Digital Politics: Pollster.com

There are several sites you should be aware of in the digital political spectrum and Pollster.com should be at the top! Pollster.com is hosted by the Huffington Post and run Mark Blumenthal and it's the absolute "last word" on polling data.

Here's their polling map:

I've Always Thought Yellow Would Be The Color of Battle!

It's nice and interactive. On the tab over you can change the colors yourself for your personal scenarios (WHERE IS THE RON PAUL BUTTON!?). The next tab has all the polls. Let's look at the Presidential Poll graph:

Not Much To See Here ...

So, yeah--it's a nice graph. The smoothing makes it look less jagged than RCP's graph and the dots showing the individual polls are a nice touch even if it looks like some kind of scatter--but that's not the cool part. Here's the cool part:

Look At ALL Those Polls! 

So let's say you want to see just Rasmussen and Gallup? You can take out all the squishy stuff and see the world The Way It Really Is (TM).

Ahh ... I Feel Much Better ...

Or you can look at just the polls from the major networks (CNN, ABC, NBC, CBS, FOX) ...

What THE WHAT?!? ALL SQUISHY SAVE  FOR FOX!!1!

And so on. Also, don't miss the Pollster Blog. It has good analytic stories that play off whatever the recent polling data is.

What Do I Think?

It's important to note that (a) Pollster.com is not making predictions with this data (at least not here, some of their analysis sometimes does, it seems) and (b) they are not "judging" or weighting the polls based on "house effects" or anything. Everyone gets the same weight (as far as I know). It is important to note that they do some things such as group all of one pollster's results for one period so as not to give them more weight for daily polling (at least, so I am told).

In any event, Pollster.com is smart. It links to all the polls so you can check their work. It shows individual states and down-ticket races (I haven't focused on that here) and is the go-to source for polling data. It should be yours.

Exit Question: What does this data look like when you try to take out house-effects or otherwise remove the bias? We'll look at that ... shortly.