Nonreactive measures in psychotherapy outcome research

Abstract

Attempts to assess outcomes of psychotherapy have relied too heavily on measures likely to be highly reactive, by which is meant that the processes of measurement affect what is being measured. Commonly employed measures of outcome such as therapist ratings or self-report are especially susceptible to effects attributable to knowledge of what is being measured. There are advantages in developing and using measures that, even though imperfect in other ways, are minimally reactive. Nonreactive measures should be considered as supplementary and not as substitutes for other measures. Problems with nonreactive measures involving ethics, validity and other psychometric limitations are troublesome but not insurmountable. Promising categoires of nonreactive measures include physical traces of past behaviors, archival records, nonverbal behaviors and observations carried out in contrived situations. Measures with dependably low levels of reactivity, including physiological measures, should also be considered for use.