The "symbol systems approach" to the study of computers in education that is outlined avoids the pitfalls of past media research--particularly research on the effects of television on children's learning and knowledge--and asserts that media can be usefully distinguished in terms of the symbol systems they present and the kinds of symbol-using skills they evoke, afford, or inculcate. Early media research questions are criticized for their naive assumptions, inadequate distinctions, and the researcher's zeal to generate answers rapidly. A summary of the implications of the symbol systems approach suggests that (1) every medium, including the microcomputer, will favor the transmission of certain symbol systems over others and is likely to require and cultivate different representational skills; (2) every medium is biased toward specific kinds of activities applied to its symbol systems, and is likely to cultivate a different set of mental operations; and (3) subjects will bring different expectations to different media. Twenty-four references are listed. (LMM)