Hi All,
The papers and final grades are now available in my mailbox, which is
outside N-354 in CGIS. Each group has a manilla envelope with its paper,
as well as separate envelopes for each individual containing final grades
for the semester. Each paper has comments from Gary and from one of the
TFs.
Off-site extension school students will receive their paper comments and
grades shortly via email.
Congratulations to all on completing the course--have a great summer.
Best,
Dan and Ian
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
Hi everyone,
I know we said we would stop pestering you about the course evaluations,
but we have been asked by the powers that be to remind you that they are
only open for another day or two. You can access them by logging into
my.harvard.edu, and then clicking on "Courses."
You really would help us out by completing them - so hopefully you can
find a few minutes tonight if you haven't already done so.
Many thanks,
Ian
our apologies for being so late.
------------
Market size, innovation and the spectre of endogeneity
Melinda Elias and Aram Harrow
Endogeneity often plagues attempts to find the effect of market size
on innovation. One possible solution (due to Acemoglu and Linn,
2004) is to measure pharmaceutical innovation by new drug approvals
and to estimate market size based on exogenous demographic trends.
We show that this approach is extremely sensitive to (possibly
endogenous) changes over time in the age profiles of markets for
different drug categories. For example, a change of less than 0.2% in
drug usage patterns could be chosen to either eliminate or triple
the estimated effects of market size on new drug approvals. This
sensitivity is a consequence of the strong colinearity between
market size and the time and drug fixed effects.
----------
the title is so that the first sentence of the paper can be "A spectre
is haunting economics---the spectre of endogeneity." if the crowd
judges this joke too corny, it can be dropped.
I'm on vacation on Friday May 11 and will not be reading my mail for a
while.
I will respond to your message when I return on Monday May 15.
Please email dataquest(a)help.hmdc.harvard.edu if you have statistical
consulting questions.
Thanks!
Lynda
Ranking Documents with Wordscores
In recent years, corporate juggernauts such as Google, have amassed enormous
fortunes by making use of relatively simple document processing algorithms. A
huge body of literature has sprung forth detailing techniques for information
retrieval, document ranking and classification of large sequences and vectors,
and countless algorithms have been developed to take advantage rapid
developments in this ever-changing realm. One such algorithm is Wordscores, a
simple document ranking algorithm recently developed by Michael Laver, Kenneth
Benoit and John Garry for extracting relative political policy positions from
documents. Despite its simplicity Wordscores performs remarkably well at
ranking political documents when compared to a variety of classic document
ranking algorithms. In this paper, we examine the performance of Wordscores
and illustrate its superiority at ranking political texts, proposing extensions
for improving its performance, detailing its assumptions, and specifying
conditions which must be met in order to use it to make substantiative claims.
Hi everyone. We're working on a new version of Amelia, one that
supposedly never breaks and works for much larger datasets too. We're
also planning to add some other features (for time series, cross-sectional
data sets) but this part isn't working yet. If anyone would like to give
it a try, the info is below. My coauthors and I would very much
appreciate hearing from you if you have an example that causes this
version of Amelia to fail. Want to give it a try?
You can download our preliminary version at:
http://www.ssc.ucla.edu/06S/polisci6-1/amelia.zip
Included in this are the source files, a windows installer (setup.exe), the
standard R package install, a .tar package for linux users, and a rough,
preliminary version of the manual.
Many thanks,
Gary
Hi All,
Thanks to everyone who has now filled out an online evaluation of Gov.
2001 at my.harvard.edu (by clicking on the "courses" tab and then on
"evaluate" next to Gov. 2001).
For those who haven't, here's our final plea: you could win an IPOD for
participating, and will certainly help us through your feedback, too. So
please take a moment to evaluate the course... Thanks.
Best,
Dan
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
in case anyone is interested...
Gary
---------- Forwarded message ----------
Date: Mon, 8 May 2006 17:02:03 -0400
From: "Small, Dylan" <dsmall(a)wharton.upenn.edu>
To: James Robins <robins(a)hsph.harvard.edu>,
Alan Zaslavsky <zaslavsk(a)hcp.med.harvard.edu>, King(a)Harvard.edu,
rubin(a)stat.harvard.edu
Subject: causal conference
Hi Alan, Don, Gary and Jamie,
We're hosting a Mid-Atlantic causal modeling conference at Penn on
June 27th. It might be a long way from Harvard for one day but I've
attached the flyer in case you'd interested be attending. Also, if any
students at Harvard would be interested in attending, note that there is
a poster session that they might want to participate in.
Best,
Dylan
------------------------------------------------------------------------
-------
Dylan Small, Ph.D.
Assistant Professor
Department of Statistics
The Wharton School, University of Pennsylvania
400 Huntsman Hall, 3730 Walnut St.
Philadephia, PA 19104
Our class party will be Saturday May 13th at noon at my house, 9 Beecher
Road, in Brookline. Please bring a guest too. The Gov Dept over the
years has called guests: 'wives', 'spouses', 'partners', 'significant
others', and believe it or not 'lovers'; anyone in any of these or other
categories of your choosing are invited! Kids are welcome too; mine's 10.
It'll be entirely informal.
Directions are at http://gking.harvard.edu/directions
Please RSVP (and tell me how many are coming) as soon as you can, so we
can get a count for food.
Many thanks. I hope to see everyone there.
Gary