Case 004 research candidate: the Mehrabian 7-38-55 rule

Source content

Case 004 research candidate: the Mehrabian 7-38-55 rule

Status: content-addressed author research packet complete; independent review pending. This note

does not admit or publish Case 004.

The candidate asks what the familiar 7% words / 38% tone / 55% face formula actually represented

in the examined editions. It does not ask for one universal percentage of communication.

The packet preserves five distinct layers:

1. P1 studied attitude inferred from inconsistent isolated words and tone.

2. P2 studied attitude inferred from facial and vocal cues carried with the neutral word maybe.

3. P2's discussion proposed .07/.38/.55 as a cross-study weighted sum; neither experiment

tested all three channels together.

4. The 1971 book repeated and narrowed the formula, while the author's public page later stated

that it does not apply outside feelings or attitudes.

5. A direct all-channel reconsideration reported equations that varied by message, while related

studies found that cue-strength and attitude context changed the verbal/nonverbal ratio.

The exact derivation of .07 remains unresolved and receives no evidential credit. The packet

also distinguishes two participant-data roots from full independence: the papers share an author,

a grant, a linked program, related methods and scales, a neutral-word stimulus, and an explicit

citation/derivation path. Participant and speaker overlap is recorded only to the extent the

sources disclose it.

Public recirculation examples establish that the broad rule traveled; they do not establish that

it is scientifically true. Their evidence credit is mechanically fixed at zero.

The author recommendation is to proceed only if a fresh independent review closes every source,

span, calculation, dependence edge, and limitation. A later dossier should teach how a narrow,

qualified proposal became a universal-sounding rule—not replace it with a different unsupported

percentage.

Full records and the pending review gate are in

research/how-we-know/mehrabian-7-38-55/.