Survey Software: R survey functions

Summary of survey software: R survey functions

Distributor

R Project (the R Foundation)

The Survey functions for R were contributed by Thomas Lumley, Department of Biostatistics, University of Washington.

Types of designs that can be accommodated.

Designs incorporating stratification, clustering, and possibly multistage sampling, allowing unequal sampling probabilities or weights.

Types of estimands and statistical analyses that can be accomodated.

mean, quantiles, variance, tables, ratios, totals
generalised linear models (e.g. linear regression, logistic regression, Poisson models, etc.)
proportional hazards models
post-stratification and raking
tests of association in two-way tables

Restrictions on number of variables or observations.

Only those due to limitations of available memory or disk capacity.

Primary methods used for variance estimation.

Taylor series linearization
Replication weighting

General description of the "feel" of the software.

Functions to be used within a programming language, with each call to be entered from the command line or included in scripts or programs. R is an open-source implementation of the popular S language (commercially distributed in the S-Plus package).

Platforms on which the software can be run.

R can be run on

Intel computers with Windows 95 or better
Mac OS 8, 9, X
Linux
Most Unix systems.

R is distributed as precompiled binaries for Windows, Mac, and as auto-configuring source code for Unix. Most parts of the code should be easy to port to work with S-Plus; some suggestions are included in the package.

Pricing and terms.

Free download (GNU LGPL). R is updated about twice per year and the survey package is updated as needed. The current (11/2004) versions are 2.0.1 for R and 2.8-1 for the survey package.

Contact information.

For information on R see http://www.r-project.org/. To download R, select a CRAN (R archive) site from http://cran.r-project.org/mirrors.html. To download the survey analysis package, visit a CRAN site, select "Package sources" from the menu, and select "survey". Then you can download source for the package and a Reference Manual in PDF. Binary packages for Windows and soon (10/2003) for Mac OS X can be installed from menu options in R.

The home page for this package, maintained by the author, Thomas Lumley, is at http://faculty.washington.edu/tlumley/survey/.

Manual pages can be read at http://finzi.psych.upenn.edu/R/library/survey/html/00Index.html.

Return to main page for survey software

DISCLAIMER