Help = Can somebody who understands maths well explain to me

[Mass]

how they got these results, particularly with Durbin a lot less liberal than Bayh or Feinstein.

I am open to new pseudo-mathematical methods to rank senators, but I cant find the method of ranking and some results seem just arbitrary.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=364&topic_id=730259&mesg_id=730259

[TayTay]

1. Funny you should ask.

I couldn't make heads nor tails of it myself. (But I am not very math-literate.)

Sen. Kerry is around 10th or so. Sounds okay to me. (Maybe they really hated that Radio Marti vote and didn't much care about the Condi vote or the Roberts vote. Who knows?)

[MH1]

2. The algorithm is explained

here:
http://voteview.com/edith.htm

I haven't completely digested it. However my first impression is that they might be picking whichever vote position 50% or more of the minority (i.e. Dems plus Jeffords) took as the "liberal" vote; if so then in some cases they will score a vote as "liberal" that really wasn't. That could explain why there are some rather unexpected results once you get away from the extreme ends of the scale.

But, I could be all wet. :shrug:

[Mass]

3. Thanks.

Apparently, from his answers, the person who posted the thread does not understand any more than I do. He/She continues to post what I can read by myself.

I guess he liked the fact that Feingold was ahead and this is what it is about.

[kerrygoddess]

4. It makes no friggin sense

There are votes that probably shouldn't even be counted in this thing. I think this was posted in a different incarnation the other day too.

[Mass]

5. I have less pb with the one that was posted last week

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=132&topic_id=2520708

You can certainly disagree with the votes that were chosen, but the method is correct and the results reasonnable.

This one has results that are very unreasonable and the method is very weird. Some votes should not even be considered because they do not reflect a Left-Right vote, but a regional vote. For some votes, very liberal and very conservatives agree for very different reasons. Counting all votes and actually dismissing votes that seem different from the average way the voter voted seems the best way to have a weird result. The only way to have a reasonnable result is to make a ranking on the same bills for all (and even that may be questionnable).

[MH1]

6. Here's an alternative

"On The Issues" seems to have a good methodology, and provides a lot of backup material. Scroll to the bottom of each page for the "VoteMatch" graph, and then click on the "explanation of political philosophy" link for the more detailed analysis.

Feingold:
http://www.ontheissues.org/Senate/Russell_Feingold.htm

Kerry:
http://www.ontheissues.org/John_Kerry.htm

It might be worthwhile to put together a post for GD (GD-P?) comparing all the high-profile Dems. Unfortunately I won't have time today - to busy at work for much more than drive-bys these days.

(If you do post a thread - Just move the graphs to your own server rather than linking to the original image - you probably know that but I'd hate to crash their server if it ends up as a popular thread.)