13 Feb
2009
13 Feb
'09
1:52 p.m.
Hi Folks,
I encourage you all to take a look at the solution set (especially the R
code) before getting too far into working on ps2. Although many of you were
able to successfully complete ps1, there are more efficient ways to answer
many of the questions in ps1 (e.g. using apply() rather than for() ). Neat
and concise code is generally always better.
Miya
--
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
13 Feb
13 Feb
8:26 p.m.
Thank you for posting the solutions. It's nice to see that some of our
kludgey workarounds are indeed unnecessary.
At the risk of sounding like a noodge, I thought that the formula for the
sigma-squared regression parameter (which we had to estimate in 2.1) had
(n-k-1) in the denominator, not (n-k). Or is it that it should only be
(n-k-1) when calculating the standard errors for the coefficient estimates?
John-Paul Ferguson
PhD Candidate, Economic Sociology
MIT Sloan School of Management
50 Memorial Drive, E52-533
Cambridge, MA 02142
http://web.mit.edu/jpferg/www
617.253.3940 (w)
617.549.8482 (c)
On Fri, Feb 13, 2009 at 1:52 PM, Miya Woolfalk <woolfalk at fas.harvard.edu>wrote:
Hi Folks,
I encourage you all to take a look at the solution set (especially the R
code) before getting too far into working on ps2. Although many of you were
able to successfully complete ps1, there are more efficient ways to answer
many of the questions in ps1 (e.g. using apply() rather than for() ). Neat
and concise code is generally always better.
Miya
--
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
8:30 p.m.
An HTML attachment was scrubbed...
URL:
http://lists.fas.harvard.edu/mailman/private/gov2001-l/attachments/20090213…
9:11 p.m.
Point well taken.
John-Paul
On Fri, Feb 13, 2009 at 8:30 PM, Gary King <king at harvard.edu> wrote:
the "-1" makes it unbiased (on average
right even in small samples);
without it its only consistent. but if your results depend on the
difference, then its surely time to go get some more data!
Gary
---
http://gking.harvard.edu
On 2/13/2009 8:26 PM, John-Paul Ferguson wrote:
Thank you for posting the solutions. It's nice to see that some of our
kludgey workarounds are indeed unnecessary.
At the risk of sounding like a noodge, I thought that the formula for the
sigma-squared regression parameter (which we had to estimate in 2.1) had
(n-k-1) in the denominator, not (n-k). Or is it that it should only be
(n-k-1) when calculating the standard errors for the coefficient estimates?
John-Paul Ferguson
PhD Candidate, Economic Sociology
MIT Sloan School of Management
50 Memorial Drive, E52-533
Cambridge, MA 02142
http://web.mit.edu/jpferg/www
617.253.3940 (w)
617.549.8482 (c)
On Fri, Feb 13, 2009 at 1:52 PM, Miya Woolfalk <woolfalk at fas.harvard.edu>wrote:
Hi Folks,
I encourage you all to take a look at the solution set (especially the R
code) before getting too far into working on ps2. Although many of you were
able to successfully complete ps1, there are more efficient ways to answer
many of the questions in ps1 (e.g. using apply() rather than for() ). Neat
and concise code is generally always better.
Miya
--
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
------------------------------
_______________________________________________
gov2001-l mailing listgov2001-l at
lists.fas.harvard.eduhttp://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
14 Feb
14 Feb
3 a.m.
Also, you need to be careful about what you mean by "k". In many textbooks,
the denominator has the formula as n-(k+1) where k is the number of
covariates so the +1 is for the intercept term. In the R code for our
solution, our X matrix already has a column of 1s so when we subtract
ncol(X), we are actually subtracting k+1 rather than k, if k is the number
of x variables. So our code is actually consistent with n-(k+1).
On Fri, Feb 13, 2009 at 9:11 PM, John-Paul Ferguson <jpferg at mit.edu> wrote:
Point well taken.
John-Paul
On Fri, Feb 13, 2009 at 8:30 PM, Gary King <king at harvard.edu> wrote:
--
Patrick Lam
Department of Government and Institute for Quantitative Social Science,
Harvard University
http://www.people.fas.harvard.edu/~plam
the "-1" makes it unbiased (on average
right even in small samples);
without it its only consistent. but if your results depend on the
difference, then its surely time to go get some more data!
Gary
---
http://gking.harvard.edu
On 2/13/2009 8:26 PM, John-Paul Ferguson wrote:
Thank you for posting the solutions. It's nice to see that some of our
kludgey workarounds are indeed unnecessary.
At the risk of sounding like a noodge, I thought that the formula for the
sigma-squared regression parameter (which we had to estimate in 2.1) had
(n-k-1) in the denominator, not (n-k). Or is it that it should only be
(n-k-1) when calculating the standard errors for the coefficient estimates?
John-Paul Ferguson
PhD Candidate, Economic Sociology
MIT Sloan School of Management
50 Memorial Drive, E52-533
Cambridge, MA 02142
http://web.mit.edu/jpferg/www
617.253.3940 (w)
617.549.8482 (c)
On Fri, Feb 13, 2009 at 1:52 PM, Miya Woolfalk <woolfalk at fas.harvard.edu>wrote:
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
Hi Folks,
I encourage you all to take a look at the solution set (especially the R
code) before getting too far into working on ps2. Although many of you were
able to successfully complete ps1, there are more efficient ways to answer
many of the questions in ps1 (e.g. using apply() rather than for() ). Neat
and concise code is generally always better.
Miya
--
Miya Woolfalk
Ph.D. Student
Harvard University
Government and Social Policy
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
------------------------------
_______________________________________________
gov2001-l mailing listgov2001-l at
lists.fas.harvard.eduhttp://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
5571
days inactive
5572
days old
gov2001@lists.gking.harvard.edu
4 comments
4 participants
participants (4)
-
jpferg＠mit.edu -
king＠harvard.edu -
plam＠fas.harvard.edu -
woolfalk＠fas.harvard.edu