Panel studies have frequently been used by mass communication researchers. While these studies allow researchers to draw inferences about mass communication effects over time, they also tend to sensitize respondents to the issue under study. A study was undertaken to investigate panel effects by examining the survey interview as a medium of communication. Unlike other models of survey research that have conceptualized the interview as an "upward flow" of communication, the study used an approach that investigated the flow of information from interviewer to respondent to other members of the social network in a "downward" two-step flow. Subjects were 81 college students, 53 of whom lived in the same dormitory. Of these, 27 were involved in the survey panel and were interviewed twice, the interviews separated by a two-week interval. The remaining 26 dorm students were interviewed only at the second interview period, as were 28 students who did not live in the dorm. The findings indicated that significant knowledge gain and diffusion effects can occur in and thus contaminate panel studies. (Author/FL)