Work

Better Understanding Whether and How Psychotherapy Works: Novel Statistical Applications

Year: 2023
Type: dissertation
Abstract: Randomized controlled trials (RCTs) provide critical empirical insights about which psychotherapies are effective and in what ways, and also can shed light on why specific psychotherapies are effective. While there are well-established and commonly-used statistical methods for analyzing whether and how psychotherapies work using RCT data, those methods are limited in terms of modeling numerous process and outcome variables simultaneously and in examining participants’ variance in treatment response as a function of RCT condition assignment. This doctoral dissertation seeks to test alternative statistical approaches that address the above limitations using data from the Valued Living trial (Arch, Mitchell, et al., 2021) an RCT of group-format Acceptance and Commitment Therapy (ACT) versus minimally enhanced usual care for anxious cancer survivors. First, comparing standard mediation analysis to network intervention analysis, I demonstrate that modeling associations between and among putative process and outcome variables when examining treatment mechanisms leads to remarkably different findings regarding the mechanisms. Next, I present the application of multiple-groups latent growth curve modeling to compare consistency in treatment responses between two RCT conditions. I demonstrate that ACT participants showed improved consistency relative to usual care on two RCT outcomes, and describe findings from Monte Carlo simulations that guide future applications of this modeling technique. I conclude by highlighting clinical and empirical implications of both the study findings regarding ACT and use of the examined statistical modeling techniques more broadly in psychological intervention research.
Language: English
Field: Psychology
Open Access status: green
OpenAlex ID: W6998938874
Cites: 147
Cited by:
Related to: 10