SABRE - Software for the Analysis of Binary Recurrent Events
SABRE is a program for the statistical analysis of binary, ordinal and
count recurrent events. Such data are common in many surveys either with
recurrent
information collected over time or with a clustered sampling scheme. It is
particularly appropriate for the analysis of work and life histories, and has
been used intensively on many longitudinal datasets. Its development has been
funded by ESRC, ALCD and Lancaster University. In 1989, SABRE 2.0 was released,
written by Jon Barry, Brian Francis and Richard Davies. SABRE 3.0, developed
by
Dave Stott, with
substantially enchanced statistical facilities, was released as freeware on
the WWW in 1996. The current release is version 3.1. As part of the ALCD
initiative,
SABRE was also incorporated as a function in the S-plus
Lancaster University OSWALD package , which provides comprehensive
facilities for the analysis of longitudinal data.
The details of using and loading OSWALD are not given here but can be found on
the OSWALD website.
Specification
- A command driven package, with over 35 commands. However a basic set of
only a few commands is needed to fit models to data.
- Fits the mover-stayer model, conventional logistic, logistic-normal and
logistic-normal with end-points models to binary data.
- Fits a continuation ratio-type generalisation of the above models to
ordinal data.
- Fits conventional log-linear, log-linear normal and log-linear normal with
end-point models to count data.
- Substantial control is available over the parameters of the algorithm for
the sophisticated user
- Can deal with very long sequences of data
- Comprehensive user manual and on-line help system.
Typical Applications
- Studies of voting behaviour, trade union membership, economic activity and
migration.
- Demographic surveys
- Studies of infertility in humans
- Animal husbandry
- Absenteeism studies
- Clustered sampling schemes
For lots more information, see the
SABRE web
pages
