[R] Survey Package with Binary Data (no Standard Errors reported)

[Thomas Lumley]

On Fri, 3 Apr 2009, Paul Jones wrote:

> Hi,
>
> I'm trying to get standard errors for some of the variables in my data frame. 
> One of the questions on my survey is whether faculty coordinate across 
> curriculum to include Arts Education as subject matter. All the responses are 
> coded in zeros and ones obviously. For some of the other variables I have a 2 
> for those that responded with "Don't Know".
>
> I'm getting NA for mean and standard deviations from svymean. Am I doing 
> something wrong of can the survey package not handle this type of data?

Are you sure you don't have any NA values in the data? If you do, the 
na.rm=TRUE option to svymean() will fix your problem. If you don't, 
something mysterious is happening. You could try svytable(~Curriculum, 
survey), which will give a tabulation and might show up what is strange 
about your data.

As a separate issue, you might want to look at svyciprop() if some of your 
proportions are close to 0 or 1, to get better confidence intervals.


> Here's my code.

Nothing obviously wrong with it.

 	-thomas

>> survey <- svydesign(id=~1, data=General, strata=~Grade.Level)
> Warning message:
> In svydesign.default(id = ~1, data = General, strata = ~Grade.Level) :
> No weights or probabilities supplied, assuming equal probability
>
>> summary(survey)
> Stratified Independent Sampling design (with replacement)
> svydesign(id = ~1, data = General, strata = ~Grade.Level)
> Probabilities:
>  Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
>     1       1       1       1       1       1
> Stratum Sizes:
>          Elementary High Middle
> obs               312  236    156
> design.PSU        312  236    156
> actual.PSU        312  236    156
> Data variables:
> [1] "Grade.Level"                  "Curriculum"                 [3] 
> "Field.Trips"                  "Residencies"                [5] "PTA.Support" 
> "Community.Open.Performances"
> [7] "Visual.Arts.Attendance"       "Literary.Arts.Attendance"   [9] 
> "Arts.Organization.Membership" "Arts.Essential"
>         > svymean(~Curriculum, survey)
>          mean SE
> Curriculum   NA NA
>
> ???
>
> PJ
>
> ______________________________________________
> 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.
>

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle