Skip to main content
ABCT 2023 Annual Convention Main banner
Final Program


Program Flipbook
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Presentation Icons
Ticketed Session
Ticketed Session
Recorded Session
Recorded Session
Back
Evaluation ViewAttendees 1

Category: Research Methods and Statistics

Symposium 37 - Caveats and recommendations in using novel analytic methods: Application to clinical psychological data

Friday, November 17, 2023
2:30 PM – 4:00 PM PST
Location: Washougal (408), Level 4
CE: 1.5

Keywords: Statistics, Longitudinal, Psychotherapy Outcome
Level of Familiarity: Basic to Moderate
Recommended Readings: Asparouhov, T., Hamaker, E. L., & Muthén, B. (2017). Dynamic structural equation models. Structural Equation Modeling: A Multidisciplinary Journal, 25(3), 359-388. https://doi.org/10.1080/10705511.2017.1406803,  

Wright, A. G. C., & Woods, W. C. (2020). Personalized models of psychopathology. Annual Review of Clinical Psychology, 16(1), 49-74. https://doi.org/10.1146/annurev-clinpsy-102419-125032

,  

Lutz, W., Schwartz, B., & Delgadillo, J. (2021). Measurement-based and data-informed psychological therapy. Annual Review of Clinical Psychology. https://doi.org/10.1146/annurev-clinpsy-071720-014821

,   Liu, Q. (2023) A Discrete Latent State-Trait Model, Multivariate Behavioral Research, DOI: 10.1080/00273171.2022.2160297,  

Jordan, D. G., Winer, E. S., & Salem, T. (2020). The current status of temporal network analysis for clinical science: Considerations as the paradigm shifts? Journal of Clinical Psychology, 76(9), 1591-1612. https://doi.org/10.1002/jclp.22957

Advanced statistical methods improve our ability to account for complex psychological processes and address novel clinical research questions. Examples include emerging analytic approaches that can handle intensive longitudinal data and capture fine-grained, within- and between-person dynamics occurring in daily life. Such methods can contribute to measurement-based care, routine outcome monitoring, and enhanced clinical outcomes (Lutz et al., 2021). Nonetheless, researchers can face challenges in applying these novel analytic methods to “real-life” clinical data, which are often messy and imperfect. Common challenges include sparse data due to low participant compliance and deviations from required statistical assumptions. Potential pitfalls also lie in adhering to conventional analytic practices without recognition of their limitations. The key to effectively applying advanced statistics to future clinical research is to consider these challenges and educate ourselves about possible solutions. To this end, this symposium presents innovative developments and improvements in clinical data analysis, with an emphasis on potential caveats and practical recommendations for analytic decisions.

 

The first presentation will discuss the use of temporal network analysis and sleep diary to examine the inter-daily relationships between sleep variables, depressed mood, and stress. This presentation will discuss the best practices in estimating temporal networks and recommendations for accommodating fewer than ideal data points. The second presentation uses dynamic structural equation modeling (DSEM) to examine within-person emotional processes and their associations with suicidal ideation among adult psychiatric inpatients following discharge. This presentation will discuss methodological challenges in collecting intensive longitudinal data in a high-risk clinical sample and analytic choices to make in applying DSEM. The third presentation will give an overview of an innovative approach to examining both state-level and trait-level latent classes using intensive longitudinal data. This study will present simulated data examples to demonstrate how heterogeneity can be modeled at momentary state levels and stable trait levels while accounting for potential trends across or unique to trait-level classes. The fourth presentation discusses how the current practices of calculating effect sizes and reliable change, two commonly used metrics of intervention effects, can be susceptible to the influence of sampling errors. This presentation will provide simulation results and recommendations for improving the comparability of the metrics across treatment trials. Lastly, the symposium discussant, a senior faculty member with expertise in multivariate statistics in psychopathology research, will discuss the broader implications of these presentations for furthering clinical psychological science.

Learning Objectives:

At the end of the session, the learner will be able to:
  • Describe the optimal approaches to determine longitudinal networks and remedying less-than-ideal counts of data points.
  • Report the caveats and recommendations for collecting ecological momentary assessment data with high-risk samples.
  • Communicate the decision points using advanced intensive longitudinal data methods such as dynamic structural equation models.
  • Verbalize the clinical applications for differentiating symptom-linked variables at the stable trait level and momentary state levels with ILD.
  • Relay the standard methods and principled recommendations for computing treatment effect sizes and reliable change based on simulation studies.