« Really Shitty American Express | Main | Rob Teilhet, Inc. to be dissolved »

April 06, 2005

Fantasyland

Only in the fantasy world of the blogosphere is someone with a 66-23 approval rating vulnerable in a primary or general election. MyDD and DailyKos are all atwitter because Joe Lieberman has a higher approval rating among Republicans in his home state than among Democrats. And yet, his approval rating among Democrats is still the aforementioned 66-23, which is not great for your own party but which is also well into the comfort zone.

The fact that his approval among D's is lower than among R's means nothing, and this highlights actually a big problem among some of the amateur statisticians of the blogosphere. First let me say that I encourage anyone who wants to crunch numbers to do so and also it is an unqualified good thing to take a look at the way things are from angles that others maybe hadn't considered.

However, it is important to remember that you can have a lot of fancy statistics but you must understand which ones mean something and which ones are just fancy statistics. This is something I've meant to discuss for a while.

MyDD has what they call a "partisan index" that compares the vote for President in a state with the national popular vote each year. If the Republican margin was better than nationally, they have a Republican partisan index and likewise for a Democrat. Here is a link to their partisan index page compiled before the 2004 election. It is kind of a neat reference to see how a state fared compared to the rest of the country, but doesn't tell you much about what you can expect in the future.

It doesn't even accurately paint a portrait of what already happened. Here is a DailyKos post that illustrates that. According to Kos, these are the states that moved from the RNC partisan index camp in 2000 to the DNC camp this year, states that are "getting bluer": Iowa, Nevada, New Hampshire, New Mexico, Ohio, Oregon and Wisconsin. Hmm, notice anything about that list? Maybe that Iowa and New Mexico were carried by Democrats in 2000 but Kerry lost them this time! Some comfort it is to know that those two states are more Democratic than the country as a whole in 2004 but Bush still won them.

On another list that includes states where the largest gains on partisan index were included you have states like Colorado (lost) and Pennsylvania (won). Gee, I wonder if that is because both states were among the most heavily targeted this time, Colorado wasn't last time and Kerry basically matched Gore's performance in Pa but improved that state's partisan index because he underperformed Gore nationwide.

To show that a measure isn't that worthwhile of a predictor one thing you can do is look at it's track record for the past. After the 2000 election, none of the states qualified for the first DailyKos list, that is none of the states saw their partisan index switch from pro-Republican to pro-Democratic. Yet, four years later New Hampshire flipped to a Democratic state. Additionally, these states would have made Kos's B list: CA, CT, DE, DC, FL, IL, KS, MD, MA, NV, NJ, NY, PA, RI, TN, VA. To suggest that Florida, Kansas, Tennessee and Virginia were on the cusp of making great gains this election cycle is laughable. Even most of the blue states on the list got redder. The one state that had the biggest swing (Vermont) actually decreased in the partisan index from 1996 to 2000.

Further still, the state that came closest to flipping the election for Kerry, Ohio, appeared to be trending the wrong way (again, using the index) and we all know how wrong that is. People in the blogosphere (myself included) have a tendency to obsess over data. It's problematic, however, when someone obsesses over something that isn't even useful.

Amusingly, many candidates obsess over trivial minutia until they talk to a consultant who tells them it doesn't matter and to get back on the phone raising money. MyDD has been very aggresive in what I'd call the "fire the consultants" clique in the blogosphere, and it is no secret that they and many like them aim to be the new consultants hired once the mass layoffs begin. And yes, as invaluable as their blog is for discussion, breaking news and occasionally analysis (I read it all the time), I would have little faith turning over the keys because for all their interest in politics, they don't seem to be able to offer much insight into what to do differently in order to win other than fear and respect the netroots and utilize the partisan index and other ultimately useless measures to determine where to spend resources and how to spend them. Interestingly they also argue that Democrats do not have a national security problem, which to me is laughable.

I apologize for the length of this post, but it's something I've been wanting to write about for some time.

Posted by Chris at April 6, 2005 05:12 PM

Comments

Post a comment

Thanks for signing in, . Now you can comment. (sign out)

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)


Remember me?