Summary of survey software: CLUSTERS


Professor Vijay K VERMA.
ESRC Research Centre
University of Essex

Types of designs that can be accommodated.

The program computes sampling errors taking into account the actual sample design. There are no specific restrictions on the designs handled, except the basic ones of 'large enough' sample size and the availability of two or more independent primary selections per stratum with replacement. The finite population correction is disregarded. Flexible facilities for specifying computing strata, primary units and sample weights are included

Types of estimands and statistical analyses that can be accomodated.

Computes sampling errors and derived statistics such as design effects and intra-cluster correlations for ratios and their differences over population subclasses.

Restrictions on number of variables or observations.

None explicitly; limited only by the available memory.

Primary methods used for variance estimation.

Taylor linearisation.

General description of the "feel" of the software.

Standalone. Designed to minimise the need for manipulation of input data files, though the data must be in text mode and appropriately sorted. Hierarchical file structures are handled, though rectangular files remain most convenient. Facilities for variable recoding and for identification and exclusion of missing values are included.

Platforms on which the software can be run.

IBM PC or compatibles (286 or higher), with math co-processor.
Distributed as pre-compiled FORTRAN programs. No special requirements.

Pricing and terms.

Distributed with documentation free of charge to organisations and individuals in developing countries. For others, a nominal handling charge is requested.

Contact information.

Professor Vijay K VERMA.
ESRC Research Centre
University of Essex, Wivenhoe Park
Colchester CO4 3SQ, U.K.
Tel (44 1206) 87 3545 Fax (44 1206) 873151
105 Park Road, Teddington
Middlesex TW11 0AW, U.K.
Tel (44 181) 977 2601 Fax (44 181) 943 4056
E-mail: 100626.3055@CompuServe.Com

Additional information.

Main limitations: The program is designed to facilitate routine, large scale computation of sampling errors in social surveys. It does not directly provide information on variance components, or handle statistics more complex than differences of ratios.

This software is discussed in the review article from The Survey Statistician.

Return to main page for survey software