This paper presents an information-processing model that is directly applicable to the investigation of how mediated messages are processed. It applies the model to the case of television viewing to demonstrate its applicability. It provides a measure for each part of the model. It presents evidence that supports the model in the television-viewing situation. Finally, it demonstrates how the model may be used to further research and understanding in well-known theoretical traditions. This model is not meant to stand in opposition to any of these theories but, rather, should work well with them by providing hypothesized mechanisms that may underlie well-known effects. This model should prove useful both to researchers and, eventually, to message producers. To the extent that we can better understand how the content and structure of messages interact with a viewer's information-processing system to determine which parts and how much of a communication message is remembered, we will make great strides in understanding how people communicate.