EMFACS is based on FACS and requires knowledge of and the ability to identify a) the action units involved in any facial movement, b) the intensity of the action units and 3) the degree of asymmetry. Action units where there is evidence they do not affect the interpretation of emotion are not noted.
EMFACS scoring is done without the aid of slowed motion viewing and the location of an expression is simplified by requiring only a single locational point near the beginning of the event. Intercoder reliability has consistently exceeded .80.
EMFACS measurement reduces scoring time for one minute of facial behavior from approximately 100 minutes of FACS scoring to approximately 10 minutes of EMFACS scoring. In one study of 24 psychiatric interviews, FACS and EMFACS scoring of the frequency of occurrence of specific emotions were moderately related (.646, p<.001). The largest source of disagreement was that FACS identified emotional expressions that EMFACS scoring missed. Since that study, the rules for EMFACS scoring have been clarified, but no systematic comparison of FACS and EMFACS has been made.
EMFACS should not be used if the study is primarily interested in blends, if the emotional expressions are likely to be disguised or highly controlled, or very subtle.
Standard FACS codes are used, but: 1) not all AUs are scored; 2) all the modifications described in the CHANGES IN FACS SCORING document are followed: intensity scored on 5 levels, cooccurence rules are not followed, AUs are scored during speech, and minimum requirement rules are dropped.