Skip to content

svykappa using the survey package

2 messages · Muhuri, Pradip (AHRQ/CFACT), Anthony Damico

Muhuri, Pradip (AHRQ/CFACT)
# 
Hello,

My goal is to calculate the weighted kappa measure of agreement between two factors  using the R  survey package.  I am getting the following error message (the console is appended below; sorry no data provided).
Error in names(probs) <- nms : 
  'names' attribute [15] must be the same length as the vector [8]

I have followed the following major steps:

1) Used the "haven" package to read the sas data set into R.
2) Used the dplyr mutate() to create 2 new variables and converted to factors [required for the svykappa()?].
3) Created an object (named design) using the survey design variables and the data file.
4) Used the svykappa() to compute the kappa measure of agreement. 

I will appreciate if someone could give me hints on how to resolve the issue.

Thanks,

Pradip Muhuri

###############  The detailed console is appended below  ####################
+   names(df) <- tolower(names(df))
+   df
+ }
+                          xbpchek53 = ifelse(bpchek53 ==1, 1,
+                             ifelse(bpchek53 %in% 2:6, 2,NA)), 
+                          xcholck53 = ifelse(cholck53 ==1, 1,
+                            ifelse(cholck53 %in% 2:6, 2,NA)))
[1] TRUE
[1] TRUE
xbpchek53
bpchek53     1     2   Sum
     -9      0     0     0
     -8      0     0     0
     -7      0     0     0
     -1      0     0     0
     1   19778     0 19778
     2       0  2652  2652
     3       0  1014  1014
     4       0   538   538
     5       0   737   737
     6       0   623   623
     Sum 19778  5564 25342
xcholck53
cholck53     1     2   Sum
     -9      0     0     0
     -8      0     0     0
     -7      0     0     0
     -1      0     0     0
     1   14850     0 14850
     2       0  3153  3153
     3       0  1170  1170
     4       0   696   696
     5       0   909   909
     6       0  3764  3764
     Sum 14850  9692 24542
xcholck53
xbpchek53     1     2   Sum
      1   14667  4379 19046
      2     163  5225  5388
      Sum 14830  9604 24434
Error in names(probs) <- nms : 
  'names' attribute [15] must be the same length as the vector [8]

#################################################################

Pradip K. Muhuri,  AHRQ/CFACT
 5600 Fishers Lane # 7N142A, Rockville, MD 20857
Tel: 301-427-1564




-----Original Message-----
From: R-help [mailto:r-help-bounces at r-project.org] On Behalf Of Muhuri, Pradip (AHRQ/CFACT)
Sent: Thursday, June 16, 2016 2:06 PM
To: David Winsemius
Cc: r-help at r-project.org
Subject: Re: [R] dplyr's arrange function - 3 solutions received - 1 New Question

Hello David,

Your revisions to the earlier code have given me desired results.

library("gtools")
mydata[ mixedorder(mydata$prevalence_c, decreasing=TRUE), c("indicator", "prevalence_c")  ]

Thanks,

Pradip


Pradip K. Muhuri,  AHRQ/CFACT
 5600 Fishers Lane # 7N142A, Rockville, MD 20857
Tel: 301-427-1564





-----Original Message-----
From: David Winsemius [mailto:dwinsemius at comcast.net] 
Sent: Thursday, June 16, 2016 12:54 PM
To: Muhuri, Pradip (AHRQ/CFACT)
Cc: r-help at r-project.org
Subject: Re: [R] dplyr's arrange function - 3 solutions received - 1 New Question
Try instead just a vector of names for the second argument to "["

 mydata[ mixedorder(mydata$prevalence_c, decreasing=TRUE), 
         c("indicator", "prevalence_c") ]
David Winsemius
Alameda, CA, USA

______________________________________________
R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
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.
Anthony Damico
# 
hi pradip, this should give you what you want


    library(foreign)
    library(survey)

    tf <- tempfile()

    download.file( "
https://meps.ahrq.gov/mepsweb/data_files/pufs/h163ssp.zip" , tf , mode =
'wb' )

    z <- unzip( tf , exdir = tempdir() )

    x <- read.xport( z )

    names( x ) <- tolower( names( x ) )

    design <- svydesign(id=~varpsu,strat=~varstr, weights=~perwt13f,
data=x, nest=TRUE)

    # include missings as "No" values here
    design <-
        update(design,
            xbpchek53 = ifelse(bpchek53 ==1,'yes','no or missing'),
            xcholck53 = ifelse(cholck53 ==1, 'yes','no or missing')
        )

    # subset out records that were missing for either variable
    svykappa( ~ xbpchek53 + xcholck53 , subset(design, bpchek53 > 0 &
cholck53 > 0 ) )


















On Mon, Jun 20, 2016 at 7:49 PM, Muhuri, Pradip (AHRQ/CFACT) <
Pradip.Muhuri at ahrq.hhs.gov> wrote: