An articulatory inventory was administered to 19 dysarthric adults and scored using two judging formats--phoneme identification and traditional testing. Results indicated that samples judged using the traditional testing format, in which the judge knew the target phoneme, were consistently scored more accurately than those that had been judged using a phoneme identification format, in which the target was not known. Although overall both judging formats were characterized by high inter-rater reliability, the traditional testing format was less reliable than phoneme identification with samples obtained from severely involved speakers. Potential uses of articulatory inventories for dysarthric adults are described.