Studies using observational measures often fail to meet statistical standards for both reliability and validity. The present study examined the psychometric properties of the Coding Interactive Behavior (CIB) System within a German sample of parent–child dyads. The sample consisted of 149 parents with and without a mental illness and their children [n experimental group (EG) = 75, n control group (CG) = 74] who participated in the larger Children of Mentally Ill Parents at Risk Evaluation (COMPARE) study. The age of the children ranged from 3 to 12 years (M = 7.99, SD = 2.5). Exploratory factor analysis supported a five-factor model of the CIB with items describing 1) parental sensitivity/reciprocity, 2) parental intrusiveness, 3) child withdrawal, 4) child involvement, and 5) parent limit setting/child compliance. Compared to international samples, the model was reduced by two independent dyadic factors. Testing for predictive validity identified seven items with predictive power to differentiate parental group membership. The CIB factors did not seem to be sufficiently sensitive to illustrate differences in interaction within a sample of parents with various mental illnesses. To apply the CIB to the described sample or similar ones in the future, additional measurement instruments may be necessary.