Thirty-two children aged 7 to 12 participated in a study to determine the reliability of the Ekwall Reading Inventory (ERI) and the Classroom Reading Inventory (CRI). The children were randomly assigned to take one of the two inventories, which were administered by four different specially trained teachers. The study used a test-retest design, in which the test was a "live" administration of one of the inventories by one teacher and the retest was the scoring of a videotape of the first teacher by a second. The test and retest for each child was compared on two dimensions of test scoring purity: (1) teacher adherence to test scoring procedures and (2) the use of teacher judgment to supersede strict scoring rules. Results showed that both the ERI and the CRI are reliable when used by professionals trained in their use. Teachers using the ERI scoring criteria and teacher judgment placed their pupils within one grade level 14 of 16 times, while teachers using the CRI with teacher judgment achieved agreement within one grade in 11 of 16 times. (FL)