Inhabitants of rural areas represent an invaluable but neglected source of information for scientists concerned with land evaluation in the tropics. In this study, site suitability ratings were scientifically developed for six uses: coffee; beans; grazing; pigeon peas; garlic and rice. These ratings were compared to farmers' opinions of site suitability. The study area was Las Cuevas watershed, a tropical steepland region in the Dominican Republic. Suitability indices were computed using a multiplicative model and converted into four ordinal ratings. Chi-square and Kendall's tau-b statistics showed a clear statistical correlation between the author's ratings and farmers' evaluations of their land, confirming the hypothesis that suitability ratings are meaningful. The majority of farmers who disagreed with the author gave a more optimistic appraisal. This finding is consistent with studies of British farmers who over-estimate poor quality land and present an over-optimistic evaluation of land for familiar uses. The validity of the suitability rating procedure is also supported by statistical correlation between current land use and index values.