20 May
2010
20 May
'10
2:04 p.m.
I'm having trouble getting Zelig to run simulations from a multinomial
logit model. The DV is called "defhow2" and has five categories, and
I'm running the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+osc
Xginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the
following error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this
error message, so I eliminated all of my missing data just to see if I
could get it to work in principle. No luck-same message. Does anyone
know what might be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org <http://www.ppic.org>
Any opinions expressed in this message are those of the author alone and
do not necessarily reflect any position of the Public Policy Institute
of California.
20 May
20 May
2:26 p.m.
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I’m having trouble getting Zelig to run simulations
from a multinomial logit
model. The DV is called “defhow2” and has five categories, and I’m running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck—same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
3:22 p.m.
Thanks Olivia. I've been poking around and it seems like the problem is connected to
the IVs I'm including. If I pare the model way down, it works. Is it possible
I'm overloading my data, and the mlogit function is just dropping values of the DV
where I don't have enough degrees of freedom? The full model runs in Stata, but
Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
3:31 p.m.
its probably collinearity. run a regression; if that works (with all
the variables included), the mlogit will usually work.
stats runs anything but it does that by dropping variables. so unless
you're watching closely, you wind up estimating a model you didn't ask
it to run.
Gary
---
http://gking.harvard.edu
On Thu, May 20, 2010 at 3:22 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
Thanks Olivia. I've been poking around and it
seems like the problem is connected to the IVs I'm including. If I pare the model
way down, it works. Is it possible I'm overloading my data, and the mlogit function
is just dropping values of the DV where I don't have enough degrees of freedom? The
full model runs in Stata, but Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
3:32 p.m.
Hi, Eric.
It looks like the model will "run" in R as well since it makes it
through the zelig step. Why don't you check to see if all of the IVs
that you include are output from the model-fitting step (e.g., do
summary on the output from zelig)? If some are being dropped, it
probably indicates that those IVs are collinear with each other (and
shouldn't be included anyway).
Yours,
Olivia
On Thu, May 20, 2010 at 3:22 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
Thanks Olivia. I've been poking around and it
seems like the problem is connected to the IVs I'm including. If I pare the model
way down, it works. Is it possible I'm overloading my data, and the mlogit function
is just dropping values of the DV where I don't have enough degrees of freedom? The
full model runs in Stata, but Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
6:25 p.m.
I just ran the full model--first as OLS, and then as mlogit--and it doesn't look like
it drops any variables. And some very simple models--for example, just ideology and party
ID as IVs--can produce the error message. I'm not wedded to this model; it was very
preliminary, and I can plausibly recode the DV to make it binary and run logit. (When I I
do that the full model works.) But I'm curious what the problem is!
Thanks for your help. If you have any further insights or ideas, let me know.
Best,
Eric
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 12:32 PM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
It looks like the model will "run" in R as well since it makes it
through the zelig step. Why don't you check to see if all of the IVs
that you include are output from the model-fitting step (e.g., do
summary on the output from zelig)? If some are being dropped, it
probably indicates that those IVs are collinear with each other (and
shouldn't be included anyway).
Yours,
Olivia
On Thu, May 20, 2010 at 3:22 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
Thanks Olivia. I've been poking around and it
seems like the problem is connected to the IVs I'm including. If I pare the model
way down, it works. Is it possible I'm overloading my data, and the mlogit function
is just dropping values of the DV where I don't have enough degrees of freedom? The
full model runs in Stata, but Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
7:59 p.m.
for logit, there are only 2 things that can cause it to fail:
collinearity and perfect discrimination.
you're using mlogit, and so you can add to that that each of these can
be true for any of the individual equations. If you have very rare
events in one of them, this can happen easily.
Gary
---
http://gking.harvard.edu
On Thu, May 20, 2010 at 6:25 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I just ran the full model--first as OLS, and then as
mlogit--and it doesn't look like it drops any variables. And some very simple
models--for example, just ideology and party ID as IVs--can produce the error message.
I'm not wedded to this model; it was very preliminary, and I can plausibly recode
the DV to make it binary and run logit. (When I I do that the full model works.) But
I'm curious what the problem is!
Thanks for your help. If you have any further insights or ideas, let me know.
Best,
Eric
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 12:32 PM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
It looks like the model will "run" in R as well since it makes it
through the zelig step. Why don't you check to see if all of the IVs
that you include are output from the model-fitting step (e.g., do
summary on the output from zelig)? If some are being dropped, it
probably indicates that those IVs are collinear with each other (and
shouldn't be included anyway).
Yours,
Olivia
On Thu, May 20, 2010 at 3:22 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
Thanks Olivia. I've been poking around and
it seems like the problem is connected to the IVs I'm including. If I pare the model
way down, it works. Is it possible I'm overloading my data, and the mlogit function
is just dropping values of the DV where I don't have enough degrees of freedom? The
full model runs in Stata, but Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
21 May
21 May
11:46 a.m.
Thanks again for all the help!
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: Gary King [mailto:king@harvard.edu]
Sent: Thursday, May 20, 2010 5:00 PM
To: Eric McGhee
Cc: Olivia Lau; zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
for logit, there are only 2 things that can cause it to fail:
collinearity and perfect discrimination.
you're using mlogit, and so you can add to that that each of these can
be true for any of the individual equations. If you have very rare
events in one of them, this can happen easily.
Gary
---
http://gking.harvard.edu
On Thu, May 20, 2010 at 6:25 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I just ran the full model--first as OLS, and then as
mlogit--and it doesn't look like it drops any variables. And some very simple
models--for example, just ideology and party ID as IVs--can produce the error message.
I'm not wedded to this model; it was very preliminary, and I can plausibly recode
the DV to make it binary and run logit. (When I I do that the full model works.) But
I'm curious what the problem is!
Thanks for your help. If you have any further insights or ideas, let me know.
Best,
Eric
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 12:32 PM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
It looks like the model will "run" in R as well since it makes it
through the zelig step. Why don't you check to see if all of the IVs
that you include are output from the model-fitting step (e.g., do
summary on the output from zelig)? If some are being dropped, it
probably indicates that those IVs are collinear with each other (and
shouldn't be included anyway).
Yours,
Olivia
On Thu, May 20, 2010 at 3:22 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
Thanks Olivia. I've been poking around and
it seems like the problem is connected to the IVs I'm including. If I pare the model
way down, it works. Is it possible I'm overloading my data, and the mlogit function
is just dropping values of the DV where I don't have enough degrees of freedom? The
full model runs in Stata, but Stata will run just about anything.
Eric McGhee | Research Fellow | PPIC | 415-291-4439
Any opinions expressed in this message are those of the author alone and do not
necessarily reflect any position of the Public Policy Institute of California.
-----Original Message-----
From: monkeykupo(a)gmail.com [mailto:monkeykupo@gmail.com] On Behalf Of Olivia Lau
Sent: Thursday, May 20, 2010 11:26 AM
To: Eric McGhee
Cc: zelig(a)lists.gking.harvard.edu
Subject: Re: [zelig] mlogit
Hi, Eric.
I think that after eliminating the missing data, that one of the
factor levels in defhow2 drops off. A factor variable can have empty
levels (a defined value, but no values of that defined value), but
this becomes a problem when the factor is the DV. I would do the
list-wise deletion before, then check to see how many levels are
actually left in defhow2, then do factor(...) as you have it below
with those levels only.
Yours,
Olivia
On Thu, May 20, 2010 at 2:04 PM, Eric McGhee <mcghee(a)ppic.org> wrote:
I'm having trouble getting Zelig to run
simulations from a multinomial logit
model. The DV is called "defhow2" and has five categories, and I'm
running
the following syntax:
ppicjan10 <- read.dta("Jan 2010.regonly.dta", convert.factors=FALSE)
ppicjan10$defhow2 <- factor(ppicjan10$defhow2, levels=c(1,2,3,4,8),
labels=c("cuts", "taxes",
"cuts+taxes", "borrow", "other/dk"))
z.out <- zelig(as.factor(defhow2) ~
age+income+educ+lat+sex2+own2+cv+ba+osc+oth+pid7+ideo5+
ageXginfo+incXginfo+educXginfo+latXginfo+ownXginfo+cvXginfo+baXginfo+oscXginfo+
othXginfo+pidXginfo+ideoXginfo+ginfo, model="mlogit",
data=ppicjan10, baseline="other/dk")
x.out <- setx(z.out, fn=NULL)
s.out <- sim(z.out, x=x.out)
Everything is fine until the simulation command, where I get the following
error:
Error in factor(pr, levels = sort(unique(pr)), labels = ynames) :
invalid labels; length 5 should be 1 or 4
Resources online suggest that missing data is often to blame for this error
message, so I eliminated all of my missing data just to see if I could get
it to work in principle. No luck-same message. Does anyone know what might
be going on?
Eric McGhee
Research Fellow
PUBLIC POLICY
INSTITUTE OF CALIFORNIA
500 Washington Street, Suite 600
San Francisco, CA 94111
tel 415 291 4439
fax 415 291 4401
web www.ppic.org
Any opinions expressed in this message are those of the author alone and do
not necessarily reflect any position of the Public Policy Institute of
California.
5112
days inactive
5113
days old
7 comments
3 participants
participants (3)
-
Eric McGhee -
Gary King -
Olivia Lau