Abstract
AbstractObjective. We developed the Nonverbal Individual-Level Coding and Observation (NICO) Manual, a concise, descriptive, and low-inference coding manual that captures observable nonverbal behavior across body regions at the individual level. The manual was designed for communicative interaction and tested across heterogeneous settings, with psychotherapy as its primary clinical application.Method. The manual and its coding rules were refined through pilot coding, coder feedback, and repeated agreement checks. The final manual contains 20 directly coded behavioral categories and one derived gaze state across seven body-region tiers. Each person was coded separately on a fixed grid of consecutive, non-overlapping two-second intervals beginning at video onset. Preliminary interrater agreement was examined across 2,952 two-second interaction intervals, each coded separately for the two individuals from three adult interaction paradigms: dyadic interaction, emotion induction, and psychotherapy setting.Results. Overall, Cohen's kappa, pooled across all binary code-by-interval decisions within each sample, was consistently high across the three paradigms (κ = .78-.81), with observed agreement of .94 and Prevalence-Adjusted Bias-Adjusted Kappa (PABAK) values of .87-.89. Code-level results were more heterogeneous: eight of 20 codes (including gaze, speaking-related mouth movement, gestures, and hand-to-head contact) reached κ ≥ .60 in every sample in which κ was estimable. Lower code-level κ values often occurred for rare behaviors despite high observed agreement. Conclusion. The findings provide initial empirical support for a novel and methodologically concise NICO manual that translates continuous video recordings into transparent, time-aligned descriptions of recognizable nonverbal behavior. Comparable overall agreement across three markedly different settings supports its cross-context potential and identifies clear priorities for independent validation and code-specific refinement.