So after loading "fearondata.dta" (using the read.dta command as someone
pointed out earlier), what command can I use to call its columns and such?
Does read.dta() load the State file into R as a matrix?
Thanks!
Hi again Iain,
I seem to be the master of asking the obvious. Let's blame the fact that I am an international student for that :)
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 08, 2010 9:40 AM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 13
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 11 (Malekovic, Nino)
2. Re: gov2001-l Digest, Vol 56, Issue 11 (Iain Osgood)
----------------------------------------------------------------------
Message: 1
Date: Sun, 7 Mar 2010 22:00:13 -0500
From: "Malekovic, Nino" <nino_malekovic at hks11.harvard.edu>
Subject: Re: [gov2001] gov2001-l Digest, Vol 56, Issue 11
To: "gov2001-l at lists.fas.harvard.edu"
<gov2001-l at lists.fas.harvard.edu>
Message-ID:
<5D4D069369CE1A4984610C041D23C9AF0129506232E3 at MAIL.hks.internal>
Content-Type: text/plain; charset="us-ascii"
Hi all,
Could someone tell me what causes R to report: "Error in family$linkfun(mustart) : Value 0 out of range (0, 1)"?
I run the command "rare <- zelig(onset ~ warl + gdpenl + lpopl1 + lmtnest + ncontig + Oil + nwstate + instab + polity2l + ethfrac + relfrac, model = "relogit", data = data1)", where data1 is the dataframe that I created from the original fearondata matrix. After I run the command, I get "error in family$linkfun(mustart) : Value 0 out of range (0, 1)".
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: Sunday, March 07, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 11
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. estimating fundamental and estimation uncertainty (Yousuf)
2. Re: estimating fundamental and estimation uncertainty
(Iain Osgood)
3. replication software (Maya Sen)
----------------------------------------------------------------------
Message: 1
Date: Sat, 6 Mar 2010 20:10:48 -0500
From: Yousuf <usuf.marvi at gmail.com>
Subject: [gov2001] estimating fundamental and estimation uncertainty
To: gov2001-l at lists.fas.harvard.edu
Message-ID:
<27c64f031003061710p21e4bea5u3bde7deea0f5312e at mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Dear Iain, Maya, and Professor King:
>From what I understand from the class lecture is that we use a chosen
parametric value, given the MLE estimate that we get from our model, to
generate some simulations (lets say 1000) of our systematic and stochastic
variables; correct?
My question, given my understanding, is that why do we need to reduce
uncertainty by increasing the number of simulations, if this uncertainty is
recorded by the standard error and the error term. Additionally, I
understand that the std.error is associated with one value of the systematic
parameters, and hence, if we get a different a better estimate of those
parameters, we would then get a different value for std. errors. However,
what is the likelihood that value of beta doesn't fall within the range of
std. error?
Regards
--
Yousuf
Thanks Iain, I corrected it. Perhaps you would care to tell me if I understood the first part of the problem 2 correctly. We are supposed to start solving it by establishing median covariate values for countries where civil war onset is equal to one only over the civil war years. In other words, in calculating our median covariate values, we use two criteria: onset equal to one and years that onset was equal to one. Could you confirm that? I appreciate your effort.
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 08, 2010 9:40 AM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 13
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 11 (Malekovic, Nino)
2. Re: gov2001-l Digest, Vol 56, Issue 11 (Iain Osgood)
----------------------------------------------------------------------
Message: 1
Date: Sun, 7 Mar 2010 22:00:13 -0500
From: "Malekovic, Nino" <nino_malekovic at hks11.harvard.edu>
Subject: Re: [gov2001] gov2001-l Digest, Vol 56, Issue 11
To: "gov2001-l at lists.fas.harvard.edu"
<gov2001-l at lists.fas.harvard.edu>
Message-ID:
<5D4D069369CE1A4984610C041D23C9AF0129506232E3 at MAIL.hks.internal>
Content-Type: text/plain; charset="us-ascii"
Hi all,
Could someone tell me what causes R to report: "Error in family$linkfun(mustart) : Value 0 out of range (0, 1)"?
I run the command "rare <- zelig(onset ~ warl + gdpenl + lpopl1 + lmtnest + ncontig + Oil + nwstate + instab + polity2l + ethfrac + relfrac, model = "relogit", data = data1)", where data1 is the dataframe that I created from the original fearondata matrix. After I run the command, I get "error in family$linkfun(mustart) : Value 0 out of range (0, 1)".
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: Sunday, March 07, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 11
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. estimating fundamental and estimation uncertainty (Yousuf)
2. Re: estimating fundamental and estimation uncertainty
(Iain Osgood)
3. replication software (Maya Sen)
----------------------------------------------------------------------
Message: 1
Date: Sat, 6 Mar 2010 20:10:48 -0500
From: Yousuf <usuf.marvi at gmail.com>
Subject: [gov2001] estimating fundamental and estimation uncertainty
To: gov2001-l at lists.fas.harvard.edu
Message-ID:
<27c64f031003061710p21e4bea5u3bde7deea0f5312e at mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Dear Iain, Maya, and Professor King:
>From what I understand from the class lecture is that we use a chosen
parametric value, given the MLE estimate that we get from our model, to
generate some simulations (lets say 1000) of our systematic and stochastic
variables; correct?
My question, given my understanding, is that why do we need to reduce
uncertainty by increasing the number of simulations, if this uncertainty is
recorded by the standard error and the error term. Additionally, I
understand that the std.error is associated with one value of the systematic
parameters, and hence, if we get a different a better estimate of those
parameters, we would then get a different value for std. errors. However,
what is the likelihood that value of beta doesn't fall within the range of
std. error?
Regards
--
Yousuf
Hi all,
Could someone tell me what causes R to report: "Error in family$linkfun(mustart) : Value 0 out of range (0, 1)"?
I run the command "rare <- zelig(onset ~ warl + gdpenl + lpopl1 + lmtnest + ncontig + Oil + nwstate + instab + polity2l + ethfrac + relfrac, model = "relogit", data = data1)", where data1 is the dataframe that I created from the original fearondata matrix. After I run the command, I get "error in family$linkfun(mustart) : Value 0 out of range (0, 1)".
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: Sunday, March 07, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 11
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. estimating fundamental and estimation uncertainty (Yousuf)
2. Re: estimating fundamental and estimation uncertainty
(Iain Osgood)
3. replication software (Maya Sen)
----------------------------------------------------------------------
Message: 1
Date: Sat, 6 Mar 2010 20:10:48 -0500
From: Yousuf <usuf.marvi at gmail.com>
Subject: [gov2001] estimating fundamental and estimation uncertainty
To: gov2001-l at lists.fas.harvard.edu
Message-ID:
<27c64f031003061710p21e4bea5u3bde7deea0f5312e at mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Dear Iain, Maya, and Professor King:
>From what I understand from the class lecture is that we use a chosen
parametric value, given the MLE estimate that we get from our model, to
generate some simulations (lets say 1000) of our systematic and stochastic
variables; correct?
My question, given my understanding, is that why do we need to reduce
uncertainty by increasing the number of simulations, if this uncertainty is
recorded by the standard error and the error term. Additionally, I
understand that the std.error is associated with one value of the systematic
parameters, and hence, if we get a different a better estimate of those
parameters, we would then get a different value for std. errors. However,
what is the likelihood that value of beta doesn't fall within the range of
std. error?
Regards
--
Yousuf
Ninoslav 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: Sunday, March 07, 2010 12:00 PM
To: gov2001-l at lists.fas.harvard.edu
Subject: gov2001-l Digest, Vol 56, Issue 11
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. estimating fundamental and estimation uncertainty (Yousuf)
2. Re: estimating fundamental and estimation uncertainty
(Iain Osgood)
3. replication software (Maya Sen)
----------------------------------------------------------------------
Message: 1
Date: Sat, 6 Mar 2010 20:10:48 -0500
From: Yousuf <usuf.marvi at gmail.com>
Subject: [gov2001] estimating fundamental and estimation uncertainty
To: gov2001-l at lists.fas.harvard.edu
Message-ID:
<27c64f031003061710p21e4bea5u3bde7deea0f5312e at mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"
Dear Iain, Maya, and Professor King:
>From what I understand from the class lecture is that we use a chosen
parametric value, given the MLE estimate that we get from our model, to
generate some simulations (lets say 1000) of our systematic and stochastic
variables; correct?
My question, given my understanding, is that why do we need to reduce
uncertainty by increasing the number of simulations, if this uncertainty is
recorded by the standard error and the error term. Additionally, I
understand that the std.error is associated with one value of the systematic
parameters, and hence, if we get a different a better estimate of those
parameters, we would then get a different value for std. errors. However,
what is the likelihood that value of beta doesn't fall within the range of
std. error?
Regards
--
Yousuf
Thank you Iain. Your explanation has almost trivialized my questions
regarding, ("what is the likelihood that value of beta doesn't fall within
the range of std. error?"). I will follow up with the rest during office
hours.
Thank you for your help.
Yousuf
On Sun, Mar 7, 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. estimating fundamental and estimation uncertainty (Yousuf)
> 2. Re: estimating fundamental and estimation uncertainty
> (Iain Osgood)
> 3. replication software (Maya Sen)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 6 Mar 2010 20:10:48 -0500
> From: Yousuf <usuf.marvi at gmail.com>
> Subject: [gov2001] estimating fundamental and estimation uncertainty
> To: gov2001-l at lists.fas.harvard.edu
> Message-ID:
> <27c64f031003061710p21e4bea5u3bde7deea0f5312e at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Dear Iain, Maya, and Professor King:
>
> >From what I understand from the class lecture is that we use a chosen
> parametric value, given the MLE estimate that we get from our model, to
> generate some simulations (lets say 1000) of our systematic and stochastic
> variables; correct?
>
> My question, given my understanding, is that why do we need to reduce
> uncertainty by increasing the number of simulations, if this uncertainty is
> recorded by the standard error and the error term. Additionally, I
> understand that the std.error is associated with one value of the
> systematic
> parameters, and hence, if we get a different a better estimate of those
> parameters, we would then get a different value for std. errors. However,
> what is the likelihood that value of beta doesn't fall within the range of
> std. error?
>
> Regards
>
> --
> Yousuf
>
Hi all,
A few students have had questions about this, so we'd thought we'd
clarify. For the strict replicating part of the assignment, you should
use R to replicate the authors' results. Since everyone will be
swapping their code with others the week after Spring Break (which is
soon!), we need everyone to use the same common software and all of
you have, at this point, used R pretty extensively.
We also recognize that you might need to use Stata or something else
to figure out what others have done. So it's ok to use Stata/something
else to explore your authors' conclusions and results (as long as you
do a replication in R for your colleagues to replicate your results
with), and also to use other programs to extend your replication into
the final paper. We personally think R is the best for this, and you
might end up using R even if you set out to use Stata or something
else.
Lastly, we should note that you need to make SOMETHING replicable, not
necessarily the entirety of the authors' article. You may, for
example, focus on replicating and extending the set of conclusions
that are the most important. You're the ones who get to choose what's
most important, not the original author.
hope that helps--
Maya
Dear Iain, Maya, and Professor King:
>From what I understand from the class lecture is that we use a chosen
parametric value, given the MLE estimate that we get from our model, to
generate some simulations (lets say 1000) of our systematic and stochastic
variables; correct?
My question, given my understanding, is that why do we need to reduce
uncertainty by increasing the number of simulations, if this uncertainty is
recorded by the standard error and the error term. Additionally, I
understand that the std.error is associated with one value of the systematic
parameters, and hence, if we get a different a better estimate of those
parameters, we would then get a different value for std. errors. However,
what is the likelihood that value of beta doesn't fall within the range of
std. error?
Regards
--
Yousuf
I seem to remember from long ago (before the land of MLE) that heteroskedasticity would affect your SEs, but not your coef. estimates. As such, I was wondering if we should expect our coefs to be the same for problem 1 and problem 3 of our psets, since the diff is only that we are adding non-constant variance . .
thoughts?
EXL
