THE MEDIA EQUATION describes research into human-media interactions from two media researchers who go on to offer design recommendations for media as diverse as word-processor software and film. The authors took a number of well-replicated psychology experiments in which human interactions were studied, and they conducted these experiments using humans interacting with various media -- from simple text interfaces to video clips. The experiments range from measuring emotional response to mediated information to measuring human ability to consider a computer part of a "team." They discovered that people's responses to media are fundamentally social and natural. The book is fascinating for what it tells us about human responses to media, and it is also a highly appealing digest of social science experience in the measurement of human interaction.

CSLI was founded in 1983 by researchers from Stanford University, SRI International, and Xerox PARC to further research and development of integrated theories of language, information, and computation. The findings presented in The Media Equation grew out of CSLI investigators Reeves and Nass's research into Social Responses to Communication Technologies (SRCT), a multiyear research program of experimental studies that demonstrate the ways in which users respond socially to information technologies. The Media Equation is divided into five sections -- treating manners, personality, emotion, social roles, and form -- and every section reflects the same methodology: Social science models were used to ask questions about humans interacting with media.

Reeves and Nass demonstrate that many unquestioned notions about media are simply false. Among their more eye-catching conclusions is that even people who state that they do not consider a PC a social actor do, in fact, respond to PCs as if they are social actors. The authors' study subjects represented many groups and many ages and included sophisticated users of computers and other media. The Media Equation shows that politeness to a computer or perception of a television's expertise are not marks of inexperience or ignorance, but simple reflexive responses. People respond to media the way they respond to people because it is simpler to respond naturally. Moreover, only minimal cues are enough to elicit these social responses, which mediate perceptions as diverse as personality of the medium and production quality.

The authors are aware of how deeply counterintuitive some of their results are to sophisticated computer users, and while they are quick to point out that even sophisticated users had fundamentally social responses to media, they offer little detail about some of the logical extensions of that. Reeves and Nass suggest, for example, that cartoon characters that appear on a computer screen will be palatable vehicles for help and friendly error messages. The key is probably configurability; novice users will find them palatable, and sophisticated users can turn them off.

People don't just use media as tools, they also react to them socially, unless they are really proficient, in which case they want them at least to act like tools. This potential wrinkle in their thesis is not amplified, although ultimately it may not even be relevant. Reeves and Nass have shown that even people who would likely dismiss cartoon help characters as silly toys can still have fundamentally social responses to computers -- sophistication doesn't disengage the brain's tendency to do the simple, social thing even if it emphasizes different parts of the computer-using experience -- and their interface suggestions are geared for the most part toward commercial products or other applications that expect a very large user base.

The Media Equation's implications range widely, from media design to management principles. Reeves and Nass are interested in asking their questions about human responses to media, but the descriptions of the models for the experiments themselves are interesting, nicely described, and extensively referenced. The social science experiments adapted for this study are certainly available, but it would be difficult to construct a more appealing tour of some of these very basic conclusions about human behavior. Reeves and Nass also offer a fine example of the strength of interdisciplinary investigation, setting this engaging stage and presenting their conclusions about media with enviable economy, clarity, and charm.