Post from May, 2010

We Remember

Monday, 31. May 2010 23:04

Category:PoliSciPundit | Comments Off | Author:

Memorial Day

Monday, 31. May 2010 17:09

Pictures and a thousand words.

Category:PoliSciPundit | Comments Off | Author:

Live Stream of the Oil Spill

Thursday, 27. May 2010 12:03

Can be seen here.

Category:PoliSciPundit | Comments Off | Author:

Wrong Turn…

Wednesday, 19. May 2010 17:18

obama/car

Category:PoliSciPundit | Comments Off | Author:

Selecting on the Dependent Variable?

Wednesday, 19. May 2010 16:53

Barry Friedman has written a book on the Supreme Court and the public called The Will of the People: How Public Opinion Has Influenced the Supreme Court and shaped the Meaning of the Constitution. His argument is that the Court follows public opinion for the most part. However, the cases he looks at are limited to salient cases. A participant in a chat on Barry’s book raises the issue of selection bias:

[Comment From Adam Liptak, NYTAdam Liptak, NYT: ]
You say your claim is limited to “salient” issues and I wonder if you might say something about what you mean by the term. Does the principle of salience allow you to cherry pick examples that work for your theory?

Tuesday May 18, 2010 2:05 Adam Liptak, NYT
2:05 Barry: what good’s a theory, adam, if its author can’t cherry pick the cases? :)

Tuesday May 18, 2010 2:05 Barry
2:06 Barry: i think though that most salient examples fit the theory. but there are exceptions and the real value of the theory is figuring out what is different about them. so, the first amendment is an area in which the court often bucks the tide; think of school prayer, flag burning . . .

Barry’s answer isn’t particularly satisfactory. The problem with accepting his evidence on the Supreme Court and its sensititivity to public opinion is that he has selected cases that…he expects the public to be very sensitive to. IOW, he has selected his cases on his dependent variable. You aren’t supposed to select cases that “fit” your theory. Rather, you’re supposed to select cases *representative* of the population from which you are estimating parameters (or qualitiatively assessing select examples). Then that sample can serve as a database with which you can *test* your thesis to see if it *does* fit the selected cases. Choosing cases because you think they fit your thesis is classically bad research design. Confirmation bias in sampling. I haven’t read the book and so I can’t speak to the rest of it, but this is a fairly bad mistake to make in designing his inquiry, and hence I would hesitate to rely on his evidence as confirmation of his thesis…a thesis I am sympathetic to.

Category:PoliSciPundit | Comments Off | Author:

Greatest Political Ad Ever?

Monday, 17. May 2010 6:36

Awesome.

Category:PoliSciPundit | Comments Off | Author:

Enviornmentalism

Friday, 7. May 2010 8:33

Errr…

Awww. Somebody needs a hug. Too bad trees aren’t very good at hugging. ;)

Category:PoliSciPundit | Comments Off | Author:

%d bloggers like this: