Back in the winter when Dean crashed and burned in Iowa, I asked:

To paraphrase an old Jewish aphorism, is this good for the blogs? Regardless of one's political stripe, the blogosphere embraced Dean's Internet campaign as a kindred spirit, emblematic of the same phenomenon that propelled blogs into prominence.

Now, the reason I asked this was obvious -- most people associated campaign blogs with Dean, and if Dean flamed out, surely that meant that having the most successful blog around didn't mean all that much even in primary campaigns.

The unstated assumption behind my question was that Blog for America was actually the most successful campaign blog out there. Even though campaign blogs are different from other kinds of blogs, and even though I had criticized its content in another venue, I certainly believed it to be the most professional.

However, I may have been in error. [Again--ed.] Gene Koprowski, UPI's telecomminications reporter, reports on an interesting study about campaigns and blogs:

Most of the information about the influence of blogs is qualitative -- anecdotal, based on what readers say about the sites. But before the Democratic primary season was completed this past spring, one software developer tested the reliability of blogs run by political candidates and gleaned some interesting results that may continue to play out in the fall.

"Blogs run by the campaign of President George W. Bush and Sen. Kerry were the most effective," said Joe Alwan, vice president of marketing at Empirix Inc., a Web applications software developer in Waltham, Mass. "They had a 100 percent reliability. The blog run by the campaign of (Sen. John Edwards, D-N.C.) had only a 50 percent reliability," he told UPI.

Former Vermont Governor Dean maintained about a 98 percent reliability rate for his Web blogs, the study demonstrated.

Empirix tested the sites by sending electronic queries to them during a set period of time, and often received error messages saying, for example, the "Apache Web server was down" for Edwards, Alwan said.

This could have been interpreted as a signal to some Internet savvy voters that the candidate just did not care enough about reaching them online, in the way in which they wanted to be reached. The reaction is similar to the way customers become turned off if an e-commerce site is not working properly, said Pete Cruz, director of Web applications management for Empirix.

"The point is that blogs are now important in politics, and they need to make sure that their sites are working," Cruz told UPI. "For Edwards, the blog became a liability to his campaign. Users who visit expect performance. A lack of performance is more likely to alienate users."

Another interesting point: Sen. Joe Lieberman, D-Conn., did not have a blog on his site and did not perform well in any of the Democratic primaries, the Empirix survey concluded.

Now, a few caveats -- first, I can't find a press release or an executive summary of this study on the Empirix web site (see below for an update). Second, the difference between Dean's 98% effectiveness and Kerry's 100% effectiveness is not huge. Third, Edwards outperformed Dean in the primary campaign even though his blog was only half as effective.

Still, this is the first (report of a) study I've seen in which Kerry's blog comes out on top by any metric.

By the way, if you read the entire UPI report, you'll find a mention of danieldrezner.com -- clearly, Greg Wythe was not the only person impressed with my ability to fold in a Kristin Davis reference to a post about Sarbanes-Oxley.

UPDATE: Drezner gets results from UPI and Empirix! After an e-mail query, the good people at Empirix were nice enough to send me their study, which was done at the behest of Baseline magazine -- though it doesn't appear to have been cited in their December 2003 package on campaign blogs. But for those who care, their study was conducted from "October 31, 2003 at 1:00pm Eastern through November 7, 2003 at 1:00pm Eastern." It looks quite proper.