Abstract
Thirty intermediate-grade subjects, who by the Informal Reading Inventory (IRI) criteria demonstrated in basal materials a fifth grade instructional level in reading, used their intuitive knowledge of language, when reading, as effectively as Biemiller and Weber found first-grade subjects using theirs. These subjects' miscues were analyzed according to concepts set forth in the Reading Miscue Inventory (RMI) which are very similar to the ideas used by Biemiller and Weber. Subjects' performances demonstrated that reading behavior is different when reading at the instructional and frustrational levels. When reading at the frustrational level, subjects tended to adhere more closely to the sound and graphic materials represented in the text than when reading at their instructional level. Miscues made at the frustrational level had a higher frequency of having the same grammatical function as that intended in the text than when reading at the instructional level. Grammatical and semantic miscues acceptable to a paragraph or the whole text are more apt to occur at the instructional level of reading. Grammatical and semantic miscues acceptable only in the sentence or phrase in which they occur are produced more often when reading at the frustrational level. The RMI concepts are very powerful for analyzing oral reading errors made within the boundary set by IRI concepts. The concepts in these two techniques should be synthesized.