A pilot study explored the nature of the relationship between test and classroom evidence and considered how these different sources of evidence were used in assessing and promoting student growth and development in reading. Test estimates of third-grade children in two public elementary programs in New York City were compared with estimates of the difficulty of the material the children were reading at the time of the testing. This was done by comparing one index of student reading abilities: Degrees of Reading Power (DRP) test scores, with a second index, i.e., the DRP readability values of books the students had actually read. Results of the comparative analysis indicated relatively close agreement between the test and classroom records for children who performed well on the test. For low-scoring students, however, there was a clear discrepancy between the two sources of evidence. According to classroom records, these students were reading and comprehending books beyond, sometimes several years beyond, the test's estimate of the reading abilities. Low-scoring students were actually reading materials that test evidence predicted would be beyond their capabilities. Several questions for future research raised by these findings are briefly discussed. (Two tables of data are included.) (RS)