A cohort analysis of data gathered in three national election surveys (1956, 1968, 1980) was used to study the effect of media use on political knowledge, which was divided into knowledge of issues, personalities, and political parties. Knowledge levels were calculated by creating indices from open-ended questions about why a person liked or disliked a candidate. Since the number of questions varied in each study, Z scores were calculated and crosstabulated with each age group. The subjects were divided into five age groups (18-29, 30-41, 42-53, 54-65, and 66 and older), and the data analyzed for age, period, and cohort effects. Results indicated that all age groups declined steadily in media use from 1956 to 1980. The youngest cohort showed a stronger increase in personality related knowledge than older cohorts; the youngest cohort also showed increased issue knowledge and political knowledge in general, while the older cohorts declined in both areas. The findings suggest that (1) the ability of the youngest cohort to increase in political knowledge while declining in media use may be explained by the existence of a "media generation," rather than a television generation; and (2) the youngest group is more efficient in its use of media than the older groups. (FL)