Imputations of Missing Information to Incomplete Facial Expressions: A Study With Functional Measurement
AbstractImputation or "filling in" might be a ubiquitous process in a world of incomplete information.Commonly held assumptions are that default values are imputed in place of absentinformation or that people infer omitted information from the one they have at hand. Theoccurrence of imputation strategies in tasks where people were made to evaluate expressionsof incomplete schematic faces was investigated using Information Integration Theory andFunctional Measurement. An algebraic averaging model found for the integration ofschematic facial features (mouth, eyes, eyebrows) provided the basis for predicting what theeffects of different imputation strategies, manifest as discrepancies from the model, wouldbe. Despite strong claims for the primarily holistic processing of face stimuli, imputationinferences do not appear to bear a significant impact in these particular tasks.