[R] Error in svychisq and svyttest with svrepdesign
Anabela Afonso
afaafonso at gmail.com
Tue Mar 10 14:31:12 CET 2015
Dear Forum,
I’m working with a complex sample and with replication weights. I defined
my design svrepdesign function. I’m trying to run svychisq and
svyttest function
from the survey package and I get the error:
Error in crossprod(x, y) :
requires numeric/complex matrix/vector arguments
I can’t understand this error. I kindly ask if someone can help me out.
Thanks in advance,
Here is my code and some output:
> library(foreign); library(survey)
> dados<-read.spss("dadosSPSS.sav", use.value.labels=T, to.data.frame=T)
> class(dados)
[1] "data.frame"
> str(dados)
'data.frame': 7624 obs. of 4 variables:
$ Sex : Factor w/ 2 levels "Male","Female": 1 1 1 1 1 1 1 1 1 1 ...
$ Computer: Factor w/ 2 levels "Yes","NO": 1 1 2 1 1 1 1 1 1 2 ...
$ Color : Factor w/ 3 levels "Red","Green",..: 1 1 1 1 1 1 1 1 1 1 ...
$ Number : num 2 1 0 2 1 2 1 2 1 0 ...
$ final.w : num 1267 596 1143 1069 542 ...
# Note: Variable Color with NA
> repdes<-svrepdesign(data=dados, repweights=rep.w, scale=1, rscales=r.sc, type="JKn", weights=~final.w, combined.weights=F)
> summary(repdes)
Call: svrepdesign.default(data = dados, repweights = rep.w, scale = 1,
rscales = r.sc, type = "JKn", weights = ~final.w, combined.weights = F)
Stratified cluster jackknife (JKn) with 428 replicates.
Variables:
[1] "Sex" "Computer" "Color" "Number" "final.w"
> svytable(~Sex+Computer, repdes)
Computer
Sex Yes NO
Male 1501598.7 1063055.3
Female 1485933.1 810557.9
> svytable(~Sex+Color, repdes) # NA are ignored
Color
Sex Red Green Yellow
Male 2060708.5 219678.4 286038.6
Female 1840511.7 229763.8 224444.0
> svychisq(~Sex+Computer, repdes)
Error in crossprod(x, y) :
requires numeric/complex matrix/vector arguments
> svychisq(~Sex+Color, repdes)
Error in crossprod(x, y) :
requires numeric/complex matrix/vector arguments
> svyttest(Number ~Sex, repdes)
Error in crossprod(x, y) :
requires numeric/complex matrix/vector arguments
[[alternative HTML version deleted]]
More information about the R-help
mailing list