Interesting News (if it's News)

Bearing in mind that I still hold to the precept that blogging is not journalism, I have to confess that I was curious about a post on Net News Publisher under the headline, "‘Long Shot’ Kucinich Leads Among Online Independent Voters." Posted by "admin," this jumble of assertions, which probably did not go through any fact-checking process, is a perfect poster child for advertising the value of good editing at a time when we are still be deluged by Web 2.0 evangelism. Nevertheless, the headline fascinated me so much that I could not resist dusting off my editor's visor (metaphorically) and trying to figure out for myself if there was actually signal in all the noise. It turns out that there is signal; but you have to read pretty closely to learn that the post is based on the results from a poll conducted by an organization called IndependentPrimary.com. There is nothing scientific about this poll. If you want to participate, you just sign up on the Web site (using the hyperlink I attached to IndependentPrimary.com); and, as far as I can tell, that is all it takes for your voice to be heard.

The result Web page for this organization provides the evidence behind the headline. Among those who declared themselves as Democrats (80,153), 76.7% (61,477 respondents) "voted" for Kucinich. There was also a Republican poll with only 25,269 participants, a whopping 93% of whom selected Ron Paul. In both parties none of the other would-be candidates could muster a double-digit standing (i.e. 10% or higher), meaning that, while the sample itself may not be statistically representative, the results are statistically significant. (If you think that what you just read is a candy-coated way of saying garbage-in-garbage-out, you probably have a good point!)

The Net News Publisher post cites three other polls, each of which may be flawed in its own way. One was conducted by Democracy for America (DFA), whose slogan is "Social Progress; Fiscal Responsibility; Grassroots Activism." Their account of their poll leaves a bit to be desired:

Last month over 154,000 of you voted in the largest primary poll of 2008. The poll made clear two striking facts: A 78% consensus for the top three progressive candidates of Edwards, Kucinich, and Obama¹. And 95% of DFA members voted for someone other than the media's frontrunner.

Once again, the question of sampling has been totally disregarded; but in this case the question of statistical significance of results is not as decisive. The superscript refers to the Web page with the numbers. Kucinich is still at the top with roughly 11,000 "votes" putting him ahead of Al Gore. The way you get that "78% consensus" if to drop Gore from the options and rebalance the percentages over the reduced sample.

Then there is the group called Progressive Democrats of America (PDA). They have provided the most thorough account of their results in a PDF file. The problem here is that their membership is only 80,685; and only 15,810 of the members voted. So we have lots of tables all with lots of relatively small numbers. With these limitations, their own account of the results is about as acceptable as one could anticipate:

Not surprisingly in a field of eight contenders, no candidate came close to gaining a majority of the total vote in PDA's recently completed presidential straw poll. But two candidates--Dennis Kucinich (41%) and John Edwards (26%)--combined for more than 2/3 of the total vote. Over 15,000 PDA activists voted in the presidential straw poll. Full results here.

After Kucinich and Edwards, only one other candidate-Barak Obama (13%)-made it into double digits. All the rest were in single digits: Hillary Clinton (9%), Bill Richardson (5%), Joe Biden (3%), Chris Dodd (1%), Mike Gravel (less than 1%). In contradiction to media reporting on the primary race, PDA's results parallel those of DFA, Daily KOS, the Texas Democratic Party and others whose polls show very weak support for Clinton among the Democratic base.

Finally, The Nation conducted its own poll, presumably among its readers. I have not yet tried to track down these numbers. I do not recall seeing them on my RSS feed. My guess is that the sample would be somewhere in the same league as that of Progressive Democrats of America. Here is the summary from Net News Publisher:

And, in a poll conducted by the progressive The Nation magazine, he [Kucinich] won with 35% of the vote. Obama came in second with 24%, and Edwards was third with 13%.

Now I personally have enough interest in Kucinich and many of the principles he is trying to promote that nothing would please me more than to find him pulled out of the slough of statistical insignificance. I only object to extricating him with methods that justify ranking statistics as worse than damn lies! What depresses me is that these are results that take place out there on the long tail, which means that one only reads about them on that same long tail. Viewed from the other end of the telescope, this means that these results do not "count as news" at the editorial desks of the mainstream media; and, critical as I am of how the mainstream media tries to shape our thoughts, rather than report on them, the statistical methods behind those polls may be sufficiently weak to justify the editorial decisions.

This brings us back to my original theme. The blogosphere is a great place for the stuff that Abraham Lincoln had in mind when he said, "This is the sort of thing that people who like that sort of think like." Call it an echo chamber or a vanity mirror; just do not pretend that it is journalism. Whatever the media business may be doing to undermine the role of journalism as a public trust, there is still a critical mass of practitioners out there trying to carry the flame. We should not abandon them just because we can find places on the Internet where we can read things that make us feel better.