This research examines farmers’ use of various sources and channels of conservation information in three midwestern US watersheds. A primary objective was to determine perceptual and farm structure factors influencing the use of particular information sources for farm-level decision-making. Data were collected from 1011 farm operators, the Maquoketa River watershed in east-central Iowa, the Lower Minnesota River watershed in southeast Minnesota, and the Darby Creek watershed in central Ohio. Respondents were asked to indicate frequency of use for 22 sources of conservation information identified from the literature and to rank the perceived importance of 11 of the most common communication channels for accessing agricultural information. Factor analysis was used to reduce the number of information sources to a smaller set of variables that explained much of the variance of the original data set. Selected elements of diffusion, risk communication, and farm structure theories were used to interpret the factor loadings and to identify predictors of information use. Regression analysis was used to test the communication source models developed for the overall data set and for each state. Descriptive findings revealed that farmers use multiple sources and channels when accessing soil and water conservation information. Substantial differences in information-use patterns were noted among the study watersheds. The results of the factor analysis showed that the 22 information sources could be categorized into six overarching groups based on their intercorrelation. The regression models were shown to vary widely in their predictive capacity, explaining from 1 to 29% of the total variance in source use. The variability noted among farmers’ information-use patterns and perceptions across the three study areas casts doubt on the value of broad-based or “shotgun” approaches for delivering agricultural information. The use of factor analysis has promise in future studies as a valuable tool for developing empirical measures of information use and improving measurement of key theoretical constructs in agricultural communication.