Oswald

Object oriented Software for the Analysis of Longitudinal Data in S-plus

24 June 1998 -- Version 3.4 available

Oswald, developed by the Statistics Group at the University of Lancaster, is a suite of S-plus functions for analysing longitudinal data. The features it currently provides include:

Data structures for single time series and associated operations
Data structures for balanced and unbalanced collections of time series, with operations for:
- Mathematical operations, selection, and comparison
- Printing and Plotting
- Interactive graphics
Analysis methods:
- Ordinary least squares analysis
- Kernel smoothing with cross-validation
- Variogram analysis
- Parametric modelling of covariance structure with REML estimation
- Informative dropout models
- GEE and ALR analysis
- Mixed-effects models

A more detailed description of the Oswald features is available.

Oswald version 3.4 is now available

It is available for Splus 4.5 for Windows, and Unix systems. See the NEWS file for more details of new features in the latest version.

This cartoon will give you some idea just how many facets Oswald has.

Oswald is still under development, however is is available for beta-testing. Oswald may be freely redistributed under the terms of the GNU General Public Licence.

More information about Oswald

Oswald README file
Slides for a 3-day lecture course on Oswald version 3.4. In Powerpoint 97 format (610Kb) and Postscript format (1.04Mb, compressed 600dpi A4 postscript). The course used S-plus 4.5 for Windows, but almost all of the material applies equally well to the Unix version.
Brief introduction (132Kb PostScript)
The Oswald manual, in A4 format (170Kb gzipped PostScript) and in US letter format (170Kb gzipped PostScript)
A draft copy of the manual in HTML format
Set of presentation slides about Oswald. These were created from LaTeX source using the seminar style.

Software Available

Oswald is available for both Windows and Unix platforms. Both distributions contain full S-plus and C source code, a PostScript version of the manual and full installation instructions, which you should read before downloading the software.

Unix Version

Requirements: A version of Splus 3 from Splus 3.4 onwards (Oswald will not work as supplied with Splus 5). An ANSI C compiler (e.g. gcc).

Oswald for Unix version 3.4 (773Kb compressed tar file).

Windows version

Requirements: S-PLUS for Windows version 4.5. At least 32Mb of RAM (more if using NT). You'll also need a relatively new version of PKUNZIP; if you don't have it then try this one, and make sure you unzip with the -d option. Note that no compiler is required (pre-compiled binaries are supplied).

Oswald version 3.4 for S-PLUS for Windows 4.5 (997Kb zip file).
Oswald version 3.1 for Splus for Windows 3.3 (997Kb zip file).
Oswald version 3.2 for Splus for Windows 4.0 (1.159Mb zip file).

Note: Version 3.2 is works with Splus 4.0. Splus 3.3 users should use Oswald version 3.1.

Bugs

Oswald is still in development and there may be missing or broken features. Bug reports and feature suggestions may be sent by email to oswald-bugs@lists.lancs.ac.uk. For general queries about using Oswald or longitudinal data analysis in general, we suggest mailing to the Oswald Discussion List (oswald@lists.lancs.ac.uk) where questions and answers may be of benefit to all Oswald users.

Please note that as free software, Oswald is supplied without any warranty; see the copying conditions for further details.

The Oswald Mailing List

If you retrieve this software, you should join the Oswald Electronic Mailing List to be informed of new releases. Send mail to oswald-request@lists.lancs.ac.uk with the word "subscribe" in the body to be added to the list.

The development of Oswald is funded by the Economic and Social Research Council under the ALCD (Analysis of Large and Complex Datasets) initiative.

Related Software

Links to other software for mixed modelling may be found at the MLN Multilevel Modelling pages.

Dave M. Smith <D.M.Smith@lancaster.ac.uk>

Last modified: Mon Dec 11 15:48:16 GMT 2000