Hi Maya and Iain,
a couple of quick questions about this week's assignment:
1)do we have to turn in our write up and R.file for our review to you in section?To the authors via-email?or both?
2) should we print out the R.code if we did any change?
Thank you.
Best,
Chiara
Dear all,
How do you specify the systematic and stochastic components of a model in
Gary's notation if you are using a multilevel/hierarchical/mixed effects
model?
The model we want to use has a three level structure, comprising both fixed
and random effects. Our independent variables are linear ability, quadratic
ability and school-average ability and dependent variable is maths
self-concept.
In standard notation our model is defined as follows:
Y_ijk = alpha_jk + beta_rjk*X_rijk + epsilon_ijk
where
alpha_jk = gamma_00 + u_0j + v_0jk (intercept)
beta_rjk = gamma_r0 + u_rj + v_rjk (slope)
gamma's represent the fixed component of the intercept and slope.
u's represent the random component of the intercept and slope at the second
level. u~N(0, SIGMA_u)
v's represent the random component of the intercept and slope at the third
level. v~N(0, SIGMA_v)
epsilon is the error in the model as a whole. epsilon~N(0, SIGMA_e)
Note SIGMA is a covariance matrix.
Many thanks,
Nicola and David
Hi again,
I found a way to run cem ( ), so please disregard the earlier question.
Nino Malekovic
MPA Candidate, Class 2011
Harvard Kennedy School
________________________________________
From: gov2001-l-bounces at lists.fas.harvard.edu [gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of gov2001-l-request at lists.fas.harvard.edu [gov2001-l-request at lists.fas.harvard.edu]
Sent: Monday, March 29, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 50
Send gov2001-l mailing list submissions to
gov2001-l at lists.fas.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
or, via email, send a message with subject or body 'help' to
gov2001-l-request at lists.fas.harvard.edu
You can reach the person managing the list at
gov2001-l-owner at lists.fas.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of gov2001-l digest..."
Today's Topics:
1. Re: gov2001-l Digest, Vol 56, Issue 49 (Malekovic, Nino)
2. Re: gov2001-l Digest, Vol 56, Issue 49 (Iain Osgood)
3. Re: gov2001-l Digest, Vol 56, Issue 49 (Yousuf)
----------------------------------------------------------------------
Message: 1
Date: Sun, 28 Mar 2010 16:41:16 -0400
From: "Malekovic, Nino" <nino_malekovic at hks11.harvard.edu>
Subject: Re: [gov2001] gov2001-l Digest, Vol 56, Issue 49
To: "gov2001-l at lists.fas.harvard.edu"
<gov2001-l at lists.fas.harvard.edu>
Message-ID:
<5D4D069369CE1A4984610C041D23C9AF012952155636 at MAIL.hks.internal>
Content-Type: text/plain; charset="us-ascii"
Hi,
I guess my question is for our TF-s.
To what extent we have to help our colleagues improve their papers? Specifically, should we focus on improving their contribution (until now that was mainly replication of original data analysis),
or should we also draw their attention to what seem questionable in author's original analysis?
Any help is appreciated.
Nino Malekovic
MPA Candidate, Class 2011
Harvard Kennedy School
________________________________________
From: gov2001-l-bounces at lists.fas.harvard.edu [gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of gov2001-l-request at lists.fas.harvard.edu [gov2001-l-request at lists.fas.harvard.edu]
Sent: Saturday, March 27, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 49
Send gov2001-l mailing list submissions to
gov2001-l at lists.fas.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
or, via email, send a message with subject or body 'help' to
gov2001-l-request at lists.fas.harvard.edu
You can reach the person managing the list at
gov2001-l-owner at lists.fas.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of gov2001-l digest..."
Today's Topics:
1. logit clustered SEs (Lin, Eric)
----------------------------------------------------------------------
Message: 1
Date: Fri, 26 Mar 2010 20:46:06 -0400
From: "Lin, Eric" <elin at hbs.edu>
Subject: [gov2001] logit clustered SEs
To: Class List for Gov 2001/E-2001 <gov2001-l at lists.fas.harvard.edu>
Message-ID: <75A246F5-57C4-4DCE-B134-DA0B18718805 at hbs.edu>
Content-Type: text/plain; charset="us-ascii"
one follow up:
Like lmr() from the Design package, `gee' doesn't like it when you give it dummies that create collinearity. run glm() first, and see what it dropped to avoid collinearity. Then, drop those from your specification. Run this using gee(), and use the id=`` for specifying what you want to cluster on.
--
ERIC LIN
Technology and Operations Management
Harvard Business School
Boston, MA 02163
elin at hbs.edu
Hi all,
Iain, Thank you for your answer.
There seems to be a delay, sometimes up to a day, in time at which I receive emails from gov2001-I-Digest.
There is another thing that either Maya or you could help me with. Following today's class and Gary's recent (2009) notes on CEM that I found online, I am writing my own code for CEM in R.
We still haven't covered the R code for it, but it makes sense to try it out on our classmates' work before I hand my assignment in on Thursday.
I use
>todrop <- c("nucass", "acquire")
>imbalance(group=In.Nucass$nucass, data=In.Nucass, drop=todrop)
in order to drop the treatment variable (nucass) and the dependent variable (acquire) from the analysis.
That is the code from Gary's notes.
I do that so to prevent matching on those two variables.
However, I get this error report:
Error in which(t1 > 0 & t2 > 0) :
binary operation on non-conformable arrays
In addition: Warning messages:
1: In reduce.var(data[[i]], breaks[[vnames[i]]]) :
NAs introduced by coercion
2: In reduce.var(data[[i]], breaks[[vnames[i]]]) :
NAs introduced by coercion
3: In min(x) : no non-missing arguments to min; returning Inf
4: In max(x) : no non-missing arguments to max; returning -Inf
Could you tell me interpretation of this error report, so I could run the command properly.
I appreciate your help.
Nino Malekovic
MPA Candidate, Class 2011
Harvard Kennedy School
________________________________________
From: gov2001-l-bounces at lists.fas.harvard.edu [gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of gov2001-l-request at lists.fas.harvard.edu [gov2001-l-request at lists.fas.harvard.edu]
Sent: Monday, March 29, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 50
Send gov2001-l mailing list submissions to
gov2001-l at lists.fas.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
or, via email, send a message with subject or body 'help' to
gov2001-l-request at lists.fas.harvard.edu
You can reach the person managing the list at
gov2001-l-owner at lists.fas.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of gov2001-l digest..."
Today's Topics:
1. Re: gov2001-l Digest, Vol 56, Issue 49 (Malekovic, Nino)
2. Re: gov2001-l Digest, Vol 56, Issue 49 (Iain Osgood)
3. Re: gov2001-l Digest, Vol 56, Issue 49 (Yousuf)
----------------------------------------------------------------------
Message: 1
Date: Sun, 28 Mar 2010 16:41:16 -0400
From: "Malekovic, Nino" <nino_malekovic at hks11.harvard.edu>
Subject: Re: [gov2001] gov2001-l Digest, Vol 56, Issue 49
To: "gov2001-l at lists.fas.harvard.edu"
<gov2001-l at lists.fas.harvard.edu>
Message-ID:
<5D4D069369CE1A4984610C041D23C9AF012952155636 at MAIL.hks.internal>
Content-Type: text/plain; charset="us-ascii"
Hi,
I guess my question is for our TF-s.
To what extent we have to help our colleagues improve their papers? Specifically, should we focus on improving their contribution (until now that was mainly replication of original data analysis),
or should we also draw their attention to what seem questionable in author's original analysis?
Any help is appreciated.
Nino Malekovic
MPA Candidate, Class 2011
Harvard Kennedy School
________________________________________
From: gov2001-l-bounces at lists.fas.harvard.edu [gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of gov2001-l-request at lists.fas.harvard.edu [gov2001-l-request at lists.fas.harvard.edu]
Sent: Saturday, March 27, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 49
Send gov2001-l mailing list submissions to
gov2001-l at lists.fas.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
or, via email, send a message with subject or body 'help' to
gov2001-l-request at lists.fas.harvard.edu
You can reach the person managing the list at
gov2001-l-owner at lists.fas.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of gov2001-l digest..."
Today's Topics:
1. logit clustered SEs (Lin, Eric)
----------------------------------------------------------------------
Message: 1
Date: Fri, 26 Mar 2010 20:46:06 -0400
From: "Lin, Eric" <elin at hbs.edu>
Subject: [gov2001] logit clustered SEs
To: Class List for Gov 2001/E-2001 <gov2001-l at lists.fas.harvard.edu>
Message-ID: <75A246F5-57C4-4DCE-B134-DA0B18718805 at hbs.edu>
Content-Type: text/plain; charset="us-ascii"
one follow up:
Like lmr() from the Design package, `gee' doesn't like it when you give it dummies that create collinearity. run glm() first, and see what it dropped to avoid collinearity. Then, drop those from your specification. Run this using gee(), and use the id=`` for specifying what you want to cluster on.
--
ERIC LIN
Technology and Operations Management
Harvard Business School
Boston, MA 02163
elin at hbs.edu
Erin:
I found your suggestions for getting the robust std. errors quite helpful;
thanks. A question: while using the glee package. we have to identify this
argument, called "id." I understand it to be the length of the rows;
correct? I am sure that is not the case since that doesn't seem to be
working. Additionally, I am assuming that for the argument "corstr", it is
best to put "independence" if the model assumes independence?
Thank you for your help
Regards
Yousuf
On Sat, Mar 27, 2010 at 12:00 PM,
<gov2001-l-request at lists.fas.harvard.edu>wrote:
> Send gov2001-l mailing list submissions to
> gov2001-l at lists.fas.harvard.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
> or, via email, send a message with subject or body 'help' to
> gov2001-l-request at lists.fas.harvard.edu
>
> You can reach the person managing the list at
> gov2001-l-owner at lists.fas.harvard.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of gov2001-l digest..."
>
>
> Today's Topics:
>
> 1. logit clustered SEs (Lin, Eric)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 26 Mar 2010 20:46:06 -0400
> From: "Lin, Eric" <elin at hbs.edu>
> Subject: [gov2001] logit clustered SEs
> To: Class List for Gov 2001/E-2001 <gov2001-l at lists.fas.harvard.edu>
> Message-ID: <75A246F5-57C4-4DCE-B134-DA0B18718805 at hbs.edu>
> Content-Type: text/plain; charset="us-ascii"
>
> one follow up:
>
> Like lmr() from the Design package, `gee' doesn't like it when you give it
> dummies that create collinearity. run glm() first, and see what it dropped
> to avoid collinearity. Then, drop those from your specification. Run this
> using gee(), and use the id=`` for specifying what you want to cluster on.
>
>
>
> --
> ERIC LIN
> Technology and Operations Management
> Harvard Business School
> Boston, MA 02163
> elin at hbs.edu
>
Hi,
I guess my question is for our TF-s.
To what extent we have to help our colleagues improve their papers? Specifically, should we focus on improving their contribution (until now that was mainly replication of original data analysis),
or should we also draw their attention to what seem questionable in author's original analysis?
Any help is appreciated.
Nino Malekovic
MPA Candidate, Class 2011
Harvard Kennedy School
________________________________________
From: gov2001-l-bounces at lists.fas.harvard.edu [gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of gov2001-l-request at lists.fas.harvard.edu [gov2001-l-request at lists.fas.harvard.edu]
Sent: Saturday, March 27, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 49
Send gov2001-l mailing list submissions to
gov2001-l at lists.fas.harvard.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
or, via email, send a message with subject or body 'help' to
gov2001-l-request at lists.fas.harvard.edu
You can reach the person managing the list at
gov2001-l-owner at lists.fas.harvard.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of gov2001-l digest..."
Today's Topics:
1. logit clustered SEs (Lin, Eric)
----------------------------------------------------------------------
Message: 1
Date: Fri, 26 Mar 2010 20:46:06 -0400
From: "Lin, Eric" <elin at hbs.edu>
Subject: [gov2001] logit clustered SEs
To: Class List for Gov 2001/E-2001 <gov2001-l at lists.fas.harvard.edu>
Message-ID: <75A246F5-57C4-4DCE-B134-DA0B18718805 at hbs.edu>
Content-Type: text/plain; charset="us-ascii"
one follow up:
Like lmr() from the Design package, `gee' doesn't like it when you give it dummies that create collinearity. run glm() first, and see what it dropped to avoid collinearity. Then, drop those from your specification. Run this using gee(), and use the id=`` for specifying what you want to cluster on.
--
ERIC LIN
Technology and Operations Management
Harvard Business School
Boston, MA 02163
elin at hbs.edu
one follow up:
Like lmr() from the Design package, `gee' doesn't like it when you give it dummies that create collinearity. run glm() first, and see what it dropped to avoid collinearity. Then, drop those from your specification. Run this using gee(), and use the id=`` for specifying what you want to cluster on.
--
ERIC LIN
Technology and Operations Management
Harvard Business School
Boston, MA 02163
elin at hbs.edu
all,
for those of you who are looking to do robust clustered SEs with a logit specification, I found that this package worked for us (finally). Wish I could have done so before thurs, but at least now we've resolved it. Hey Mark, Matt and Alonso, we'll make sure to add this to our existing work . . .
documentation attached. specify a binomial family, just like in glm().
this will create an object called robust.variances. This is a varcov matrix, so just grab what you need along the diag and take the sqrt().
good luck!!!
Hi all,
By now, you should have heard from the group whose paper you are replicating
(and, vice versa, you should have already contacted the folks who are
replicating your paper).
If you haven't, will you please let me know?
Maya
Hey folks,
As promised, I've got your replication-of-replication assignments. (Attached
to this email; I also posted this to the course website.) In almost all
cases, I was able to honor exactly everyone's preferences. (If you didn't
submit preferences, you got matched more randomly.)
IMPORTANT: In lieu of a problem set, your assignment this upcoming week is
to replicate the paper you've been assigned and to draft a memo to the group
with constructive feedback and suggestions. Remember, these are your
colleagues, so you should be respectful and helpful. You should give
feedback on both the replication itself (code as well as substantive
content) as well as suggestion on possible extensions and improvements
(ditto). This memo will be due one week from today, April 1, by 6pm.
ALSO: You should plan on getting in touch via email with the people who are
replicating your paper this evening and sending them the *code, data, and
PDF write up of your replication so far*. Feel free to CC me and Iain on
these emails.
best, Maya
