The proximity coefficient was developed as a way to examine the dependencies among behaviours within a sequence (e.g., in a dyadic interaction; Taylor, 2006). It is promising approach because it avoids several extrinsic statistical assumptions that are common among existing methodologies. You can read about the conceptual and statistical basis of the coefficient in Taylor (2006).
Uptake of the coefficient has been hampered by the fact that users rely on hand calculations supported with some inadequate Perl script (Yes we did realise!). This inaccessibility has not only led external researchers to shy away from using the method, but it also limits our capacity to conduct analyses internally. However, as of August 2010, a new Java based program for calculating the coefficient is available. It runs through a web browser so it works regardless of your operating system.
This software contains a number of exciting new features, as well as more robust implementations of the features associated with the old Perl based ProxCalc (for those brave enough to have tried it!). Highlights of this software include:
• Easy to manage case selection and editing tools
• Graphing of state transition diagrams that allows interactive movement of nodes and connectors (hurrah!)
• A more comprehensive set of features for comparing across groups, allowing up to 3-way ANOVA designs and allowing for contingency selection
• More comprehensive (and nicer looking!) reporting of the results
Please use the links at the top to get the software and user guide.
Worst comes to worst, you can always email…