MyDD Poll: The Structure of the Voter Universe, Part 2

From the diaries--Chris

This is a long post and a very, very important one. Last time I introduced the psychographics borne out of this poll, a mini-history of the method, and a tree-tops view of the themes and groups in action. This time, we're going to dive deeply into them and go through the demographics and detail relative to specific questions. Next time, I'm going to provide my view of what political communications strategy would look like when based on our findings. Then, the final primary post of this analysis will be my strategic recommendations, including action items. After that, I'll be cleaning up some loose ends, including an experiment designed to test the impact of the OBL tape and probably some detailed pretest/posttest Bush approval analysis.

My goal with this post is for us to more fully understand the groups, on a personal level, so they become familiar to us. I hope you'll be able to recognize real people, friends, neighbors, others, in them, even though we couldn't flesh the groups out more thoroughly this time because of budget and interview time limitations. But, it was the first time out on the MyDD Poll and repeated research will solve that problem nicely, thank you very much. Join me on the flip...

A couple of housekeeping points. First, I usually write up a table, discussing findings, and then show those tabular data following. We're talking about seven separate groups in this post. If I used the usual approach, I'd end up showing the same tables seven times. The post would be a mile longer than it already is. To avoid that, I'm going to show the tabular data for demographics and questions one time, at the end of the post. So that we don't lose any reference or context in the text of the post, I'm adopting a reporting convention I'll insert directly into the text analysis. It's like this: "yadda, yadda (US% / Group%)." (Remember, the margin of error of the data is +/- 3.1 points overall.) Thus, it shakes like this: "It is not surprising the Red Core is strongly dominated by Republicans (30%/62%) but it's important to note it also includes some Independents/Others (38%/28%) and even a few Democrats (33%/10%)." So, in the first notation, we see I'm showing the 30% US Republican number compared to the 62% Red Core Republican number, a difference of 32 points and clearly significant given a 3.1 point margin of error of the data. All data notations in the post work like that.

Second, I'm very, very appreciative of your comments regarding each post. Whatever you want to say or ask, do it. I'll do my level best to respond to questions, issues of each commenter. Given my regular work sched, and it's a bitch of a sched these days, it may take me some time before I do. There are comments to the last post I haven't read yet and others I haven't responded to yet. I ask simply for your patience. I'll be circling back and responding as soon as I possibly can. Thanks, in advance.

Now, let's get to the real work at hand. First up, I need to introduce three more themes found in the analysis. That's a total of five found, including the Bush Meme and fear. As I said last time, the themes analysis is interpretive, so what follows is my view of the meaning of each, given the limited data we have at hand. (I'll provide my interpretive thinking in parentheses.) Further, I'll hit each one quick because you'll see how the themes actually work in the groups analysis below. And the groups analysis is today's analytical target.

The first new theme is religious tolerance. It is a pattern found through two key demographics asked in the survey: religious affiliation (e.g., faith/denomination) and religious orientation (fundamentalist/evangelical, mainstream and liberal). This is a theme I've found repeatedly over many years and when I had the opportunity to measure religiosity more fully. So, while we didn't have a lot of detailed questions that might input more breadth and depth into this theme in this project, finding it here jibes with previous findings and that's why I interpret it in the fuller sense of religious tolerance.

Next up is what I call `trust in external solutions' theme. This is a pattern consisting of the rating of the federal government in protecting the US since 9/11 (`they're protecting me'), the perception of job availability (`the `market' will provide me a good job, decent wages') and whether one has followed Homeland Security's recommendations for disaster preparation (`I'll do my part to help them cope.').

The last one is what I call the Bubba theme. It's a pattern combining the belief right-wing extremists are protecting society (as opposed to pushing too far on policy) and confidence the feds will respond in a timely and effective manner in a disaster, natural or man-made. So there's an aspect of buying in to social conservative political messaging in addition to a (highly naïve) view the feds have disasters covered. FYI, my notes written on the original statistical printout include these: clueless, sheep, falafel (Bill O'Reilly) audience, bubba.

Now, we're locked and loaded. I'm going to go through the groups in-depth, profiling statistically significant findings for each one, discussing: 1) demographic characteristics, 2) psychology and 3) survey questions. You might be surprised at how much we already know, even with limited data. Here `tis:

The Blue Core comprises 13% of the US electorate. It is dominated by Democrats (33%/58%) but it's important to understand Independents/Others (38%/36%) are as likely to be found here as they are nationally. Tellingly, few Republicans are found in this group (29%/6%). Blue Core is more likely to be found in the Rockies (7%/10%) and the West Coast (15%/22%). Further, they tend to live in urbanized areas (27%/30%)  Minority voters are a substantial force (20%/29%) as are those who self-identify as religiously liberal (38%/65%). There is no significant difference between US and Blue Core numbers regarding gender or income.

Psychologically, Blue Core voters do not feel personally threatened, as indicated by the Fear theme (41%/79%) and the Terrorism Threat Index (39% `Elevated'/59%).Their  negative reaction to the Bush Meme almost reaches unity, or 100% (41%/94%). They are religiously tolerant (27%/41%) and react very negatively to the Bubba theme (42%/70%). This group is average on the external trust theme, no significant difference from the sample norm.

In terms of opinion, they say the US is off on the wrong track (48%/71%), they disapprove of Bush job performance both pretest (50%/87%) and posttest (48%/84%), they recognize serious problems in the job market (25% `jobs rare'/33%), are critical of fed terrorism protection efforts (18%/31%), are neutral on personal safety/security ratings (34%/47%), are not worried about a terrorist attack soon (36%/65%), or that OBL hasn't been captured (48%/66%), are not confident in fed disaster response (45%/59%), opposed the Iraq invasion big time (48%/91%), oppose staying in Iraq for years big time (44%/74%), oppose Murtha's plan (34%/53%), are tuned in to the NSA gig (49% `heard a lot'/55%), believe the government should not have the right to bypass courts on citizen spying (41%/72%), don't trust the NSA to monitor only citizen threats (32%/61%), believe Congress should investigate Dubya (50%/90%), support impeachment if he broke the law (50%/83%) and believe right-wing extremists are pushing public policy too far to the right (42%/53%).

Whew. There they are, based on what we've got right now. The big picture I see with this group is not surprising, to me anyway. They're blue voters straight down the line. Aware, tuned in. Here's a thought, though: Even without a coherent, consistent Blue Meme being communicated out there in the zeitgeist, they've found their own spot and it's a strong one. As I see it, the center of internal gravity of this group is: an Anti-Bush Meme. In other words, they oppose Dubs on every single dimension tested in this study. They are the Blue Core and they oppose the guy. Oppose. Hello Beltway insiders, especially consultants! What a helluva statement this is winging your way from your strongest supporters: "We don't need you to figure things out; we've done it ourselves. Unlike you, we stand in direct opposition. Thanks very fucking much. Not."

Next up is the Urban Blue, comprising 17% of the US electorate. This group is dominated by Democrats, (33%/65%), even more so than Blue Core. It also includes a good chunk of Independents/Others (38%/27%, or a quarter). And Republicans are scarce here, too (29%/8%). A big demographic difference is gender. Women are the strong majority here (53%/64%), which was not so with Blue Core. While Urban Blue are more likely to be found on the West Coast (15%/22%), they are also found throughout the country, particularly in, what a shocker, urban areas (27%/31%). They are more likely minority (20%/24%) and religiously liberal (38%/55%). There is no significant difference from the sample norm relative to this group on income. They're average.

Psychologically, they share a lot with the Blue Core but there are some really huge, and strategically crucial, differences. For example, Urban Blues feel threatened and fearful. Big Time. The Blue Core was opposite of this. Urban Blues, in fact, feel the most threatened of all groups, as shown in the Threat Index (27% High threat/78%). The Fear theme confirms the point (40% fearful/82%). Additionally, these guys are strongly negative toward the Bush Meme (41%/65%), are religiously tolerant (27%/39%) and have the strongest negative reaction to the Bubba theme (42%/86%). Yet they are also positive toward the external trust theme (37%/43%), unlike Blue Core, who was average on it. (The external trust theme data don't surprise me, especially given the urban aspect of this group. They need to trust in external solutions, particularly terrorism protection. It's a logically consistent position for them to take.)

Opinion-wise, they are even stronger on some measures than Blue Core, compared to sample norms. They're most likely to say the US is on the wrong track (48%/86%), to disapprove of Dubs' job performance pretest (50%/94%) and posttest (48%/93%), to say jobs are not available at all (9%/19%) or are rare (25%/32%), are very negative in their rating of fed terrorism protection efforts (18%/43%), feel less safe/secure since 9/11 (24%/48%), are most worried the US will be attacked (34%/69%) and that OBL hasn't been captured (36%/73%), second most likely to have followed DHS disaster prep recommendations (34%/48%), are literally the least confident in fed disaster response (45%/75%), opposed the 2003 Iraq invasion (48%/89%, oppose staying there for years (44%/76%), oppose Murtha's plan (34%/50%), have heard a lot about the NSA gig (49%/65%), are most likely to oppose bypassing the courts on citizen spying (41%/82%), are least likely to trust NSA (32%/71%), strongly believe Congress should investigate Bush lies (50%/90%) and, if he lied, they'll boot his sorry ass in heartbeat (50%/89%). And they really get the point on extremists pushing policy too far right (42%/72%). Luv `em.

So Urban Blue is a real piece `o work group. They are savv-eee politically. And strapped in the shuttle cabin on the launch pad, figuratively speaking, ready to go. They are very, very worried, fearful, and they do trust in external solutions yet they don't believe this government is anywhere close to handling protection or recovery. Or the economy/jobs. Or foreign policy/Iraq. Or domestic policy/NSA. Or, basically, anything. Shit, hit the launch button, now. And, finally, these guys' perception of political reality is also miles and miles ahead of the Beltway insiders, even more so than Blue Core, for chrissakes. These are not `can we all get along' people. Or `Together America Can Do Better' people, the 2006 communications umbrella proposed by the Beltway insiders.

So I imagine you've been waiting for the next group since you saw the last post, outlining the groups. Yep, let's do the Progressives, shall we? And spend some time looking at ourselves, too. I think you'll find this really, really interesting. I hope so. Probably a few surprises, as Progressives, as a group, are more than online progressives.

Let's start with a May Be a surprise: Half (50%) of Progressives are not Democrats. One in six (16%) are Republicans. A third (34%) are Independent/Other. Are they more likely to be Dem than the sample norm? Sure (33%/50%). But this ain't your `normal' blue group, either.

Try this on for size: while they're found all across the country, of course, they're also more likely to be found in the South (30%/35%). A shocker? Not to me, not one whit. I'm a Southern native. I get that right off the bat. It's socially rigid in the South, so either you're going to support the status quo or oppose it. Polarization, folks. A natural outgrowth given the social situation, in my view. And experience. Even more so with a massively dominant, elite class, (Southern Dems or Southern Reps make no difference here) hell-bent to maintain their `superiority' and power over the vast majority of everyone else. Simple. Painful still, but simple.

Here's additional demography: Progressives are more likely to be found in urban (27%/37%) and suburban (35%/38%) areas, are the second-youngest group (14% 35 years of age or less/22%) but also are more likely to be in the oldest age category (20% 65+ years of age/32%). Surprised? Shouldn't be, progressivism is as old as the hills in America. Of course some of our people have been progressives for decades and decades. So, at one level, we're just now catching up to them. More demography: the single most likely group to include minorities (20%/40%), they are also religiously liberal (38%/56%) and the least affluent (26% less than $25K per year/49%). And finally, with an average split on gender, this group is not more likely to be male nor female.

Onward to the psychology. Progressives feel moderately threatened, as indicated by the Threat Index (27% High threat/36%; 39% Medium threat/53%) and the Fear theme (40% fearful/53%). Their negative reaction to the Bush Meme is second strongest of all groups (41%/80%), they're religiously tolerant (27%/40%), they're neutral on the Bubba theme (20%/53%) and are the second least-trusting of external solutions (46%/77%). In short, they're worried, concerned, fearful, despise the Bush Meme (and I would hypothesize it's the power structure behind the Meme they actually despise), are fine with different types of people, religious and Bubbas, and don't give a rat's ass this government says it will `fix this' or `take care of' that.

Opinion-wise, Progressives have yet to gel as a group on several issues. Some, however, are no sweat for them, including the US is off on the wrong track (48%/70%), disapproval of Dubs' job performance pretest (50%/83%) and posttest (48%/79%), a more skeptical view of jobs and wages (37% jobs available, not easy/40%; 25% jobs rare/32%), worry the US will be attacked (34%/47%), that OBL hasn't been captured (36%/54%), they have not followed DHS disaster prep recommendations (61%/77%), they opposed the 2003 Iraq invasion (48%/85%), oppose staying there (44%/82%), oppose Murtha's plan (34%/52%), believe Congress should investigate Dubs' lies (50%/80%) and support booting his butt if he broke the law (50%/89%).

Here's where they're more murky on issues and that murkiness is a key finding I'll discuss in the Applications post. Specifically, Progressives are neutral in their ratings of fed protection efforts since 9/11 (31%/55%), they're neutral in ratings of personal safety/security (34%/50%) but do show a slight lean to `less safe' (12%/20%) and are neutral in confidence in the feds' future disaster response (30%/35%).

The NSA spy gig is an entire area where they're murky, compared to other groups. For example, over half have heard nothing at all about the issue (16%/54%). They don't believe NSA should have the right to bypass the courts and spy on citizens (41%/51%), but they're also more likely to be neutral on that issue (15%/21%). And while Progressives are less likely to trust NSA to spy only on citizens who are potential national security threats (32%/41%), they're also more likely to be neutral (30%/43%) on that one, too. Finally, and most tellingly in my book, Progressives are more likely to be neutral on the `global' measure of whether right-wing extremists are pushing public policy too far to the right (26%/31%).

So, our first, albeit limited, view of Progressives as a national group indicates they are open-minded, vehemently anti-Bush but also in need of more pointed communications, information. Again, more about this later, but in my view these data indicate a crucial strategic issue for the online community: we've got important work to do in our own house. We need to reach, and bring up to speed, our own people if we expect to successfully mobilize them. And these are people who may not be connected online at all, or as thoroughly as we are. We'll find out next time.

Next, we focus on the Heartland Blue. Now, we'll start moving out of partisan party registration and see how political opinions and attitudes determine orientation. For example, this group, 12% of the electorate, is dominated by Independents/Other parties (38%/72%), with only a fifth (33%/21%) Democrats and a smattering of Republicans (29%/8%).

Men (47%/55%) comprise the majority of this group and they are most likely to be found in the Northeast (22%/27%) and Midwest (26%/34%). You might think they're more likely to live in rural areas, but they're not. The urban/suburban/rural split on them reflects the sample norm. They do tend to be older (27% 65+ years of age/34%) and they are heavily Anglo (80%/85%). Half consider themselves religiously mainstream (39%/49%) and they are split between middle income (27% $25K-$49.9K/32%) and affluent (25% $75K+/32%).

In terms of psychology, this is where the rubber meets the road for these guys. They have a very interesting, and completely logical, twist when it comes to threats and fear. They score Medium (39%/57%) and High (27%/30%) on the Terrorism Threat Index, so they perceive national threats above the sample norm. However, those national-level concerns do not translate into personal safety/security worries. They're neutral on the Fear theme (19%/28%). So, for them, threats appear to be more external and less internal. Fascinating. A tantalizingly sophisticated view, I think. I suspect (hypothesize) they likely went bonkers over the Dubai Ports gig because, for them, it was an obvious boneheaded compromise of national security. (Reminiscent of Cleavon Little holding a gun to his own head in Blazing Saddles. "Don't move, or I'll shoot!") It wasn't that a ports disaster would affect them personally in the heartland, in short. It was plainly stupid regarding national security, generally. Logical as the day is long, in my view. Be interesting to test that hypothesis down the road.

Heartland Blue is also neutral on the Bush Meme (16%/29%). Unlike other blue groups, this one is not, by definition, anti-Bush. They're open to him. Hello? They're also neutral on religious tolerance (42%/52%), negative on the Bubba theme (42%/66%) and very negative on trust in external solutions (46%/82%).

So let's re-cap this part of psychology before we go to attitudes. Heartland Blue is, as Bob Seeger sings, `like a rock'. They aren't ruffled by much. Concerned about national security, yeah, but not crazy concerned about it. Not worried Osama's hiding in their flower garden, either. They'll listen to Bush and the Vast Right Wing Conspiracy (VRWC). And the Dems. Pretty steady on religious tolerance, not loading one way or the other. Their hot spots in this research are: a) negative on external solutions, that is, believing the `guvmint' will take care of national security, jobs, and such and b) negative on the Bubba theme, that is, buying into the belief right-wingers are protecting society and the `guvmint' will take care of natural disasters, no sweat.

In terms of attitudes and opinions, Heartland Blue voters say the US is off on the wrong track (48%/68%) and disapprove of Bush both pretest (50%/81%) and posttest (48%/71%) --  (note the 10-point drop between pretest and posttest, folks. Very important, shows they can swing on key issues). They're more likely to say good jobs at decent wages are rare (25%/33%), they're fair/neutral (31%/48%) in the rating of fed protection efforts since 9/11, neutral (34%/47%) on the personal safety rating since 9/11, neutral (30%/44%) to not worried (18%/22%) about terror attacks, or Osama not being captured (neutral: 26%/37%; not worried: 15%/18%). They haven't followed Homeland Security recommendations for attack prep (61%/90%) and are neutral (30%/36%) to not confident (45%/59%) on whether the fed response to a disaster will be timely and effective.

The `blue' in Heartland Blue really jumps out on Iraq. They were big time opponents of the 2003 invasion (48%/69%), oppose staying there with 100K troops (44%/58%) and are big time supporters of Murtha's plan (57%/74%).

Their `blueness' is there, but more subdued, on the NSA gig and investigation/impeachment. They've heard a lot about it (49%/54%), believe the feds should not have the right to bypass courts and spy on citizens (41%/53%) and are neutral on trust in the NSA to monitor only real national security threats (32%/41%). A majority (50%/56%) believe Congress should investigate Dubs and are neutral (12%/17%) to supportive (50%/53%) of impeaching him if he broke the law.

Finally, and I believe most tellingly regarding attitudes, they react strongly to right-wingers pushing public policy too far to the right (42%/61%). In my view, this is the spot, the position, that taps directly into their `blueness', like a T1 Internet connection, and should be a top priority messaging point for any and all Dem campaigns in 2006. The Heartland Blues swing, folks. I believe you can swing them back home, back to blue with this basic startup position: the right wing has gone too far. More about this next time.

It is a great, Grand Canyon-sized understatement to say I have severe doubts whether the DLC/DCCC basic position, `Together America Can Do Better', brings these folks home this year. For Heartland Blue, that line is more like what they hear at the nearest comedy club, I think. (Heartland Blue to Rahm: "Hey, thanks for the yuks and don't forget that syntax problem, buddy".) You get the drift.

Now let's tee up Heartland Red, 8% of the electorate. This is another psychographic group where  Independents/Others (38%/63%) dominate, but with about equal numbers of Reps (30%/20%) and  Dems (33%/17%). Further, women (53%/67%) dominate this group, interestingly, and they are most likely to be found in the Midwest (26%/37%). Unlike Heartland Blue, almost half of these reds live in rural areas (36%/48%). Further, they're younger (14%/26%). It's also interesting to note minorities are more likely to be in this group (20%/26%), a phenomenon we see in the Southwest: conservative-leaning minorities. These reds are split on religious orientation, more likely to self-identify as fundamentalist/evangelical (23%/26%) and liberal (38%/50%). Finally on demographics, Heartland Red really struggle economically: 26% less than $25K/32%; 27% $25K to $49.9K/32%).

If you were looking only at demographics, you'd think these guys were blue. You'd be wrong.

Psychologically, while they score almost identically to Heartland Blue on the Threat Index (39% medium/58%; 27% high/30%), it is the Fear theme where they appear to differ (40% fearful/61%). More on this below . Further, through the Bush Meme, we know they love, love, love Dubs (43% positive/83%). They are fairly tolerant religiously (27% positive/33%) and strongly positive toward the Bubba theme (38%/76%). Like Heartland Blue, though, they are also non-trusting in external solutions (46% negative/71%).

So the basic psychological profiles show some real similarities and differences with Heartland Blue. Similar on the Threat Index, religious tolerance and distrust of the `guvmint', as shown in the external solutions theme. They're very, very different on personal fear/safety/security, their view of Dubs and the Bush Meme and the Bubba theme.

Attitudinally, the red-ness of this group really comes out. It is the first group to say the US is headed in the right direction, albeit by a slim majority (37%/51%) and to actually approve of Dubs' job performance pretest (43%/77%) and posttest (48%/72%). (Note the five point slippage pre- and post-. They swing on Dubs, like Heartland Blue, just less so.)

However, they are pessimistic on jobs at decent wages (9% unavailable/13%; 25% rare/37%) but they rate fed protection efforts since 9/11 positively (51%/70%).

On the components/questions comprising the Fear theme, we see they're split and gravitate to the middle on the personal safety measure (12% less safe/16%; 34% neutral/38%; 18% safer/28%). It's terror attacks and Osama where they light up (34% worried about attacks/47%; 36% worried about OBL/58%). So, similar to their Heartland Blue brethren, threats and fear are less internally- than externally-based.

Like Heartland Blue, they haven't prepared for an attack (61%/84%) but, unlike Heartland Blue, they are neutral on whether the feds will provide a timely and effective disaster response (30%/46%). Heartland Blue is negative on this issue.

Like Heartland Blue, Iraq is a crucial defining issue for this group. They supported the 2003 invasion (46%/79%), support staying there with 100K troops (50%/69%) and are big time supporters of Murtha's plan (57%/80%).

Both heartland groups are big supporters of Murtha's plan. I'm screaming `hello!' at the top of my lungs at the Beltway boobs. Silence. What a shocker.

These reds are not keyed in on the NSA gig. They've heard a little (34%/51%) or nothing at all (16%/29%). However, their red orientation jumps out on whether the feds should have the right to spy on citizens (44%/62%) even though they're lackadaisical on whether they trust the NSA (30% neutral/45%; 38% positive/43%).

They again show red colors on investigation and impeachment. But their red is muted. Even though it's a major red issue, less than half say Congress should not investigate whether Dubs broke the law or lied (42%/48%). (Red-to-the-bones groups view it differently, as we'll see.) Opposition to impeachment if Dubs broke the law barely cracks half in this group (38%/54%). That's important. There's no groundswell among these guys to keep Dubs if he broke the law.

And finally, they are split on whether right-wing extremists are pushing public policy too far right or are protecting society. They lean protecting society (32%/39%), as one would expect, but over 60% of this group can't bring themselves to sign on to that. Thus, about a third (26%/30%) are neutral and another third (42%/31%) say extremists are pushing too far right. A split on this, a crucial, positioning issue. Among red groups, I'll take that any day.

Next, we start treading into real red territory with Trusting Boomers, 6% of the electorate. This group may be more accurately portrayed as social conservatives, but we didn't have the broad base of measures in this initial MyDD Poll to tease that out. Let's see how they shake out as is and, next time, we'll get additional dimensions of psychology that focus more on this issue.

Demographically, this is the first group with a plurality of Republicans (30%/38%) but it also includes chunks of Independents/Others (38%/28%) and Democrats (33%/34%). The group is dominated by women (53%/70%) and, while found throughout the country, it is the Northeast (22%/26%) and West Coast (15%/20%) where they're most likely found. They are suburban (35%/40%) and rural (36%/44%) residents. They tend to be middle age (46 to 55 years of age 23%/28%) and older (65+ years of age 27%/31%). Anglos (80%/86%) dominate the group and they are religiously mainstream (39%/46%) and, importantly, fundamentalist (23%/33%). Economically, they are fairly well off ($50K to $74.9K 22%/31%), especially considering their suburban and rural residence.

The Terrorism Threat Index shows this group feels threatened in a big way (second only to Urban Blue), with almost half scoring medium (39%/47%) on the scale and half scoring high on it (27%/51%). The Fear theme reiterates the point: fearful 40%/92%. The Bush Meme is a big time hit with these guys (positive 43%/95%), more than any other group (only Heartland Red even came close to them, at 83%). And they are religiously intolerant (31%/49%), a crucial point particularly as it relates to social conservatism. They are neutral on the Bubba theme (20%/53%) and they are very trusting in external solutions (37%/69%), more than any other group. In short, the big moving threads with these guys are fear and threats, the Bush meme/support, religious intolerance and trust in others/external solutions.

Given that, in terms of attitudes and opinions, there's not going to be any big surprises. In fact, get ready for those who buy Dubs' `party line' big time. For example, Trusting Boomers say the US is headed in the right direction (37%/45%) and they approve of Dubs' job performance pretest (43%/66%) and posttest (45%/70%). (Note the pretest/posttest increase in approval: +4 points.) To them the jobs issue is not a problem. They're more likely to say jobs are available, not easy to find (37%/41%) or widely available (18%/25%). They're more likely to rate the feds positively in protecting America since 9/11 (good 38%/52%). As expected, they're neutral (34%/47%) to negative (13%/17%) on rating their personal safety since 9/11. They're highly worried about terror attacks in the near future (34%/77%) and that OBL hasn't been captured (41%/66%). They're much more likely to have followed DHS attack prep guidelines (34%/64%) but they're less than confident (neutral 30%/42%) in a timely and effective response from the feds in the event of a disaster.

Like others, the Iraq issue is a big one for this group, clearly showing their cohesiveness. They were highly supportive of the invasion (46%/74%), are supportive of staying there (50%/84%) and of Murtha's plan (57%/83%). (Another hello! to the Beltway.)

Almost two-thirds (49%/64%) have heard a lot about the NSA gig. They're big supporters of the government having the right to spy on citizens (44%/72%) and they trust NSA to spy only on potential national security threats (38%/46%).
They buy the party line on investigation and impeachment. Congress should not investigate Dubs (42%/67%) and they oppose impeachment and removal if he did break the law (38%/66%). And finally, surprisingly, they're neutral (26%/33%) on whether right-wing extremists are protecting the moral and social fabric of society. Hmm. Food for thought, no doubt.

In the end, the plethora of ways in which this group buys the Gooper line is a hallmark of them. And it's important to note the Gooper strategy plays directly into the underlying dynamics operative in this group: fear, religious intolerance, trust in the `guvmint' and the overall Bush meme. In my view, they are the exemplar, the touchstone, Oberleutnant Rove uses to build the communications/framing strategy of this administration. Look here to understand what the Goopers are trying to do, folks.

Now comes the final group: Red Core. The Big Kahuna. The Big Slice, at 33% of the electorate. This is the group that is dominated by Republicans (30%/62%), with some Independents/Others (38%/28%) and a few Dems (33%/10%). They're male (47%/57%) and they're white (Anglo 80%/90%). They're found all over the country, but most pronounced in to take a guess?...South (30%/35%). They don't pop up regarding residence type; they live in urban, suburban and rural areas pretty much at the sample norms. They don't pop regarding age groups, again, they're found pretty much at sample norms. And I'm sure you will be shocked (not) to find out they're more likely to be fundamentalists (23%/37%) or mainstream (39%/47%) religiously. And I'm sure you'll also be shocked (not) to find out they're affluent ($50K to $74.9K 22%/26%; $75K+ 25%/29%).

Psychologically, Red Core voters score low, the lowest of all groups, on the Terrorism Threat Index (34%/74%) and they are not fearful (41%/67%). They buy the Bush Meme in a big way (43%/70%), although not as strongly as Heartland Red and Trusting Boomers. Which is a very interesting finding in and of itself. They're very positive toward the Bubba theme (38%/57%). They're neutral (42%/46%) to negative (31%/40%), on religious tolerance, showing this is a key connection between them and Trusting Boomers. And they show another key connection with them by also being positively-oriented toward trust in external solutions (37%/54%).

In terms of attitudes and opinions, we'll also find no surprises here, with this group. The US is headed in the right direction (37%/75%) and they approve of Dubs' job performance pretest (43%/89%) and posttest (45%/91%). Jobs are no problem in Dubs' economy; they're either available, not easy to find (37%/40%) or widely available (18%/35%). The feds are doing great in protecting the US since 9/11 (positive 52%/86%) and they certainly feel safer since then (42%/76%). They're not worried about terrorism attacks in the near future (36%/56%) or that OBL hasn't been captured (38%/65%). And they're more likely to have taken DHS disaster prep guidelines to heart (34%/42%), in addition to being confident in the feds' timely and effective response should a disaster occur (25%/53%).

Iraq is no problem for these guys. Almost all supported the invasion (46%/90%), they support staying there (50%/83%) and Murtha's plan (57%/63%). (I'm tired of saying hello! to the Beltway on this. Jeebus, jiminy crickets, people.)

The NSA spy gig is no problem. They've heard a lot about it (49%/54%). NSA spying on citizens is just fine with them (44%/79%), thank you very much, and they, of course, trust NSA to spy only on the wannabe OBLs hiding in the flower garden (38%/71%).

No question what to do on investigation and impeachment. Don't investigate (42%/80%) and don't impeach (oppose 38%/68%). Simple, you ignorant liberals. And, oh by the way, those right-wingers are out there are protecting the moral and social fabric of society (32%/56%), in case you were wondering about that. As Falafel Bill O'Reilly says: "Now, shut up!"

Yeah, right. The certainty of the ignorant and unscrupulous is a dangerous thing, in my view. It's why we must understand voters from a more nuanced, more sophisticated and, most importantly, more human perspective than run-of-the-mill  demography, as our Beltway Bums are wont to do. Humans are so much more than their characteristics, like age, gender, income and such. Not only is the standard demographic approach an insult to people, voters, it's also incredibly stupid. It leads to political failure because our campaigns deal with voters from the incomplete, mechanical view of demography. Republicans use psychographics, data mining techniques and they construct, and most importantly, effectively use sophisticated targeting databases.

This we must do. We, the netroots. We must change how we learn, what we learn and how we wage campaigns in order to survive. We must. This is what Jerome, Markos and I talked extensively about when they were here. In my view, we have no choice and we have few allies in this effort, especially inside The Gate.

So, get ready. Next time I'll take the knowledge we've gained through the MyDD Poll and apply it to communications strategy. We'll see how it measures up to `Together America Can Do Better', or whatever embarrassing crap the Beltway is offering these days. In the meantime, keep the faith.

Below are the crosstab tables for your review. Hope you've enjoyed the post and found it useful. Hasta.


                                                                                       Heart-   Heart-
                                             Blue    Urban   Pro-            land      land     Trusting      Red
                                  US     Core     Blue     gressives     Blue     Red     Boomers     Core

Democrat                  33%    58%     65%      50%           21%      17%     34%           10%
Republican               29          6          8         16                 8          20        38              62    
Indep/Other              38        36        27         34               71         63         28              28

Male                         47        47        36         47               55          32         30             57
Female                     53        53        64         53               45          68         70             43

Northeast                 22        21        22         22               27          16          26             21
South                       30        22        26         35               24          30          32             36
Midwest                  26        24        23         23               34          37          18              26
Rockies                     7        10          7           8                 5            4            3               7
West Coast              15        22        22         12               10          13          20             10

Urban                      27        30        31         37                29          20          16             24
Suburban                 35        34        33         38               33          31           40            37
Rural                       36        32        33         23                37          48          44             38

35 yrs. or less          14          9        11         22                12          26            6             15
36 to 45 yrs.            15         16       14           9                16          14           14            17
46 to 55 yrs.            23         24       27         24                23          19           28            21
56 to 65 yrs.            21         25       26         13                15          14           20            21
65+ yrs.                   27         26       22         32                34          27           31            26

Anglo                      80         71        76         60                85          74           86           90
Minority                  20         29        24         40               15           26           14           10

Fund/Evan.              23          5         14         10                17          26           33           37
Mainstream              39        30        31         34                49          24           46           47
Liberal                     38         65        55         56                34          50           21           16

Less than $25K        26        25         28         49                20         32            25           17
$25K to $49.9K       27        28         23         27                32         33            21           28
$50K to $74.9K       22        24         23         12                16         18            31           26
$75K +                    25         23         26         12                32         17            23           29

Terrorism Threat Index and Themes

Terrorism Threat Index:

                                                         High               Medium            Low

US                                                     27%                 39%                 34%

Blue Core                                            5                     59                    37
Urban Blue                                        78                     21                      1
Progressives                                      36                     53                     11
Heartland Blue                                  30                     57                     13
Heartland Red                                   30                     58                     12
Trusting Boomers                             51                     47                       2
Red Core                                             1                     25                     74

Fear theme:
                                                         Fearful            Neutral               Not fearful

US                                                    40%                 19%                    41%

Blue Core                                           9                    12                       79
Urban Blue                                       82                    14                         4
Progressives                                     53                    33                       14
Heartland Blue                                 29                    28                       43
Heartland Red                                  61                    22                       17
Trusting Boomers                            92                      8                         0
Red Core                                          15                    18                       67

Bush Meme:
                                                         Positive            Neutral               Negative

US                                                    43%                 16%                    41%

Blue Core                                           1                      5                       94
Urban Blue                                       13                    22                       65
Progressives                                       5                    15                       80
Heartland Blue                                 41                    29                       30
Heartland Red                                  83                      7                       10
Trusting Boomers                            95                      3                         2
Red Core                                          70                    16                       14

Religious Tolerance theme:

                                                        Positive            Neutral               Negative

US                                                    27%                 42%                    31%

Blue Core                                         41                     44                      15
Urban Blue                                       39                     32                      29
Progressives                                     40                     39                      21
Heartland Blue                                 22                     52                      26
Heartland Red                                  33                     41                      26
Trusting Boomers                            17                     34                      49
Red Core                                          13                     47                      40

Bubba theme:

                                                        Positive            Neutral               Negative

US                                                    38%                 20%                    42%

Blue Core                                         18                     12                      70
Urban Blue                                         5                       9                      86
Progressives                                     66                     19                      15
Heartland Blue                                 12                     12                      76
Heartland Red                                  76                     14                      10
Trusting Boomers                            15                      53                      32
Red Core                                          57                      29                      14

Trust in External Solutions theme:

                                                        Trust                 Neutral               Distrust

US                                                    37%                  17%                    46%

Blue Core                                         36                     16                       47
Urban Blue                                       43                     16                       41
Progressives                                     12                     11                       77
Heartland Blue                                   8                     10                       82
Heartland Red                                  10                     19                       71
Trusting Boomers                            69                       9                       22
Red Core                                          54                     23                       23

Survey Questions:

Direction US is headed:

                                                         Right Direction       Wrong Track       Not Sure

US                                                     37%                          48%                    15%

Blue Core                                          13                             71                        16
Urban Blue                                          6                             86                          8
Progressives                                      11                             70                         19
Heartland Blue                                  10                             68                         22
Heartland Red                                   51                             32                         17
Trusting Boomers                             45                             31                          24
Red Core                         &n

Tags: Crashing the Gate, extended analysis, MyDD Poll, poll, psychographic analysis, strategy, tactics (all tags)



What a post!
Thank you. This is an enormous amount of work. In order to help me, and perhaps others, understand a little bit better, please define the following:
What is Bubba?
What is Mainstream religious v. Liberal religious?
Also, I seem to be confused on the term "Murtha plan." To me, that means out of Iraq ASAP. But support for the plan appears strong with people who support the war and who want plenty of troops in Iraq. What am I missing?
Finally, ignoring the final two groups, who appear to be happily ignorant, what are the three positions you would take if you were a Democratic candidate in 2006 that would resonate with the broadest number of remaining voters?
by grayslady 2006-03-14 08:29AM | 0 recs
Re: What a post!

See below ( 2/979/2#2), my post below was meant as a reply to your entry.

by bedobe 2006-03-14 11:32AM | 0 recs
Re: What a post!


Thanks very much for your questions. It appears you've gotten the info on the Murtha plan, so I'll not deal with that.

The Bubba theme is a pattern of responses within the data that link together two very important questions: 1) are extremists pushing public policy too far to the right or are they protecting society?, and 2) how confident are voters in a timely and effective disaster response from the feds, if needed?

Like I said, the analysis indicates there is a link, something underlying, these two questions together. A theme connecting the two questions. That theme is what I call the Bubba theme. What I interpret that to mean is this: voters where this theme is prominent, that is positive, are naively buying the Dubs line that 'everything's OK', the feds will be there in a disaster and extremists are protecting society. They believe it's all OK, and naively trust Dubs and the gang on this stuff. They're not observing reality nor thinking much about such stuff. They're approaching things like good old boy Bubbas, in my view. That's why I call it the Bubba theme.

I'll be the first to say we need more information on this. But we didn't have the opportunity in this first poll because of very limited budget and interview time. So, in the future, I hope to gather much more information to flesh this out in much greater detail.

The Mainstream vs. Liberal issue you cite is based on a question in demographics. We first asked people what faith/denomination they are and, of those that cited a faith/denomination, we asked the follow-up of "do you consider yourself fundamentalist, evangelical, mainstream or liberal in your beliefs?" Thus, the data you see is those who self-identify, and I emphasize self-identify, as either fundamentalist/evangelical, mainstream or liberal in their religious beliefs.

On your final question, what three positions would I take, I'm going to ask you to keep checking the MyDD site because my very next post deals with your question exclusively. I don't want to let the cat out of the bag yet! :) In sweeping terms and also in fairly extensive detail. Please watch for that post, it's sort of the climax of my whole analytical effort. It will be labeled MyDD Poll: Strategic Applications, or something very similar. I expect to have that post up late next week, after I've taken care of some regular biz that's very pressing right now. Please keep checking and I'll be extremely interested in your thoughts! Thanks, XOX.

by Sun Tzu 2006-03-15 06:53AM | 0 recs
Re: What a post!

Thanks for the explanation. Again, a TERRIFIC job, and I look forward to the third segment.

by grayslady 2006-03-15 08:42AM | 0 recs
Re: MyDD Poll: The Structure of the Voter Universe

This is quite amazing to me:

I seem to be confused on the term "Murtha plan." To me, that means out of Iraq ASAP. But support for the plan appears strong with people who support the war and who want plenty of troops in Iraq. What am I missing?

Because it illustrates how republican'ts truly dominate the public discourse. Now, I presume that you follow current events, particularly politics, more closely than the average person; yet, the republican noise machine is so effective and dominant that they manage to throw enough dust up to confuse even the closest of observers. The fact is that the Murtha plandoes not mean get out of Iraq ASAP; rather, the plan seeks to "redeploy troops just over the horizon," so to continue to project influence over the region. The key to Murtha's plan is "redeploy[ing] U.S. troops consistent with the safety of U.S. forces." Basically, that gives the Pentagon plenty of leeway to do so in a measured fashion. However, the plan sets a clear mandate to begin moving towards the removing forces out of Iraq.

by bedobe 2006-03-14 11:30AM | 0 recs
Re: MyDD Poll: The Structure of the Voter Universe

Thanks. You're correct, of course. I had forgotten the tag line of "keeping troops on the horizon."

by grayslady 2006-03-14 11:55AM | 0 recs
I agree

Obviously, the blue-group folks who opposed the Murtha plan wanted U.S. troops out completely. Still, I do wonder how exactly the question was framed in the poll. I can't find the info anywhere right now.

by brainwave 2006-03-14 08:24PM | 0 recs
Re: I agree


Here's the exact question wording:

"A new troop deployment plan was proposed recently. The plan is to withdraw US troops from Iraq but to keep them close by in neighboring countries like Kuwait to be sent back in if they are needed to maintain civil order. Do you, strongly support, support, oppose or strongly oppose this new troop deployment plan for Iraq?"

by Sun Tzu 2006-03-15 06:30AM | 0 recs
Dude! What Are These Groups???

You've told us oodles about them, but still haven't explained how someone qualifies as a member.  What makes the difference between "Blue Core," "Urban" and "Progressive"?

I'm not talking about the statistical differences you've explained above.  How do you get the groups in the first place?

by Paul Rosenberg 2006-03-14 12:38PM | 0 recs
Multivariate statistics

At least that's my guess. One of a bunch of different algorithms (cluster analysis, multi-dimensional scaling, factor analysis and whatnot) that identify events or individuals or whatever who/that pattern together in a sea of data points. You then need to interpret these clusters/patterns, which is what Sun Tzu does when he calls them stuff like "heartland blue" and "red core" and such. Iow. there is no criterion for what it means to be "blue core" or "red core", etc. - these are just make-shift labels for patterns in the data.

by brainwave 2006-03-14 08:04PM | 0 recs
Factor analysis

I think you are right.

Check these out:

cluster analysis

Cluster analysis is a class of statistical techniques that can be applied to data that exhibits "natural" groupings. Cluster analysis sorts through the raw data and groups them into clusters. A cluster is a group of relatively homogeneous cases or observations. Objects in a cluster are similar to each other. They are also dissimilar to objects outside the cluster, particularly objects in other clusters.

data clustering

Data clustering is a common technique for statistical data analysis . . . Clustering is the classification of similar objects into different groups, or more precisely, the partitioning of a data set into subsets (clusters), so that the data in each subset (ideally) share some common trait - often proximity according to some defined distance measure.

factor analysis

Factor analysis is a statistical technique that originated in psychometrics. It is used in . . . applied sciences that deal with large quantities of data. The objective is to explain most of the variability among a number of observable random variables in terms of a smaller number of unobservable random variables called factors.

by tgeraghty 2006-03-14 09:03PM | 0 recs
Re: Factor analysis

Brainwave and tgeraghty,

You two are spot on. Multivariate, factor and cluster. Discriminant analysis also is applicable when the method is extended to build targeting databases. The latter is something Jerome and I are talking about now, in fact.

I'm not sure what Rosenberg expects. If he expects me to lay out all the ideas, operations and how-to's about psychographics, it ain't gonna happen for a variety of reasons, but most importantly because you really have to know how these stat routines work, which switches to throw and how to interpret the ouptput. There's a whole lot of different options and methods within these routines that must be understood in order to know which to use and when. Some of this is situational, for example. That's why there are professionals like me that do this stuff.

And if he expects me to lay out the precise steps and switches I throw, nope, that's proprietary and I've spent 20 years developing and tweaking my method for real-world application. Similarly, I wouldn't expect him to throw the book he's writing out on the Internet for free download.

He seems to think psychographic groups are expanded demographic characteristics. That's exactly, 180 degrees, incorrect. Psychographics is a three-dimensional, holistic, representation of human groups. Groups are anchor points, ways to understand people. Once you do, then you know how to talk to them, persuade and, yes, manipulate through communications. In the original Sun Tzu tradition, psychographics is simply the most sophisticated and effective way to understand the battlefield, so that you know how to win. Simple. In my view, anyway. I hope this clears it up for him.

by Sun Tzu 2006-03-15 05:40AM | 0 recs
What I'm Expecting

What I'm expecting is a discussion of what seems salient--such as what loads most significantly.  Like a dictionary definition before you give us the encyclopedia entry--which is what you've done in this post.  Plus an over-arching contrast/comparison.

I've done some reading in factor analysis, and some of the more significant books and articles I've read that rely on it do a fairly good job of discussing what the factors represent in a concise manner.  Fred Kerlinger's book Liberalism and Conservatism is one example. James Simson's Public Opinion In America: Moods, Cycles, And Swings is another.  (Okay, only two factors, but still...)

While I realize it's different when you're doing factor analysis of attitudes and cluster analysis of people, the the math tools are the same, and similar sorts of comments should be possible.

I'm certaily not looking for you to reveal the fine-tuned proprietary stuff.  Just the sort of gloss that you might give if you were making a presentation to a group of your peers. A quick sketch of the art involved, not the blueprints.  And maybe a word or two about how this clustering compared to alternate ones.  

I guess this comes from having different backgrounds.  I'm sure your clients could usually care less about this stuff.  But me, well, I did a lot of undergrad math--how much algebraic topology do I remember? not much!--and I've done some nosing through different literature on attitudes, plus I've reviewed several hundred serious non-fiction books, and been a political reporter.    Put them all together, and I just tend to want a clearer definition (knowing full well about the inherent fuziness of cluster analysis) with a bit of the researcher's rationale thrown in.  A mid-level thing between the cluser name and the sort of analysis you've presented here, with a bit of your commentary to hold it together.

Now, I realize that this may seem odd to you. I think that academics probably reflect on this more than someone like you might.  If no one usually asks this sort of thing, then why worry about it?  If the clusters pop and make sense in an encyclopedic entry sort of way that your clients can use, why even think of a dictionary definition?  I can understand that, if that's what's happening here.

But I don't think I'm the only one here at MyDD who would like to hear you take a stab at what I'm talking about.  What's more, I think there's a whole lot more who would like to hear it, out there in the wider blogosphere, and I'm hoping this is just the first of an ongoing series of polls you'll be doing, so that more and more of them are going to be paying attention.

It's not just that we'd like it, either.  The sort of synoptic intro I'm asking for could help set up some clear signposts that could help keep folks from going off on mistaken tangents.  

(Sorry for not responding earlier.  My life's a bit chaotic just now, not helped by intermittent loss of my DSL connection.)

by Paul Rosenberg 2006-03-15 07:10AM | 0 recs
Re: What I'm Expecting


Thanks, I think I understand more of what you're looking for. And I definitely understand chaotic life right now. Because of that, I'll keep this kind of short, if it still doesn't help on your issue, let's be sure and keep on keeping on on it (tortured phrasing there :)).

I believe what you're looking for is part of what's coming next, the next post. I'll have to move to a shorthand style with the groups in order to discuss how to use them in communications strategy. As usual, you're ahead of the curve and getting in front of my rollout strategy.

Being able to move to a shorthand on the groups requires everyone understanding detail first (this post). Being able to understand detail requires everyone seeing the big picture on psychographics first (my last post). You appear to want it all now. That's not going to work in this blog context and given everyone doesn't see research or do it regularly. Please be patient. This is a step process with course prerequisites. I'm covering the prerequisites because the primary gig, Applications, is what's most important in my view. And, if you haven't noticed, every single post of mine has been within a step process of rollout. I'm literally bringing everyone up to speed in order to get to Applications.

Further, there's nothing barring you from positing your own dictionary entries on these groups now. Go right ahead, I've no problem with anyone taking the existing analyses, and the future Applications one when it's available, and commenting, tweaking, doing whatever they want. I'm hoping they do. I'm hoping people really get involved with this analysis because I know this approach is what we need in communications.

You want factor loads? I'll publish them if there's a desire to see them. It's geeky thing, though. And I'm not woofy about it, either. Publishing the loads makes no diff to me, in all seriousness. All the questions loading on each factor have been published in this or the last post. All of them, Paul. They're all significant. By definition. The precise definition of significance re the loads is I used an eigenvalue of 1.0 as the cutoff. It's the SPSS default and I usually use it. Does knowing that add anything to the analysis? Not that I see, but there you go.

So, talk back here and/or see if the next post includes that dictionary type feel for the groups. Take your own stab at it. I encourage you and everyone. Get involved, get inside, live and breathe this research. I'll be thrilled.

by Sun Tzu 2006-03-16 08:02AM | 0 recs
Clusters & Typologies

I'm kind of with Paul Rosenberg on this. I don't need Sun-tzu's proprietary work, but a little more explanation of his categories would be helpful.

It seems clear that you have put some evocative labels on your clusters, but please give us a little more understanding. For example, I still don't know what the term "progressive" means to you. To me it means "left of liberal"; to some it means "right of liberal" to others it means "liberal but can't call it that because I've accepted the right-wing framing".

The PEW Typologies report "Beyond Red and Blue" identifies some clusters, more or less similar to yours. The utility is that they can ask two or three questions and peg which typology fits that respondent.

To Paul: This kind of data has a lot of colinearity in the measurables. Identifying new clusters gives them a new set of dependent variables, perhaps more orthogonal than demographic variables alone. This makes the predictions (regressions) more stable.

Also, relying on demographics alone for your primary dependent variables misses important stuff. For example income/class alone doesn't help predict the likelihood of a person to vote Democrat or Republican. Combining it with Religion and race (for example) and suddenly you see structure in the data.

What fascinates me is how these categories break down geographically. People tend to live near people like themselves. Political boundaries may (or may not) capture a lot of similar groups. Despite gerrymandered districts, there may be more structure in the voters than the Republican-Democrat voting record would show.

Another fascinating thing is how misleading it can be to look just at the aggregates. This leads national dems to be captured by certain conventional wisdoms and take a cautious "can't offend anyone strategy". If they examined the dis-aggregations, they would have a lot more flexibility of ideas.

This is one part of the "failure of leadership" of the beltway consultants. Why are they so f'n scared of Murpha's proposal? It splits a couple of the red groups.

I can see Sun-tzu jumping up and down.

by MetaData 2006-03-15 09:10AM | 0 recs
Re: Clusters & Typologies

I don't pretend to understand any of the math stuff.  but it seems like you are asking this question because the names evoke a meaning the analysis doesn't jibe with and raises questions instead of clarifying them.

i believe the name of the grouping is more arbitrary but the groupings are based on folks who answer questions the same and have certain similar characteristics.  so the progressives are similar backgrounds and answer questions the same way but if the title progressive is what throws you - the title is less important than that there are a block of people who you can make predictions about in this survey titled "progressive".  

the discussion about the bubba name is a similar tangent.  why bubba - why not.  it's not the bubba part that has any meaning but the link between answers that has meaning.  

by kevin19611 2006-03-15 02:54PM | 0 recs
Re: Clusters & Typologies

IIRC, with factor analysis it is often hard to interpret what the clusters mean.  Which group would a pro choice, pro 2nd amendment, pro environment, anti gay marriage believer belong in?  Pick your bundle of seeming contradictions and that person will exist somewhere and will be put into a cluster by factor analysis with some more or less degree of statistical success.  But it can be very difficult to tell why that person was placed into a given cluster.

by Skjellifetti 2006-03-15 06:15PM | 0 recs
What I'm Asking About

The first goal of cluster analysis is to get distinct clusters that partition the entire population. Without that, you've got nothing. Doing this can often produce clusters without particularly strong loadings. I understand that.  I just want to know which ones are like that, which ones aren't, a little bit about how they relate to one another, and why this partition was chosen over alternatives.

Like I said before, I want a sketch of this overview, not the blueprints.

by Paul Rosenberg 2006-03-16 05:09AM | 0 recs
Re: What I'm Asking About


I understand a little more about why I've been confused about your request. Please note that clusters don't load, my friend. There are no loads associated with clusters. Period. Loads are in factor analysis. It appears you're confusing the two. Cluster produces groups and there are no significance stats associated with those groups. They simply are, they exist, as a function of case by case question by question proximity. You need to be clear on that, please.

by Sun Tzu 2006-03-17 02:15AM | 0 recs

You've fleshed out some more of my unspoken concerns.  (When the DLC calls itself "progressive"!)

I know different sorts of things appeal to different analysts.  Everyone likes having a sharp elbow in the eigenvalues.  But from there on, folks seems to fragment. Some really like their factors to have just a few strong loadings--sort of the way you describe the Pew topologies--and would pass over what others might consider a better fit just to get that.  Others don't seem to care that much about strong loadings, seeing them as a bonus.  I'd just like a better feel for how these were arrived at.

by Paul Rosenberg 2006-03-15 05:45PM | 0 recs
Local Maxima or "Hot Spots"

Sun-tzu has offered a lot of food for thought. First there is the fact that he clearly has a lot of educated intuition, and practical experience.

But, I'm also beginning to appreciate the flexibility of his approach.

It is clear that to be effective at understanding which "knobs to twiddle" you need to combine several things: Demographics, Clusters/Typologies plus Polling Data on specific questions (which can be changed as needed).

Rotate the n-dimensional space to the right angle, and interesting structures appear. Also, you get localize "hot spots", places where a particular issue is very significant to a certain cluster.

by MetaData 2006-03-15 09:03PM | 0 recs
It's An Excellent Methodology

Which is why I'm pleased as punch that he's the pollster we hooked up with.  

The sophistication gap (in results, not just analytic effort) between this approach and just looking at crosstabs is greater than that between polling whole populations and using crosstabes.  This is clearly the way to go.  

I just want a little more specific understanding of where these clusters come from.

by Paul Rosenberg 2006-03-16 05:18AM | 0 recs
Re: Right!

Paul and Metadata,

I'll talk about how I see/definition of Progressives, and other groups, in the next post. In the meantime, here are the factor loadings for each factor. The ones I took into account, that is. I've given a shorthand on the questions. Let me know if there's a question about which question is which.

Just to refresh, factor analysis produces loads. The loads 'substitute', roughly, for significance. Cluster analysis produces groups. There are no significance stats for the groups. They simply are a product of the clustering process and are based on case by case question by quesion proximity.

Bush Meme:
OK for NSA to spy              .614
Support staying in Iraq       .593
Support impeachment            -.582
Pretest Bush approval          .577
Support '03 invasion           .576
Investigate Dubs               -.573
Posttest Bush approval         .562
Support Murtha plan            .522
Trust the NSA                  .506
US headed in right direction   .447

Religious tolerance:
Summary measure of religious
affiliation (demographic of
faith, denomination)          .981
Summary of religious
orientation (fund./evan.,
mainstream, liberal
demographic)                  .980

Bubba theme:
NSA spy gig awareness        -.758
Extremists pushing too far   -.511
Confidence in fed in a
disaster                      .509

Fear theme:
Worry US will be attacked     .823
Worry OBL hasn't been
captured                      .784
Personal safety/security      .611

Trust in external solutions theme:
HH followed DHS disaster
recommendations               .791
Availability of jobs locally  .576
Rating of fed protecttion
since 9/11                    .412

by Sun Tzu 2006-03-17 02:09AM | 0 recs
Re: MyDD Poll: The Structure of the Voter Universe

This is excellent, intriguing work.

I look forward to the next installment.

A note on formatting:

It helps, with a large block of text like this, to break up your sections with topic headins in bold typing.

Helps the reader navigate, find anchors for the eye, chunk the cognitive processing of sections and return to the text later for a second reading.

by Pachacutec 2006-03-14 05:31PM | 0 recs
Re: MyDD Poll: The Structure of the Voter Universe


10-4, excellent input. Love it. Thank you very much. Will do, for sure, for sure.

by Sun Tzu 2006-03-15 05:15AM | 0 recs
Wow. Can you Poll my district?

Seriously, this is great stuff. One concern. You say that trusting reds feel safer than before 9/11, but do you think it might be more accurate to say that they feel now that they are safer than before?

Your current phrasing implies that they felt unsafe before 9/11.

How would you recommend a congressional campaign utilize this sort of polling on a smaller scale? Do you think it would be cost-effective for campaigns in certain areas or states to pool resources and commision a combined poll, or do you think the diluted psychographic information would mitigate the benefit of the cheaper polling operation?


by msnook 2006-03-14 08:57PM | 0 recs
Re: Wow. Can you Poll my district?


Thank you, your observation on clarity re safety is excellent. You are absolutely right: they feel safer now than before, and not that they felt unsafe before. I suspect they didn't feel unsafe before. I appreciate the clarification very much!

Yes I can poll your district. Psychographics is applicable in large- or small-scale universes and certainly works in a congressional district context. Example: Arizona CD 6 in 1992 (now CD 1, Rep. Rick Renzi). In that race, Dem Karan English (my client) defeated the Dem establishment candidate in the primary and then went on to defeat Republican Doug Wead in the general to become only the second woman in the state's history elected to Congress. Our campaign was dirt poor and I, basically, gave the research to Karan, who's my friend and had been my client for three state lege races before. Our communications people, local people and also friends, were pretty much pro bono, too. We all pulled together, I broke the district down into six key groups, we targeted four, went after them like crazy banshees and won.

And Wead, whom you might recall from last year as the 'friend' of Dubs who secretely audio-taped their conversations (see this WaPo story) is a lunatic fundamentalist right-winger and he was then, too. In short, we thumped the right-wing fundamentalists before anybody nationally had a clue about them as a threat. Very gratifying campaign. (Karan lost her re-elect in '94. With Beltway consultants, not us locals. Pffft.)

I think it would be most effective in a CD race for the campaign to have its own CD profile and applications system. However, I break states down regularly into smaller geography and then crosstab the psychographics. Works great. So I'd see no problem in campaigns pooling resources to get the research. In fact, now that I think about it, there's probably some real strategic and messaging synergy that could be generated across campaigns that have the same view/situation assessment through research. If they cooperate, they could leverage key messages to get better reach and to more effectively penetrate communications clutter, especially in an election year. No question.

by Sun Tzu 2006-03-15 05:13AM | 0 recs
I'm not Al. I'm Michael.. er, msnook

not to nitpick, but you know, we can't have people thinking that the things said by private citizen msnook are coming from congressional candidate Al Weed.

by msnook 2006-03-15 02:04PM | 0 recs
You're not in the "contact us" menu

so how can I contact you? I'd like to talk about that idea (mostly I need an idea of how much it would cost before I start talking to local candidates about it.) You can email me if you like: michael.snook <at>

by msnook 2006-03-15 02:17PM | 0 recs
Re: You're not in the &quot;contact us&quot; menu

I'm not sure why my public email doesn't show up. I'll check it out. Thanks for letting me know. FYI, mine is Thanks.

by Sun Tzu 2006-03-16 08:32AM | 0 recs
Re: MyDD Poll: The Structure of the Voter Universe

This is wonderful!

I had to go back and read your previous post to interpret some of the memes, in particular the fear meme.  My own identification is clearly Blue Core, but I have strong fears:  of the Bushoid presidency and the erosion of liberty, of another aggressive war.  But your definition is of external fears, most specifically fear of another 9/11-scale terrorist attack.

What I wonder is whether this fear has replaced what used to be the common fear denominator among suburban whites, in particular:  crime, specifically race-based and drug-based crime.  And if so, how much does this represent a shift in the rhetoric of the fearmongers, having found a new and more productive target?

by Feh 2006-03-15 04:31AM | 0 recs

An excellent area of discussion, imo ...

If you look at the polling on the issues that matter, there's a clear inverse correlation, iirc, with the frequency with which "crime" and "terrorism" are mentioned. So as "terrorism" gets mentioned more, "crime" is mentioned less. Fear of the "predator" is the common dynamic there, but it's transferrable. I wonder if that's actually tapping into something more fundamental in those people ... they're just freakin' scared of personal violation and attack in some root area in their psyche.

However, the fear of loss of freedom is a different fear ... I don't want to imply too much, but it's more akin to a fear of confinement. I don't know near enough to try to imply that there's some root causes of what we're "afraid" of in a political sense. But it's a fascinating question ...

And, as you point out, only one kind of fear is really covered and/or labeled in the data here. I know I'm afraid of the path the government is on right now, but I wouldn't show up on the "fear index" because I have nearly no fear for my personal safety vis-a-vis a terrorist attack.

by BriVT 2006-03-15 05:26AM | 0 recs
Re: Fear

Interesting, that inverse correlation is much what I suspected.

I think the fear meme in both cases taps into a sense that people are entitled to perfect security, that their world can be made entirely safe. And anything that threatens this is an easy target.

So they are willing to endure the loss of personal freedom, which is what I fear, in order to feel free of that fear.

I tend to suspect, w/o much evidence that this fear/safety thing is strongest in people who actually are safest, who have very little in reality to fear.  They don't really know how to deal with present dangers, being inexperienced, have never learned how to cope.  

by Feh 2006-03-15 08:25AM | 0 recs
5,000 fingers of Dr W.

Observing Sun-tzu's deconstruction of the US voters helps us understand the attentiveness and skill of the Cheney-Rove political machine. Nowhere is this more apparent than in the republican use of the fear theme.

I'm thinking of the way they pull the emotional triggers. They are willing to go extremist when they need to: TREASON! if you criticize the president. OUTRAGE! over Terry Schiavo. TERRORISM!

Clearly they have identified the psychological clusters and know where the levers are. They also know WHEN it matters, i.e. when to bring out the big guns, and when to put the repetition machine into action.

Also, they know when to drop a stone quietly. I'm thinking of code words slipped into a speech. For example a religious phrase that would fly right by me, but triggers knowing looks among evangelicals: "He's one of us'.

by MetaData 2006-03-15 09:32AM | 0 recs
But Not Always

OUTRAGE! over Terry Schiavo.
didn't play so well, now did it?

by Paul Rosenberg 2006-03-15 05:45PM | 0 recs
Fear of a predator vs. fear of confinement

I think you're exactly right about this split. These externally fearful people are worried about black folks with guns and arabs with boxcutters. They are perfectly happy to hide in their caves, under their covers, or in the warmth and comfort of daddy Bush's arms.

I suspect these people really like their gun rights, and don't care that a handgun in the home is 3 times as likely to injure a family member than an intruder.

"The path of justice and honor involves one in danger" - Thucydides

by msnook 2006-03-15 02:25PM | 0 recs
Re: Fear of a predator vs. fear of confinement

Ya know, the more I think of this, the more likely it gets. GOP fears are always based on a predation fear. There's always a stalker in the woods element to it. It's terrorists looking to kill us all, gays coming to "recruit" to their lifestyle, illegal immigrants "flooding" the border to take all our jobs. Remember the laughably inept "Wolves" ad during the 2004 race? It was poorly done--it looked like a combination Sierra Club video and Alpo ad--but it's the distillation of the base fear those guys are tweaking. Predator comin' to get ya!

Like you said, guns are a part of this, the racially charged stuff in the South ... even the "predatory taxes" are a part of it.

Progressives always talk about the fact the GOP whips up "enemies," but it's more subtle than that. It's predators they're creating, scary stories of the wolves in the forest they create so that only the strong chief can protect the clan. So, the key question is not "which enemy will the GOP create." That's not a good way to predict what they will do. You have to think of what predator can they create and exploit. Gay adoption? It's like the Rosetta Stone of the GOP message strategy.

Liberals/Progressives, otoh ... I'm not sure there. It's not as cohesive a narrative, not as primal a fear.

I find this fascinating.

So the question is ... should Democrats look to create their own predators (NSA spies, etc)? Or work against the predator narrative? I think the liberal instinct is to do the latter, but that may be like bringing a knife to a gun fight ...

by BriVT 2006-03-16 03:55AM | 0 recs
Re: Fear of a predator vs. fear of confinement

What does the Red Core fear?

The survey, in defining fear as the external threat, didn't address this question.  The numbers might suggest falsely that they don't fear, but this isn't so.  It just didn't identify and measure what they do fear.

by Feh 2006-03-16 06:56AM | 0 recs

Still seems to me that Iraq redeployment and NSA are a winning combination. The Heartland Reds look like they can be brought around on NSA because of their lukewarm trust of the NSA and their willingness to consider the word "impeachment" of Bush (which, imo, shows a softness in their support that's not there for the Red Core). I'd bet that this group bailed on Bush big-time over the Dubai Ports deal ... if the motives of the Bush Administration stay in question, the NSA probe has a lot more salience.

And they love them some Murtha Plan.

One helpful thing would be to tie in the Murtha Plan with Osama and terrorism. "We need to redeploy our troops so that they can search out the true terrorists. We can't have Osama running around free while our troops are caught in the middle of a civil war in Iraq." 'Cause those folks are mighty afraid of Osama.

The economic factors aren't really fleshed out in this poll, so it's hard to know how domestic factors play with these groups  ...

by BriVT 2006-03-15 05:40AM | 0 recs
Definitions of Rural, etc.

The overall percentage of rural voters, 36, is much higher than I would expect.  The Census Bureau says that 79% of the population is "urban" (presumably includes suburban) ao 21% must be rural as of 2000 (Statistical Abstract, online data).

What this says to me is that either voters are not typical of the population disbursement as a whole, your definition of rural is different from the Census Bureau's definition, or a combination of both factors.  I expected a number here like 20% so I am blown away.

Does rural, for example, include the much discussed exurban vote?  Would Jean Schmidt and OH-2, be a rural or suburban district or a mix?
I can start naming specidfic areas but that is really getting off topic.  Clearly inner suburbs and outer suburbs vote differently at this point.  Maybe you are clustering them in different groups.


by David Kowalski 2006-03-15 09:10AM | 0 recs
Re: Definitions of Rural, etc.


Thanks for your question. Rural (and urban and suburban) are defined by the respondents themselves. Self-identification. Same with religiosity and religious orientation. It's what people say they are.

So while the rural number might seem large to you, it is where people told us they live. And I would say that where one feels they live is a better measure than assigning such a variable through geographic mapping or whatever. If one feels one is a rural resident, then that has an impact on how they perceive other issues, regardless of the zip code or tract number.

Make sense? Thanks, again.

by Sun Tzu 2006-03-16 08:23AM | 0 recs
Comprehensive Database

Would it be possible/ how much would it cost to create a comprehensive data base organized like this and subdivided by state, congressional district and media market? It seems like this would be incredibly useful for targeting rhetoric. For example, a canidate could bring up specific issue that would play well among one group and then tweak the message so it plays better when campaigning among another group. What is the viability of such a program?

I'd also be interested in the positions these groups take on other issues- if you know them from past work you've done great, otherwise maybe questions for the next poll?

Also, regarding the "progressive" group. Is it just a label used for convenience or is there some similarity between that group and all of us online that I'm missing. Is it possible that all of us who call ourselves progressives are actually part of one of the other Blue groups?

by js noble 2006-03-15 05:40PM | 0 recs
Re: Comprehensive Database


You can make a first cut by looking at Sun-tzu's demographic cross-tabs. They give you a first-order approximation of how your region or district differs from national. This is simple matrix arithmetic when compared to the subtleties of Sun-tzu's polling and statistics. The primary drivers are geography, race. Religion (with race break-down) is also a fairly significant.

For example, the West has a surprisingly high percentage of non-religious and "lifestyle liberals". Thus, a Western politician could be less beholden to the theo-cons, meaning they could support something like gay-rights. Ken Salazar of Colorado is either principled or demographically-aware when he takes on Dobson's "Focus on the Family".

This is useful as far as it goes. Sun-tzu would surely point out that designing a messaging campaign requires additional, more specific information relevant to your district. His clusters may hold some truth locally, but probably shift enough that you want localized polling. Let's say your district has a large population of African-american or lot of rural farmers. Then you would want to understand more about the issues of importance to a locally-significant community.

Still, you could make a broad-based cut first, which would inform you about what questions to pursue with further polling. Sun-tzu probably already has a hunch if not specific data about that.

by MetaData 2006-03-15 08:35PM | 0 recs
Re: Comprehensive Database

i think the "progressive" label is confusing.  lots of folks use it.  i would be surprised if a lot of folks on this site fell into the blue core instead.

by kevin19611 2006-03-16 12:13PM | 0 recs
Constructing Future Polls

I should think that you could find a pretty small number of questions that allow the voters in a survey to be psychographically categorized well.  That would be useful in constructing future polls, sort of the psychographic equivalent of the demographic questions that are always asked up front.  Then one could ask entirely new questions but still be able to correlate them back to the original groups.

Just which questions those are would be super-fascinating to post.  (As would some measure of the degree of clumpiness in this lower-dimensional space.)

by Professor Foland 2006-03-16 11:41AM | 0 recs
Re: Constructing Future Polls


You're exactly correct. What an excellent observation. The VALS program at SRI International had a strict set of questions, about 30, that they used to classify people into their system. The Pew Center has a set they use, too.

This survey wasn't technically a psychographic survey, it was a poll that I applied the method to. Over the years, though, I've probed areas of political perception involving patriotism, views of government, emotion regarding issues and such. I rotate some questions in and out and design new ones based on current events. While not a rigorous set system, it 'floats' a bit, I concur wholeheartedly that a set of questions to get the psychographic data is apropos and allows asking new questions to roll back through the psychographics. Right 'tone, Professor.

by Sun Tzu 2006-03-17 01:40AM | 0 recs


Advertise Blogads