Feeds: 
symbol identifies content for paid subscribers only.
LIGHTNING ROUND: MAYBE EVERYONE'S TOO BUSY ENJOYING SPRING TO COMMIT NEWS.
- The ever-awesome Dahlia Lithwick has a great column in Newsweek about Obama and affirmative action.
- Ben Smith has some interesting commentary on two maps aggregating Obama and Clinton's chances against McCain in all 50 states.
- Bill Clinton gets more than a bit upset when asked about Bill Richardson.
- Clinton has a new ad on the air in Pennsylvania attacking McCain on the economy (which is a good idea), but the reuse of the 3 a.m. format is just distracting. Still, Obama should follow suit. After all, the best way to determine who would be better at running against McCain has got to be seeing both candidates actually do it.
- Chris Cillizza looks at what kind of win Clinton needs in Pennsylvania to stay viable. Meanwhile the Real Clear Politics poll average shows her sliding precipitously although the Pollster.com average (which aggregates more polls for greater accuracy and less sensitivity) does not show nearly as dramatic a trend.
- Finally, just for fun, Mike Gravel performs a Hillary-bashing version of "Helter Skelter." 'Cause why not?
--Sam Boyd
COMMENTS (4)
Let me first be clear that I'm neither a polling expert nor a statistician. But regarding your comment on the relative trends shown by RCP and Pollster, the two graphs are on totally different scales, so they really aren't showing the same thing.
Pollster shows data going way back to early last year, while RCP just picks up the line from the end of February.
So the Pollster chart shows Obama trending upward gradually, then more quickly, while Hillary has a gentle upward trajectory.
But if you mentally zoom in to about the same period shown on the RCP slot, you'll be looking at more or less the right half of the 2008 portion of the graph. And -- well, it's basically just a cloud of data points, with the only real obvious conclusion that Hillary's points are mostly above Obama's.
But the closer we get to the present, the more of Hillary's points are falling below the orange line, and the more of Obama's are reaching above his line.
I bet if you did a regression of just the last 6 week or so of data on Pollster, the graph would look more like RCPs. Though ultimately I agree that it would likely be less dramatic.
But still, I think that the extremely long tail on the Pollster data is obscuring more recent trends (which are toward Obama).
Posted by: ResumeMan | April 2, 2008 8:02 PM
I've got a problem with the whole idea of trendlines in polls. Unless there's some strong reason to believe that last week's trend will also be next week's trend, the best guide, IMHO, is to look at the most current group of polls by reliable pollsters.
Which means that the problem with RCP isn't that they looked at too few polls - it's that they considered one too many, with the incredibly erratic PPP poll being the one that should have been tossed.
A week or so ago, PPP released a poll in NC that had Obama increasing his margin from +1 to +21 in a week. No freakin' way.
Now, PPP releases another poll that has Hillary's margin going from +26 to -2 in just over 2 weeks. That's just about infinitely improbable - I'd be astounded if her real lead diminished by 10%, let alone 28.
Toss out that, and Hillary's ahead 50.0 - 41.3, which is not dramatically different from Pollster.com's 50.9 - 41.0.
Posted by: low-tech cyclist | April 2, 2008 9:43 PM
poor Richardson; sometimes switching sides leaves you between bases.
Posted by: Anonymous | April 2, 2008 11:10 PM
Interesting graph over at Krugman's blog:
http://krugman.blogs.nytimes.com/
It jibes with everything I've ever known.
Posted by: parsec | April 3, 2008 12:44 AM