ABSTRACTA converging approach is taken to study the construct of “success as a farmer” and to develop measures of success. It is assumed at the outset that “success” means more than one thing to experts in the area. Thus it is necessary to determine the definitions precisely in operational terms. This was accomplished by first asking experts in the area to provide examples of observable behavior that they felt were relevant to success or failure of farmers. A large list of behaviors (items) was obtained which provided the basis for the “relevant behaviors” approach.A “schools‐of‐thought” procedure required that another group of experts assign numerical values to each item on the basis of their evaluation of the behavior's relevance to success in farming. A principal component analysis of their relevance judgments indicated high agreement by the experts of the relevance of the majority of behaviors on the list and a disagreement between certain experts with regard to other behaviors on the list. The meaning of the school‐of‐thought on which there is disagreement is provided by examining the relevance judgments of the items. Component scores were then calculated for all items. Those with high scores on one component and low scores on the other defined each checklist for measuring the school of thought.A “scaled‐behavioral‐expectation” procedure required that an expert group, working together, categorize the behaviors into dimensions relevant to success as a farmer and then that other experts, working independently, place the behaviors in the categories. Those items which are consistently placed within a given category are then assigned scale positions on a rating scale by other groups of experts, working independently. Behaviors that are consistently given the same or nearly the same scale position are then used as “anchor points” on that rating scale.Farm Management Service fieldmen rated 116 cooperating farmers using the checklists and rating scales. These measures constituted the rating scale criterion variables. Data were collected on the total sample of 195 farmers for two other sets of criterion variables; aggregate farm record measures and ratio farm record measures. A factor analysis of each set of criterion variables revealed that each criterion space was multidimensional. The relationships between the rating space and each of the farm record spaces were determined by canonical correlation analysis and factor analysis. One dimension was found to be common to both spaces in each analysis.Biographical information, interest scores and ability test scores were obtained on 108 farmers in the sample. The relationship between the predictors and each criterion space was analyzed by canonical correlation analysis and factor analysis.Results indicated that an aggregate record variate, “land quality plus returns above costs and investment,” was related to the Overall Success, Technical Skill, and Business Skill rating scales. The aggregate record variate could also be predicted by Years of Formal Schooling, Years on Farm in Youth, Scaled Goal, Interest Maturity, Mechanical Comprehension, and Numerical Reasoning, with Specialization Level and Farmer Interest receiving negative weights. Other criterion variates were found and discussed. Exploration of the meaning of a criterion construct through the relevant behavior approach is suggested in areas where the criterion construct may have several definitions.