Thank you for your help. I did try Anthony's recommendation of removing the 'na.action=na.exclude' ; I thought I needed that argument as the data set includes NA values. I found it interesting that without the 'na.action=na.exclude' argument, the baseline level of two of my predictor variables (BYINCOME & F1HIMATH) were changed.
Courtney Benjamin Broome-Tioga BOCES Automotive Technology II Teacher Located at Gault Toyota Doctoral Candidate-Educational Theory & Practice State University of New York at Binghamton cbenj...@btboces.org<mailto:cbenj...@btboces.org> 607-763-8633 ________________________________ From: William Dunlap <wdun...@tibco.com> Sent: Sunday, October 23, 2016 2:24 PM To: Anthony Damico Cc: Courtney Benjamin; r-help@r-project.org; Thomas Lumley Subject: Re: [R] Significance of Svyrepdesign Object Warning The immediate problem could be solved by changing the following lines in survey:::summary.svrepglm from presid <- resid(object, "pearson") dispersion <- sum(object$survey.design$pweights * presid^2, na.rm = TRUE)/sum(object$survey.design$pweights) to presid <- resid(object, "pearson") pweights <- naresid(object$na.action, object$survey.design$pweights) dispersion <- sum(pweights * presid^2, na.rm = TRUE)/sum(pweights, na.rm = TRUE) 'naresid' uses the information from na.exclude to match up the residuals with the row in the data that they correspond to. resid() calls it so it should also be applied to pweights so they line up correctly. Bill Dunlap TIBCO Software wdunlap tibco.com<http://tibco.com> On Sun, Oct 23, 2016 at 11:17 AM, Anthony Damico <ajdam...@gmail.com<mailto:ajdam...@gmail.com>> wrote: hi, great example. i am ccing survey package author/maintainer dr. lumley. why do you have `na.action=na.exclude`? if you remove it, things work as expected-- library(RCurl) library(survey) data <- getURL(" https://raw.githubusercontent.com/cbenjamin1821/careertech-ed/master/elsq1adj.csv ") elsq1ch <- read.csv(text = data) #Specifying the svyrepdesign object which applies the BRR weights elsq1ch_brr<-svrepdesign(variables = elsq1ch[,1:16], repweights = elsq1ch[,18:217], weights = elsq1ch[,17], combined.weights = TRUE, type = "BRR") elsq1ch_brr #Logistic regression call which yields a warning regarding svyrepdesign object # your warning a <- svyglm(formula=F3ATTAINB~F1PARED+BYINCOME+F1RACE+F1SEX+F1RGPP2+F1HIMATH+F1RTRCC,family="binomial",design=elsq1ch_brr,subset=BYSCTRL==1&G10COHRT==1,na.action=na.exclude) summary(a) # works fine a <- svyglm(formula=F3ATTAINB~F1PARED+BYINCOME+F1RACE+F1SEX+F1RGPP2+F1HIMATH+F1RTRCC,family="binomial",design=elsq1ch_brr,subset=BYSCTRL==1&G10COHRT==1) summary(a) the mismatch of vectors generating that warning happens inside debug(survey:::summary.svrepglm) [..snip..] Browse[2]> length(presid) [1] 12614 Browse[2]> length(object$survey.design$pweights) [1] 8397 and including vs excluding the na.action=na.exclude gives you a slightly different dispersion parameter calculation (Dispersion parameter for binomial family taken to be 0.7756235) (Dispersion parameter for binomial family taken to be 0.7849244) not sure if the two survey:::residuals.sv<http://residuals.sv>* methods should deal with the na.action= parameter? thanks On Sun, Oct 23, 2016 at 11:56 AM, Courtney Benjamin <cbenj...@btboces.org<mailto:cbenj...@btboces.org>> wrote: > Hello R Users, > > I am using Lumley's Survey Package in R to analyze complex survey data > that involves 200 balanced repeated replicate (BRR) weight variables. I > have ensured that my svyrepdesign object that specifies the application of > the BRR weights to the data set is accurate and I have matched the > published standard errors of the data set. > > When doing a logistic regression through the svyglm call, I receive the > following warning: > > In object$survey.design$pweights * presid^2 : > longer object length is not a multiple of shorter object length? > I have search around quite a bit online and have not been able to find any > good interpretation of its meaning. I want to be sure that I am not making > some type of mistake that is causing this warning to be produced. Any > advisement is greatly appreciated. > The following is an MRE that can be pasted into the R console: > library(RCurl) > library(survey) > data <- getURL("https://raw.githubusercontent.com/ > cbenjamin1821/careertech-ed/master/elsq1adj.csv") > elsq1ch <- read.csv(text = data) > #Specifying the svyrepdesign object which applies the BRR weights > elsq1ch_brr<-svrepdesign(variables = elsq1ch[,1:16], repweights = > elsq1ch[,18:217], weights = elsq1ch[,17], combined.weights = TRUE, type = > "BRR") > elsq1ch_brr > #Logistic regression call which yields a warning regarding svyrepdesign > object > svyglm(formula=F3ATTAINB~F1PARED+BYINCOME+F1RACE+F1SEX+ > F1RGPP2+F1HIMATH+F1RTRCC,family="binomial",design= > elsq1ch_brr,subset=BYSCTRL==1&G10COHRT==1,na.action=na.exclude) > allCC <- summary(svyglm(formula=F3ATTAINB~F1PARED+BYINCOME+ > F1RACE+F1SEX+F1RGPP2+F1HIMATH+F1RTRCC,family="binomial", > design=elsq1ch_brr,subset=BYSCTRL==1&G10COHRT==1,na.action=na.exclude)) > allCC > > #Session Info > #R version 3.3.1 (2016-06-21) > #Platform: x86_64-w64-mingw32/x64 (64-bit) > #Running under: Windows >= 8 x64 (build 9200) > > #locale: > # [1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United > States.1252 > #[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C > #[5] LC_TIME=English_United States.1252 > #attached base packages: > # [1] grid stats graphics grDevices utils datasets > methods base > #other attached packages: > #[1] survey_3.31-2 survival_2.39-4 Matrix_1.2-6 RCurl_1.95-4.8 > bitops_1.0-6 > #loaded via a namespace (and not attached): > #[1] tools_3.3.1 splines_3.3.1 knitr_1.14 lattice_0.20-33 > > > Courtney Benjamin > > Broome-Tioga BOCES > > Automotive Technology II Teacher > > Located at Gault Toyota > > Doctoral Candidate-Educational Theory & Practice > > State University of New York at Binghamton > > cbenj...@btboces.org<mailto:cbenj...@btboces.org><mailto:cbenj...@btboces.org<mailto:cbenj...@btboces.org>> > > 607-763-8633<tel:607-763-8633> > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org<mailto:R-help@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. > [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org<mailto:R-help@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. [[alternative HTML version deleted]] ______________________________________________ R-help@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.
