Interrater and test-retest reliability of the Behavioral Assessment Scale of Oral Functions in Feeding were computed for two separate samples of severely and profoundly handicapped students having oral-motor and feeding disorders. Reliability was assessed using the intraclass correlation-generalizability theory approach and resulted in coefficients ranging from .68 to .84. These values were considered only marginally acceptable by current standards of instrument development advocated in the behavioral sciences. This paper discusses the difficulty in developing psychometrically satisfactory assessment instruments for use with the severely and profoundly developmentally disabled and gives implications for practice and future research.