Demonstration of Two Traditional Statistical Techniques for Use with Small Sample, within-Person Experiments: Unified Structural Equations Modeling and Mixed Model Trajectory Analysis

Idiographic (intensive, within-individual) research has played critical roles in sciences which inform etiology of psychopathology (e.g., neurology, psychology, biology, treatment research such as smoking cessation).  Some features of idiographic studies to fill research gaps which traditional research designs are incapable of addressing.  Prevention research, per se, has yet to develop consolidated lines of idiographic research.  Recent advances have adapted widely-used statistical techniques for idiographic analysis that control for sources of bias such as serial dependency.  By adapting statistical techniques from nomothetic approaches (large sample studies and traditional randomized clinical trials), identical effect sizes and parameters can be obtained, implications can be drawn across nomothetic and idiographic studies, and clinicians can determine how an individual patient compares to results reported for groups.  This presentation will demonstrate use of mixed model trajectory analysis and unified structural equations modeling to experimental idiographic tests of intervention efficacy and models of within-person processes that lead to those outcomes.  Three recent, distinct within-person experiments will be used to demonstrate these techniques.  The first program provided wheel-chair-bound military veterans (n=16) with mindfulness-oriented reminders to monitor their comfort level and accordingly adjust their wheelchair positioning.  This program improved compliance with the intervention regiment more than twofold of instruction alone; in turn, patient discomfort was reduced by large effect sizes.  The second tested a novel smoking cessation intervention for a difficult-to-treat and retain population – homeless men recovering from drug abuse (n=24).  The direct effect of self-reported emotional state prior to smoking was demonstrated in addition to intervention efficacy.  The third study tested a program to improve communication skills of a child with autism in terms of (a) how well intervention impact generalizes from a speech pathologist delivery to delivery by family members and (b) the degree to which reduction in the patient’s level of distress corresponds to intervention outcomes in communication.  Outcomes of family member-led intervention were equivalent to or better than the speech pathologist-led outcomes; patient distress directly predicted response to intervention.  These studies highlight multiple strengths of idiographic research that complement traditional prevention research, including the ability to test hypothesized “active ingredients” of interventions within highly specified populations (e.g., persons with a particular genotype), experimental testing of person-centered intervention, quantifying short-term within-individual change, application of experimental methods in clinical settings, use of N-of-1 designs (regardless of sample size), and research questions that can be answered.  How these methods can be used for “back-translation” research will be discussed.