Evaluating post-intervention effects by comparing alternative structural equation modeling: decision-trees (DT) based on formal model comparisons This paper focuses on common group invariance assumptions behind tests of post-intervention differences in outcomes. A decision-tree (DT) comparison approach is proposed for structural equation models (SEM) which evaluate preventive intervention effects according to model fit to the data and statistical power to detect the outcome effects. DT for SEM rank-orders competing models and illustrates potential different outcome evaluation conclusions. The method is exemplified with the Youth Action Research for Prevention (YARP) project, a risk prevention youth development program for low-income inner city youth (average age 15.3 years) in Hartford, Connecticut. Eight well-fitting models had enough statistical power and demonstrated positive intervention effects on internal locus of control (ILC) in the intervention YARP youth group, while six alternative well-fitting models with insufficient power showed no effects. DT for SEM proves valuable for constituent policy makers and organizations in nuanced decision-making on funding or replicating preventive interventions based on most plausible evidence.#QuantitativeMethods-TheoryandDesign #Youth #fit #prevention #power #Alcohol,DrugAbuse,andMentalHealth #Control #locus #action #of #participatory #research #2010Conference #risk