CBS news national poll

Some interesting results, but smallish sample.

Prefered Candidates among Registerd Dem primary voters:

Clark 12%
Gephardt 10%
Lieberman 9%
Dean 9%
Kerry 8%
Moseley-Braun 4%
Sharpton 3%
Graham 3%
Edwards 2%
Kucinich 1%

the link is here. The full results in pdf format are on the right.

Not that I don't think she has 4% -- I bet she has much more!

But I suspect there's something going on there besides honest sampling, don't you?

Basically it looks tilted to the right a bit by sampling. What's your suspicion? Are you thinking pushpolling or something else?

Curious that "Congressman Dennis Kucinich" was regarded less favorably than "Dennis Kucinich." That raises an eyebrow, but I don't know what to conclude.

It's alleged to be of likely Dem primary voters, who are typically characterised as being more politically aware. But it basically looks like it represents pure name recognition based on Media Inc coverage du jour. It doesn't map to placement on the alleged spectrum (except for the righward tilt you remarked), it doesn't map to fundraising,...it just looks hokey for what it claims to represent.

negative connotations that the word "Sentator" doesn't have, because Kerry goes UP when prefaced with "Senator".

It doesn't make sense.

A negative reaction too much alliteration? Seems implausible.

The Senator Kerry jump could be for other reasons, like "Ah hah, I remember him." The effect isn't uniform.

And what's up with pollingreports summation of the same poll? That just makes no sense. I mean, intuitively, those results would support your idea, and "Former Senator Moseley Braun" could be a huge negative, as could "General," But that's not what cbs reported. Something's not matching up here.

The sample size is rather small. Hmm. I'm inclined to let it slide for the moment as there are other good national polls. The idea of polling with different titles is good. Hopefully somebody will repeat it with better methods and reporting.

with 5 guys in it.

Clark 12%
Gephardt 10%
Lieberman 9%
Dean 9%
Kerry 8%

I would prefer that the five guys on this list included Edwards and dropped Lieberman, but I think that difference is explained by name recognition.

I've never met a Lieberman supporter. Not that they don't exist, I just wonder how much of his numbers are name recognition. I'd like to see some polling that leaves out some people. Where do Lieberman supporters go? Or Gephardt? Or Edwards? And so on.

If a poll doesn't cover likely voters, it's worthless. Period.

People who give an opinion when asked are not the same as people who actually bother to vote. Especially for Democrats, unfortunately.

- C.D.

1. I suspect their identification of "primary voters" was similar to a "likely voters" filter.

2. "Likely voters" filters out the no-shows, sure, but it also filters out people who are the targets of voter registration drives. Also both "registered voters" and "likely voters" will tend to skew rightward, as I believe there is a correlation between tending to move often and holding less conservative views. So if for example the first filtering question is "Are you registered to vote at your current address?" before you even get to "likely voters" you have ruled out people who may be likely to vote for reasons you haven't considered.

Or if you filter by asking "Did you vote in the last primary?" you will be ignoring young voters who may not have been eligible. If for instance there's an emerging trend of generational conflict, your filter won't adequately reflect that.

You could design better filters, ask for instance, "Have you ever failed to vote in an election in which you were eligible to vote?" And then finetune it from there, because of course people have various reasons for failing to vote, and some of those reasons say very little about whether a person is likely to vote in future elections. But that's kind of moot. Most of the likely voter filters I've seen aren't nearly that sensitive or objective.

So we're really talking about assumptions here, not facts. As the election draws closer, the likely voters filter becomes more meaningful. Then again, the ease of registration has not been constant over the years or across state lines. And we have known instances of targeting certain groups for registration on the one hand, and expunging or obstructing on the other. I just don't think in light of all that it makes sense to assign an absolute value to such filters.

3. I think all kinds of polling data are at once valuable and biased. Your particular prejudice risks throwing the baby out with the bathwater.

...if this is a poll for *Registered Dem primary voters* to choose their preferred candidate, it totals only 61%! Just who might the other 39% have chosen if not one of the ten candidates of their own party? Or am I missing something?

Lots of little things to say here. Dean for instance will have really dedicated support in his percentages, just because his apparatus being Internet based, will be very solid. He is also ahead in the first three primaries and will get huge recognition if he wins those, and particularly New Hampshire looks sewed up, the other two he's ahead in. Of course this will help him gain national recognition if he wins them, I didn't have any idea who Clinton was before he won a couple of primaries.

Black voting, and recent massive registration. This factor alone will bring huge numbers of new voters into the process, who will still likely vote in large numbers for Democrats. I would admit that it may not be 92% like in the past, though I could never see why most blacks would choose a Republican president, some do. Sharpton and Braun will have a similar effect as Jackson did in 1988, where though Bush was elected, Senate and Representative seats went more Democratic due to his registration and black voting factors. This is one factor that most Republicans are ignoring, for the most part.

Bush has been running a continuing assault on labor, trying to destroy overtime, force comp-time, and he's done nothing to staunch the flow of jobs, but like Clinton and the Bush before him, just continued the destruction of America's lifestyle. Americans are really pissed about that, and that amounts to a higher percentage of Democrats voting in the general election.

People will be dropping out, in this huge field. Who will get Kucinich's votes, Braun and Sharpton's, and Graham's votes, the most likely people to exit the race? For my money, they will go to Dean, as his positions are most like theirs. This along with early primary wins by Dean, is just another factor that will catapult Dean even further ahead of the pack.

Simple anger at Bush will make normal Democratic base turnout really good too.

One thing is for sure, the polls are all over the place, Dean leads in MSNBC's, Clark in Newsweek, and I forget who was winning in this one, but it seems like it is highly questionable, I agree with the rest of you. We'll see, but there are many more factors than usual, I am sure a few that I haven't covered here.

About whose votes will go to whom, you can get a sense of that from some of the instant runoff polls, like the potus poll. Lot's of biases, and if you want to extrapolate you have to make some assumptions, but it does tell you something. Look at it in table form, you'll know what I mean.

Carol, btw, is in it for the long haul. I'm sure supporters of the other candidates feel the same way. Perhaps some will drop, but not all that you mentioned. Many on the left want a candidate who is unequivocally anti-war and liberal on a broad range of social and economic issues. Dean doesn't quite fit that bill--not that he's a bad candidate mind you, or that he would be unaceptable to the vast majority of leftist Dems in a head to head against Bush. He's just not a liberal.