Balancing internal and external validity.

Abstract

Posted 03/21/1999. Designing studies of treatment effectiveness in public health settings requires balancing internal against external validity. Although these two dimensions of validity are often pitted against each other, there may be ways to maximize each in carefully crafted treatment trials. Several research recommendations are offered: (a) basic studies of predictor-outcome relationships that examine whether the theoretical models underlying different psychosocial treatment approaches apply equally in diverse public health settings, (b) hybrid treatment study designs in which, in one arm, a randomized study is undertaken and in another arm, treatments are selected on the basis of clinician and patient choice, and (c) studies of therapists-as-subjects, in which the investigator identifies community clinicians who can effectively administer the manual-based treatments and reliably administer the assessment instruments relevant to the population under study. [Comments are directed to articles by G. Norquist et al (see record 1999-11093-001), by D. F. Klein and L. B. Smith (see record 1999-11093-002), and by G. Niederehe et al (see record 1999-11093-003).] (PsycINFO Database Record (c) 2002 APA, all rights reserved)