Note that some people (read: Mostly me and Tiffany once) have sent some
messages to the old mailing list by mistake. They go there but they don't get
sent out. You can see them at:
http://www.fas.harvard.edu/pipermail/gov1760-l/2002-November/date.html#start
and look at the last 5 messages.
If you have advice for Tiffany or an interest in lunch with me or questions
about the paper, then please respond.
Dave
--
David Kane
Lecturer In Government
617-563-0122
dkane(a)latte.harvard.edu
1d:
When you say "something like 'democratic' and 'republican'", i take this to mean
changing 0 to -1? Is there a different interpretation?
1f:
When we set up the factor variable, should we just use year - ie 1910, 1920,
etc.; or should we use numbers like t = 1, 2, 3, 4...? 1d implies there won't
be a difference, but just to make sure. I recall trying this for the midterm
and getting really weird values when I used the year numbers, esp. for the
interaction effect. Is the ordinal ranking all that matters?
2b:
What exactly are you looking for here? The question seems to answer itself in
the parentheses. I'm not sure I understand the wording.
Thanks,
Phillip.
-------------------------------------------------
Phillip Y. Lipscy
Perkins Hall Room #129
35 Oxford Street
Cambridge, MA 02138
(617)493-4893
lipscy(a)fas.harvard.edu
Ph.D. Candidate
Harvard University, FAS, Department of Government
-------------------------------------------------
Thanks, Nirmala. This does help. But what exactly is this "pooled" data that
we rbind together? Is it *only* the data from 1920 and 1918, 1930 and
1928, ..., 1990 and 1988, or should we somehow be including the data from 1912,
1914, 1916, 1922, ..., 1992? If we should be including the 2, 4, 6 -ending
years (Ryan's earlier email implied that maybe we should), I have no idea how
to go about this.
Anna
> By "pooling" I think he means that we combine all the years rather than
> finding the incumbency effect for each election year. In other words,
> instead of running the regression within the loop that merges current year
> with lag year, rbind the merged data into a single dataframe within the
> loop and then run the regression after you have the "pooled" data for all
> the
> years. I hope this helps.....
>
> - Nirmala
I load and clean my data for problem 1(a), and then I'd like to rename the
columns on my data frame. (Just to keep things sorted out, I've made a data
frame that contains only the columns of data I need for the regression, without
extra info like incumbency in years -8 or party of the winner in years -0.)
So, I'd really like to name my columns, to remember what is what. But the
command isn't working.
> clean <- clean.data(hdata)
> dim(clean)
[1] 1250 7
> names(clean) <- c
("year", "state", "district", "dpct0", "dpct8", "dwin8", "incumb0")
Error in "names<-.default"(*tmp*, value = c("year", "state", "district", :
names attribute must be the same length as the vector
I guess I don't have to name my columns, but I'd like to figure out why this
isn't working. It seems like it should, since the dimensions have 7 columns
and I've concatenated 7 names. Any thoughts?
Anna
--
Anna Lorien Nelson
Department of Government,
Harvard University
alnelson(a)fas.harvard.edu
Anna,
I think the "right" should actually be "left," i.e. the dependent variable
is percentage of Dem votes in years ending in "0," while the independent
variables are the percentage of Dem votes and party of the winner two
years prior and incumbency in the current year.
By "pooling" I think he means that we combine all the years rather than
finding the incumbency effect for each election year. In other words,
instead of running the regression within the loop that merges current year
with lag year, rbind the merged data into a single dataframe within the
loop and then run the regression after you have the "pooled" data for all the
years. I hope this helps.....
- Nirmala
On Sat, 30 Nov 2002, Anna Lorien Nelson wrote:
> In 1(a), you ask for a model where the *right-hand* side variable is a vector
> of percentage Dem. votes for districts in 1910, 1920, ..., 1990.
>
> But looking at the equation on the first page, don't you want the right-hand
> side variable to be a vector of percentage Dem. votes in 1918, 1928, ..., 1988
> in order to estimate a vector of predicted values for percentage Dem. vote in
> 1910, 1920, ..., 1990 as the *left-hand* side, dependent variable?
>
> Or am I confused? Just trying to set up the problem correctly.
>
> Anna
>
> --
> Anna Lorien Nelson
> Department of Government,
> Harvard University
> alnelson(a)fas.harvard.edu
>
> _______________________________________________
> gov1000-list mailing list
> gov1000-list(a)fas.harvard.edu
> http://www.fas.harvard.edu/mailman/listinfo/gov1000-list
>
Do I understand correctly that by "pooling" the data in problem 1, we will not
actually be using the data from years ending in 2, 4, or 6 at all?
(or should I be using the data from these years too -- if so, where?)
Anna
--
Anna Lorien Nelson
Department of Government,
Harvard University
alnelson(a)fas.harvard.edu
In 1(a), you ask for a model where the *right-hand* side variable is a vector
of percentage Dem. votes for districts in 1910, 1920, ..., 1990.
But looking at the equation on the first page, don't you want the right-hand
side variable to be a vector of percentage Dem. votes in 1918, 1928, ..., 1988
in order to estimate a vector of predicted values for percentage Dem. vote in
1910, 1920, ..., 1990 as the *left-hand* side, dependent variable?
Or am I confused? Just trying to set up the problem correctly.
Anna
--
Anna Lorien Nelson
Department of Government,
Harvard University
alnelson(a)fas.harvard.edu
Tao has kindly posted many of the midterms here.
http://www.courses.fas.harvard.edu/~gov1000/Students_Midterms_Online/
Although I have not finalized the grades yet, I have no problem in saying that
the midterms by "pink" and "dog" are both excellent.
If, for some reason, you failed to provide Tao with a pdf file of your exam,
please do so at your earliest convenience.
Dave
--
David Kane
Lecturer In Government
617-563-0122
dkane(a)latte.harvard.edu
Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
Dear Dave,
The first challenge I'm finding with homework 7 is that I have no idea if I'm
approaching problems 1 and 2 correctly.
When you commented earlier this week that "more than one person was unable to
do *any* empirical analysis on the midterm," it struck fear in the hearts of
more than one student! :-)
For those who did not tackle the midterm properly, I am concerned that they
will make similar, unwitting mistakes with the homework. Just as each homework
has built on the last one (e.g., requiring correct answers from homework 4 to
make sense of homework 5), so it seems that homework 7 builds on the midterm.
Would you consider you posting a correct answer to the midterm, so the class
can learn from it to more fruitfully approach homework 7?
Thanks,
Anna
P.S. I realize grading our midterms takes time. But if you could just pick one
or two disks with good results to post, that could be a big help.
--
Anna Lorien Nelson
Department of Government,
Harvard University
alnelson(a)fas.harvard.edu