A study was conducted to provide insight into the consistency of reading achievement scores from four standardized tests. Several sets of data were compared to assess the accuracy of grade equivalents or instructional reading levels obtained on standardized tests. Each test was randomly assigned for administration to 23 third graders in a group setting during one of four consecutive mornings. The four tests were: (1) the Iowa Tests of Basic Skills (Form 4); (2) Stanford Achievement Test (Primary Level III--Form A); (3) Gates-MacGinitie Reading Test (Level C--Form 2); and (4) Metropolian Achievement Test (Elementary--Form F)--Metro. The resulting grade levels revealed significant discrepancies across the different tests. Seventy percent of the children received grade scores ranging over more than one year. A comparison of the instructional reading levels obtained from the Metro with those obtained from a subsequently administered informal reading inventory indicated that more than 50% of instructional reading levels from the Metro varied as much as two to five reader levels from those of the informal reading inventories. The results did not support the use of standardized test scores as adequate measures of reading achievement or as a substitute for individually administered informal reading inventories. (HTH)