Here is the introduction of his talk at the 2009 PAS Symposium in Toronto.
(Here is his YouTube Account - of particular note is Part 4 that deals with his arguments in favour of determining the "Scientific Validity" of PA within the Court Room.)
Do qualitative methods and PAS meet the Daubert criteria? You know, whats considered to be science? Well what are they? Testability, falsifiability, error rates, peer review, general acceptance, validity, reliability.
Well, whats testability? Its proof by disproof. Can we test ideas about parental alienation? Yeah, sure we can. PAS is certainly testable. Gardner described eight factors, we can test whether they exist or not. In fact Bricklin and Elliot are trying to develop inventories that have some inter-rater reliability and some test-retest reliability that can be used in family conflict circumstances. They can certainly test whether those are accurate or not. They can test whether or not they speak to a phenomenon of interest accurately or not. They can test whether or not inter-raters or test-retests find them to be reliable or not.