Processing math: 100%

Variance Component Estimation in Multistage Sampling

Richard Valliant, Jill A. Dever, and Frauke Kreuter

2024-06-24

To allocate a sample among different stages of sampling, the contributions of the different stages to the variance of an estimator must be considered. These components of variance generally depend on the analysis variable and also on the form of the estimator. This vignette covers some basic variance results for linear estimators in two-stage and three-stage sampling and how the components can be estimated with functions in PracTools. Technical background is in Valliant, Dever, and Kreuter (2018b), ch.9. First, the package must be loaded with

library(PracTools)

Alternatively, require(PracTools) can be used.

Two-stage Sampling

Consider a two-stage sample design in which the first-stage units are selected using πps sampling, i.e., with varying probabilities and without replacement. We will also refer to this as ppswor sampling. Elements are selected at the second stage via simple random sampling without replacement (srswor). Quite a bit of notation is needed, even in this fairly simple case:

U = universe of PSUs

M = number of PSUs in universe

Ui = universe of elements in PSU i

Ni = number of elements in the population for PSU i

N=iUNi is the total number of elements in the population

πi = selection probability of PSU i

πij = joint selection probability of PSUs i and j

m = number of sample PSUs

ni = number of sample elements in PSU i

s = set of sample PSUs

si = set of sample elements in PSU i

yk = analysis variable for element k in PSU i (subscript i is implied)

ˉyU = mean per element in the population

ˉyUi = mean per element in the population in PSU i

The π-estimator of the population total, tU=iUkUiyk, of an analysis variable y is

ˆtπ=isˆtiπi

where ˆti=(Ni/ni)ksiyk, which is the estimate of the total for PSU i with a simple random sample. The design variance of the estimated total can be written as the sum of two components: V(ˆtπ)=iUjU(πijπiπj)tiπitjπj+iUN2iπini(1niNi)S2U2i where

S2U2i=kUi(ykˉyUi)2/(Ni1)

is the unit variance of y among the elements in PSU i.

Formula (1) is difficult or impossible to use for sample size computations because the number of PSUs in the sample is not exposed. Another is to analyze srswor sampling of PSUs and SSUs as in Example 1 below. Determining sample sizes this way does not mean that you are necessarily locked into selecting PSUs and elements within PSUs via srswor or srswr. Basing sample sizes on a design that is less complicated than the one that will actually be used is a common approach, although it can be deceptive for some analysis variables.

Special case: srswor at first and second stages

Suppose the first stage is an srswor of m out of M PSUs and the second stage is a sample of ni elements selected by srswor from the population of Ni. The π-estimator is

ˆtπ=MmisNiniksiyk

Its variance is equal to V(ˆtπ)=M2mMmMS2U1+MmiUN2iniNiniNiS2U2i where S2U1=iU(tiˉtU)2M1 with ti being the population total of y in PSU i and ˉtU=iUti/M is the mean total per PSU.

If ˉn elements are selected in each PSU and the sampling fractions of PSUs and elements within PSUs are all small, then the relvariance can be written as V(ˆtπ)t2U=B2m+W2mˉn

where B2=S2U1/ˉt2U=M2S2U1/t2U is the unit relvariance among PSU totals and W2=MiUN2iS2U2i/t2U. The term B2 is called the “between (PSU) component” while W2 is the “within component”. Expression (3) is the form used in the R function, BW2stageSRS. Textbooks often list a specialized form of (3) that requires that all PSUs have the same size, NiˉN, and that ˉn elements are selected in each. In that case, the second-stage sampling fraction is ˉn/ˉN. This implies that the sample is self-weighting: πiπk|i=mˉn/MˉN. The relvariance based on (2) then simplifies to the less general form V(ˆtπ)t2U=1mMmMB2+1mˉnˉNˉnˉNW2

where W2=1Mˉy2UiUS2U2i.

Assuming that ˉn elements are selected in each sample PSU, and m/M and ˉn/Ni are both small, the more general form of the relvariance in (3) can also be written in terms of a measure of homogeneity δ as follows:

V(ˆtπ)t2U˜Vmˉnk[1+δ(ˉn1)] where ˜V=S2U/ˉy2U, k=(B2+W2)/˜V, and

δ=B2B2+W2. With some effort, it can be shown that when Ni=ˉN and both M and ˉN are large,

S2Uˉy2U=1ˉy2UiUkUi(ykˉyU)2(N1)B2+W2 i.e., the population relvariance can be written as the sum of between and within relvariances. If k=1, (4) equals the expression found in many textbooks. However, when the population count of elements per cluster varies, k may be far from 1, as will be illustrated in an example below. In those cases, (4) with an estimate of the actual k should be used for determining sample sizes and computing advance estimates of coefficients of variation.

Expressions (3) and (4) are useful for sample size calculation since the number of sample PSUs and sample units per PSU are explicit in the formula. Equation (4) also connects the variance of the estimated total to the variance that would be obtained from a simple random sample since ˜V/mˉn is the relvariance of the estimated total in an srswor of size mˉn when the sampling fraction is small. The product k[1+δ(ˉn1)] is a type of design effect. When k=1, the term 1+δ(ˉn1) is the approximate design effect found in many textbooks.

The next example uses the MDarea.popA from PracTools. This dataset is based on the U.S. Census counts from the year 2000 for Anne Arundel County in the US state of Maryland. The geographic divisions used in this dataset are called tracts and block groups. Tracts are constructed by the US Census Bureau to have a desired population size of 4,000 people. Block groups (BGs) are smaller with a target size of 1,500 people. Counts of persons in the dataset are the same for most tracts and block groups as in the 2000 Census.

 require(PracTools)
 data(MDarea.popA)
 BW2stageSRS(MDarea.popA$y1, psuID=MDarea.popA$PSU)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#> 0.007979563 1.455575239 1.463117997 1.463554802 1.000298544 0.005452179
 BW2stageSRS(MDarea.popA$y1, psuID=MDarea.popA$SSU)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#>  0.03661379  1.42824599  1.46311800  1.46485978  1.00119046  0.02499474

The values of δ are 0.005 for PSU and 0.025 for SSU. Next, to illustrate the dramatic effect that varying sizes of clusters can have, we compute the same statistics as above using tracts and block groups (BGs) within tracts as clusters. These vary substantially in the number of persons in each cluster. A new variable called trtBG is computed since the values of the variable, BLKGROUP, are nested within each tract:

   trtBG <- 10*MDarea.popA$TRACT + MDarea.popA$BLKGROUP
   BW2stageSRS(MDarea.popA$y1, psuID=MDarea.popA$TRACT)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#>   0.2603356   1.8405833   1.4631180   2.1009189   1.4359190   0.1239151
   BW2stageSRS(MDarea.popA$y1, psuID=trtBG)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#>   0.3495410   1.9488250   1.4631180   2.2983660   1.5708685   0.1520824

The value of δ is 0.124 TRACTs are clusters and 0.152 when trtBG defines clusters. The measures of homogeneity increase substantially when tracts or BGs are clusters compared to the PSU and SSU results. This is entirely due to the increase in B2 when units with highly variable sizes are used and an srs is selected. For example, B2=0.0079 for y1 when PSU is a cluster but is 0.2605 when TRACT is a cluster.

More General Two-stage Designs

Variances of estimators in two-stage designs more complicated than simple random sampling at each stage can be written as a sum of components. However, these have limited usefulness in determining sample sizes for the same reason that (1) is not. A more convenient formulation is the case where PSUs are selected with varying probabilities but with replacement, and the sample within each PSU is selected by srswor. With-replacement designs may not often be used in practice but have simple variance formulae. The pwr-estimator of a total (Särndal, Swensson, and Wretman 1992) is ˆtpwr=1misˆtipi where ˆti=Niniksiyk is the estimated total for PSU i from a simple random sample and pi is the one-draw selection probability of PSU i. The variance of ˆtpwr is

V(ˆtpwr)=1miUpi(tipitU)2+iUN2impini(1niNi)S2U2i.

Making the assumption that ˉn elements are selected in each PSU, the variance reduces to V(ˆtpwr)=S2U1(pwr)m+1mˉniU(1ˉnNi)N2iS2U2ipi where, in this case, S2U1(pwr)=iUpi(tipitU)2. Dividing this by t2U and assuming that the within-PSU sampling fraction, ˉn/Ni, is negligible, we obtain the relvariance of ˆtpwr as, approximately, V(ˆtpwr)t2UB2m+W2mˉn=˜Vmˉnk[1+δ(ˉn1)]

with ˜V=S2U/ˉy2U, k=(B2+W2)/˜V, B2=S2U1(pwr)t2U,

W2=1t2UiUN2iS2U2ipi,

δ=B2/(B2+W2)

Expression (7) has the same form as (4) but with different definitions of B2 and W2. Expression (7) also has the interpretation of an srs variance of an unclustered variance, ˜V/mˉn, times a design effect, k[1+δ(ˉn1)], in the same way that (4) did.

 pp.PSU <- table(MDarea.popA$PSU) / nrow(MDarea.popA)
 pp.SSU <- table(MDarea.popA$SSU) / nrow(MDarea.popA)
 BW2stagePPS(MDarea.popA$y1, pp=pp.PSU, psuID=MDarea.popA$PSU)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#> 0.007878320 1.455574516 1.463117997 1.463452835 1.000228853 0.005383378
 BW2stagePPS(MDarea.popA$y1, pp=pp.SSU, psuID=MDarea.popA$SSU)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#>  0.03652265  1.42825477  1.46311800  1.46477742  1.00113417  0.02493392

The code for PSUs that are tracts and block groups is

 pp.trt <- table(MDarea.popA$TRACT) / nrow(MDarea.popA)
 pp.BG <- table(trtBG) / nrow(MDarea.popA)
 BW2stagePPS(MDarea.popA$y1, pp=pp.trt, psuID=MDarea.popA$TRACT)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#> 0.009458290 1.454058071 1.463117997 1.463516361 1.000272271 0.006462716
 BW2stagePPS(MDarea.popA$y1, pp=pp.BG, psuID=trtBG)
#>          B2          W2 unit relvar       B2+W2           k       delta 
#>  0.01643324  1.44797251  1.46311800  1.46440575  1.00088014  0.01122178

The between term when clusters are defined by PSU is about the same as when clusters are selected by srs because PSU’s all have the same size. With PSUs being either tracts or block groups in the ppswr/srswor design, the between term is much smaller than the within, compared to the results in the srs/srs example. For example, with y1 and srs sampling of tracts, B2=0.2604 but for pps sampling of tracts B2=0.0091.

When clusters are selected by srs, S2U1 is the variance of the cluster totals around the average cluster total. In contrast, with pps sampling of clusters, S2U1(pwr) is the variance of the estimated population totals, ti/pi around the population total, tU. When clusters are selected with probability proportional to Ni, then ti/pi=NiˉyUi. If these one-cluster estimates of the population total are fairly accurate, as they are here, the B2 term can be quite small. This leads to much smaller values of δ in pps sampling of clusters. This implies that the negative effect of clustering on the variance is lessened for a design that selects clusters with pp(Ni). This kind of comparison explains most practitioners’ preference for pps sampling of clusters, especially when the clusters vary in population size.

General Three-stage Designs

In the case of with-replacement sampling of PSUs with varying probabilities and srswor at the second and third stages, the relvariance can be written (with a few assumptions) in a form useful for sample size calculations. Treating the case where SSUs are selected via srs (either with or without replacement) is not too unrealistic since SSUs (like block groups) are often created to have about the same population sizes.

The variance formulae for a three-stage design with ppswor selection of first-stage units is complex enough that it is not useful for sample size planning. See Valliant, Dever, and Kreuter (2018b), sec. 9.2.4 for details. To obtain a simpler formula, suppose that ˉn SSUs are sampled in each sample PSU, the sampling fractions of SSUs in each PSU, ˉn/Ni, are small, and ˉˉq elements are selected in each sample SSU. The relvariance of the pwr-estimator is then

V(ˆtpwr)t2U=B2m+W22mˉn+W23mˉnˉˉq,

where B2=S2U1(pwr)/t2U is given by (8),

W22=1t2UiUN2iS2U2i/pi;

W23=1t2UiUNipijUiQ2ijS2U3ij.

The relvariance can also be written in terms of two measures of homogeneity:

V(ˆtpwr)t2U=˜Vmˉnˉˉq{k1δ1ˉnˉˉq+k2[1+δ2(ˉˉq1)]} where

k1=(B2+W2)/˜V with ˜V=1Q1iUjUikUij(ykˉyU)2/ˉy2U is the unit relvariance of y in the population.

k2=(W22+W23)/˜V

δ1=B2/(B2+W2)

W2=1t2UiUQ2iS2U3i/pi with S2U3i=1Qi1jUikUij(ykˉyUi)2 and ˉyUi=jUikUijyk/Qi, i.e., S2U3i is the element-level variance among all elements in PSU i

δ2=W22/(W22+W23)

Note that the term W2 in δ1 does not enter the variance in (11) but is defined by analogy to the term in two-stage sampling. If elements were selected directly from the sample PSUs (instead of first sampling SSUs), then W2 above would be the appropriate within-PSU component.

The term δ1 is a measure of the homogeneity among the PSU totals. If the estimate of the population total from each PSU total, ti/pi, was exactly equal to the population total, tU, then B2=0 and δ1=0. That is, if the variation within PSUs is much larger than the variation among PSU totals, then δ1 will be small; this is the typical situation in household surveys if PSUs all have about the same number of elements. As we saw in the earlier example, the condition of equal-sized PSUs can be critically important to insure that B2 is small.

If the SSUs all have about the same totals, tij, then W22 will be small and δ20. Although attempts may be made to create SSUs that have about the same number of elements Qij, the totals tij of other variables tend to vary, leading to values of δ2 that are larger than those of δ1.

The R function, BW3stagePPS, will calculate B2, W2, W22, W23, δ1, and δ2 defined above for ppswr/srs/srs and srswr/srs/srs sampling. The function is appropriate if an entire frame is available and takes the following parameters:

Parameter Description
X data vector; length is the number of elements in the population.
pp vector of one-draw probabilities for the PSUs; length is number of PSUs in population.
psuID vector of PSU identification numbers. This vector must be as long as X. Each element in a given PSU should have the same value in psuID. PSUs must be in the same order as in X.
ssuID vector of SSU identification numbers. This vector must be as long as X. Each element in a given SSU should have the same value in ssuID. PSUs and SSUs must be in the same order as in X. ssuID should have the form psuID||(ssuID within PSU).
 M <- length(unique(MDarea.popA$TRACT))
 trtBG <- 10*MDarea.popA$TRACT + MDarea.popA$BLKGROUP
 pp.trt <- rep(1/M,M)
 BW3stagePPS(X=MDarea.popA$y1, pp=pp.trt,
       psuID=MDarea.popA$TRACT, ssuID=trtBG)
#>           B           W          W2          W3 unit relvar          k1 
#>   0.2575952   1.8405833   0.2714429   2.1074463   1.4631180   1.4340460 
#>          k2      delta1      delta2 
#>   1.6259039   0.1227709   0.1141049

We repeat the calculation but assuming ppswr sampling of PSUs. The calculation for y1 using tracts and block groups as the first- and second-stage sampling units is done via this call:

 trtBG <- 10*MDarea.popA$TRACT + MDarea.popA$BLKGROUP
 pp.trt <- table(MDarea.popA$TRACT) / nrow(MDarea.popA)
 BW3stagePPS(X=MDarea.popA$y1, pp=pp.trt,
       psuID=MDarea.popA$TRACT, ssuID=trtBG)
#>           B           W          W2          W3 unit relvar          k1 
#> 0.009458290 1.454058071 0.272308839 1.685246592 1.463117997 1.000272271 
#>          k2      delta1      delta2 
#> 1.337934080 0.006462716 0.139106579

Notice that δ1=0.123 with srs sampling of tracts but is 0.006 when tracts are sampled proportional to their population sizes.

An important practical, sample design problem that we do not cover in this vignette is how to estimate variance components and measures of homogeneity from a complex, multistage sample. This topic is covered in detail in section 9.4 of Valliant, Dever, and Kreuter (2018b). The PracTools package includes a variety of other functions relevant to two- and three-stage sampling that are also not discussed in this vignette:

Function Description
BW2stagePPSe Estimate components of relvariance for a sample design where primary sampling units (PSUs) are selected with pps and elements are selected via srs. The input is a sample selected in this way.
BW3stagePPSe Estimate components of relvariance for a sample design where primary sampling units (PSUs) are selected with probability proportional to size with replacement (ppswr) and secondary sampling units (SSUs) and elements within SSUs are selected via simple random sampling (srs). The input is a sample selected in this way.
clusOpt2 Compute the sample sizes that minimize the variance of the pwr-estimator of a total in a two-stage sample.
clusOpt2fixedPSU Compute the optimum number of sample elements per primary sampling unit (PSU) for a fixed set of PSUs.
clusOpt3 Compute the sample sizes that minimize the variance of the pwr-estimator of a total in a three-stage sample.
clusOpt3fixedPSU Compute the sample sizes that minimize the variance of the pwr-estimator of a total in a three-stage sample when the PSU sample is fixed.
CVcalc2 Compute the coefficient of variation of an estimated total in a two-stage design. Primary sampling units (PSUs) can be selected either with probability proportional to size (pps) or with equal probability. Elements are selected via simple random sampling (srs).
CVcalc3 Compute the coefficient of variation of an estimated total in a three-stage design. Primary sampling units (PSUs) can be selected either with probability proportional to size (pps) or with equal probability. Secondary units and elements within SSUs are selected via simple random sampling (srs).
deff Compute the Kish, Henry, Spencer, or Chen-Rust design effects.

References

Särndal, C.-E., B. Swensson, and J. Wretman. 1992. Model Assisted Survey Sampling. New York: Springer-Verlag.
Valliant, R., J. A. Dever, and F. Kreuter. 2018a. Files for 2nd Edition of Practical Tools for Designing and Weighting Survey Samples. https://umd.app.box.com/v/PracTools2ndEdition.
———. 2018b. Practical Tools for Designing and Weighting Survey Samples. 2nd ed. New York: Springer-Verlag.