Vmat: Compute covariance matrix of residuals for general linear...

In svydiags: Linear Regression Model Diagnostics for Survey Data

Compute covariance matrix of residuals for general linear models fitted with complex survey data

Description

Compute a covariance matrix using residuals from a fixed effects, general linear regression model fitted with data collected from one- and two-stage complex survey designs.

Usage

Vmat(mobj, stvar = NULL, clvar = NULL)

Arguments

mobj	model object produced by svyglm
stvar	field in mobj that contains the stratum variable in the complex sample design; use stvar = NULL if there are no strata
clvar	field in mobj that contains the cluster variable in the complex sample design; use clvar = NULL if there are no clusters

Details

Vmat computes a covariance matrix among the residuals returned from svyglm in the survey package. Vmat is called by svyvif when computing variance inflation factors. The matrix that is computed by Vmat is appropriate under these model assumptions: (1) in single-stage, unclustered sampling, units are assumed to be uncorrelated but can have different model variances, (2) in single-stage, stratified sampling, units are assumed to be uncorrelated within strata and between strata but can have different model variances; (3) in unstratified, clustered samples, units in different clusters are assumed to be uncorrelated but units within clusters are correlated; (3) in stratified, clustered samples, units in different strata or clusters are assumed to be uncorrelated but units within clusters are correlated.

Value

n \times n matrix where n is the number of cases used in the linear regression model

Author(s)

Richard Valliant

References

Liao, D, and Valliant, R. (2012). Variance inflation factors in the analysis of complex survey data. Survey Methodology, 38, 53-62.

Lumley, T. (2010). Complex Surveys. New York: John Wiley & Sons.

Lumley, T. (2021). survey: analysis of complex survey samples. R package version 4.1-1.

See Also

svyvif

Examples

require(Matrix)
require(survey)
data(nhanes2007)
black <- nhanes2007$RIDRETH1 == 4
X <- nhanes2007
X <-  cbind(X, black)
X1 <- X[order(X$SDMVSTRA, X$SDMVPSU),]

    # unstratified, unclustered design
nhanes.dsgn <- svydesign(ids = 1:nrow(X1),
                         strata = NULL,
                         weights = ~WTDRD1, data=X1)
m1 <- svyglm(BMXWT ~ RIDAGEYR + as.factor(black) + DR1TKCAL, design=nhanes.dsgn)
summary(m1)

V <- Vmat(mobj = m1,
          stvar = NULL,
          clvar = NULL)

    # stratified, clustered design
nhanes.dsgn <- svydesign(ids = ~SDMVPSU,
                         strata = ~SDMVSTRA,
                         weights = ~WTDRD1, nest=TRUE, data=X1)
m1 <- svyglm(BMXWT ~ RIDAGEYR + as.factor(black) + DR1TKCAL, design=nhanes.dsgn)
summary(m1)
V <- Vmat(mobj = m1,
          stvar = "SDMVSTRA",
          clvar = "SDMVPSU")

svydiags documentation built on April 28, 2022, 1:07 a.m.