Tuesday, March 09, 2010

Andrew Sullivan Doesn't Understand Polling Data *sigh*

Really, this is unacceptable in a political journalist: Health Reform Gaining

As this poll of polls shows, if Obama keeps up his performances like yesterday and campaigns like this day after day after day, the first poll - from YouGov/The Economist - showing a clear 53- 47 majority in favor of his reform will become a harbinger of the future.

Got that? A "clear majority," so says Sullivan. The trouble is when you actually look at the poll you discover it has an error term of +/- 3.6%, which means that the poll numbers of 53% and 47% are not statistically different from one another. Oops.

Well, if you cannot say Obama's plan has a majority backing it, you can at least say it is as popular as not, right? Well, yes and no. When you look at the data you discover that support for the plan is "soft." Of those people asked, only 17% of respondents said they "Strongly Support" the plan. 32% "Strongly Oppose" Obama's scheme. What support the Obama plan has is far shakier than the opposition.

Well, maybe the plan is increasing in support from where it had been polling? That would be something we measure with a tracking poll. Sadly, the YouGov poll was not tracking the Obama health plan.

Gallup, on the other hand, has been doing tracking polls on the question, and, Lo! and Behold!, they too released results today. (Sully seems to have missed this tracking poll. Maybe he has never heard of Gallup.)

March 2010 - Support for Obama Plan

For: 45%
Against: 48%

January 2010 - Support for Obama Plan

For: 49%
Against: 46%

The +/- on the Gallup polls is 4 points, which means the numbers are essentially and statistically unchanged.

Really, none of this is very difficult to understand. So why can't Sullivan understand it?


Larry Sheldon said...

I don't care what Andrew Sullivan says, and I don't think he understands anything that doesn't involve KY Jelly.

I don't continue to read blogs that continue to obsess over him and to drive traffic to him.

Ain't going there.

The Iconic Midwesterner said...


Hmmm...let's see... a quick search of the IMW archives shows:

2010: This post on Sully.
2009: Zero Sully posts.
2008: Two Sully posts.
2007: One post where Sully is mentioned. (I thank him for linking to a piece of mine he liked.)
2006: Zero posts.
2005: One post where I use Sully as an example of someone trying to make "Catholicism safe for liberal Democrats." He's linked to, not discussed.
2004: Zero posts.

So that makes five posts in 6+ years (or five posts out of 1666) that even mention Sully.

I think our definitions of "obsess" are a little different.