R help - May 2010 - Error messages with psm and not cph in Hmisc

While

sm4.6ll<-fit.mult.impute(Surv(agesi, si)~partner+ in.love+ pubty+ FPA+
strat(gender),fitter = cph, xtrans = dated.sexrisk2.i, data dated.sexrisk2,
x=T,y=T,surv=T, time.inc=16)

runs perfectly using Hmisc, Design and mice under R11 run via Sciviews-K,
with

library(Design)
library(mice)
ds2d<-datadist(dated.sexrisk2)
options(datadist="ds2d")
options(contrasts=c("contr.treatment","contr.treatment"))

the equivalent

sm4.6ll<-fit.mult.impute(Surv(agesi, si)~partner+ in.love+ pubty+ FPA+
strat(gender),fitter = *psm*, xtrans = dated.sexrisk2.i, data dated.sexrisk2,
x=T,y=T,surv=T, time.inc=16)

returns the error message

Error in  dimnames(X) <- list(rnam, c("(Intercept)", atr$colnames))
:
  length of 'dimnames' [2] not equal to array extent


Using survreg{survival} for which psm is a wrapper, also runs perfectly on
the unimputed dataset.

Is this a bug, or am I doing something wrong?  I particularly want to make
use of Design's validation and calibration utilities on multiply imputed
data.

With many thanks for everyone's help

David Foreman
Consultant and Visiting Professor in Child and Adolescent Psychiatry

	[[alternative HTML version deleted]]

On 05/05/2010 11:29 AM, David Foreman wrote:
> While
>
> sm4.6ll<-fit.mult.impute(Surv(agesi, si)~partner+ in.love+ pubty+ FPA+
> strat(gender),fitter = cph, xtrans = dated.sexrisk2.i, data >
dated.sexrisk2, x=T,y=T,surv=T, time.inc=16)
>
> runs perfectly using Hmisc, Design and mice under R11 run via Sciviews-K,
> with
>
> library(Design)
> library(mice)
> ds2d<-datadist(dated.sexrisk2)
> options(datadist="ds2d")
>
options(contrasts=c("contr.treatment","contr.treatment"))
>
> the equivalent
>
> sm4.6ll<-fit.mult.impute(Surv(agesi, si)~partner+ in.love+ pubty+ FPA+
> strat(gender),fitter = *psm*, xtrans = dated.sexrisk2.i, data >
dated.sexrisk2, x=T,y=T,surv=T, time.inc=16)
>
> returns the error message
>
> Error in  dimnames(X)<- list(rnam, c("(Intercept)",
atr$colnames)) :
>    length of 'dimnames' [2] not equal to array extent
>
>
> Using survreg{survival} for which psm is a wrapper, also runs perfectly on
> the unimputed dataset.
>
> Is this a bug, or am I doing something wrong?  I particularly want to make
> use of Design's validation and calibration utilities on multiply
imputed
> data.
>
> With many thanks for everyone's help
>
> David Foreman
The rms and Design packages do not support the user specifying a 
contrast option.

validate and calibrate function do not explicitly support multiple 
imputation.

Think about converting the the rms package.

Frank
> Consultant and Visiting Professor in Child and Adolescent Psychiatry
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Frank E Harrell Jr   Professor and Chairman        School of Medicine
                      Department of Biostatistics   Vanderbilt University