Noting that most classroom teachers use informal reading inventories (IRI) as diagnostic instruments, a study examined what oral reading accuracy level is most appropriate for the instructional level and whether repetitions should be counted as oral reading errors. Randomly selected students from the second through fifth grades at two elementary schools in Central Florida were tested and 136 test protocols were analyzed. The findings indicated that criteria of 90% word recognition accuracy with 70% comprehension for passage level one and 94% word recognition accuracy (70% comprehension) for levels two through four should be used with the Joels, Anderson, and Thompson (1983) IRI. The limited data available did not seem to support the inclusion of repetition among scorable errors nor is repetition an important criterion of oral reading accuracy for determining the instructional level on an IRI. (Tables indicating word recognition accuracy at each level of an IRI and percentage of oral reading accuracy with and without repetitions are included.) (SRT)