OK, when I looked at Valdis Krebs post on the overlap between bundlers, I felt that he hadn’t done a very good job on the numbers. I still feel that way, and spent some time at breakfast today downloading bundler data from the Public Citizen website, and doing some fast analysis on it (honestly, I spent more time converting the HTML to csv). Here is the Excel spreadsheet, so you can play with the data yourself.
Here’s some basic data:
506 unique Obama bundlers
528 unique Kerry bundlers
60 overlaps
1094 total bundlers
5.5% pct overlap/total
10.2% pct Kerry that overlaps
10.6% pct Obama that overlaps
725 unique McCain bundlers
432 unique Bush04 bundlers
125 overlaps (not 128 as Valdis found – don’t know where his data came from)
1282 total bundlers
9.8% pct overlap/total
22.4% pct Bush04 that overlaps
14.7% pct McCain that overlaps
So roughly twice as many GOP bundlers overlapped from Bush04 to McCain. Draw your own conclusions – I’m not sure what I think of that, and whether it’s useful information.
But I’ll add to the dataset over the next week and we’ll play with it some more.
Here’s the simple way to look at it. If you line up 10 McCain bundlers, 8 of them weren’t doing it for Bush and 2 were.
If you line up 10 Obama bundlers,9 of them weren’t doing it for Kerry and 1 was.
I think that’s only meaningful if you’re really looking hard for meaning.
Both ratios are much smaller than I would have expected.