Category: Research Methods and Statistics
Ki Eun (Kay) Shin, Ph.D.
Assistant Professor
Long Island University - Post
Brookville, New York
Gemma Wallace, M.S.
Graduate Student
Colorado State University
Fort Collins, Colorado
Craig Henderson, Ph.D.
Professor
Sam Houston State University
Huntsville, Texas
Ki Eun (Kay) Shin, Ph.D.
Assistant Professor
Long Island University - Post
Brookville, New York
Patrice Arkfeld, M.S.
Graduate Student
Colorado State University
Fort Collins, Colorado
Alessandro De Nadai, Ph.D.
Assistant Professor
Texas State University
San Marcos, Texas
Steven Brunwasser, Ph.D.
Rowan University
Glassboro, New Jersey
Recent research on the transdiagnostic and comorbid features of mental health has emphasized the need for individualized psychological research and treatment. Advanced statistical methods that can handle high-dimensional data and account for complex relationships between variables improve our ability to capture people’s lived experiences in psychopathology research. Thus, advanced statistics have been increasingly applied to clinical science and have the potential to significantly improve individualized prediction and treatment of psychopathology. The key to effectively applying advanced statistics to future clinical research is to examine their applications in context. To this end, this symposium includes two presentations that provide overviews of innovative analyses for longitudinal psychopathology data, and two presentations that discuss the application of innovative modeling approaches to inform individualized intervention recommendations. Presentations will highlight statistical approaches that can address common challenges in psychopathology research.
The first presentation will discuss the use of mixture modeling to examine heterogeneity in interpersonal problems and generalized anxiety symptoms, discussing how unique profiles of interpersonal difficulties can inform individualized treatment plans for anxiety. This presentation will also discuss the integration of multi-informant data into mixture models via multilevel modeling. The second presentation will give an overview of an innovative approach to evaluating longitudinal data, including examples from a randomized controlled trial and a prospective cohort study examining behaviors that increase risk for infection during the COVID-19 pandemic. Specifically, this presentation will discuss how restricted cubic splines can address common problems in growth models of mental health constructs, including flexibly modeling non-linear relationships and avoiding model overfit. The third presentation will discuss unsupervised machine learning (ML), an increasingly popular approach for predicting psychopathology outcomes. This presentation will discuss best practices and common pitfalls of using machine learning in behavioral sciences, and will show two examples of applying unsupervised ML to develop personalized treatment recommendations in response to health emergencies such as the opioid crisis. The fourth presentation provides an example of using ML to elucidate the nuanced role of intersectional minoritized identities in adolescent suicide attempts. This project employs classification trees and count regressions to examine how interactions between risk factors for suicide attempts vary across intersecting gender and sexual identities. Lastly, the symposium discussant, a senior faculty member with expertise in multivariate statistics in evidence-based treatment research, will discuss the broader implications of these presentations for furthering clinical science.
Presenter: Ki Eun (Kay) Shin, Ph.D. – Long Island University - Post
Co-author: Michelle G. Newman, Ph.D. – Penn State University
Presenter: Patrice A. Arkfeld, M.S. – Colorado State University
Co-author: Patrice A. Arkfeld, M.S. – Colorado State University
Co-author: Gemma T. Wallace, M.S. – Colorado State University
Co-author: Maggie Mataczynski, B.S. – Colorado State University
Co-author: Noah Emery, Ph.D. – Colorado State University
Co-author: Bradley T. Conner, Ph.D. – Colorado State University
Co-author: Mark A. Prince, Ph.D. – Colorado State University
Presenter: Alessandro S. De Nadai, Ph.D. – Texas State University
Presenter: Steven M. Brunwasser, Ph.D. – Rowan University