Viewer Recall of Television Commercials: Prediction from the Propositional Structure of Commercial Scripts

Abstract

The authors investigate a model for using what is said in a television commercial to predict what viewers recall of it. The psycholinguistically based ad language model (ALM) was applied to analysis of the propositional structure of 19 commercial scripts and 941 associated viewer recall protocols. The protocols came from either day-after telephone interviews by market research agencies or immediate free recall after forced exposure. Stepwise regressions with ALM script measures as predictors accounted for significant portions of the variance in ALM measures of how many ideas viewers would generate when recalling, how much information about execution and product characteristics they would recall, and their use of brand name and descriptive words. ALM script measures also predicted the day-after recall scores available for 16 of the commercials. Linking of the propositional structures of commercials with their recall leads to an improved foundation for use of recall in the evaluation of advertising messages.