26 Apr
2006
26 Apr
'06
6:39 a.m.
Hi,
What does exactly "all their multiplicative interactions" mean? Is is
all the possible interactions of pairs of covariates or are we supposed
to allow for higher order intercations of the type: X1*X2*X3*X4?
Juan
26 Apr
26 Apr
7:07 a.m.
A good question. Clearly, you could just keep going--what about
age^4*voted86^2? When we wrote the problem, we envisioned:
* all the covariates
* all of their squared terms
* all of the multiplicative interactions of the original 4 variables
I hope that clarifies things--obviously, it's OK if people also included
cubic terms.
Best,
Dan
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
On Wed, 26 Apr 2006, Juan Vargas wrote:
Hi,
What does exactly "all their multiplicative interactions" mean? Is is
all the possible interactions of pairs of covariates or are we supposed
to allow for higher order intercations of the type: X1*X2*X3*X4?
Juan
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
7:09 a.m.
Dan,
I think Juan was referring to three-way interactions and
four-way interactions, i.e. age * household * democrat.
Holger
Dan Hopkins wrote:
A good question. Clearly, you could just keep
going--what about
age^4*voted86^2? When we wrote the problem, we envisioned:
* all the covariates
* all of their squared terms
* all of the multiplicative interactions of the original 4 variables
I hope that clarifies things--obviously, it's OK if people also included
cubic terms.
Best,
Dan
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
On Wed, 26 Apr 2006, Juan Vargas wrote:
--
Holger Lutz Kern
Graduate Student
Department of Government
Cornell University
Institute for Quantitative Social Science
Harvard University
1737 Cambridge Street N350
Cambridge, MA 02138
www.people.cornell.edu/pages/hlk23
Hi,
What does exactly "all their multiplicative interactions" mean? Is is
all the possible interactions of pairs of covariates or are we supposed
to allow for higher order intercations of the type: X1*X2*X3*X4?
Juan
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
7:13 a.m.
Yes, sorry to be imprecise again, by "all of the multiplicative
interactions of the original 4 variables," I mean the one four-way
interaction and all of the three-way interactions. See Clayton's post.
Best,
Dan
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
On Wed, 26 Apr 2006, Holger Lutz Kern wrote:
Dan,
I think Juan was referring to three-way interactions and
four-way interactions, i.e. age * household * democrat.
Holger
Dan Hopkins wrote:
A good question. Clearly, you could just keep
going--what about
age^4*voted86^2? When we wrote the problem, we envisioned:
* all the covariates
* all of their squared terms
* all of the multiplicative interactions of the original 4 variables
I hope that clarifies things--obviously, it's OK if people also included
cubic terms.
Best,
Dan
----
Ph.D. Student
Department of Government
Harvard University
Tutor, Currier House
dhopkins(a)fas.harvard.edu
http://www.danhopkins.org
On Wed, 26 Apr 2006, Juan Vargas wrote:
--
Holger Lutz Kern
Graduate Student
Department of Government
Cornell University
Institute for Quantitative Social Science
Harvard University
1737 Cambridge Street N350
Cambridge, MA 02138
www.people.cornell.edu/pages/hlk23
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
Hi,
What does exactly "all their multiplicative interactions" mean? Is is
all the possible interactions of pairs of covariates or are we supposed
to allow for higher order intercations of the type: X1*X2*X3*X4?
Juan
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l 
7:11 a.m.
A number of us have interpreted this to mean X1*X2*X3*X4 (including the
lower-order interaction terms, of course). If you run this in MatchIt it
appears to automatically include the lower-order terms.
C
On 4/26/06, Juan Vargas <jfvargas(a)fas.harvard.edu > wrote:
Hi,
What does exactly "all their multiplicative interactions" mean? Is is
all the possible interactions of pairs of covariates or are we supposed
to allow for higher order intercations of the type: X1*X2*X3*X4?
Juan
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
29 Apr
29 Apr
6:14 a.m.
New subject: [gov2001-l] preliminary abstract
Re-evaluating the effect of bonding social capital on the quality of
local representation.
Social capital is widely viewed as the cornerstone of functioning
democratic governance. Countering this consensus, Hill and
Matsubayashi (2005) argues that membership in voluntary organizations
which foster 'bonding relations' is negatively associated with the
quality of democratic representation in a community. Our paper
examines these findings by simulating the sampling variability
inherent in the community-level indices that Hill and Matsubayashi
(2005) aggregate from individual-level data. We find that when taking
into account this measurement error of the community level
information, the models do not allow conclusive findings as claimed.
7:44 a.m.
New subject: [gov2001-l] preliminary abstract
Sounds like it will be a cool paper. Here are my recommendations:
1) The title doesn't give a clear indication of exactly what you
demonstrate--once we have re-evaulated, what conclusion should we draw?
Your title is your first chance to state your contribution, not just your
topic.
You also might be able to 2) consolidate the first two sentences, giving
you more room to focus on your own contribution. Also, 3) you might take
an extra sentence to explain more fully and more generally how exactly you
are contributing--those unfamiliar with the structure of the data might
not understand why the sampling variability is so important. And you
might consider 4) stating exactly what you find in that final sentence--is
it that you find a small/null effect, or that you find the confidence
bounds are too wide to make any concrete statements at all? It sounds
like the latter, but I wasn't sure.
Best,
Dan
On Sat, 29 Apr 2006, Thomas Soehl wrote:
Re-evaluating the effect of bonding social capital on
the quality of
local representation.
Social capital is widely viewed as the cornerstone of functioning
democratic governance. Countering this consensus, Hill and
Matsubayashi (2005) argues that membership in voluntary organizations
which foster 'bonding relations' is negatively associated with the
quality of democratic representation in a community. Our paper
examines these findings by simulating the sampling variability
inherent in the community-level indices that Hill and Matsubayashi
(2005) aggregate from individual-level data. We find that when taking
into account this measurement error of the community level
information, the models do not allow conclusive findings as claimed.
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
30 Apr
30 Apr
4:14 a.m.
New subject: [gov2001-l] preliminary abstract
i take it what you mean is that they compute the se's wrong, you do it
right, when when you do their results go away. in addition to dan's
points, you might try to say something productive at the end, such as more
data are needed to figure out whether the hypotheses are valid. Gary
On Sat, 29 Apr 2006, Dan Hopkins wrote:
Sounds like it will be a cool paper. Here are my
recommendations:
1) The title doesn't give a clear indication of exactly what you
demonstrate--once we have re-evaulated, what conclusion should we draw?
Your title is your first chance to state your contribution, not just your
topic.
You also might be able to 2) consolidate the first two sentences, giving
you more room to focus on your own contribution. Also, 3) you might take
an extra sentence to explain more fully and more generally how exactly you
are contributing--those unfamiliar with the structure of the data might
not understand why the sampling variability is so important. And you
might consider 4) stating exactly what you find in that final sentence--is
it that you find a small/null effect, or that you find the confidence
bounds are too wide to make any concrete statements at all? It sounds
like the latter, but I wasn't sure.
Best,
Dan
On Sat, 29 Apr 2006, Thomas Soehl wrote:
Re-evaluating the effect of bonding social
capital on the quality of
local representation.
Social capital is widely viewed as the cornerstone of functioning
democratic governance. Countering this consensus, Hill and
Matsubayashi (2005) argues that membership in voluntary organizations
which foster 'bonding relations' is negatively associated with the
quality of democratic representation in a community. Our paper
examines these findings by simulating the sampling variability
inherent in the community-level indices that Hill and Matsubayashi
(2005) aggregate from individual-level data. We find that when taking
into account this measurement error of the community level
information, the models do not allow conclusive findings as claimed.
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
6574
days inactive
6578
days old
gov2001@lists.gking.harvard.edu
7 comments
6 participants
participants (6)
-
dhopkins＠fas.harvard.edu -
hlk23＠cornell.edu -
jfvargas＠fas.harvard.edu -
king＠harvard.edu -
nall＠fas.harvard.edu -
thomas_soehl＠ksg07.harvard.edu