A study examined the shift of the United States public's primary news source from newspapers to radio and then to television between 1937 and 1987 to determine the validity of the generally accepted explanation that people have shifted media (displacement). Two alternative processes were discovered that could also account for this phenomenon: (1) the wording of the question has changed, and (2) people have grown up with different media--a "cohort" effect. A cohort analysis of the original Roper Poll data collected between 1937 and 1987 revealed that question wording differences, differences across cohorts, and displacement effects within cohorts were each responsible for some of this change in primary news source. The results also suggest that television has displaced radio more than newspapers, compatible with the concept of a functional equivalence between the two electronic media. (Five figures, 7 tables of data are included; 38 references are attached.) (Author/KEH)