Description Usage Arguments Details Value Author(s) References See Also Examples
msvytotal
computes robust Horvitz-Thompson estimates of the total or robust weighted total estimates for complex samples by means of M-estimation.
1 2 |
y |
a formula object (only one variable) |
design |
a |
k |
robustness tuning constant |
type |
either |
na.rm |
should cases with missing values be dropped? (default |
control |
control object; see |
... |
(additional specifications which are delivered to |
msvytotal
performs (inverse probability-) weighted M-estimation (Huber psi-function or asymmetric Huber psi-function; asymmetric=TRUE
). The msvytotal
methods supports the following two methods (depending on the underlying survey design)
robust Horvitz-Thompson total estimator (type="rht"
),
robust weighted total estimator (type="rwt"
).
If y
is positively correlated with the inclusion probabilities a "rht" type estimator should be used, and "rwm" otherwise. The initial value is a weighted median or a ratio of weighted medians. You may set steps
equal to one in order to get a one-step estimation. Variance estimates are computed as first-order linearization using the design-based estimation facilities in the survey package.
msvytotal
allows also the estimation for domains. Use the command subset
and a design subset expression instead of the original survey.design
object in msvytotal
(see examples for more details).
Users may set exact=TRUE
to compute an "exact" linearization-variance estimate, which takes into account that the MAD has been used as preliminary scale estimate. However, the estimates may become very unstable.
Object of class "svystat.rob"
The following (S3) methods are defined for objects of the class "svystat.rob"
:
print
method,
summary
method,
coef
method,
vcov
method,
residuals
method,
robweights
method.
Beat Hulliger and Tobias Schoch
Hulliger, B. (1995): Outlier robust Horvitz-Thompson estimators, Survey Methodology 21 (1), pp. 79-87.
Hulliger, B. (1999): Simple and robust estimators for sampling, Proceedings of the Survey Research Methods Section, American Statistical Association, 1999, pp. 54-63.
Hulliger, B. and T. Schoch (2011): Elementary robust estimators. In: Robust methodology for Laeken indicators: AMELI Deliverable D4.2, ed. by B. Hulliger, A. Alfons, P. Filzmoser, A. Meraner, T. Schoch and M. Templ. AMELI Project.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 | ## load "api" data set from "survey" package (a description of the data
## set can be found there)
data(api)
## define "survey.design" for stratified sampling
dstrat <- svydesign(id=~1,strata=~stype, weights=~pw, data=apistrat,
fpc=~fpc)
## compute a robust Horvitz-Thompson estimate for the total of the
## variable "api00" (Academic Performance Index in 2000)
rht1 <- msvytotal(~api00, dstrat, k=1.2)
# get a summary of the estimation
summary(rht1)
## robust Horvitz-Thompson estimates for a domain of the variable. Here
## we are interessted in the robust total for api00 for
## (sch.wide == "Yes"). That is the average of the academic performance
## in 2000 only for the schools that met the school-wide growth target.
msvytotal(~api00, subset(dstrat, sch.wide == "Yes"), k=1.2)
## to extract the estimate from the object
coef(rht1)
## to extract the variance from the object
vcov(rht1)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.