In plain language
When researchers study whether a therapy works, they almost always report group averages — the typical patient improves by a certain amount. But a fundamental assumption hides inside that practice: that what is true of the group average is also true of each individual. This paper argues that assumption is often wrong, and that averaging can hide patients whose change over treatment looks nothing like the group trend.
The authors tested an "idionomic" alternative: analyze each person's data on its own terms first, then look for patterns across people, rather than starting from the group average. They followed 51 inpatients receiving Acceptance and Commitment Therapy (ACT), measuring psychological processes (such as defusion and values-based action) and symptom severity every week for eight weeks. They compared what a standard group-average (nomothetic) analysis showed with what a bottom-up approach — modeling each individual and then clustering similar people together — revealed. The outcome of interest was patients' well-being at the end of treatment.
The two approaches led to different conclusions. Individuals varied widely in how their psychological processes related to their symptoms — for one person a process like defusion tracked with fewer symptoms, for another with more, and for others not at all — so the average trend line did not represent most individuals well. The idionomic method readily identified subgroups of patients with distinct change pathways, and those subgroups predicted well-being at post-treatment. The authors conclude that characterizing data individually before generalizing is more scientifically rigorous and clinically useful, and could help advance genuinely individualized psychotherapy.
Key findings
- 51 inpatients receiving Acceptance and Commitment Therapy completed weekly measures of psychological processes and symptom severity over eight weeks, with well-being at post-treatment as the outcome.
- The same data analyzed with a standard group-average (nomothetic) method versus an idionomic (individual-first) method led to different clinical conclusions.
- Individuals differed in how underlying processes were linked to symptoms — e.g., defusion was associated with fewer symptoms for some, more symptoms for others, and no relationship for others.
- Average trend lines did not represent most patients' actual intraindividual change; the data did not meet the "ergodic" assumption that the group model applies to every individual.
- The bottom-up idionomic approach identified meaningful subgroups of patients that differentially predicted the distal outcome of well-being.
- Relying only on averages can obscure important individual pathways; starting with idiographic analysis yields more refined, individualized, and clinically useful conclusions.
How to cite
APA
Gloster, A. T., Nadler, M., Block, V., Haller, E., Rubel, J., Benoy, C., Villanueva, J., Bader, K., Walter, M., Lang, U., Hofmann, S. G., Ciarrochi, J., & Hayes, S. C. (2024). When average isn't good enough: Identifying meaningful subgroups in clinical data. Cognitive Therapy and Research, 48(3), 537-551. https://doi.org/10.1007/s10608-023-10453-x
BibTeX
@article{gloster2024when,
author = {Gloster, Andrew T. and Nadler, Matthias and Block, Victoria and Haller, Elisa and Rubel, Julian and Benoy, Charles and Villanueva, Jeanette and Bader, Klaus and Walter, Marc and Lang, Undine and Hofmann, Stefan G. and Ciarrochi, Joseph and Hayes, Steven C.},
title = {When average isn't good enough: Identifying meaningful subgroups in clinical data},
journal = {Cognitive Therapy and Research},
year = {2024},
volume = {48},
number = {3},
pages = {537--551},
doi = {10.1007/s10608-023-10453-x}
}
Related work
- All publications by Joseph Ciarrochi (searchable, with free PDFs)
- Process-Based Therapy & Idionomic Analysis
- The Process-Based Assessment Tool (free download)
Author: Joseph Ciarrochi (ORCID 0000-0003-0471-8100). Free copy hosted with permission for scholarly use. Please cite the published version via the DOI above.