To improve the accuracy of the informal reading inventory (IRI), a differential set of criteria is necessary for both word recognition and comprehension scores for different levels and reading conditions. In initial evaluation, word recognition scores should reflect only errors of insertions, omissions, mispronunciations, substitiutions, unkown words, and transpositions; symptomatic behavior should not be considered. After the student has read, comprehension questions should be asked on a literal level, on implicit understanding, on vocabulary, and on evaluative skill; all should be wholly context dependent. Baseline criteria for determining unsatisfactory reading should be established in comprehension, word recognition, and symptomatic behavior; comprehension is the most significant in determining placement. The initial task of the IRI is to place the student at his reading level; this decision is first made on the basis of quantitative data, then an error analysis gives qualitative information. Placement precedes analysis, but both are necessary for effective reading diagnosis and placement. (DF)