of Tools)

Obama was communicating using a K13 {C,B/e,M} while Romney, Biden and Ryan all used a V13 {C,B/e,M}. While neither K13 nor V13 is a dominant communication form in the United States, V13 is closer to how the majority of US citizens talk to each other than K13 by 22%. This could be another reason for the Romney Resurgence.

Bolton Bluntness

John Bolton for President. For the run-of-the-mill conservative, such a concept is interesting — almost startling. What is also interesting, but not startling, are the results of the NSE analysis of the interview with Bolton. These results are very much in line with the public persona of Bolton as he has presented himself for some years, now.

The fact that he has a very low BS factor would suggest he’d had a chance to prepare responses for the interview ahead of time, but having seen quite a few interviews with him through the years, I doubt it was necessary. Bolton has always displayed the trained, practiced diplomat’s control of language to choose the precise word to convey the precise meaning for the precise situation. The biggest difference is he is always very fluid with the language, never slowing down the sentence to be sure of selecting each word properly. Some of the less precise formulations of the interviewer suggest this is a transcription of a real, off-the-cuff back and forth interview, and I would be surprised to learn otherwise.

While his might not inspire media venom like a Palin candidacy would, Bolton does not have a reputation for being a smooth politician. His appointment as UN ambassador was a recess appointment, a foreshadowing of the bloody battle he would face against livid liberals. And his reputation among some coworkers at the State Department was that of an “abrasive” character. (Some might find that a sterling recommendation, actually, in that environment).

The abrasiveness among fellow diplomats and his performance as UN ambassador, not to mention his public pronouncements since then, show his focus is clearly on the foreign policy challenges the country faces, not on going along to get along. And this fits well with the “vision” measures where his vision for the country has a much stronger reading than his vision for himself. And given his feisty, sometimes contrarian positions, even while in office, it’s also no surprise he gets relatively low ratings in the “man of the people” category, compared to some of the more experienced elected officials who may run.

Bolton is a realist, and I feel confident he does not expect a sudden movement to draft him into the candidacy. And Nolte, the interviewer, is an astute analyst and also sees little prospect of such an event. I suspect any dancing around that shows up in the Compatibility Gauge has to do with Bolton’s well-known bluntness. A blunt, unpredictable respondent is always a terrifying prospect in an interview.

Aside from what the Political Analyzer Tool shows, I suspect Bolton’s coming out publicly has more to do with wanting to put pressure on the field to keep foreign policy relevant in the 2012 elections, and perhaps a bit of wistful longing to work with a successful candidate to finish what he did not get a chance to finish while Bush was in office.