Kevin et al.,
I've been ruminating over your lecture comment that it's a bad idea to
normalize your slope coefficients by dividing variable xi by the
standard dev of xi. Could you reiterate your argument in a bit more
detail? Are there times when doing such a thing would be a good idea?
Thanks,
andy
Hi, everyone. We're trying to get things up and running again in the
server computing environment. If anyone's on right now, and could try
typing loading a couple of the following libraries, and sending a note
describing whether they loaded, I'd be much appreciative. If you do so,
please note whether you started a new R session since 5.30pm and loaded
the libraries there, or whether you loaded them into a session extant
prior to 5.30pm. Obviously, knowing about both successes and failures
will be helpfu. The relavant libraries are car, coda, MCMCpack,
multinomRob, mvtnorm, rgenoud, VGAM, Zelig. Thanks!
Ryan
------------------------------------------
Ryan T. Moore ~ Government & Social Policy
Ph.D. Candidate ~ Harvard University
Homepage: http://www.people.fas.harvard.edu/~rtmoore/
Gov1000: http://www.courses.fas.harvard.edu/~gov1000/
For Problem 6(a), we were supposed to use the following (per the
handout),
grsh56
bmp5
invsh55
prights5
for the Barro data, but when I called up the names, I got the following:
> barro <- read.table("Barro_data.txt", row.names=1, header=T)
> names(barro)
[1] "region" "grsh45" "govsh45" "oecd" "civlibb"
"prightssb"
[7] "revcoup" "assassp5" "human85" "humanm85" "humanf85"
anyone know what is going on? Does this have to do with the 2.0.0
installation?
Thanks!
Brodi
Hi all,
So I'm having trouble getting to the car library to get the data to do
this week's problem set; when I try to load the library (using R on my X
desktop) I get the message:
Error in library(car): 'car' is not a valid package -- installed < 2.0.0 ?
which i have never seen before, and I don't remember ever having done
anything different ot get to the library than just typing library(car)
before. I think it may be the result of an error the message for which I
get immediately upon starting R, namely:
Error in library(coda) : 'coda' is not a valid package -- installed <
2.0.0?
Am I the only person having this problem? Can I fix it without downloading
the car library from CRAN and then Secure FXing it into my x desktop (i've
never had to do this before)?
Thanks
Lucy
Hi,
I'm having trouble getting the Barro data into R, I think. Is aved it to
my desktop, dragged it to secureFX and dropped it into my gov1000 folder.
I selected non-text when it copied. And here's my code
> barro <- read.table("barro.dat", row.names=1)
But I don't think the data loaded because:
> barro$grsh56
NULL
Any ideas????
has anyone figured out what's going on with the black market premium
variable in problem 6? the codebook says that it's the black market value
divided by the official value, less one. however, i can't find how either
the black market value or the official value are calculated, and that
seems like important information for substantively interpreting the
results.
any ideas?
thanks,
michael
Hi Everyone,
Some of you had some questions about good sources of background
reading on causal diagrams, direct vs. total effects, etc. Before I
give you some things to look at, let me stress that NONE OF THE
FOLLOWING ARTICLES OR SLIDES HAVE TO BE READ TO DO THE PROBLEM SETS OR
TO DO WELL IN THE COURSE. These articles are just for those students
who are interested in these issues and who want to do some additional
reading.
OK, that being said, the best place to start are with some slides from
talks on causal inference given by Judea Pearl. These slides are
linked off the WWW Links heading on the course website under the
heading Causal Inference.
If you want to know more after reading these slides take a look at the
Spirtes et al. article in Sociological Methods and Research linked off
the Handouts section of the course website. This is a bit more
technical but it is pitched at a social science audience. The next
papers to look at are the Greenland et al. paper in Epidemiology and
the Pearl 2003 paper in Test Journal. Finally, a more technical paper
is the Pearl 1995 Biometrika piece. All of these papers are under the
Handouts section of the course website.
We aren't going to spend much time on causal diagrams in this class so
don't worry if you don't have the time to read these.
Enjoy.
Best,
Kevin
------------------------------------------------------
Kevin Quinn
Assistant Professor
Department of Government and
Center for Basic Research in the Social Sciences
34 Kirkland Street
Harvard University
Cambridge, MA 02138
Hi, everyone. Problem set 6 is now available on the course website. It's
due one week from today, Monday, 22 November.
Best of luck,
Ryan
------------------------------------------
Ryan T. Moore ~ Government & Social Policy
Ph.D. Candidate ~ Harvard University
Homepage: http://www.people.fas.harvard.edu/~rtmoore/
Gov1000: http://www.courses.fas.harvard.edu/~gov1000/