Two Methodological Problems in Psychotherapy Research

University Of Memphis, Memphis TN

Investigators

Linked publications & trials

Abstract

DESCRIPTION: Evidence-based psychotherapy practice is hampered unless psychotherapy research is methodologically sound. The objective of this proposal is to highlight two common methodological problems in psychotherapy research and to develop and disseminate methods to cope with those problems. First is publication bias. The proposed research will review psychotherapy meta-analyses to assess whether and how meta-analysts attend to publication bias. This research will then describe and apply modern methods for detecting publication bias to psychotherapy meta-analyses. Second, this research will help psychotherapy researchers to recognize and appropriately deal with statistical dependencies in their data. This effort will begin by developing a database of Intraclass Correlations (ICC) for psychotherapy outcomes, by surveying the psychotherapy literature and by asking researchers to provide their data so that ICCs can be computed. Researchers will be able to use this database to properly plan studies or when they use statistical methods that require external estimates of the ICC. Consequently, this research will also describe, and in some cases develop, statistical methods for properly analyzing such data, with applications to psychotherapy [unreadable] [unreadable]