great. here are some comments below...
On Sun, 23 Apr 2006, Justin Ryan Grimmer wrote:
Hey Gov 2001,
Here is our abstract. Please comment/trash/suggest at will!
Cheers,
Justin and Clayton
Turnout Does Not Matter: A Robust Statistical Model for Descriptive
Representation on City Councils
try to make the title reflect your point, your contribution. if i don't
read the abstract i should be able to understand what your contribution
is. its not entirely clear from the current version
An emerging consensus in political science concludes that little connection
exists between voter turnout and representation. Renewing this debate by
examining municipal-level data, Hajnal and Trounstine (2005) find that higher
citywide turnout increases the representation of Latinos and Asian Americans
on city councils.
i think i'd first build up their results before you knock them down. so
explain that given this emerging concensuis, this article has a contrary
result that if correct is quite important, etc...
We overturn these findings using two different
approaches.
you don't get to decide that you're 'overturning' these results. you
only get to say what you did. the rest is an implication. so you can say
all that below for example, but no reason to draw the conclusion
explicitly since that's for readers to judge.
First, using their methods, we show that their result
is an artifact of data
processing errors and the influence of two discrepant observations in their
ordinary least squares regression. When these problems are corrected, higher
turnout is associated with less Asian American representation and has no
relationship to Latino representation, contradicting the authors' main
findings. We then adopt a robust statistical model that avoids the
methodological and theoretical shortcomings of OLS regression and can be
applied to compositional data containing zeroes. Applying our model to an
imputed data set produced according to Hajnal and Trounstine's
specifications, we find that there is no relationship between aggregate voter
turnout and representation of specific racial groups on municipal councils.
all the above is fine from a technical perspective, but i would try to
find a way to do it more subtly. don't make it like you're on the attack
and after them. make yourself the dispassionate scientist who looks for
the facts and reports them just as they are, with no preference over
outcomes. e.g., it just happens that the result contradicts their
results, etc...
Gary
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l