“…The changes in ratings reflected in the then-test procedure were more strongly associated with changes in observed behavioural indicators of interviewing skill (median r = .33) and in judges' ratings of interviewing skills (median r = .43), than were changes reflected in the pretest/post-test procedure (median r values of ¡.05 and ¡.06, respectively). Studies using this methodology have examined the effectiveness of interventions in a variety of contexts, including teacher perceptions of effectiveness (Bray & Howard, 1980), level of job satisfaction (Gutek & Winter, 1992), perceptions of quality of life in cancer patients (Breetvelt & van Dam, 1991), effectiveness of a counsellor education programme (Manthei, 1997), and changing attitudes towards persons with disabilities (Lee, Paterson, & Chan, 1994). As just reported for Howard and Dailey (1979), a number of comparative studies have shown that results from retrospective analyses of change are more similar to objective ratings of change in behaviour or performance than are analyses using the pre/post procedure (Hoogstraten, 1982;Pohl, 1982;Sprangers & Hoogstraten, 1989).…”