The logic of observation in psychotherapy research
An analytical study of methodology in psychotherapy research was performed. Although the study selected clinical research for analysis it addressed basic, quite general methodological issues. Thus, the outcomes of the study--analytical tools, systems of observation, and approaches to mathematization of data--apply not only to psychotherapy research, but are broadly relevant to experimental psychology in general.
The study began with a comprehensive review of psychotherapy research methodology, emphasizing traditional approaches, problems, and recommended solutions. As a result, it was concluded that two fundamental processes in which research is formally grounded--observation and mathematization of acquired data--had so far failed to receive adequate analytical treatment congruent with their importance.
Consequently these two principal processes were analyzed through formal models. Observation was examined in terms of a state process model; 14 observational systems prominent in psychotherapy research were analyzed. A parallel analysis of mathematical operations was based on relational systems concepts familiar from measurement theory.
The dual analysis led (1) to a design philosophy for observational systems and (2) to the recommendation to return for the present to a logically primitive mathematical system, namely, the relation scale. The design philosophy suggested an ideal of complete, pragmatically conceived observational systems: A family of multimodal, multilevel, modular systems was envisioned.
A prototype system was constructed to illustrate the recommended approach to experimental observation. It consisted of loosely-knit diverse subsystems drawn mostly from the literature and of newlyproposed subsystems which would allow idiographic representation of events. This system appeared to be highly flexible, and easily capable of improvement and modification.
Future research areas were identified and discussed. These areas included completeness, redundancy, reliability, as well as the use of relation scales to mathematize data. The topic of other, alternative observational systems in the proposed family was also considered.
Thesis74b.B473.pdf
6.83 MB
Adobe PDF
485e392006f5edb75e5f802fd54a1456