Tea leaves
Daily Kos Straw Poll |
|
July 16 |
August 1 |
Warren |
19,975 (35%) |
19,463 (33%) |
Sanders |
11,514 (20%) |
14,945 (25%) |
Harris |
8,230 (14%) |
3,973 (7%) |
Biden |
6,560 (11%) |
6,962 (12%) |
Every two weeks Markos Moulitsas reads the tea leaves of the Daily Kos straw poll. In his article about the July 16 poll, Markos wrote:
Elizabeth Warren wins her third straight straw poll, further pulling away from fading Bernie Sanders. What was a 9-point lead a month ago, and a 4-point lead two weeks ago, is now a 15-point lead. It also marks Warren’s highest vote percentage ever in this straw poll, even if it’s by a single point.
But hold the phone! In his article about the August 1 poll, the “fading” Sanders — who was the only top-tier candidate in the poll with good-sized gains — is now described as “a strong second.”
Clearly the numbers in the straw polls will bounce around for the top-tier candidates, as they will in scientific random-sample polls. Coincidentally, Sanders also bounced up 5 points in the most recent Emerson poll and he was, in that poll too, the only candidate with a good-sized increase.
Being human, we all naturally have positive/negative emotional reactions to upward/downward movement of our favorite candidate’s poll numbers.
It’s a reflex reaction. In reality, I don’t believe (and I think most people don’t believe) these polls are any better than astrology charts at predicting which top-tier candidate will win the nomination. Even in a race with just two top-tier candidates (and this race has more), we can’t draw conclusions from random-sample polls of the electorate this early, and, obviously, drawing conclusions from Internet straw polls is even more a flight of fancy.
Here, for example, is a September, 1979 Harris poll showing Democrats favoring Ted Kennedy over Jimmy Carter for the Democratic nomination by 61 to 34 percent. Carter of course won the nomination. And in February, 1980, a Gallup poll had Carter beating Reagan 60 to 31 percent. In November, Reagan beat Carter by a landslide, winning 44 states.
The early states are still very important
In that same article about the August 1 straw poll, Markos reiterated his thesis that the Iowa and New Hampshire contests have lost their significance:
[W]e, the small-dollar-donor activist class has determined who gets to play in the Big Leagues, and everyone else is essentially ignored. …
[T]he days in which Iowa alone got to decide who the top players were is long past. This won’t be up to Iowa or New Hampshire. South Carolina will be the first contest that really matters (it will determine whether Harris has consolidated black support behind her, whether Biden will keep his, and whether Warren will compete with that demographic), and then it’s on to the delegate count starting Super Tuesday. About a third of all delegates will be awarded that day, and the race will be whittled down to the finalist 2-3 candidates for the duration of the contest.
...
[T]he first primary? It happened—debate rules that weeded out five of the announced candidates. The second one is taking place right now—it’s who will make it to the September debate, one which requires 130,000 donors and four polls at 2% (or better). That’s a ludicrously low bar to meet, yet only seven candidates have qualified so far—Biden, Booker, Buttigieg, Harris, O’Rourke, Sanders, and Warren. Julian Castro, Andrew Yang, and Amy Klobuchar are within striking distance.
That said, the public polls are far more generous to the overall field than the Daily Kos straw poll. Among those of us paying the most attention, this is a pretty compact field.
I don’t doubt that the top tier will already have been established by the time the Iowa and New Hampshire contests arrive. But these contests are huge national media events. That’s why they’ll always matter. They establish media narratives. They can impart or impede momentum. They can boost a candidate’s chances or burst their bubble. Regardless of his phenomenal crowd sizes and small-dollar haul, I don’t think the dark horse Bernie Sanders could possibly have done so well in the 2016 cycle (winning 23 contests) if he hadn’t first proved himself in Iowa (virtually tying Hillary Clinton, a really astonishing climb) and then scored a knockout in New Hampshire. These contests proved Sanders wasn’t just an insurgent, but a contender. By contrast, there’s Howard Dean, the other famous dark horse insurgent from Vermont who enjoyed big crowds and a big small-dollar donor base — who never recovered after placing third in the 2004 Iowa caucus. By showing his chops in Iowa — coming in ahead of Hillary Clinton and John Edwards — Barack Obama in 2008 surely boosted his chances in the next contests, including South Carolina. And, given the bombshell scandal he was hiding, thank goodness Edwards didn’t score a victory in Iowa or New Hampshire! His platform made him my favorite candidate at that time, as he was of many Kossacks. For example, here are the top candidates in the December 20, 2007 Daily Kos straw poll, just two weeks before the Iowa Caucus:
Edwards |
41% |
Obama |
27% |
Dodd |
11% |
Clinton |
6% |
While Edwards may have been very popular among the straw poll participants — that is, among the folks who Markos suggests are “paying the most attention” — ultimately our preference for Edwards was not shared by the folks whose opinions really mattered — the early state voters.
Who is participating in the Daily Kos straw polls?
It seems that in the comment thread of every Daily Kos straw poll this cycle, some folks are quick to point out a distinction between the votes of the engaged readership of Daily Kos — the Community — and the votes of folks who aren’t regular participants at this site, but just come to vote in the poll for their favorite candidate when their fellow supporters put out the word that the poll is up. (Markos also has made this distinction in some of his articles.) In particular, it is the votes for Bernie Sanders which seem always to prompt these sorts of comments, and the implication of many of these comments is that the votes from non-Community members are somehow sullying the results of the poll, or at least obscuring the results such that they don’t give a true reflection of the preferences of the Community.
Markos, who of course founded this website and this straw poll, has repeatedly said that there’s nothing wrong with offsite supporters encouraging fellow supporters to go vote in this poll. He’s pointed out that every candidate’s supporters are free to do so, and that they’re welcome to. He’s noted that having lots of participants in the poll is helpful to grow the site’s mailing list which is used for numerous progressive campaigns.
I would point out, moreover, that folks who are encouraging fellow supporters to go vote in the straw poll are modelling exactly what campaigns and supporters will do and should do in the real primaries and the general election: get out the vote. So it seems to me these supporters are modelling a good and appropriate behavior.
But if some folks prefer to focus on the preferences of the engaged participants of this site — the Community — I think we should give some consideration to what those demographics look like.
The demographics of the entire Daily Kos readership (as described on this Daily Kos advertising page) aren’t necessarily the same as the demographics of the engaged readers who participate in diaries here, though they might give us a hint:
Readers with a college education |
81% |
Readers with graduate degrees |
34% |
Readers with household income > $100K |
26% |
How many engaged readers do we have in this Community — folks who regularly participate in the diaries? In this recent diary, I estimated (based on the maximum number of Recommendations we see on extremely popular, noncontroversial diaries) that the number is likely in the low thousands, i.e. a number very much smaller than the entire Daily Kos readership and the number of participants in the typical straw poll (50,000+). Another way to estimate the number of engaged readers is by using the Search tool. For example, when searching for the number of users whose most recent comment was in the past week, I found it was 5000—6000; many of these are recent accounts and already banned (spam I assume), but most are users who joined more than a few months ago.
By sampling just a few hundred of the engaged readers, we might get a pretty good gauge of the whole group. Clearly this must be the reasoning of Daily Kos users who take snapshots of the early results of the straw polls. For example, in the August 1 straw poll’s comment thread, Daily Kos contributing editor Adam B posted a snapshot 10 minutes after the poll opened, capturing the first roughly 400 votes. He joked in the comment that his snapshot is “avant le déluge” — in other words, prior to an influx of votes from outside the Community.
In 2017, I ran a poll to gauge Community demographics; 89% of the 264 Kossacks who responded said they were white and 91% said they were 45 or older. Another user posted a poll in 2018 in which 81% of 1130 respondents said they were 51 or older and 66% said they were 61 or older. Very recently I posted a poll in which 92% of 242 respondents said they were white.
Arguably, these side page polls are capturing enough responses to estimate the demographics of engaged readers, though the results may be impacted by factors such as the time of day and day of week (which might also impact the front page straw poll and snapshots of it). The side page polls may also be affected by authorship (i.e., if you recognize a diarist’s name, you might be more likely to read the diary, or you might be less likely to.)
But if these side page polls do fairly accurately reflect the demographics of folks who regularly participate in diaries here — and I wouldn’t be surprised if they do — then it seems our little Community skews very white. And skews quite old, too.
So I think this should give us pause if we have a strong desire to focus on just the voter preferences of the folks who comprise our Community, and we have a reluctance to have the straw poll results “tainted” by an influx of votes from outside our Community.
In the comment thread of Markos’s article about the August 1 straw poll, Daily Kos contributing editor Denise Oliver-Velez remarked:
Not surprised at Harris numbers here. … I’ll stick with Kamala. She’s picking up support outside of the blogosphere.
I’m a Sanders supporter but I think the implication of that comment — that we shouldn’t be insular in our focus — is very sensible.