The magnitude of the cultivation effect for perceptual estimates of social reality has been shown to be affected by a number of contextual factors such as source priming and motivation to process information during judgment construction, and these contextual factors have been linked to the use of heuristic processing strategies when constructing judgments of frequency and probability (L. J. Shrum, in press). An experiment that manipulated data collection method explored the implications of these findings. A random sample of general population respondents were randomly assigned to either telephone or mail survey conditions. Because telephone surveys generally result in greater heuristic processing than mail surveys, telephone surveys were expected to produce larger cultivation effects than mail surveys. Results showed that not only were the magnitude of the estimates of affluence, crime, vice, marital discord, and occupations generally greater in the telephone than in the mail survey but the correlation of the estimates with amount of viewing was also greater in the telephone than in the mail survey. The implications for measuring the cultivation effect are discussed.