1 May
2008
1 May
'08
3:12 p.m.
I am trying to use setx and am running into the following error.
gen.low <- setx(z.out2, nkrace = median, nkgender =
1, nkhourly = median, age
= median, soc.desir = median, nkorient = median, nativity1
= median, language1
= median, nkedu = median, pi= median, site.type = median, union = median,
nkclass = median, nkcogender1= median, nkcorace = median)
Error in as.vector(x, "double") : cannot coerce to vector
z.out2 was computed using Zelig and has a bunch of dummy variables, which I
suspect could be part ot this problem. Does anyone have any ideas? thanks!
1 May
1 May
3:31 p.m.
You might try the following:
1) gen.low <- setx(z.out2, numeric=median, ordered=median, others=mode, nkgender
= 1)
That should simplify the function and account for everything, since it might be
getting hung up in the number of functions it's running.
Another thing that gave us the same error was trying to set just the value of
one variable to fixed value. We used a logit.survey, and we could only run the
function if we set two things to specific values or vectors i.e.,
2) ruler=seq(x:y, by=z) ##Assuming gender is a sequence for the moment...
gen.low <- setx(z.out2, numeric=median, ordered=median, others=mode, nkgender
=ruler, nkclass=1)
This worked for setting a "high" and "low" value of the second
variable to run
our simulations against. We couldn't have run something like the first function
with logit.survey, but the first function worked just fine when the model was a
simple logit.
Hope this helps,
-Colin
Citeren johns18 at fas.harvard.edu:
I am trying to use setx and am running into the
following error.
gen.low <- setx(z.out2, nkrace = median,
nkgender = 1, nkhourly = median,
age
= median, soc.desir = median, nkorient = median, nativity1 = median,
language1
= median, nkedu = median, pi= median, site.type = median, union = median,
nkclass = median, nkcogender1= median, nkcorace = median)
Error in as.vector(x, "double") : cannot coerce to vector
z.out2 was computed using Zelig and has a bunch of dummy variables, which I
suspect could be part ot this problem. Does anyone have any ideas? thanks!
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
2 May
2 May
11:01 a.m.
New subject: [gov2001-l] change in duration model
Hi everyone,
My partners and I replicated a paper examining various reasons why the duration
of some civil wars is longer than that of others. They used a Weibull model to
estimate these relationships.
We are interested in finding a relationship between the change of foreign aid
(increase, decrease) on an immediate change of the duration (i.e whether
duration increases or decreases when aid increases). We are wondering what
statistical model to use for such an analysis. If anyone has an idea, please
let me know!
Johannes
5861
days inactive
5862
days old
gov2001@lists.gking.harvard.edu
2 comments
3 participants
participants (3)
-
brown4＠fas.harvard.edu -
jcastner＠fas.harvard.edu -
johns18＠fas.harvard.edu