In-session client involvement (i.e., participation in in-session therapeutic tasks) is hypothesized to be a necessary component of youth therapy and associated with positive outcomes. However, research on in-session client involvement has been slowed by definitional problems. At present, the field has not yet adopted a single definition of client involvement that is applicable across different theoretical orientations, which has impacted the measurement of this construct. To remedy this problem, the field needs to adopt a definition of in-session client involvement that includes important components (i.e., behavioral, affective, and cognitive) of this construct that applies across different theoretical orientations and use this definition to guide instrument development. The current study reports on the development and initial psychometric assessment of the Therapy Process Observational Coding Scale – Involvement (TPOCS-I), an observational measure designed to capture in-session involvement for youth therapy. Treatment sessions (N = 895) were drawn from (a) 55 youth (ages 7-13 years; M = 9.89, SD = 1.71; 51.5% Caucasian; 58.8% male) who received standard cognitive-behavioral therapy, modular therapy, or usual care for youth anxiety; and (b) 51 youth (ages 7-14; M = 10.35, SD = 1.89; 86.3% Caucasian; 60.8% male) receiving standard cognitive behavioral therapy for youth anxiety. Sessions were independently scored by seven coders using observational instruments designed to assess involvement, alliance, therapist competence, and therapist interventions. Interrater reliability – intraclass correlation coefficients (2,2)—for the item scores averaged 0.73 (SD = 0.08) and 0.82 (SD = 0.08) for the Kendall and Child STEPS samples, respectively. The TPOCS-I scale and subscale (Behavioral, Affective, Cognitive, Positive, Negative) scores failed to demonstrate discriminant validity from the alliance. The use of two subscale configurations (i.e., Behavioral, Affective, Cognitive; and Positive, Negative) was not supported. These findings are discussed and future directions including measure development in a sample of youth with diverse diagnoses and the use of different perspectives in the measurement of in-session involvement.