Four types of video surrogates were compared under two tasks. Multidimensional scaling was used to map dimensional dispersions of users' judgments of similarity between videos and surrogates. Congruence between these maps was used to evaluate representativeness of each surrogate type. Congruence was greater for image-based than for text-based surrogates overall, while congruence between text-based surrogates and videos was greatest when a specific task was introduced. (Contains 25 references.) (AEF)