Message impact: What percentage is nonverbal?

Mehrabian's formula for message impact was reconsidered. Multiple regression analyses of four messages were undertaken in which the all‐channel condition was the criterion variable and the verbal, vocal, and facial channels predictor variables. Regression equations for each message were quite different from Mehrabian's formula and from each other. A “contextual”; rather than an “additive”; model of nonverbal communication was supported.